Mirrors/pdfcraft
1
0 Fork 0
mirror of https://github.com/storytold/pdfcraft.git synced 2026-11-13 01:41:18 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge upstream/main into fix/stroke-degenerate-and-dash

Browse Source 
vendor/README.md: main's rows kept; this branch's hayro-interpret patch is (61), the number its
code comments use, as main now uses (64).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Pq7HyDBXygTf7jcqdsjcfn
This commit is contained in:
mtlcl410andClaude Opus 5.5 committed 2026-11-06 01:26:22 +08:00
2 parents b1ba2111ab 496e37d72b
commit ba30b18ddf
328 files changed
+34737 -97433
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

.gitattributes
+3
View File
@@ -3,6 +3,9 @@
# signatures) and assets whose SHA-256 ATTRIBUTION.toml records (cargo xtask assets).
* -text
# The craft-i18n pin: CI and the build read it as one SHA, so it stays LF everywhere (readers trim it too).
craft-i18n.rev text eol=lf
# Not text at all: no textual diffs or merges either.
*.pdf binary
*.p12 binary
.github/workflows/release.yml
+1
View File
@@ -42,3 +42,4 @@ jobs:
AZURE_SIGNING_ENDPOINT: ${{ secrets.AZURE_SIGNING_ENDPOINT }}
AZURE_SIGNING_ACCOUNT: ${{ secrets.AZURE_SIGNING_ACCOUNT }}
AZURE_CERT_PROFILE: ${{ secrets.AZURE_CERT_PROFILE }}
LINUX_SIGNING_KEY: ${{ secrets.LINUX_SIGNING_KEY }}
AGENTS.md
+3 -3
View File
@@ -52,14 +52,14 @@ Anything else is forbidden, including assets whose licence you cannot prove from
- path or URL, title, author;
- source (URL, with version or commit);
- SPDX licence, and the licence file in the repo;
- SHA-256, except for translation catalogs (see below);
- SHA-256;
- how it is used.
- **Translation catalogs** (`crates/ui-egui/src/i18n/*.tsv`, `kind = "translation"`) are project files under the project licence. Their entries carry no `sha256`: git history records every change and who made it. Every other kind (fonts, images, data) is hashed.
- **Translations** live in craft-i18n (`locales/<code>/common.tsv` and `pdfcraft.tsv`), embedded at build time from the checkout pinned in `craft-i18n.rev`; PdfCraft keeps no catalogs. `ATTRIBUTION.toml` records them as one `[[build_input]]` (`kind = "translation"`); craft-i18n's own `ATTRIBUTION.md` credits the translators. Every `[[asset]]` (fonts, images, data) is hashed.
- **`ATTRIBUTION.md`** is generated from it (`cargo xtask assets --write`). Never edit it by hand.
- **`NOTICE`** keeps the licence notices that the licences require.
- **`cargo xtask assets`** runs in `cargo xtask ci`. It fails if:
- an asset file in the repository has no entry;
- an entry's hash is missing or does not match (translation catalogs carry none);
- an entry's hash is missing or does not match;
- a licence is not on the allowlist;
- a licence file is missing;
- an image, icon or font names Adobe as author or source, except a font licensed under OFL-1.1 alone.
ATTRIBUTION.md
+6 -37
View File
@@ -2,40 +2,12 @@
<!-- Generated from ATTRIBUTION.toml by `cargo xtask assets --write`. Do not edit by hand. -->
Every asset PdfCraft includes, bundles or uses to build its published material, with its author, source and licence. The policy is in [AGENTS.md](AGENTS.md) §1. The machine-readable list, with the SHA-256 of every asset but the translation catalogs, is [ATTRIBUTION.toml](ATTRIBUTION.toml). Licence texts are kept beside the assets and summarised in [NOTICE](NOTICE).
Every asset PdfCraft includes, bundles or uses to build its published material, with its author, source and licence. The policy is in [AGENTS.md](AGENTS.md) §1. The machine-readable list, with the SHA-256 of every asset, is [ATTRIBUTION.toml](ATTRIBUTION.toml). Licence texts are kept beside the assets and summarised in [NOTICE](NOTICE).
## In this repository (289)
## In this repository (257)
| Asset | Title | Author | Licence | Source | Used for |
|---|---|---|---|---|---|
| `crates/ui-egui/src/i18n/el.tsv` | Greek interface translation catalog | Camthalion77 and contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft English UI labels; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Greek UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/zh-hans.tsv` | Simplified Chinese interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Simplified Chinese UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/zh-hant.tsv` | Traditional Chinese (Taiwan) interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Traditional Chinese UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/cs.tsv` | Czech interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 signature-check and standard-identification wordings are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Czech UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/pt-br.tsv` | Brazilian Portuguese interface translation catalog | André Cardoso (andrecards), PdfCraft contributors | MIT OR Apache-2.0 | Clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources: the ~1,826-entry gist posted in issue #288 (https://gist.github.com/andrecards/ac388c25fba17312ce83207f8b7b467e), completed and reviewed by PdfCraft contributors; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table limit messages are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Brazilian Portuguese UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/es.tsv` | Spanish interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; the 2026-10-20 coverage additions are AI-assisted and not separately native-reviewed; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 signature-check wordings are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Spanish UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/fr.tsv` | French interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | French UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/de.tsv` | German interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | German UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/lo.tsv` | Lao interface translation catalog | @bounkirdni-2025 | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Lao UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/te.tsv` | Telugu interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Telugu UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/it.tsv` | Italian interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Italian UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/pl.tsv` | Polish interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 signature-check wordings are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Polish UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/ru.tsv` | Russian interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Russian UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/hu.tsv` | Hungarian interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Hungarian UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/bg.tsv` | Bulgarian interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Bulgarian UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/kk.tsv` | Kazakh interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of the English UI labels into Kazakh, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Kazakh UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/ja.tsv` | Japanese interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Japanese UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/uk.tsv` | Ukrainian interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Ukrainian UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/ar.tsv` | Arabic interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Arabic UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/ckb.tsv` | Central Kurdish (Sorani) interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Central Kurdish UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/fa.tsv` | Persian (Farsi) interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-25 initial catalog covering the current 2,776-entry key union, AI-assisted translations from English, not separately native-reviewed | Persian UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/tr.tsv` | Turkish interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Turkish UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/lt.tsv` | Lithuanian interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels and user-facing messages; original catalog reviewed by a native Lithuanian-speaking contributor; 2026-10-15 coverage additions are AI-assisted and not separately native-reviewed; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Lithuanian UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/hi.tsv` | Hindi interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Hindi UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/id.tsv` | Indonesian interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Indonesian UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/ko.tsv` | Korean interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Korean UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/ro.tsv` | Romanian interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels and user-facing messages; AI-assisted and not yet reviewed line by line by a native Romanian speaker; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Romanian UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/fi.tsv` | Finnish interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: AI-assisted (Claude Code) clean-room translations of PdfCraft's English UI labels and user-facing messages; not yet proofread by a native Finnish speaker; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 calculation-order, text-format, print, thumbnail, layer, comment-reshape and signature labels (previously English placeholders) are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Finnish UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `assets/icons/accessibility.svg` | Lucide icon "accessibility" | Lucide Contributors (portions Cole Bemis, Feather) | ISC AND MIT | https://github.com/lucide-icons/lucide (npm lucide-static 1.49.0, icons/accessibility.svg) | UI icon, embedded via crates/ui-egui/src/icon_data.rs |
| `assets/icons/align-left.svg` | Lucide icon "align-left" | Lucide Contributors (portions Cole Bemis, Feather) | ISC AND MIT | https://github.com/lucide-icons/lucide (npm lucide-static 1.49.0, icons/align-left.svg) | UI icon, embedded via crates/ui-egui/src/icon_data.rs |
| `assets/icons/align-center.svg` | Lucide icon "align-center" | Lucide Contributors (portions Cole Bemis, Feather) | ISC AND MIT | https://github.com/lucide-icons/lucide (npm lucide-static 1.49.0, icons/align-center.svg) | UI icon, embedded via crates/ui-egui/src/icon_data.rs |
@@ -231,6 +203,7 @@ Every asset PdfCraft includes, bundles or uses to build its published material,
| `vendor/hayro-interpret/assets/FoxitSymbol.pfb` | FoxitSymbol (standard-14 substitute font) | Foxit Software Inc. / PDFium Authors | BSD-3-Clause | https://pdfium.googlesource.com/pdfium (via hayro-interpret 0.7.0 assets/) | Renders non-embedded standard-14 fonts (vendored hayro-interpret) |
| `vendor/hayro-interpret/assets/CGATS001Compat-v2-micro.icc` | CGATS001Compat v2 micro ICC profile | Clinton Ingram (saucecontrol/Compact-ICC-Profiles) | CC0-1.0 | https://github.com/saucecontrol/Compact-ICC-Profiles | Default CMYK colour conversion (vendored hayro-interpret) |
| `crates/fonts/src/std14_metrics.rs` | Standard-14 font advance widths and encodings (as Rust tables) | Adobe (Core14 AFM metrics, PDF specification encodings); tables generated by the PdfCraft contributors | MIT OR Apache-2.0 | ISO 32000-2 Annex D and the Adobe Core14 AFM metrics. Generated by joining the published WinAnsi, Symbol and ZapfDingbats encodings (code to glyph name) with the per-glyph advance widths, then spot-checked against the specification by hand in the tests of crates/fonts/src/std14.rs. Data only: the code that reads these tables lives in std14.rs, so that ordinary edits do not churn this file's hash. | Advance widths for the fourteen standard fonts, which a PDF may name without embedding. Non-visual data, not a typeface: the item AGENTS.md 1.1 permits, held first-party so generated appearance streams and the renderer agree. |
| `crates/prc/src/acof/table.rs` | PRC Double prefix-code table (as a Rust table) | Kristof Ralovich (prc-rs); the code table itself is defined by the PRC format (ISO 14739-1, working draft clause 11.17.2) | MIT | Rows taken from prc-rs src/double.rs at commit 287b671bb5b31b05efb5ec1b98f7c359b9e4dd7d (https://github.com/ralovich/prc-rs), mechanically reformatted into (code length, code, entry) tuples; checked to be a complete prefix code by the tests in crates/prc/src/acof.rs. Data only: the code that reads the table lives in acof.rs. The MIT notice is kept in the file header. | Decodes every PRC Double (coordinates, transforms, colours) in 3D annotations' PRC streams. Non-visual data. |
| `vendor/hayro-cmap/assets/cmaps.brotli` | Adobe CMap resources (predefined CJK CMaps) | Adobe | BSD-3-Clause | https://github.com/adobe-type-tools/cmap-resources (via hayro-cmap 0.1.0 assets/) | Character-code mapping tables required to read CJK PDFs. Non-visual data, allowed by AGENTS.md §1.1 |
| `docs/brand/artcraft-logo-white.png` | ArtCraft brand: artcraft-logo-white | ArtCraft Team | LicenseRef-Storyteller-Trademark | ArtCraft Team brand assets (provided by the project owner, 2026-10-01; identical copies in storytold/photocraft) | ArtCraft branding next to the community links (About dialog, Help menu, home screen) and in docs. ArtCraft Team trademark, not open source; may be used only unmodified within PdfCraft; forks and modified versions must remove it. |
| `docs/brand/artcraft-logo-white.svg` | ArtCraft brand: artcraft-logo-white | ArtCraft Team | LicenseRef-Storyteller-Trademark | ArtCraft Team brand assets (provided by the project owner, 2026-10-01; identical copies in storytold/photocraft) | ArtCraft branding next to the community links (About dialog, Help menu, home screen) and in docs. ArtCraft Team trademark, not open source; may be used only unmodified within PdfCraft; forks and modified versions must remove it. |
@@ -286,17 +259,12 @@ Every asset PdfCraft includes, bundles or uses to build its published material,
| `packaging/macos/dmg/background.tiff` | PdfCraft macOS DMG window background (rendered) | @XusBadia | MIT OR Apache-2.0 | Rendered from packaging/macos/dmg/background.svg by packaging/macos/dmg/generate.py | 1x + 2x HiDPI 16-colour palette TIFF copied into the macOS DMG by packaging/macos/package.sh; includes the app icon and the free LittleCMS built-in sRGB profile |
| `crates/sign/tests/data/x509-rsa-sha1.pdf` | One-page PDF with a legacy adbe.x509.rsa_sha1 signature by the RSA test key | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: hand-written PDF objects; signature made with `openssl dgst -sha1 -sign` over its byte ranges (see crates/sign/tests/data/README.md) | Signature validation tests (crates/sign/tests/it/pdf.rs): the legacy adbe.x509.rsa_sha1 format |
| `assets/icons/monitor.svg` | Lucide icon "monitor" | Lucide Contributors (portions Cole Bemis, Feather) | ISC AND MIT | https://github.com/lucide-icons/lucide (npm lucide-static 1.49.0, icons/monitor.svg) | UI icon, embedded via crates/ui-egui/src/icon_data.rs |
| `crates/ui-egui/src/i18n/vi.tsv` | Vietnamese interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels; PhotoCraft and VectorCraft Vietnamese catalogs used only as translation memory when English source and semantic context matched; original catalog reviewed by a native Vietnamese-speaking contributor; 2026-10-15 coverage additions are AI-assisted and not separately native-reviewed; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 signature-check wordings are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Vietnamese UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/sl.tsv` | Slovene interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Slovene UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `vendor/vello_common/assets/probe.png` | vello_common probe reference image (rendered from probe.rgba) | Linebender vello contributors | MIT OR Apache-2.0 | vello_common 0.0.8 (crates.io), vendored in vendor/vello_common; regenerated by its vello_sparse_tests regenerate_probe_reference binary | Reference picture of the vendored vello_common probe module, for people only; the build reads probe.rgba, not this file |
| `crates/ui-egui/src/i18n/nb.tsv` | Norwegian Bokmål interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Norwegian Bokmål UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `assets/icons/superscript.svg` | Lucide icon "superscript" | Lucide Contributors (portions Cole Bemis, Feather) | ISC AND MIT | https://github.com/lucide-icons/lucide (npm lucide-static 1.49.0, icons/superscript.svg) | UI icon, embedded via crates/ui-egui/src/icon_data.rs |
| `assets/icons/subscript.svg` | Lucide icon "subscript" | Lucide Contributors (portions Cole Bemis, Feather) | ISC AND MIT | https://github.com/lucide-icons/lucide (npm lucide-static 1.49.0, icons/subscript.svg) | UI icon, embedded via crates/ui-egui/src/icon_data.rs |
| `assets/icons/pilcrow.svg` | Lucide icon "pilcrow" | Lucide Contributors (portions Cole Bemis, Feather) | ISC AND MIT | https://github.com/lucide-icons/lucide (npm lucide-static 1.49.0, icons/pilcrow.svg) | UI icon, embedded via crates/ui-egui/src/icon_data.rs |
| `assets/icons/a-large-small.svg` | Lucide icon "a-large-small" | Lucide Contributors (portions Cole Bemis, Feather) | ISC AND MIT | https://github.com/lucide-icons/lucide (npm lucide-static 1.49.0, icons/a-large-small.svg) | UI icon, embedded via crates/ui-egui/src/widgets/icon_data.rs |
| `assets/icons/move-horizontal.svg` | Lucide icon "move-horizontal" | Lucide Contributors (portions Cole Bemis, Feather) | ISC AND MIT | https://github.com/lucide-icons/lucide (npm lucide-static 1.49.0, icons/move-horizontal.svg) | UI icon, embedded via crates/ui-egui/src/icon_data.rs |
| `crates/ui-egui/src/i18n/sq.tsv` | Albanian interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 signature-appearance labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Albanian UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/th.tsv` | Thai interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Thai UI strings, compiled in by crates/ui-egui/src/i18n.rs |
| `crates/ui-egui/src/i18n/ca.tsv` | Catalan interface translation catalog | PdfCraft contributors | MIT OR Apache-2.0 | Contributor-original: clean-room translations of PdfCraft's English UI labels, written from their meaning; no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed | Catalan UI strings, compiled in by crates/ui-egui/src/i18n.rs |
## Compiled in through dependencies (6)
@@ -338,10 +306,11 @@ Fonts are fetched by `cargo xtask demo-pdf` into `target/demo-fonts/`, OCR model
| `text-detection.onnx` | Ocrs text detection model | Robert Knight (ocrs-models; trained on HierText, CC-BY-SA-4.0) | CC-BY-SA-4.0 | https://github.com/robertknight/ocrs/blob/0c826441362a701c55560fefc17bf356b6d1037e/ocrs/examples/download-models.sh (author-published ONNX models; model card: https://huggingface.co/robertknight/ocrs/blob/df0edd170279ab971b53e094c627255a87e1a503/README.md, cc-by-sa-4.0) | Scan & OCR: finds words on page images (cargo xtask models) |
| `text-recognition.onnx` | Ocrs text recognition model | Robert Knight (ocrs-models; trained on HierText, CC-BY-SA-4.0) | CC-BY-SA-4.0 | https://github.com/robertknight/ocrs/blob/0c826441362a701c55560fefc17bf356b6d1037e/ocrs/examples/download-models.sh (author-published ONNX models; model card: https://huggingface.co/robertknight/ocrs/blob/df0edd170279ab971b53e094c627255a87e1a503/README.md, cc-by-sa-4.0) | Scan & OCR: reads the words it finds (cargo xtask models) |
## Optional build inputs (1)
## Optional build inputs (2)
Not in this repository and never downloaded by it: compiled in only when the build sets the option (official releases do). Each input attributes its own files.
| Input | Option | Title | Author | Licence | Source | Attribution | Used for |
|---|---|---|---|---|---|---|---|
| `craft-i18n` | `CRAFT_I18N_DIR` | Interface translations: the craft-i18n catalogs (locales/<code>/common.tsv and pdfcraft.tsv) of the 34 languages PdfCraft offers besides English | The craft-i18n translators, including every translator of PdfCraft's former in-repo catalogs (crates/ui-egui/src/i18n/*.tsv, migrated at pdfcraft@ecdd1e6) | MIT OR Apache-2.0 | https://github.com/storytold/craft-i18n (pinned by commit in craft-i18n.rev, which CI, releases and the flatpak manifests read) | https://github.com/storytold/craft-i18n/blob/main/ATTRIBUTION.md | Interface text in languages other than English, embedded by crates/ui-egui/build.rs and looked up through craft-libs' craft-i18n crate; a build without CRAFT_I18N_DIR is English only. Clean-room translations from PdfCraft's English interface labels |
| `craft-fonts` | `CRAFT_FONTS_DIR` | Japanese fonts: BIZ UDPGothic (Regular, Bold), Shippori Mincho, BIZ UDMincho; native Simplified Chinese font: LXGW WenKai Lite; Arabic font: Noto Sans Arabic | The BIZ UDGothic and BIZ UDMincho Project Authors (Morisawa Inc.), The Shippori Mincho Project Authors (FONTDASU), LXGW and the original Klee Project Authors / Fontworks (LXGW WenKai Lite), The Noto Project Authors (Noto Sans Arabic) | OFL-1.1 | https://github.com/storytold/craft-fonts (pinned by commit in .github/workflows/release.yml and .github/workflows/ci.yml); LXGW WenKai Lite from https://github.com/lxgw/LxgwWenKai-Lite/releases/tag/v1.530 (commit b712c83d98acf977bacfb7fb010207802808bda9) | https://github.com/storytold/craft-fonts/blob/main/ATTRIBUTION.md | Japanese fallback in the interface (BIZ UDPGothic first) and Japanese glyphs in edited PDF text (Shippori Mincho); Simplified Chinese interface fallback on native builds only (LXGW WenKai Lite Regular, font SHA-256 f799a538a3355877e597a0ae04050e897932c9ae7689eb9227c2953c5dd549ea, OFL SHA-256 0333a85fe077b6849f8e417acc44e8bd879442ceb681e271acf93e2d57fd4e2e, read from fonts/simplified-chinese-manifest.txt and never embedded on wasm); Arabic interface text, file names and document titles (Noto Sans Arabic); the web build embeds BIZ UDPGothic Regular and the pinned small Arabic, Hebrew, Thai and Devanagari faces; the selected craft-fonts revision contains no Telugu face. Other non-Japanese CJK faces retain the general 4 MiB cap; the reviewed native-only LXGW fallback has a separate 16 MiB cap. Open-source OFL releases are permitted under current AGENTS.md §1.1 |
ATTRIBUTION.toml
+34 -341
View File
@@ -1,296 +1,15 @@
[[asset]]
path = "crates/ui-egui/src/i18n/el.tsv"
title = "Greek interface translation catalog"
author = "Camthalion77 and contributors"
source = "Contributor-original: clean-room translations of PdfCraft English UI labels; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Greek UI strings, compiled in by crates/ui-egui/src/i18n.rs"
# ATTRIBUTION.toml: one entry per asset (AGENTS.md §1). The source of truth for provenance and licences.
# ATTRIBUTION.md is generated from this file: `cargo xtask assets --write`. Checked by `cargo xtask assets` (part of `xtask ci`).
#
# [[asset]] a file in this repository (committed or vendored); `sha256` must match the file.
# Translation catalogs (kind = "translation", crates/ui-egui/src/i18n/*.tsv) are project files
# under the project licence and carry no `sha256`: git history records their provenance.
# [[bundled]] a file compiled into PdfCraft by a Cargo dependency (its crate version must match Cargo.lock).
# [[fetched]] downloaded at build time by xtask (never committed), pinned by `sha256`.
# [[build_input]] an optional build input outside this repository (e.g. craft-fonts via CRAFT_FONTS_DIR);
# [[build_input]] an optional build input outside this repository (craft-fonts via CRAFT_FONTS_DIR, craft-i18n via CRAFT_I18N_DIR);
# never committed or downloaded here, its own repository attributes each file.
#
# licence: SPDX expression from the allowlist in AGENTS.md §1.2. adobe_data = true marks the closed list of
# Adobe-authored non-visual technical data allowed by AGENTS.md §1.1.
[[asset]]
path = "crates/ui-egui/src/i18n/zh-hans.tsv"
title = "Simplified Chinese interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Simplified Chinese UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/zh-hant.tsv"
title = "Traditional Chinese (Taiwan) interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Traditional Chinese UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/cs.tsv"
title = "Czech interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 signature-check and standard-identification wordings are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Czech UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/pt-br.tsv"
title = "Brazilian Portuguese interface translation catalog"
author = "André Cardoso (andrecards), PdfCraft contributors"
source = "Clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources: the ~1,826-entry gist posted in issue #288 (https://gist.github.com/andrecards/ac388c25fba17312ce83207f8b7b467e), completed and reviewed by PdfCraft contributors; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table limit messages are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Brazilian Portuguese UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/es.tsv"
title = "Spanish interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; the 2026-10-20 coverage additions are AI-assisted and not separately native-reviewed; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 signature-check wordings are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Spanish UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/fr.tsv"
title = "French interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "French UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/de.tsv"
title = "German interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "German UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/lo.tsv"
title = "Lao interface translation catalog"
author = "@bounkirdni-2025"
source = "Contributor-original: clean-room translations of PdfCraft English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Lao UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/te.tsv"
title = "Telugu interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Telugu UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/it.tsv"
title = "Italian interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Italian UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/pl.tsv"
title = "Polish interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 signature-check wordings are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Polish UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/ru.tsv"
title = "Russian interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Russian UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/hu.tsv"
title = "Hungarian interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Hungarian UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/bg.tsv"
title = "Bulgarian interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Bulgarian UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/kk.tsv"
title = "Kazakh interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of the English UI labels into Kazakh, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Kazakh UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/ja.tsv"
title = "Japanese interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Japanese UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/uk.tsv"
title = "Ukrainian interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Ukrainian UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/ar.tsv"
title = "Arabic interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Arabic UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/ckb.tsv"
title = "Central Kurdish (Sorani) interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Central Kurdish UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/fa.tsv"
title = "Persian (Farsi) interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-25 initial catalog covering the current 2,776-entry key union, AI-assisted translations from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Persian UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/tr.tsv"
title = "Turkish interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Turkish UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/lt.tsv"
title = "Lithuanian interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels and user-facing messages; original catalog reviewed by a native Lithuanian-speaking contributor; 2026-10-15 coverage additions are AI-assisted and not separately native-reviewed; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Lithuanian UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/hi.tsv"
title = "Hindi interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Hindi UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/id.tsv"
title = "Indonesian interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Indonesian UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/ko.tsv"
title = "Korean interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Korean UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/ro.tsv"
title = "Romanian interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels and user-facing messages; AI-assisted and not yet reviewed line by line by a native Romanian speaker; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Romanian UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/fi.tsv"
title = "Finnish interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: AI-assisted (Claude Code) clean-room translations of PdfCraft's English UI labels and user-facing messages; not yet proofread by a native Finnish speaker; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 calculation-order, text-format, print, thumbnail, layer, comment-reshape and signature labels (previously English placeholders) are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Finnish UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "assets/icons/accessibility.svg"
title = "Lucide icon \"accessibility\""
@@ -2437,6 +2156,17 @@ adobe_data = true
usage = "Advance widths for the fourteen standard fonts, which a PDF may name without embedding. Non-visual data, not a typeface: the item AGENTS.md 1.1 permits, held first-party so generated appearance streams and the renderer agree."
sha256 = "f225e0b044875b8fa6375b6b5ad935f0d99fcc1ad03ecf9f0baae2cedbfca4d2"
[[asset]]
path = "crates/prc/src/acof/table.rs"
title = "PRC Double prefix-code table (as a Rust table)"
author = "Kristof Ralovich (prc-rs); the code table itself is defined by the PRC format (ISO 14739-1, working draft clause 11.17.2)"
source = "Rows taken from prc-rs src/double.rs at commit 287b671bb5b31b05efb5ec1b98f7c359b9e4dd7d (https://github.com/ralovich/prc-rs), mechanically reformatted into (code length, code, entry) tuples; checked to be a complete prefix code by the tests in crates/prc/src/acof.rs. Data only: the code that reads the table lives in acof.rs. The MIT notice is kept in the file header."
licence = "MIT"
licence_file = "licenses/prc-rs-LICENSE.txt"
kind = "data"
usage = "Decodes every PRC Double (coordinates, transforms, colours) in 3D annotations' PRC streams. Non-visual data."
sha256 = "8e705c5cde6a3ef778e8aca69759e38845b228b3f7f12949e587b19855d2114c"
[[bundled]]
crate = "epaint_default_fonts"
version = "0.37.0"
@@ -3381,6 +3111,17 @@ kind = "model"
usage = "Scan & OCR: reads the words it finds (cargo xtask models)"
sha256 = "86c145c2edb96c8caed5b1ebb8f44d706408922211451c309c625157dd6061c5"
[[build_input]]
name = "craft-i18n"
option = "CRAFT_I18N_DIR"
title = "Interface translations: the craft-i18n catalogs (locales/<code>/common.tsv and pdfcraft.tsv) of the 34 languages PdfCraft offers besides English"
author = "The craft-i18n translators, including every translator of PdfCraft's former in-repo catalogs (crates/ui-egui/src/i18n/*.tsv, migrated at pdfcraft@ecdd1e6)"
source = "https://github.com/storytold/craft-i18n (pinned by commit in craft-i18n.rev, which CI, releases and the flatpak manifests read)"
licence = "MIT OR Apache-2.0"
attribution = "https://github.com/storytold/craft-i18n/blob/main/ATTRIBUTION.md"
kind = "translation"
usage = "Interface text in languages other than English, embedded by crates/ui-egui/build.rs and looked up through craft-libs' craft-i18n crate; a build without CRAFT_I18N_DIR is English only. Clean-room translations from PdfCraft's English interface labels"
[[build_input]]
name = "craft-fonts"
option = "CRAFT_FONTS_DIR"
@@ -3414,26 +3155,6 @@ kind = "icon"
usage = "UI icon, embedded via crates/ui-egui/src/icon_data.rs"
sha256 = "d1a443233345724859de8e4fb48968ba91e885e12ddbd710701548bd91b39428"
[[asset]]
path = "crates/ui-egui/src/i18n/vi.tsv"
title = "Vietnamese interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels; PhotoCraft and VectorCraft Vietnamese catalogs used only as translation memory when English source and semantic context matched; original catalog reviewed by a native Vietnamese-speaking contributor; 2026-10-15 coverage additions are AI-assisted and not separately native-reviewed; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 signature-check wordings are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Vietnamese UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/sl.tsv"
title = "Slovene interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-18 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Slovene UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "vendor/vello_common/assets/probe.png"
@@ -3446,16 +3167,6 @@ kind = "image"
usage = "Reference picture of the vendored vello_common probe module, for people only; the build reads probe.rgba, not this file"
sha256 = "78c59a0cfd2c829909d77006f94e34dac1ae5d1f90e7f10f72e84de541fef697"
[[asset]]
path = "crates/ui-egui/src/i18n/nb.tsv"
title = "Norwegian Bokmål interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Norwegian Bokmål UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "assets/icons/superscript.svg"
title = "Lucide icon \"superscript\""
@@ -3489,6 +3200,17 @@ kind = "icon"
usage = "UI icon, embedded via crates/ui-egui/src/icon_data.rs"
sha256 = "d344d9f9222f6c1b698c0c82a753a2212d9bd7785f06c00cd6e4910fe2fee8f8"
[[asset]]
path = "assets/icons/a-large-small.svg"
title = "Lucide icon \"a-large-small\""
author = "Lucide Contributors (portions Cole Bemis, Feather)"
source = "https://github.com/lucide-icons/lucide (npm lucide-static 1.49.0, icons/a-large-small.svg)"
licence = "ISC AND MIT"
licence_file = "assets/icons/LICENSE-lucide.txt"
kind = "icon"
usage = "UI icon, embedded via crates/ui-egui/src/widgets/icon_data.rs"
sha256 = "d4985fc71174d108e322c3b25cb6abe2837a115fe3f239a1918645db436743ec"
[[asset]]
path = "assets/icons/move-horizontal.svg"
title = "Lucide icon \"move-horizontal\""
@@ -3500,32 +3222,3 @@ kind = "icon"
usage = "UI icon, embedded via crates/ui-egui/src/icon_data.rs"
sha256 = "864ab2f19254e6ac7240bc6a98323197332e8f714d4d3f94831f41309e5ef2f7"
[[asset]]
path = "crates/ui-egui/src/i18n/sq.tsv"
title = "Albanian interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 signature-appearance labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Albanian UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/th.tsv"
title = "Thai interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Thai UI strings, compiled in by crates/ui-egui/src/i18n.rs"
[[asset]]
path = "crates/ui-egui/src/i18n/ca.tsv"
title = "Catalan interface translation catalog"
author = "PdfCraft contributors"
source = "Contributor-original: clean-room translations of PdfCraft's English UI labels, written from their meaning; no proprietary localisation resources; 2026-10-22 page-gap labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 text-field-table labels are contributor-original AI-assisted translations from English, not separately native-reviewed; 2026-10-22 secured-document notice labels are contributor-original AI-assisted translations from English, not separately native-reviewed; the 2026-10-22 reply-count label is an AI-assisted translation from English, not separately native-reviewed"
licence = "MIT OR Apache-2.0"
licence_file = "LICENSE-MIT"
kind = "translation"
usage = "Catalan UI strings, compiled in by crates/ui-egui/src/i18n.rs"
CLAUDE.md
+1 -1
View File
@@ -23,7 +23,7 @@ PdfCraft is a clean-room, open-source, Rust-native PDF application targeting Ado
- **Layering.** Nothing below L7 depends on egui/winit/eframe/rfd (`plan/architecture.md` §3). `cos`/`filters`/`crypt`/`arlington` stay standalone.
- **Page input.** Page modules read `PagePointer` (`crates/ui-egui/src/page_pointer.rs`), never egui's raw pointer state (`hover_pos`, `interact_pos`, `latest_pos`, `press_origin`, button reads); `page_pointer::tests::page_modules_read_the_pointer_only_through_page_pointer` enforces it.
- **Page cursor.** Page modules claim the cursor for what is under the pointer (`view.cursor.claim(Layer::…, icon)`, `crates/ui-egui/src/canvas/page_cursor.rs`), never set it (`set_cursor_icon`, `cursor_icon =`, `on_hover_cursor`); `document_area` sets it once, from the claim on the highest layer the active tool reaches. `page_pointer::tests::page_modules_claim_the_cursor_and_never_set_it` enforces it.
- **Page presses.** A page module decides what of its own is under a point in one hit test, which returns a `Hit` (owner, layer, cursor, tip; `crates/ui-egui/src/canvas/page_router.rs`) and gives its cursor, its tip and its press. It acts on a press only when the router says the press is its own (`view.router.owner(page)`), never by testing the press point again or by asking another module whether it wants it. Escape cancels any press the pages hold (`cancel_on_escape`), and that press is ignored until the button is up. `page_modules_take_presses_from_the_router_without_repeating_hit_tests` guards the centralized hit calls and captured-owner handlers; runtime routing tests cover gesture behavior.
- **Page presses.** A page module decides what of its own is under a point in one hit test, which returns a `Hit` (owner, layer, cursor, tip; `crates/ui-egui/src/canvas/page_router.rs`) and gives its cursor, its tip and its press. It acts on a press only when the press is its own or, for the active tool, nobody's: the page view tells it which (`PagePress`: its own, nobody's, or another page's), and it never tests the press point again or asks another module whether it wants it. Escape cancels any press the pages hold (`cancel_on_escape`), and that press is ignored until the button is up. `page_pointer::tests::page_modules_act_on_a_press_only_through_the_router` enforces it.
- **Fidelity.** The PDF object graph is the model. Preserve unknown data. Saves are incremental unless a full rewrite is required. Never silently drop or repair data without recording it.
- **Never panic.** Every PDF, script, MCP call, settings file and keystroke is untrusted input, and none of it may crash the app or lose the user's work. This outranks feature work. Return `Result` (the crate's error enum) and propagate with `?`, or fall back leniently and record it. In non-test code:
- No `unwrap`/`expect`/`panic!`/`unreachable!`/`todo!` unless it is provably infallible, with a comment saying why.
Cargo.lock
Generated
+67 -36
View File
@@ -1303,6 +1303,15 @@ dependencies = [
"zip",
]
[[package]]
name = "craft-web-host"
version = "0.1.0"
source = "git+https://github.com/storytold/craft-libs?rev=2e2312ddaf21bc3d85cd3395d0955c7d239306f6#2e2312ddaf21bc3d85cd3395d0955c7d239306f6"
dependencies = [
"wasm-bindgen",
"web-sys",
]
[[package]]
name = "craft-winprint-gdi"
version = "0.1.0"
@@ -4208,7 +4217,7 @@ dependencies = [
[[package]]
name = "pdfcraft"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"chrono",
"craft-update",
@@ -4232,7 +4241,7 @@ dependencies = [
[[package]]
name = "pdfcraft-a11y"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"pdfcraft-content",
"pdfcraft-cos",
@@ -4244,7 +4253,7 @@ dependencies = [
[[package]]
name = "pdfcraft-annot"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"hayro-syntax",
"pdfcraft-cos",
@@ -4255,7 +4264,7 @@ dependencies = [
[[package]]
name = "pdfcraft-audit"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"pdfcraft-content",
"pdfcraft-cos",
@@ -4266,7 +4275,7 @@ dependencies = [
[[package]]
name = "pdfcraft-automation"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"base64 0.23.1",
"flate2",
@@ -4277,6 +4286,8 @@ dependencies = [
"pdfcraft-model",
"pdfcraft-organize",
"pdfcraft-platform",
"pdfcraft-prc",
"pdfcraft-prc-render",
"pdfcraft-render",
"pdfcraft-xfa",
"png 0.18.1",
@@ -4288,7 +4299,7 @@ dependencies = [
[[package]]
name = "pdfcraft-cli"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"base64 0.23.1",
"pdfcraft-audit",
@@ -4306,11 +4317,11 @@ dependencies = [
[[package]]
name = "pdfcraft-compare"
version = "0.18.0"
version = "0.18.1"
[[package]]
name = "pdfcraft-content"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"hayro-syntax",
"pdfcraft-cos",
@@ -4318,7 +4329,7 @@ dependencies = [
[[package]]
name = "pdfcraft-cos"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"chrono",
"hayro-syntax",
@@ -4333,7 +4344,7 @@ dependencies = [
[[package]]
name = "pdfcraft-create"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"encoding_rs",
"hayro-ccitt",
@@ -4352,7 +4363,7 @@ dependencies = [
[[package]]
name = "pdfcraft-crypt"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"aes",
"md-5",
@@ -4364,7 +4375,7 @@ dependencies = [
[[package]]
name = "pdfcraft-edit"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"hayro-syntax",
"kurbo 0.13.1",
@@ -4382,7 +4393,7 @@ dependencies = [
[[package]]
name = "pdfcraft-engine"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"chrono",
"craft-docpath",
@@ -4426,14 +4437,14 @@ dependencies = [
[[package]]
name = "pdfcraft-export"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"flate2",
]
[[package]]
name = "pdfcraft-filters"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"fax",
"flate2",
@@ -4446,7 +4457,7 @@ dependencies = [
[[package]]
name = "pdfcraft-fonts"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"harfrust",
"pdfcraft-cos",
@@ -4456,7 +4467,7 @@ dependencies = [
[[package]]
name = "pdfcraft-forms"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"hayro-syntax",
"pdfcraft-content",
@@ -4469,11 +4480,11 @@ dependencies = [
[[package]]
name = "pdfcraft-geom"
version = "0.18.0"
version = "0.18.1"
[[package]]
name = "pdfcraft-js"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"boa_engine",
"chrono",
@@ -4481,7 +4492,7 @@ dependencies = [
[[package]]
name = "pdfcraft-measure"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"pdfcraft-annot",
"pdfcraft-content",
@@ -4494,14 +4505,14 @@ dependencies = [
[[package]]
name = "pdfcraft-model"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"pdfcraft-cos",
]
[[package]]
name = "pdfcraft-ocr"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"ocrs",
"pdfcraft-fonts",
@@ -4512,7 +4523,7 @@ dependencies = [
[[package]]
name = "pdfcraft-optimize"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"image",
"pdfcraft-annot",
@@ -4528,11 +4539,11 @@ dependencies = [
[[package]]
name = "pdfcraft-oracle"
version = "0.18.0"
version = "0.18.1"
[[package]]
name = "pdfcraft-organize"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"hayro-syntax",
"pdfcraft-content",
@@ -4543,7 +4554,7 @@ dependencies = [
[[package]]
name = "pdfcraft-platform"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"async-channel",
"futures-lite",
@@ -4551,9 +4562,27 @@ dependencies = [
"zbus",
]
[[package]]
name = "pdfcraft-prc"
version = "0.18.1"
dependencies = [
"flate2",
"proptest",
"thiserror 2.0.21",
]
[[package]]
name = "pdfcraft-prc-render"
version = "0.18.1"
dependencies = [
"flate2",
"pdfcraft-prc",
"thiserror 2.0.21",
]
[[package]]
name = "pdfcraft-preflight"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"flate2",
"pdfcraft-content",
@@ -4564,7 +4593,7 @@ dependencies = [
[[package]]
name = "pdfcraft-print"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"craft-print",
"pdfcraft-content",
@@ -4576,7 +4605,7 @@ dependencies = [
[[package]]
name = "pdfcraft-redact"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"pdfcraft-annot",
"pdfcraft-content",
@@ -4593,7 +4622,7 @@ dependencies = [
[[package]]
name = "pdfcraft-render"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"bytemuck",
"flate2",
@@ -4612,7 +4641,7 @@ dependencies = [
[[package]]
name = "pdfcraft-sign"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"aes",
"aws-lc-rs",
@@ -4648,17 +4677,18 @@ dependencies = [
[[package]]
name = "pdfcraft-testkit"
version = "0.18.0"
version = "0.18.1"
[[package]]
name = "pdfcraft-ui-egui"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"arboard",
"base64 0.23.1",
"block2 0.6.2",
"bytemuck",
"chrono",
"craft-i18n",
"craft-keymap",
"craft-menu",
"craft-quit",
@@ -4706,8 +4736,9 @@ dependencies = [
[[package]]
name = "pdfcraft-web"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"craft-web-host",
"eframe",
"egui",
"js-sys",
@@ -4718,7 +4749,7 @@ dependencies = [
[[package]]
name = "pdfcraft-xfa"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"base64 0.23.1",
"image",
@@ -4732,7 +4763,7 @@ dependencies = [
[[package]]
name = "pdfcraft-xfdf"
version = "0.18.0"
version = "0.18.1"
dependencies = [
"pdfcraft-annot",
"pdfcraft-cos",
Cargo.toml
+40 -34
View File
@@ -5,7 +5,7 @@ exclude = ["vendor"]
default-members = ["crates/*", "apps/pdfcraft", "apps/pdfcraft-cli", "xtask"]
[workspace.package]
version = "0.18.0"
version = "0.18.1"
edition = "2024"
license = "MIT OR Apache-2.0"
rust-version = "1.95"
@@ -19,6 +19,10 @@ craft-quit = { git = "https://github.com/storytold/craft-libs", version = "0.1.0
# One menu definition for the native macOS menu bar and the in-window bar elsewhere.
craft-menu = { git = "https://github.com/storytold/craft-libs", version = "0.1.0", rev = "787a7a0f972fd1d7a7618fdbeca3a5267a0407ff" }
craft-keymap = { git = "https://github.com/storytold/craft-libs", version = "0.1.0", rev = "787a7a0f972fd1d7a7618fdbeca3a5267a0407ff", default-features = false }
# Interface translations: lookup, fallback, CLDR plurals and language names over the craft-i18n
# catalogs that `crates/ui-egui/build.rs` embeds from `CRAFT_I18N_DIR` (storytold/craft-i18n, at
# the commit in `craft-i18n.rev`). Already in the tree through craft-ui.
craft-i18n = { git = "https://github.com/storytold/craft-libs", version = "0.1.0", rev = "787a7a0f972fd1d7a7618fdbeca3a5267a0407ff" }
# Shared system spooler; document composition remains in pdfcraft-print.
craft-print = { git = "https://github.com/storytold/craft-libs", version = "0.1.0", rev = "787a7a0f972fd1d7a7618fdbeca3a5267a0407ff" }
# Whether a path stored in a document may be opened (GoToR link targets).
@@ -26,39 +30,41 @@ craft-docpath = { version = "0.1.0", git = "https://github.com/storytold/craft-l
# craft-update: app-specific pretend versions apply to startup and manual checks.
craft-update = { git = "https://github.com/storytold/craft-libs", version = "=0.1.0", rev = "787a7a0f972fd1d7a7618fdbeca3a5267a0407ff", features = ["egui"] }
# internal
pdfcraft-geom = { path = "crates/geom", version = "0.18.0" }
pdfcraft-filters = { path = "crates/filters", version = "0.18.0" }
pdfcraft-cos = { path = "crates/cos", version = "0.18.0" }
pdfcraft-crypt = { path = "crates/crypt", version = "0.18.0" }
pdfcraft-organize = { path = "crates/organize", version = "0.18.0" }
pdfcraft-measure = { path = "crates/measure", version = "0.18.0" }
pdfcraft-annot = { path = "crates/annot", version = "0.18.0" }
pdfcraft-audit = { path = "crates/audit", version = "0.18.0" }
pdfcraft-fonts = { path = "crates/fonts", version = "0.18.0" }
pdfcraft-model = { path = "crates/model", version = "0.18.0" }
pdfcraft-content = { path = "crates/content", version = "0.18.0" }
pdfcraft-redact = { path = "crates/redact", version = "0.18.0" }
pdfcraft-print = { path = "crates/print", version = "0.18.0" }
pdfcraft-xfdf = { path = "crates/xfdf", version = "0.18.0" }
pdfcraft-sign = { path = "crates/sign", version = "0.18.0" }
pdfcraft-optimize = { path = "crates/optimize", version = "0.18.0" }
pdfcraft-a11y = { path = "crates/a11y", version = "0.18.0" }
pdfcraft-oracle = { path = "crates/oracle", version = "0.18.0" }
pdfcraft-testkit = { path = "crates/testkit", version = "0.18.0" }
pdfcraft-ocr = { path = "crates/ocr", version = "0.18.0" }
pdfcraft-js = { path = "crates/js", version = "0.18.0" }
pdfcraft-compare = { path = "crates/compare", version = "0.18.0" }
pdfcraft-preflight = { path = "crates/preflight", version = "0.18.0" }
pdfcraft-export = { path = "crates/export", version = "0.18.0" }
pdfcraft-edit = { path = "crates/edit", version = "0.18.0" }
pdfcraft-create = { path = "crates/create", version = "0.18.0" }
pdfcraft-forms = { path = "crates/forms", version = "0.18.0" }
pdfcraft-xfa = { path = "crates/xfa", version = "0.18.0" }
pdfcraft-render = { path = "crates/render", version = "0.18.0" }
pdfcraft-engine = { path = "crates/engine", version = "0.18.0" }
pdfcraft-platform = { path = "crates/platform", version = "0.18.0" }
pdfcraft-ui-egui = { path = "crates/ui-egui", version = "0.18.0" }
pdfcraft-automation = { path = "crates/automation", version = "0.18.0", default-features = false }
pdfcraft-geom = { path = "crates/geom", version = "0.18.1" }
pdfcraft-filters = { path = "crates/filters", version = "0.18.1" }
pdfcraft-cos = { path = "crates/cos", version = "0.18.1" }
pdfcraft-crypt = { path = "crates/crypt", version = "0.18.1" }
pdfcraft-organize = { path = "crates/organize", version = "0.18.1" }
pdfcraft-measure = { path = "crates/measure", version = "0.18.1" }
pdfcraft-annot = { path = "crates/annot", version = "0.18.1" }
pdfcraft-audit = { path = "crates/audit", version = "0.18.1" }
pdfcraft-fonts = { path = "crates/fonts", version = "0.18.1" }
pdfcraft-model = { path = "crates/model", version = "0.18.1" }
pdfcraft-content = { path = "crates/content", version = "0.18.1" }
pdfcraft-redact = { path = "crates/redact", version = "0.18.1" }
pdfcraft-print = { path = "crates/print", version = "0.18.1" }
pdfcraft-xfdf = { path = "crates/xfdf", version = "0.18.1" }
pdfcraft-sign = { path = "crates/sign", version = "0.18.1" }
pdfcraft-optimize = { path = "crates/optimize", version = "0.18.1" }
pdfcraft-a11y = { path = "crates/a11y", version = "0.18.1" }
pdfcraft-oracle = { path = "crates/oracle", version = "0.18.1" }
pdfcraft-testkit = { path = "crates/testkit", version = "0.18.1" }
pdfcraft-ocr = { path = "crates/ocr", version = "0.18.1" }
pdfcraft-js = { path = "crates/js", version = "0.18.1" }
pdfcraft-compare = { path = "crates/compare", version = "0.18.1" }
pdfcraft-preflight = { path = "crates/preflight", version = "0.18.1" }
pdfcraft-export = { path = "crates/export", version = "0.18.1" }
pdfcraft-edit = { path = "crates/edit", version = "0.18.1" }
pdfcraft-create = { path = "crates/create", version = "0.18.1" }
pdfcraft-forms = { path = "crates/forms", version = "0.18.1" }
pdfcraft-xfa = { path = "crates/xfa", version = "0.18.1" }
pdfcraft-render = { path = "crates/render", version = "0.18.1" }
pdfcraft-engine = { path = "crates/engine", version = "0.18.1" }
pdfcraft-platform = { path = "crates/platform", version = "0.18.1" }
pdfcraft-ui-egui = { path = "crates/ui-egui", version = "0.18.1" }
pdfcraft-automation = { path = "crates/automation", version = "0.18.1", default-features = false }
pdfcraft-prc = { path = "crates/prc", version = "0.18.1" }
pdfcraft-prc-render = { path = "crates/prc-render", version = "0.18.1" }
# pdf
hayro = "0.7.1"
LICENSES.toml
+7
View File
@@ -88,6 +88,13 @@ destination = "licenses/egui-emoji-icon-font-mit-license.txt"
sha256 = "b9d2c1d909aa149996fd4c91dcb92b2362a04431640c1d200959da94caf8cde1"
origin = "https://raw.githubusercontent.com/emilk/egui/0.36.2/crates/epaint_default_fonts/fonts/emoji-icon-font-mit-license.txt"
[[document]]
source = "licenses/prc-rs-LICENSE.txt"
licences = ["MIT"]
destination = "licenses/prc-rs-LICENSE.txt"
sha256 = "ddb846513d92153d8fa16b18ed3e7c86953a582f0d2f49f02b29f1bd569dac58"
origin = "https://raw.githubusercontent.com/ralovich/prc-rs/287b671bb5b31b05efb5ec1b98f7c359b9e4dd7d/LICENSE"
[[document]]
source = "vendor/hayro-interpret/LICENSE-APACHE"
licences = ["Apache-2.0"]
README.md
+1 -1
View File
@@ -94,7 +94,7 @@ Real-world typography: world scripts, vertical Japanese, colour emoji, gradients
<td width="33%" valign="top">
### Fearless
Every save appends your changes and leaves the original bytes untouched. Writes are atomic, undo runs deep, and nothing is lost if you close by mistake.
Ordinary saves append your changes and leave the original bytes untouched. Edits to damaged objects whose xref and header generations disagree use a full rewrite, keeping both accepted references connected to the same repaired object. Writes are atomic, undo runs deep, and nothing is lost if you close by mistake.
</td>
<td width="33%" valign="top">
ROADMAP.md
+30 -1
View File
@@ -21,10 +21,17 @@
> **Last reviewed:** 2026-10-23 · **Last updated:** 2026-10-23 · **Change:** minor (M2, PR #4172: extend on-screen CJK fallback to Combine thumbnails and figure previews; recognise FangSong and subset-prefixed Founder names; estimates unchanged)
> **Last reviewed:** 2026-10-15 · **Last updated:** 2026-10-15 · **Change:** minor (typed small Edit font sizes and Fill & Sign text/date colour, #2444/#2496; estimates unchanged)
> **Last reviewed:** 2026-10-20 · **Last updated:** 2026-10-20 · **Change:** patch (M1, #1143: repaired aliases survive preflight, snapshots and normalized saves; retained materialized revisions report generation disagreements; estimates unchanged)
> **Last reviewed:** 2026-10-18 · **Last updated:** 2026-10-18 · **Change:** minor (M1, #1143: explicit live xref/header generation repair policy and bounded diagnostics; estimates unchanged)
> **Last reviewed:** 2026-10-15 · **Last updated:** 2026-10-15 · **Change:** minor (M1: refresh object-generation reference identity against current upstream; estimates unchanged)
**Stage: alpha** · next: beta, ~25% (ready for real work 50% → 75%) and ~400–700 h away
> **Last reviewed:** 2026-10-29 | **Last updated:** 2026-10-29 | **Change:** minor (M7 #2517: a narrowed then widened paragraph keeps the typed spaces the wrap consumed, after save and reopen; estimates unchanged)
> **Last reviewed:** 2026-10-29 | **Last updated:** 2026-10-29 | **Change:** minor (M2: scans keep their strokes at screen size and are smoothed when zoomed; M7 #2517: a narrowed then widened paragraph keeps the typed spaces the wrap consumed, after save and reopen; estimates unchanged)
> **Last reviewed:** 2026-10-28 | **Last updated:** 2026-10-28 | **Change:** minor (M10, #4119 partial: Enhance scanned file cleans scanned pages in place: background removal, edge shadow removal, despeckle; estimates unchanged)
> **Last reviewed:** 2026-10-27 | **Last updated:** 2026-10-27 | **Change:** minor (M10: OCR model downloads use verified ONNX pairs with legacy discovery fallback; estimates unchanged)
> **Last reviewed:** 2026-10-25 | **Last updated:** 2026-10-25 | **Change:** minor (web: same-document host mode on the canvas, `data-host="page"`; estimates unchanged)
> **Last reviewed:** 2026-10-25 | **Last updated:** 2026-10-25 | **Change:** minor (M14 localization current state: Arabic logical catalog, explicit RTL label direction and mirrored stock chrome; native/browser RTL workflows unqualified. Native release Hans face via LXGW WenKai Lite; estimates unchanged)
> **Last reviewed:** 2026-10-23 · **Last updated:** 2026-10-23 · **Change:** minor (M7, #2636 partial: pasting added text boxes and images in the engine and through `content_copy`; estimates unchanged)
> **Last reviewed:** 2026-10-23 · **Last updated:** 2026-10-23 · **Change:** minor (M11: image optimization accounts for page-local UserUnit; estimates unchanged)
@@ -223,14 +230,21 @@ Ranked; detail and the full milestone table in [docs/roadmap.md](docs/roadmap.md
## Progress log
- **2026-10-28 (M10, #4119 partial):** Scan & OCR ▸ Enhance scanned file cleans scanned pages in place: background removal (level 0–100, as the camera clean-up), edge shadow removal (book gutters, copier borders) and despeckle. Each scan image is decoded, cleaned and written back into its own object at the same size and colour space (JPEG stays JPEG, others lossless, 1-bit stays 1-bit), so the OCR text layer, comments and links stay; one undo step, shared images once, unsupported images (CMYK, indexed, JPEG 2000, JBIG2, CCITT, broken data) reported. `page_enhance_scan` drives it headlessly. Not yet an option of Recognize text. Estimates unchanged.
- **2026-11-03 (M11, shared JBIG2 dictionary):** With Monochrome Images set to JBIG2, Optimize has up to 16 JBIG2 images share one symbol dictionary (`/JBIG2Globals`), the letters found on more than one page stored once, when that is smaller (`pdfcraft_filters::encode_jbig2_shared`). Eight pages of 10 pt text rendered at 300 dpi take 41% less in Helvetica and 42% in Times than coded alone, 10% in Courier; groups of 16 save 45%, of 32 only 47%. Only images redaction can store again without the dictionary share one (image masks and DeviceGray named directly, at most 64 Mi pixels), and redacting one stores the rest of its group so, which the group bounds to 15.
- **2026-11-03 (M5, #5850 review):** Resolve indirect IT when recognizing typewriter contents and Fill & Sign marks. Comment listing and property lookup retain their existing direct-Subtype behavior. A real double-click Edit composer preserves the typewriter’s text as written; ordinary comment edits keep trimming. Regressions cover direct Subtype with indirect intent, save/reopen, raw posted contents and same-index comments on different pages. The existing live-movement test now commits its setup with an outside click rather than Enter. Estimates unchanged.
- **2026-11-03 (M5, #5850):** Fill & Sign Enter inserts a line break in a growing editor; clicking outside commits and Escape cancels. Nonempty text keeps leading, internal and trailing line breaks. The headless fill_sign_add path sizes its typewriter box for the same wrapped lines so saved appearances show every row. Production-order keyboard and engine save/reopen/render regressions exercise the UI and headless callers. Estimates unchanged.
- **2026-11-03 (M7, #5835):** In image crop mode, a right-click on the image, its crop box or its handles keeps the crop and opens no menu, so the menu no longer switches as the pointer moves. A right-click elsewhere clears the selection and opens the page menu. UI tests cover all four page rotations, overlapping images, a 45° image, and a check that the Done/Cancel/Remove Crop pills at a page edge still keep the crop.
- **2026-11-02 (M12, #1297):** `model_add` embeds a U3D or PRC file, recognised by its header, as a 3D annotation with a chosen activation and a poster (an image or PDF page fitted in the box, or a labelled frame); an older file's catalog `/Version` is raised to what the model needs. Views, a rendered poster and a UI command are later steps. Estimates unchanged.
- **2026-11-02 (M10, #5743):** macOS print jobs name their title and owner (`com.apple.print.JobInfo.PMJobName`, `PMJobOwner`), as the system print dialog does, because a Kyocera filter crashes on a job without them. Checked by a test and through `doc_print` against a stand-in `lp`; not yet on a physical printer.
- **2026-11-02 (M14, #5737, #5460):** Preferences can turn off the update check at start. It stays on by default, and Help ▸ Check for updates still asks.
- **2026-11-01 (M7, #5657):** Header/footer, watermark and background additions and replacements stage the complete call before committing. Synthetic later-page failures preserve exact incremental bytes, object count, dirty state and allocation cursor; success controls retain other marks and shared/inherited resources. Estimates unchanged.
- **2026-11-01 (M14, #5572):** Preferences ▸ Documents and view sets how many recent files to keep, from 0 to 50 with 12 by default. At 0 Home has no Recent section and File ▸ Open Recent is disabled. The list is cut when Preferences closes, so the number can pass through 0 on its way to another.
@@ -257,8 +271,10 @@ Ranked; detail and the full milestone table in [docs/roadmap.md](docs/roadmap.md
- **2026-10-26 (M2, PR #4172):** Reconciled current main, preserving its Japanese serif/weight selection and both branches' progress records. Installed CJK viewer and preview fallback remains display-only; estimates unchanged.
- **2026-10-29 (M7, #2517):** Widening a paragraph rejoins a wrap with the number of spaces that wrap consumed, including a typed run and none for a Chinese or Japanese word split. The count survives an incremental save and a full rewrite, and a width-only `text_edit` uses it. Estimates unchanged.
- **2026-10-29 (M2):** Scanned pages look like the paper again. An image that doesn't ask for /Interpolate was shrunk by nearest-neighbour (since #635), which kept one sample in N per axis, so the text of a 300 dpi scan at 100% lost most of its strokes; it is now area-averaged to whole device pixels (on a 13-page 300 dpi scan every sample at 96–192 dpi is within 1 of the exact area average; it was up to 255 off). Enlarged, an image of page-scan size is smoothed bilinearly, so a zoomed scan has no staircases; icons, QR codes and pixel art keep hard pixels. Estimates unchanged.
- **2026-10-26 (M11, JBIG2 symbols):** Lossless JBIG2 now stores each distinct mark of a black-and-white image once, as a symbol, and places its exact copies, when that is smaller than the generic region: on six pages of rendered 300 dpi text 28% of the generic region's size and 17% of Group 4's, and the same bits back through two independent decoders (hayro-jbig2 and Poppler). Scans, whose letters are seldom identical, gain little until refinement coding; a dictionary shared between pages (`/JBIG2Globals`) is next. Estimates unchanged.
- **2026-10-25 (M5, comment replies):** One reply can be deleted from the Comments panel without its thread. On the selected card each reply has its own "…" menu (Edit, Copy text, Delete); Delete takes that reply with its pop-up and the replies to it, as `comment_delete` already did headlessly, and the comment and its other replies stay. The card stays selected, and the panel's editor stays open, on what they were on though the delete moved it in `/Annots`. Undo and Redo close the editor, since either can move what it was open on. A locked reply can't be deleted. Estimates unchanged.
- **2026-10-25 (web, same-document host mode):** A page that loads the PdfCraft web module into its own document (no iframe) can opt into the host bridge with `data-host="page"` on the app's canvas — the exact value `page`; any other value keeps the normal app with downloads, unless a same-origin framed `?host=parent` selects the parent bridge (unframed or cross-origin, `?host=parent` also keeps the normal app). It is the `?host=parent` bridge with the transport swapped: the same five messages travel as non-bubbling `CustomEvent`s on the canvas (`detail` holds the parent messages' fields without `type`; the host answers with a `pdfcraft:open` event), so saves reach the host instead of downloading and dirty and start-up-failure reporting are unchanged. Unit tests cover the opt-in rule; no parity or estimate changes.
- **2026-10-24 (M10, #4158 partial):** Word, HTML, RTF and PowerPoint export keep a block that continues another on its line (an italic case name, a link, a superscript) as a run in its own style inside that paragraph, with the source's word space, instead of a paragraph of its own. Style changes that wrap onto the next line still split paragraphs (the reading blocks keep style boundaries). Estimates unchanged.
- **2026-10-24 (M3, window move on Wayland, #3970 follow-up):** After a move or resize by PdfCraft's own title bar or edges, the button release given back to egui goes before the pointer leaving the window, so hover no longer sticks at the old place until the pointer moves. Found by CodeRabbit in review of #3970; a unit test covers it. Estimates unchanged.
- **2026-10-24 (M14, #4253):** The annotation quick bar remembers Keep tool selected across restarts, including unpinning. Old and malformed settings keep safe defaults; the opt-in UI control channel reads and changes the same preference. Estimates unchanged.
@@ -287,6 +303,11 @@ Ranked; detail and the full milestone table in [docs/roadmap.md](docs/roadmap.md
Estimates unchanged.
- **2026-10-21 (M3, window resize and move on Wayland):** With PdfCraft's own title bar on GNOME/Wayland, a second resize by a window edge, or a second move by the empty tab strip, works straight after the first; it waited for a click inside the window, because the compositor keeps the button's release from egui. A unit test covers the release given back; checked by hand in a test build. Estimates unchanged.
- **2026-10-24 (M5, move preview):** A comment being moved is drawn where it goes while it is dragged (a rectangle or oval showed only its blue frame): the comment is rendered alone once as the drag starts and laid over the page, which shows it where it was until the move is applied. Engine `Document::comment_layer`; synthetic engine and UI regressions. Estimates unchanged.
- **2026-10-24 (M5, border shape):** Text boxes and Fill & Sign's typed text take an oval or a circle for their border (Comment Properties ▸ Border shape, Fill & Sign's toolbar): the text box keeps its place and the shape goes around it, the text centred, a circle snug around the text (its box no wider than the text needs); stored as /RD and /PCBorderShape and drawn in the appearance. comment_add / comment_edit / fill_sign_add `border_shape`. One new label in every catalog. Synthetic annot, UI and automation regressions. Estimates unchanged.
- **2026-10-24 (M5, Escape in Fill & Sign):** Escape leaves a Fill & Sign tool as it leaves a comment tool: the first press closes the open type box, the next goes back to Select, so the Text tool no longer opens a box at every click with no way out by the keyboard. A synthetic UI regression covers the Text and Checkmark tools. Estimates unchanged.
- **2026-10-23 (M5, ortho):** Shift while drawing holds a line or arrow level or plumb, whichever is nearer (a CAD drawing's ortho mode), and each new side of connected lines, a polygon or a cloud from the corner before; dragging an end or corner with Shift is held the same way (it turned to 45° steps). `measure::snap::ortho`; `comment_add ortho`. Estimates unchanged.
- **2026-10-24 (M3, arrow keys on a wide page):** → / ← scroll a page wider than the window (a plan zoomed in) sideways a line at a time and turn the page at its edge, as ↓ / ↑ do in single-page view; ⌘→ / ⌘← turn it at once, and a page that fits across turns as before (#185). A synthetic viewer regression covers both. Estimates unchanged.
- **2026-10-24 (M5, part opacity):** Comment Properties sets a text box's text, border and fill, and a shape's line and fill, each at an opacity of its own (a see-through fill with an opaque line on a plan); stored as /CA, /ca (PDF 2.0) and /PCTextOpacity and drawn with separate graphics states, so other viewers show it. Make Properties Default and the settings carry it; comment_add / comment_edit `fill_opacity`, `text_opacity`. Four new labels in every catalog. Synthetic annot, UI and automation regressions. Estimates unchanged.
- **2026-10-22 (M4, #3916 partial):** Document edits and Undo/Redo cancel an Organize page drag. Pointer-release regressions check exact page order and save/reopen.
- **2026-10-22 (M10, #3752 paper):** CUPS jobs ask for the chosen paper in every size. The 13 sizes the Print dialog sent no paper for (A6, A0–A2, ARCH, ANSI C–E) use PWG 5101.1 names, checked against a synthetic PPD with `cupsfilter`, and `doc_print` to a printer sends the paper too. The macOS tray half (Auto Select instead of the PPD's default tray) is a craft-print change that needs a pin bump here. Estimates unchanged.
@@ -329,9 +350,13 @@ Ranked; detail and the full milestone table in [docs/roadmap.md](docs/roadmap.md
- **2026-10-20 (M3, #3541/#3620):** Comment hover follows the individual marked lines instead of the enclosing multiline rectangle, and overlapping comments follow paint order. Synthetic UI regressions cover all four text-markup kinds, page/view rotations, neighboring highlights, legacy rectangle fallback and unchanged PDF bytes. Estimates unchanged.
- **2026-10-20 (M10, #3532 partial):** Vector printing retains existing widget artwork when its appearance lacks Subtype. A print-owned copy adds Form dispatch metadata. The source stream and all explicit subtypes stay unchanged. Synthetic source/output raster checks cover resources, selected/shared appearances, placement, flags and content modes. Appearance regeneration and physical printer validation remain outside this fix. Estimates unchanged.
- **2026-10-20 (M5, several comments):** Comment Properties opened on one of several selected comments, the context menu's Colour and the quick bar's colour change all of them as one undo step, each where the property applies (a font size to text boxes, a fill to shapes, a width to lines and shapes); one that stays locked keeps its look. Copied Fill & Sign ticks, crosses, dots and lines hang on the pointer as themselves rather than as empty boxes, and copied text boxes show their text (`FillMark::strokes` is shared by the appearance and the outline). Estimates unchanged.
- **2026-10-19 (M5, comments between documents):** Comments copy and paste, within and between documents: ⌘C with comments selected copies them as XFDF (each with its replies) to the app's and the system clipboard, ⌘V / Edit ▸ Paste comments / the page menu pastes them onto the current page as new comments under names of their own (`Edit::PasteComments`, `pdfcraft_xfdf::import_comments` with a target page and fresh names, so nothing is replaced and pasting twice gives two sets; `export_xfdf_comments` writes a selection with its replies). Several comments select at once (⌘-click, ⇧-click, on the page and in the panel); Delete removes them as one step and a drag moves them together. Connected lines, polygons and clouds finish from the right-click menu's Complete (Undo last point, Cancel drawing beside it), as other editors do. Tools `comments_copy` / `comments_paste`. Estimates unchanged.
- **2026-10-20 (M5, arrow keys move comments):** With a comment selected, the arrow keys move it a screen pixel and ⌘/Ctrl or Shift and an arrow ten, instead of scrolling the view or turning the page; up is up on screen at any zoom and rotation, and each press is one undo step. Text markup and locked comments stay put. Estimates unchanged.
- **2026-10-20 (M1, PR #1143 review):** Repair-aware dictionary/preflight edits and snapshot restoration preserve both accepted generations of a damaged live slot. Such edits use the existing full-save fallback; classic and packed rewrites coalesce aliases so page-tree parent links retain one object identity. Explicit generation replacement remains single-generation. Materialized xref diagnostics inspect retained primary and hybrid rows before folding, retaining the 256-object cap plus summary and discarding abandoned-index notes. Synthetic red/green controls cover preflight, undo restoration, equal values, explicit narrowing, encryption packing and four save variants. Estimates unchanged.
- **2026-10-19 (M14, #3373):** The main menu and its long submenus scroll in short windows, keeping their commands reachable. Estimates unchanged.
- **2026-10-19 (M3, #3387):** Save-prompt shortcuts yield to later consent and utility dialogs; a synthetic UI regression checks that an edited document stays open. Estimates unchanged.
@@ -395,6 +420,8 @@ Ranked; detail and the full milestone table in [docs/roadmap.md](docs/roadmap.md
- **2026-10-15 (community, M6, #2608):** In Prepare a form the selected field moves with the arrow keys: a point per press, ten with Shift, as the object nudge and dragging it do. Each press is one undo step ("Change field properties"), and the keys no longer turn the page while a field is selected. Estimates unchanged.
- **2026-10-15 (M3, closable community card):** The ArtCraft community card on Home has a close button that hides it for the rest of the run; it returns on the next launch, and its links stay in the top bar's Discord button and the Help menu. A UI regression covers closing and the fresh-launch default. Estimates unchanged.
- **2026-10-13 (community, M3, #1569 review):** Pixel-only Wayland wheel, wheel-tilt and continuous input use standalone moved events, so Ctrl zoom does not remain latched after the modifier is released (#991); vendored winit patch (13), with a policy test. Estimates unchanged.
- **2026-10-18 (M1, PR #1143 repair policy):** A live in-file reference resolves when its generation matches the xref row or a same-number header at that indexed offset. Neither-match and free slots remain null, including when old object bytes remain. Opening records bounded xref/header disagreements without parsing object bodies or expanding compact xref rows. Synthetic controls cover read order, clones, overlays, dictionary edits, full classic/packed saves and an independent reader of incremental output. **Correction to the 2026-10-15 note below:** its async-opening-test sentence was not evidenced by this COS change; no fresh async-opening or GUI validation is claimed. The full-reference stream-length recursion key is defensive hardening, not an independently isolated regression fix. Estimates unchanged.
- **2026-10-15 (M1, PR #1143 refresh):** COS reference lookups and full saves distinguish complete object references, so stale generations cannot expose replacement content. Snapshot restoration also keeps edits to objects with nonzero generations before incremental saves. Synthetic regressions cover caches, overlays, snapshots, stream lengths and classic/packed saves. Current upstream async-opening test corrections are included. Estimates unchanged.
- **2026-10-15 (M14, #2383 candidate):** Device-limit requests now clamp unsupported maxima and minimum alignments before enabling the adapter's full 2D texture extent. This addresses the reported V3D request of 16 shader variables on an adapter supporting 15. Synthetic regressions fail before the clamp and pass afterward, along with the existing HiDPI case and Windows GPU-required application tests. Raspberry Pi startup remains unverified; milestone and overall estimates are unchanged.
- **2026-10-12 (M6):** XFDF field import and data-file extraction join all text nodes in a value, preserving text after XML comments and processing instructions. Synthetic regressions cover Unicode, CDATA, empty values and PDF save/reopen. Estimates unchanged.
- **2026-10-12 (M7, picture right-click menu):** The Edit text & images menu of a picture gains Cut Image, Redact Image (a redaction mark over the picture, applied from Redact a PDF), Create Link… and Recognize Text… (for that page), in all 21 catalogs. The menu was drawn only while the pointer was over the picture, so its items outside the picture (most of a logo's menu, and Delete below a page) could not be clicked; an open menu now stays drawn. Still missing from Acrobat's menu: Arrange (bring to front, send backward), Remove crop, Draw freehand and Edit image using… (opening an external editor).
@@ -773,6 +800,7 @@ Newest first: the date, what moved, and the new overall estimate.
| 2026-10-20 | minor | M6, #3502: form detection preserves comments across content streams; estimates unchanged. |
| 2026-10-20 | minor | M10, #3532 partial: vector printing retains existing appearances with absent Subtype; estimates unchanged. |
| 2026-10-20 | minor | M11 #3507: image merging preserves comments and state across content streams. Estimates unchanged. |
| 2026-10-18 | minor | M1, #1143: a live reference may match the xref generation or a same-number header at its indexed offset; neither-match and free slots remain null. Bounded repair diagnostics and synthetic cache, free-slot, edit and classic/packed save regressions; estimates unchanged. |
| 2026-10-18 | minor | M5, PR #1971: nested source `/Page` and `/Pages` dictionaries export as null without traversing page content or the page tree. Synthetic saved/reopened three-page regression retains unrelated metadata and aliases. FDF import semantics, copied-graph limits and estimates unchanged. |
| 2026-10-15 | minor | M5: budget retained FDF entries before cloning; discovery keeps lazy array owners and annotation positions, with alias/cache-eviction stress fixtures. Limits cover the copied graph; source parsing/page discovery retain existing COS/parser limits, with no total source-memory bound. FDF import semantics and estimates unchanged. |
| 2026-10-15 | minor | M7: the Edit text caret is as tall as the row it is on in a region with several sizes; estimates unchanged. |
@@ -783,6 +811,7 @@ Newest first: the date, what moved, and the new overall estimate.
| 2026-10-14 | minor | M1: raw-deflate fallback preserves decoding-limit errors in strict and tolerant decoding; synthetic filter and COS stream regressions. Estimates unchanged. |
| 2026-10-16 | minor | M14: Czech catalog completed (2,674 entries) and "tag" terminology unified. Estimates unchanged. |
| 2026-10-13 | minor | M9: separate page validation overlays preserve signature artwork, and snapshot tabs reuse instead of multiplying; baked captions remain unchanged (#79 open); estimates unchanged |
| 2026-10-15 | minor | M1: refresh object-generation reference identity, retaining current upstream progress and regression tests. Estimates unchanged. |
| 2026-10-16 | minor | M10 #1201 partial: lossless image export applies component Decode pairs and Indexed mapping. Estimates unchanged. |
| 2026-10-15 | minor | M6: CSV export and spreadsheet merging quote tabs, and import distinguishes quoted tabs from TSV delimiters without changing field identities. Estimates unchanged. |
| 2026-10-13 | minor | M4: rotation queries and edits resolve indirect page-tree values; inherited, negative and large rotations verified through both saves. Estimates unchanged. |
apps/pdfcraft-web/Cargo.toml
+3 -1
View File
@@ -11,10 +11,12 @@ workspace = true
[dependencies]
pdfcraft-ui-egui = { workspace = true }
# The host messages (storytold/craft-libs): one contract and one bridge for the web builds.
craft-web-host = { git = "https://github.com/storytold/craft-libs", rev = "2e2312ddaf21bc3d85cd3395d0955c7d239306f6", version = "0.1.0" }
eframe = { workspace = true, features = ["default_fonts", "glow", "persistence", "accesskit"] }
egui = { workspace = true, features = ["default_fonts", "monochrome_emoji_fonts"] }
[target.'cfg(target_arch = "wasm32")'.dependencies]
wasm-bindgen-futures = "0.4"
web-sys = { version = "0.3", features = ["Window", "Document", "Element", "HtmlCanvasElement", "HtmlElement", "Location", "Url", "UrlSearchParams", "Response", "Event", "MessageEvent", "EventTarget"] }
web-sys = { version = "0.3", features = ["Window", "Document", "Element", "HtmlCanvasElement", "HtmlElement", "Location", "Url", "UrlSearchParams", "Response", "Event", "MessageEvent", "EventTarget", "CustomEvent", "CustomEventInit"] }
js-sys = "0.3"
apps/pdfcraft-web/src/main.rs
+81 -110
View File
@@ -5,18 +5,11 @@
#[cfg(target_arch = "wasm32")]
fn main() {
use eframe::wasm_bindgen::JsCast;
use pdfcraft_ui_egui::PdfCraftApp;
let options = eframe::WebOptions { renderer: eframe::Renderer::Glow, ..Default::default() };
wasm_bindgen_futures::spawn_local(async move {
let Some(canvas) = web_sys::window()
.and_then(|w| w.document())
.and_then(|d| d.get_element_by_id("pdfcraft"))
.and_then(|e| e.dyn_into::<web_sys::HtmlCanvasElement>().ok())
else {
return;
};
let Some(canvas) = app_canvas() else { return };
let started = eframe::WebRunner::new()
.start(
canvas,
@@ -54,11 +47,11 @@ fn main() {
// eframe never raises a close request on a browser reload or tab close, so `guard_quit` never
// prompts; ask the browser to confirm instead while a document has unsaved work (#812).
warn_before_unload(app.unsaved_flag.clone());
// `?host=parent`: a same-origin page embedding PdfCraft in an iframe (such
// as the Nextcloud app) passes files in and takes saves back (`host` below).
if query_param("host").as_deref() == Some("parent")
&& let Some(bridge) = host::Bridge::new()
{
// Host modes (`host` below, over `craft-web-host`): `?host=parent`, a
// same-origin page embedding PdfCraft in an iframe (such as the Nextcloud
// app), or `data-host="page"` on the canvas, a host that runs the module in
// its own document.
if let Some(bridge) = app_canvas().as_ref().and_then(host::Bridge::connect) {
// `?author=`: the host's name for the user signs their comments,
// unless they chose a name in PdfCraft (the web default is "Guest").
if let Some(author) = query_param("author").map(|a| a.trim().to_string()).filter(|a| !a.is_empty())
@@ -81,9 +74,7 @@ fn main() {
.await;
if let Err(e) = started {
eframe::web_sys::console::error_1(&e);
if query_param("host").as_deref() == Some("parent")
&& let Some(bridge) = host::Bridge::new()
{
if let Some(bridge) = app_canvas().as_ref().and_then(host::Bridge::connect) {
bridge.post_failed(&e.as_string().unwrap_or_else(|| format!("{e:?}")));
}
}
@@ -123,136 +114,116 @@ async fn fetch_bytes(url: &str) -> Result<Vec<u8>, String> {
Ok(js_sys::Uint8Array::new(&buf).to_vec())
}
/// The `postMessage` bridge to a same-origin parent page. Messages are plain objects with a
/// `type`:
/// The app's canvas: the one element `index.html` and every embedding page must provide, with
/// the id the app looks up.
#[cfg(target_arch = "wasm32")]
fn app_canvas() -> Option<web_sys::HtmlCanvasElement> {
use eframe::wasm_bindgen::JsCast;
web_sys::window()?.document()?.get_element_by_id("pdfcraft").and_then(|e| e.dyn_into::<web_sys::HtmlCanvasElement>().ok())
}
/// Host modes: how the page embedding PdfCraft passes files in and takes saves back. The messages
/// and the two transports are `craft-web-host`'s (storytold/craft-libs), the same contract as the
/// other Craft apps' bridges, with the `pdfcraft:` prefix. This module is PdfCraft's part of it.
///
/// - `pdfcraft:ready` (to the parent): the app is listening; send files now.
/// - `pdfcraft:open` (from the parent): `{ name, bytes }`, where `bytes` is an `ArrayBuffer` or
/// `Uint8Array` holding a PDF. It opens in a tab, like a `?file=` URL.
/// - `pdfcraft:save` (to the parent): `{ name, saveAs, bytes }` when the user saves; `bytes` is
/// a transferred `ArrayBuffer`. The parent stores the file (and reports its own errors).
/// - `pdfcraft:dirty` (to the parent): `{ dirty }` whenever unsaved work appears or goes away.
/// - `pdfcraft:failed` (to the parent): `{ error }` when the app couldn't start.
///
/// Only the parent window on the app's own origin is heard or answered. With a parent on
/// another origin there is no bridge: the app runs as if unframed and saves download.
/// - `?host=parent` is the `postMessage` bridge for a same-origin page that embeds PdfCraft in an
/// iframe (such as the Nextcloud app). With a parent on another origin there is no bridge: the
/// app runs as if unframed and saves download (#2167). `data-host="page"` on the app's canvas
/// is for a host that loads the module into its own document, with no iframe: the same messages
/// travel as DOM `CustomEvent`s on that canvas. `?host=parent` wins when both are present.
/// - `pdfcraft:ready` once the app is listening; send files now. PdfCraft reads bytes, so it
/// says `file: false` and a host sends `bytes`.
/// - `pdfcraft:open` (from the host): `{ name, bytes }` holding a PDF. It opens in a tab, like a
/// `?file=` URL; a file over [`host::MAX_OPEN_BYTES`] is refused before it is copied.
/// - `pdfcraft:save` (to the host): `{ name, saveAs, bytes }` when the user saves. The host stores
/// the file (and reports its own errors).
/// - `pdfcraft:dirty` whenever unsaved work appears or goes away, `pdfcraft:failed` when the app
/// couldn't start.
#[cfg(target_arch = "wasm32")]
mod host {
use eframe::wasm_bindgen::{JsCast, JsValue, closure::Closure};
use std::rc::Rc;
use craft_web_host::web::{FileData, Host, Message};
/// The app id in the host messages (`pdfcraft:ready`).
const APP: &str = "pdfcraft";
/// The name a file gets when the host sent none (the crate says `document` for a blank name).
const DEFAULT_NAME: &str = "document.pdf";
/// The largest file the host may hand over to open, checked before it is copied (the app's own readers have their own limits).
const MAX_OPEN_BYTES: u64 = 512 << 20;
#[derive(Clone)]
pub struct Bridge {
window: web_sys::Window,
parent: web_sys::Window,
origin: String,
host: Rc<Host>,
}
impl Bridge {
/// `None` when the page isn't framed (there is no parent to talk to), or when the parent
/// is on another origin: messages only go to this origin, so the browser would drop
/// every save without an error while the document showed as saved (#2167).
pub fn new() -> Option<Self> {
let window = web_sys::window()?;
let parent = window.parent().ok().flatten()?;
if js_sys::Object::is(&parent, &window) {
return None;
}
let origin = window.location().origin().ok()?;
// Reading another origin's location throws, which is `None` too.
if parent.location().origin().ok()? != origin {
return None;
}
Some(Self { window, parent, origin })
}
fn post(&self, message: &js_sys::Object, transfer: Option<&JsValue>) -> Result<(), String> {
let sent = match transfer {
Some(t) => self.parent.post_message_with_transfer(message, &self.origin, &js_sys::Array::of1(t)),
None => self.parent.post_message(message, &self.origin),
};
sent.map_err(|e| e.as_string().unwrap_or_else(|| format!("{e:?}")))
/// The bridge for this page, or `None` when PdfCraft isn't hosted: no `?host=parent` and
/// no `data-host="page"` on the canvas, an unframed `?host=parent` page, or a parent on
/// another origin (messages only go to this origin, so the browser would drop every save
/// without an error while the document showed as saved, #2167).
pub fn connect(canvas: &web_sys::HtmlCanvasElement) -> Option<Self> {
Some(Self { host: Rc::new(Host::connect(APP, canvas)?) })
}
pub fn post_ready(&self) {
if let Some(m) = message(&[("type", "pdfcraft:ready".into()), ("version", env!("CARGO_PKG_VERSION").into())]) {
let _ = self.post(&m, None);
}
let _ = self.host.post_ready(env!("CARGO_PKG_VERSION"), false);
}
/// The app couldn't start (no WebGL, for example): let the parent say so.
/// The app couldn't start (no WebGL, for example): let the host say so.
pub fn post_failed(&self, error: &str) {
if let Some(m) = message(&[("type", "pdfcraft:failed".into()), ("error", error.into())]) {
let _ = self.post(&m, None);
}
let _ = self.host.post_failed(error, None);
}
/// Hear `pdfcraft:open` from the parent and queue the file like a `?file=` download.
/// Hear `pdfcraft:open` and queue the file like a `?file=` download.
pub fn listen(&self, inbox: pdfcraft_ui_egui::Inbox, ctx: egui::Context) {
let parent = self.parent.clone();
let origin = self.origin.clone();
let on_message = Closure::<dyn FnMut(web_sys::MessageEvent)>::new(move |event: web_sys::MessageEvent| {
if event.origin() != origin || !event.source().is_some_and(|s| js_sys::Object::is(&s, &parent)) {
return;
self.host.listen(move |message| {
if let Message::Open { name, data } = message {
queue_open(&name, &data, &inbox, &ctx);
}
let data = event.data();
if string(&data, "type").as_deref() != Some("pdfcraft:open") {
return;
}
let name = string(&data, "name").filter(|n| !n.trim().is_empty()).unwrap_or_else(|| "document.pdf".to_string());
let Ok(bytes) = js_sys::Reflect::get(&data, &JsValue::from_str("bytes")) else { return };
let bytes = if let Some(array) = bytes.dyn_ref::<js_sys::Uint8Array>() {
array.to_vec()
} else if bytes.is_instance_of::<js_sys::ArrayBuffer>() {
js_sys::Uint8Array::new(&bytes).to_vec()
} else {
return;
};
if let Ok(mut q) = inbox.lock() {
q.push((name, bytes));
}
ctx.request_repaint();
});
if self.window.add_event_listener_with_callback("message", on_message.as_ref().unchecked_ref()).is_ok() {
// The listener lives as long as the page.
on_message.forget();
}
}
pub fn save_fn(&self) -> pdfcraft_ui_egui::HostSaveFn {
let bridge = self.clone();
Box::new(move |name: &str, bytes: &[u8], save_as: bool| {
let buffer = js_sys::Uint8Array::from(bytes).buffer();
let m =
message(&[("type", "pdfcraft:save".into()), ("name", name.into()), ("saveAs", save_as.into()), ("bytes", buffer.clone().into())])
.ok_or_else(|| "couldn't build the message".to_string())?;
bridge.post(&m, Some(&buffer))
})
let host = self.host.clone();
Box::new(move |name: &str, bytes: &[u8], save_as: bool| host.post_save(name, bytes, None, save_as))
}
pub fn dirty_fn(&self) -> pdfcraft_ui_egui::HostDirtyFn {
let bridge = self.clone();
let host = self.host.clone();
// A host starts out clean: the first report is a change only when it is dirty.
let mut last = false;
Box::new(move |dirty: bool| {
if dirty != last {
last = dirty;
if let Some(m) = message(&[("type", "pdfcraft:dirty".into()), ("dirty", dirty.into())]) {
let _ = bridge.post(&m, None);
}
host.post_dirty(dirty);
}
})
}
}
fn message(fields: &[(&str, JsValue)]) -> Option<js_sys::Object> {
let m = js_sys::Object::new();
for (key, value) in fields {
js_sys::Reflect::set(&m, &JsValue::from_str(key), value).ok()?;
/// The `pdfcraft:open` payload, `{ name, bytes }` where `bytes` holds a PDF: queue it like a
/// `?file=` download and wake the app. The crate has cleaned the name (and cut it to 255
/// characters); a blank one comes as `document` and becomes `document.pdf`.
fn queue_open(name: &str, data: &FileData, inbox: &pdfcraft_ui_egui::Inbox, ctx: &egui::Context) {
let name = if name == "document" { DEFAULT_NAME } else { name }.to_string();
let FileData::Bytes(view) = data else {
eframe::web_sys::console::warn_1(&format!("PdfCraft: host file {name} came as a File; PdfCraft reads bytes only; not opened").into());
return;
};
// A file from the host is untrusted input: refuse an oversized one before copying it into the app's memory.
if data.size() > MAX_OPEN_BYTES as f64 {
eframe::web_sys::console::warn_1(
&format!("PdfCraft: host file is {} bytes, over the {} MiB host open limit; not opened", data.size(), MAX_OPEN_BYTES >> 20).into(),
);
return;
}
Some(m)
}
fn string(data: &JsValue, key: &str) -> Option<String> {
js_sys::Reflect::get(data, &JsValue::from_str(key)).ok()?.as_string()
if let Ok(mut q) = inbox.lock() {
q.push((name, view.to_vec()));
}
ctx.request_repaint();
}
}
apps/pdfcraft/src/tests.rs
+23
View File
@@ -663,6 +663,29 @@ fn the_desktop_entry_names_the_window_class() {
assert_eq!(class, Some(super::APP_ID));
}
/// Both flatpak manifests build with the craft-i18n commit `craft-i18n.rev` pins: CI and
/// releases read the file, and the manifests have to name the commit themselves. The pin may
/// have a trailing newline or CRLF (a Windows checkout), so it is trimmed, as every reader does.
#[test]
fn the_flatpak_manifests_pin_the_craft_i18n_rev() {
let root = std::path::Path::new(env!("CARGO_MANIFEST_DIR")).join("../..");
let rev = std::fs::read_to_string(root.join("craft-i18n.rev")).unwrap();
let rev = rev.trim();
assert!(rev.len() == 40 && rev.bytes().all(|b| b.is_ascii_hexdigit()), "craft-i18n.rev holds one full SHA: {rev:?}");
for manifest in ["ai.storyteller.pdfcraft.yml", "ai.storyteller.pdfcraft.bundle.yml"] {
let text = std::fs::read_to_string(root.join("packaging/linux/flatpak").join(manifest)).unwrap();
let source =
text.split("url: https://github.com/storytold/craft-i18n").nth(1).unwrap_or_else(|| panic!("{manifest}: a craft-i18n git source"));
let source = source.split("- type:").next().unwrap_or(source);
let commit = source
.lines()
.find_map(|l| l.trim().strip_prefix("commit:"))
.unwrap_or_else(|| panic!("{manifest}: the craft-i18n source pins a commit"));
assert_eq!(commit.trim().trim_matches(['\'', '"']), rev, "{manifest}");
assert!(text.contains("CRAFT_I18N_DIR: /run/build/pdfcraft/craft-i18n"), "{manifest}: the build names the catalogs");
}
}
// Regression test for #246: `eprintln!` panics when standard error is a broken pipe (a GUI
// launch on Windows), and a panic inside the panic hook aborts the process. The start-up
// path and the hook must write to standard error without it.
assets/icons/a-large-small.svg
+18
View File
@@ -0,0 +1,18 @@
<!-- @license lucide-static v1.49.0 - ISC -->
<svg
class="lucide lucide-a-large-small"
xmlns="http://www.w3.org/2000/svg"
width="24"
height="24"
viewBox="0 0 24 24"
fill="none"
stroke="currentColor"
stroke-width="2"
stroke-linecap="round"
stroke-linejoin="round"
>
<path d="m15 16 2.536-7.328a1.02 1.02 1 0 1 1.928 0L22 16" />
<path d="M15.697 14h5.606" />
<path d="m2 16 4.039-9.69a.5.5 0 0 1 .923 0L11 16" />
<path d="M3.304 13h6.392" />
</svg>
Side by Side

After

Width:  |  Height:  |  Size: 475 B

contributors/people.toml
+3
View File
@@ -183,3 +183,6 @@ sync_github_name = true
[people.Combiler01]
sync_github_name = true
[people.helenkwok]
real_name = "Helen Kwok"
craft-i18n.rev
+1
View File
@@ -0,0 +1 @@
575cfdef63227b14daf173b8cc87928bfdfa325e
crates/annot/src/appearance.rs
+189 -74
View File
@@ -232,6 +232,31 @@ fn rg_stroke(c: Rgb) -> String {
format!("{} {} {} RG\n", n(c[0]), n(c[1]), n(c[2]))
}
/// A FreeText's border colour where it has its own: the stroking colour in `/DA` (`RG`, or `G`
/// for grey). `None`: the border takes the text colour.
pub fn da_border(d: &Dict) -> Option<Rgb> {
let da = d.get(b"DA").and_then(|o| o.as_string()).map(|s| String::from_utf8_lossy(&s.bytes).into_owned())?;
let toks: Vec<&str> = da.split_whitespace().collect();
let num = |i: usize| toks.get(i).and_then(|t| t.parse::<f64>().ok()).filter(|x| x.is_finite());
let mut col = None;
for (i, t) in toks.iter().enumerate() {
match *t {
"RG" if i >= 3 => {
if let (Some(r), Some(g), Some(b)) = (num(i - 3), num(i - 2), num(i - 1)) {
col = Some([r, g, b].map(|x| x.clamp(0.0, 1.0)));
}
}
"G" if i >= 1 => {
if let Some(g) = num(i - 1) {
col = Some([g.clamp(0.0, 1.0); 3]);
}
}
_ => {}
}
}
col
}
/// `/DA` of a text box: (text colour, font size). Defaults to black 12 pt.
pub fn parse_da(d: &Dict) -> (Rgb, f64) {
let da = d.get(b"DA").and_then(|o| o.as_string()).map(|s| String::from_utf8_lossy(&s.bytes).into_owned()).unwrap_or_default();
@@ -262,23 +287,51 @@ pub fn parse_da(d: &Dict) -> (Rgb, f64) {
(col.map(|x| x.clamp(0.0, 1.0)), size.min(400.0))
}
fn ext_gstate(opacity: f64, multiply: bool) -> Option<Dict> {
if opacity >= 1.0 && !multiply {
/// How opaque a comment's parts are drawn, 0–1: its line or border (`/CA`); its fill or
/// background (`/ca`, which PDF 2.0 gives the non-stroking operations; without it, as the line);
/// and a text box's text (`/PCTextOpacity`, PdfCraft's own; without it, as the fill, since text
/// is filled too). A comment with only `/CA` is drawn at that opacity throughout, as before.
#[derive(Clone, Copy, Debug, PartialEq)]
pub struct PartOpacity {
pub line: f64,
pub fill: f64,
pub text: f64,
}
/// The parts' opacities of the comment `d` (see [`PartOpacity`]).
pub fn part_opacity(d: &Dict) -> PartOpacity {
let get = |key: &[u8]| d.get(key).and_then(|o| o.as_f64()).filter(|x| x.is_finite()).map(|x| x.clamp(0.0, 1.0));
let line = get(b"CA").unwrap_or(1.0);
let fill = get(b"ca").unwrap_or(line);
let text = get(b"PCTextOpacity").unwrap_or(fill);
PartOpacity { line, fill, text }
}
/// `/GS0`: the line's opacity for strokes and the fill's for everything filled; `/GS1`, where the
/// text's differs from the fill's: the text's for its glyphs.
fn ext_gstate(parts: PartOpacity, multiply: bool) -> Option<Dict> {
if parts.line >= 1.0 && parts.fill >= 1.0 && parts.text >= 1.0 && !multiply {
return None;
}
let mut gs = Dict::new();
gs.set(b"Type".to_vec(), Object::name("ExtGState"));
if multiply {
gs.set(b"BM".to_vec(), Object::name("Multiply"));
}
gs.set(b"CA".to_vec(), Object::Real(opacity));
gs.set(b"ca".to_vec(), Object::Real(opacity));
let gs = |stroke: f64, fill: f64| {
let mut gs = Dict::new();
gs.set(b"Type".to_vec(), Object::name("ExtGState"));
if multiply {
gs.set(b"BM".to_vec(), Object::name("Multiply"));
}
gs.set(b"CA".to_vec(), Object::Real(stroke));
gs.set(b"ca".to_vec(), Object::Real(fill));
Object::Dict(gs)
};
let mut res = Dict::new();
res.set(b"GS0".to_vec(), Object::Dict(gs));
res.set(b"GS0".to_vec(), gs(parts.line, parts.fill));
if parts.text != parts.fill {
res.set(b"GS1".to_vec(), gs(parts.line, parts.text));
}
Some(res)
}
fn form(bbox: [f64; 4], content: &[u8], resources: Dict) -> Stream {
pub(crate) fn form(bbox: [f64; 4], content: &[u8], resources: Dict) -> Stream {
let mut d = Dict::new();
d.set(b"Type".to_vec(), Object::name("XObject"));
d.set(b"Subtype".to_vec(), Object::name("Form"));
@@ -288,6 +341,18 @@ fn form(bbox: [f64; 4], content: &[u8], resources: Dict) -> Stream {
Stream::flate(d, content)
}
/// The `cm` that draws a picture into `at`: an image (its unit square) or a form XObject whose
/// `/Matrix` maps it to `size` points (`None`, or a side that is not positive, draws nothing).
pub(crate) fn picture_place(at: [f64; 4], image: bool, size: Option<(f64, f64)>) -> Option<String> {
let [x0, y0, x1, y1] = at;
let (w, h) = (x1 - x0, y1 - y0);
if image {
return Some(format!("{} 0 0 {} {} {} cm", n(w), n(h), n(x0), n(y0)));
}
let (sw, sh) = size.filter(|(sw, sh)| *sw > 0.0 && *sh > 0.0)?;
Some(format!("{} 0 0 {} {} {} cm", n(w / sw), n(h / sh), n(x0), n(y0)))
}
/// The page `/Rotate` an image signature's picture is turned back by (`/PCPictureRotate`, written
/// when the signature is added to a turned page): 0, 90, 180 or 270.
pub(crate) fn picture_turn(d: &Dict) -> i64 {
@@ -301,10 +366,18 @@ fn stamp_turn(d: &Dict) -> i64 {
/// The page `/Rotate` a text box's or Fill & Sign text's lines are turned back by
/// (`/PCTextRotate`, written when it is added to a turned page).
fn text_turn(d: &Dict) -> i64 {
pub(crate) fn text_turn(d: &Dict) -> i64 {
d.get(b"PCTextRotate").and_then(Object::as_int).unwrap_or(0).rem_euclid(360) / 90 * 90
}
/// A FreeText's `/RD`: how far its text box lies inside `/Rect` from the left, bottom, right and
/// top. A distance a hair below 0 (rounding) reads as 0; `None` when one is negative or they can't
/// be read.
pub(crate) fn inset(d: &Dict) -> Option<[f64; 4]> {
let rd: [f64; 4] = nums(d, b"RD")?.try_into().ok()?;
rd.iter().all(|x| x.is_finite() && *x > -0.001).then(|| rd.map(|x| x.max(0.0)))
}
/// The first text [`build`] would draw in the standard fonts (Helvetica, WinAnsiEncoding) that
/// they can't, which would come out as `?`: what it is (for the message) and the character. The
/// text of a text box, callout or Fill & Sign text, a line's caption, a measurement's label, a
@@ -325,7 +398,7 @@ pub fn undrawable_text(d: &Dict) -> Option<(&'static str, char)> {
}
/// `/Font` resources naming `/Helv`: standard Helvetica with WinAnsiEncoding.
fn helvetica_resource() -> Object {
pub(crate) fn helvetica_resource() -> Object {
let mut font = Dict::new();
font.set(b"Type".to_vec(), Object::name("Font"));
font.set(b"Subtype".to_vec(), Object::name("Type1"));
@@ -379,7 +452,9 @@ struct AppearanceContext<'a> {
subtype: &'a [u8],
rect: [f64; 4],
cloudy: bool,
/// The line's opacity (`parts.line`).
opacity: f64,
parts: PartOpacity,
stroke: Option<Rgb>,
width: f64,
}
@@ -396,7 +471,7 @@ impl AppearanceDrawing {
fn new(context: &AppearanceContext<'_>) -> Self {
let mut drawing = Self::default();
if let Some(gs) = ext_gstate(
context.opacity,
context.parts,
context.subtype == b"Highlight" || (context.subtype == b"Ink" && context.dict.name(b"IT") == Some(b"FreehandHighlight")),
) {
drawing.resources.set(b"ExtGState".to_vec(), Object::Dict(gs));
@@ -443,7 +518,8 @@ fn appearance_context(d: &Dict) -> Option<AppearanceContext<'_>> {
subtype,
rect,
cloudy,
opacity: d.get(b"CA").and_then(|o| o.as_f64()).unwrap_or(1.0).clamp(0.0, 1.0),
opacity: part_opacity(d).line,
parts: part_opacity(d),
stroke: color(d, b"C")?,
width: border_width(d),
})
@@ -786,15 +862,12 @@ fn draw_stamp(context: &AppearanceContext<'_>, mut drawing: AppearanceDrawing) -
// the standard counter-rotation as its form `/Matrix`: viewers fit the turned
// bounding box to `/Rect` (ISO 32000-2 §12.5.5), so it reads upright as displayed.
if let Some(pic) = d.get(b"PCPicture").and_then(Object::as_ref) {
let [x0, y0, x1, y1] = rect;
let (w, h) = (x1 - x0, y1 - y0);
let image = matches!(d.get(b"PCPictureImage"), Some(Object::Bool(true)));
let place = if image {
format!("{} 0 0 {} {} {} cm", n(w), n(h), n(x0), n(y0))
} else {
let size = nums(d, b"PCPictureSize").filter(|s| s.len() == 2 && s[0] > 0.0 && s[1] > 0.0)?;
format!("{} 0 0 {} {} {} cm", n(w / size[0]), n(h / size[1]), n(x0), n(y0))
};
let size = nums(d, b"PCPictureSize").and_then(|s| match s.as_slice() {
[w, h] if *w > 0.0 && *h > 0.0 => Some((*w, *h)),
_ => None,
});
let place = picture_place(rect, image, size)?;
c.push_str(&format!("q {place} /Pic Do Q\n"));
let mut xo = Dict::new();
xo.set(b"Pic".to_vec(), Object::Ref(pic));
@@ -836,46 +909,25 @@ fn draw_stamp(context: &AppearanceContext<'_>, mut drawing: AppearanceDrawing) -
let [x0, y0, x1, y1] = draw;
let (w, h) = (x1 - x0, y1 - y0);
let lw = w.min(h) * 0.12;
// The strokes come from `FillMark::strokes`, which a copy's outline draws too.
let path = |mark: crate::FillMark| {
let ops: Vec<String> = mark
.strokes(draw)
.iter()
.flat_map(|s| s.iter().enumerate().map(|(i, p)| format!("{} {} {}", n(p[0]), n(p[1]), if i == 0 { "m" } else { "l" })))
.collect();
format!("{} S\n", ops.join(" "))
};
match name {
b"PCCheck" => c.push_str(&format!(
"{}{} w 1 J 1 j\n{} {} m {} {} l {} {} l S\n",
rg_stroke(col),
n(lw),
n(x0 + w * 0.15),
n(y0 + h * 0.5),
n(x0 + w * 0.4),
n(y0 + h * 0.2),
n(x0 + w * 0.88),
n(y0 + h * 0.85)
)),
b"PCCross" => c.push_str(&format!(
"{}{} w 1 J\n{} {} m {} {} l {} {} m {} {} l S\n",
rg_stroke(col),
n(lw),
n(x0 + w * 0.18),
n(y0 + h * 0.18),
n(x1 - w * 0.18),
n(y1 - h * 0.18),
n(x0 + w * 0.18),
n(y1 - h * 0.18),
n(x1 - w * 0.18),
n(y0 + h * 0.18)
)),
b"PCCheck" => c.push_str(&format!("{}{} w 1 J 1 j\n{}", rg_stroke(col), n(lw), path(crate::FillMark::Check))),
b"PCCross" => c.push_str(&format!("{}{} w 1 J\n{}", rg_stroke(col), n(lw), path(crate::FillMark::Cross))),
b"PCDot" => {
let r = w.min(h) * 0.3;
c.push_str(&rg(col));
c.push_str(&ellipse(x0 + w / 2.0 - r, y0 + h / 2.0 - r, x0 + w / 2.0 + r, y0 + h / 2.0 + r));
c.push_str("f\n");
}
b"PCLine" => c.push_str(&format!(
"{}{} w 1 J\n{} {} m {} {} l S\n",
rg_stroke(col),
n(h.clamp(0.5, 2.0)),
n(x0),
n(y0 + h / 2.0),
n(x1),
n(y0 + h / 2.0)
)),
b"PCLine" => c.push_str(&format!("{}{} w 1 J\n{}", rg_stroke(col), n(h.clamp(0.5, 2.0)), path(crate::FillMark::Line))),
other => {
// Only stamps PdfCraft made: others keep their own artwork.
if !matches!(d.get(b"PCStamp"), Some(Object::Bool(true))) {
@@ -893,14 +945,59 @@ fn draw_stamp(context: &AppearanceContext<'_>, mut drawing: AppearanceDrawing) -
Some(form(rect, c.as_bytes(), res))
}
/// The oval or circle (`shape`) around a text box whose whole box is `full`: its background
/// and border go into `c`; returns the text box inside it.
#[allow(clippy::too_many_arguments)]
fn shaped_border(
d: &Dict,
shape: crate::BorderShape,
full: [f64; 4],
turn: i64,
bw: f64,
bg: Option<Rgb>,
line_color: Rgb,
c: &mut String,
) -> Option<[f64; 4]> {
// Laid out as displayed: a quarter turn swaps the sides' differences.
let rect = match inset(d) {
Some(rd) => {
let (dx, dy) = if matches!(turn, 90 | 270) { (rd[1], rd[0]) } else { (rd[0], rd[1]) };
[full[0] + dx, full[1] + dy, full[2] - dx, full[3] - dy]
}
None => shape.inside(full, bw, full),
};
if rect[2] - rect[0] < 1.0 || rect[3] - rect[1] < 1.0 {
return None;
}
let h = bw / 2.0;
let [x0, y0, x1, y1] = [full[0] + h, full[1] + h, full[2] - h, full[3] - h];
let path = match shape {
crate::BorderShape::Circle => {
let (cx, cy, r) = ((x0 + x1) / 2.0, (y0 + y1) / 2.0, (x1 - x0).min(y1 - y0) / 2.0);
ellipse(cx - r, cy - r, cx + r, cy + r)
}
_ => ellipse(x0, y0, x1, y1),
};
if let Some(bg) = bg {
c.push_str(&format!("{}{path}f\n", rg(bg)));
}
if bw > 0.0 {
c.push_str(&format!("{}{} w\n{}{path}S\n[] 0 d\n", rg_stroke(line_color), n(bw), dash(d)));
}
Some(rect)
}
fn draw_free_text(context: &AppearanceContext<'_>, mut drawing: AppearanceDrawing) -> Option<Stream> {
let d = context.dict;
let parts = context.parts;
let rect = context.rect;
let stroke = context.stroke;
let mut c = std::mem::take(&mut drawing.content);
let mut res = std::mem::take(&mut drawing.resources);
let _tail = std::mem::take(&mut drawing.tail);
let (text_color, size) = parse_da(d);
// The box's border and a callout's leader line: their own colour, else the text's.
let line_color = da_border(d).unwrap_or(text_color);
let bg = stroke;
let bw = if d.contains(b"BS") || d.contains(b"Border") { border_width(d) } else { 0.0 };
// A callout: the leader line (arrowhead at its first point), and the text box inside
@@ -939,7 +1036,7 @@ fn draw_free_text(context: &AppearanceContext<'_>, mut drawing: AppearanceDrawin
if pts.len() < 2 || pts.iter().any(|p| !p.0.is_finite() || !p.1.is_finite()) {
return None;
}
c.push_str(&format!("{}{}{} w 1 J 1 j\n", rg_stroke(text_color), rg(bg.unwrap_or([1.0; 3])), n(lw)));
c.push_str(&format!("{}{}{} w 1 J 1 j\n", rg_stroke(line_color), rg(bg.unwrap_or([1.0; 3])), n(lw)));
for (i, p) in pts.iter().enumerate() {
c.push_str(&format!("{} {} {}\n", n(p.0), n(p.1), if i == 0 { "m" } else { "l" }));
}
@@ -948,26 +1045,40 @@ fn draw_free_text(context: &AppearanceContext<'_>, mut drawing: AppearanceDrawin
return None;
}
}
if let Some(bg) = bg {
c.push_str(&format!("{}{} {} {} {} re f\n", rg(bg), n(rect[0]), n(rect[1]), n(rect[2] - rect[0]), n(rect[3] - rect[1])));
}
if bw > 0.0 {
let h = bw / 2.0;
c.push_str(&format!(
"{}{} w\n{}{} {} {} {} re S\n[] 0 d\n",
rg_stroke(text_color),
n(bw),
dash(d),
n(rect[0] + h),
n(rect[1] + h),
n(rect[2] - rect[0] - bw),
n(rect[3] - rect[1] - bw)
));
// An oval or a circle (`/PCBorderShape`) around the text box, which is `/Rect` inset by
// `/RD` and keeps its place: background and border take the shape, the text is laid out
// in the text box as in a rectangle.
let shape = crate::BorderShape::of(d);
if shape != crate::BorderShape::Rectangle {
rect = shaped_border(d, shape, full, turn, bw, bg, line_color, &mut c)?;
} else {
if let Some(bg) = bg {
c.push_str(&format!("{}{} {} {} {} re f\n", rg(bg), n(rect[0]), n(rect[1]), n(rect[2] - rect[0]), n(rect[3] - rect[1])));
}
if bw > 0.0 {
let h = bw / 2.0;
c.push_str(&format!(
"{}{} w\n{}{} {} {} {} re S\n[] 0 d\n",
rg_stroke(line_color),
n(bw),
dash(d),
n(rect[0] + h),
n(rect[1] + h),
n(rect[2] - rect[0] - bw),
n(rect[3] - rect[1] - bw)
));
}
}
let text = d.get(b"Contents").and_then(|o| o.as_string()).map(PdfString::to_text).unwrap_or_default();
let pad = 2.0 + bw;
let width = (rect[2] - rect[0] - 2.0 * pad).max(1.0);
let q = d.int(b"Q").unwrap_or(0);
// In an oval or a circle the lines are centred in the text box, up and down too.
let round = shape != crate::BorderShape::Rectangle;
let q = if round { 1 } else { d.int(b"Q").unwrap_or(0) };
// The text at its own opacity, where it has one.
if parts.text != parts.fill {
c.push_str("/GS1 gs\n");
}
c.push_str(&format!(
"{} {} {} {} re W n\nBT\n/Helv {} Tf\n{}",
n(rect[0]),
@@ -978,8 +1089,12 @@ fn draw_free_text(context: &AppearanceContext<'_>, mut drawing: AppearanceDrawin
rg(text_color)
));
let mut out = c.into_bytes();
let lines = wrap(&text, size, width);
let mut y = rect[3] - pad - size * 0.9;
for line in wrap(&text, size, width) {
if round {
y -= ((rect[3] - rect[1] - 2.0 * pad - lines.len() as f64 * size * 1.2) / 2.0).max(0.0);
}
for line in lines {
if y < rect[1] - size {
break;
}
crates/annot/src/appearance/symbols.rs
+2 -2
View File
@@ -1,4 +1,4 @@
use super::{Dict, Object, PdfString, Rgb, Stream, color, ext_gstate, form, literal, n, nums, rg, rg_stroke, text_width, win_ansi};
use super::{Dict, Object, PartOpacity, PdfString, Rgb, Stream, color, ext_gstate, form, literal, n, nums, rg, rg_stroke, text_width, win_ansi};
/// The label and colour of a standard rubber stamp name (ISO 32000-2 Table 184), for a stamp
/// another program wrote without an appearance. Names are matched byte for byte, as PDF names
@@ -51,7 +51,7 @@ pub(crate) fn display_stamp(d: &Dict) -> Option<Stream> {
// A malformed `/C` falls back to the stamp's own colour rather than hiding the stamp.
let stroke = color(d, b"C").flatten();
let mut res = Dict::new();
if let Some(gs) = ext_gstate(opacity, false) {
if let Some(gs) = ext_gstate(PartOpacity { line: opacity, fill: opacity, text: opacity }, false) {
res.set(b"ExtGState".to_vec(), Object::Dict(gs));
}
// A stamp whose intent (`/IT`, PDF 2.0) is an image or a snapshot has no `/Name` to default.
crates/annot/src/create.rs
+14 -13
View File
@@ -1,6 +1,6 @@
use super::draw::draw_appearance;
use super::lookup::{annot_dict, annot_ref, annots, fill_sign_shape, page_ref, set_annots};
use super::modify::{n, rich_text};
use super::modify::{free_text_da, n, rich_text, set_part_keys, shape_text_box};
use super::types::{
FLAG_NO_ROTATE, FLAG_NO_ZOOM, FLAG_PRINT, LineEnding, Markup, Meta, NOTE_SIZE, NewAnnotation, OverlayLook, Rgb, Shape, StampGroup, Style,
};
@@ -214,8 +214,13 @@ pub fn add_annotation(doc: &mut Document, new: &NewAnnotation, meta: &Meta) -> R
let rect = rect_for(&new.shape, style)?;
let mut d = base_dict(new.shape.subtype(), rect, page, &new.contents, &new.author, meta);
d.set(b"Subj".to_vec(), PdfString::text(subject(&new.shape)));
let opacity = style.opacity.clamp(0.0, 1.0);
set_non_opaque_style(&mut d, opacity);
if style.fill_opacity.into_iter().chain(style.text_opacity).any(|o| !o.is_finite()) {
return Err(AnnotError::Invalid("invalid opacity".into()));
}
let line = style.opacity.clamp(0.0, 1.0);
let fill = style.fill_opacity.map_or(line, |o| o.clamp(0.0, 1.0));
let text = style.text_opacity.map_or(fill, |o| o.clamp(0.0, 1.0));
set_part_keys(&mut d, appearance::PartOpacity { line, fill, text });
let border = |d: &mut Dict| {
let mut bs = Dict::new();
bs.set(b"W".to_vec(), Object::Real(style.width.max(0.0)));
@@ -380,20 +385,22 @@ pub fn add_annotation(doc: &mut Document, new: &NewAnnotation, meta: &Meta) -> R
d.set(b"IT".to_vec(), Object::name("FreeTextTypeWriter"));
}
let size = if font_size.is_finite() && *font_size > 0.0 { font_size.clamp(1.0, 400.0) } else { 12.0 };
let [r, g, b] = style.color.map(|x| x.clamp(0.0, 1.0));
d.set(b"DA".to_vec(), PdfString::literal(format!("{} {} {} rg /Helv {} Tf", n(r), n(g), n(b), n(size)).into_bytes()));
d.set(b"DA".to_vec(), free_text_da(style.color, style.border_color, size));
d.set(b"Q".to_vec(), Object::Int(0));
if let Some(f) = style.fill {
d.set(b"C".to_vec(), rgb(f));
}
border(&mut d);
rich_text(&mut d);
// The text is laid out for the page as displayed, like a stamp's label.
if !matches!(new.shape, Shape::Callout { .. }) {
// The text is laid out for the page as displayed, like a stamp's label.
let turn = pdfcraft_model::pages(doc).get(new.page).map_or(0, |p| p.rotation(doc));
if matches!(turn, 90 | 180 | 270) {
d.set(b"PCTextRotate".to_vec(), Object::Int(turn));
}
// An oval or a circle goes around the text box, its text wrapped as displayed:
// after `/PCTextRotate`, which swaps the box's sides on a quarter-turned page.
shape_text_box(&mut d, rect, style.border_shape, style.width);
}
}
_ => {}
@@ -421,7 +428,7 @@ pub fn add_annotation(doc: &mut Document, new: &NewAnnotation, meta: &Meta) -> R
}
fn validate_style(style: &Style) -> Result<(), AnnotError> {
if !finite(&style.color) || !style.opacity.is_finite() || !style.width.is_finite() {
if !finite(&style.color) || style.border_color.is_some_and(|c| !finite(&c)) || !style.opacity.is_finite() || !style.width.is_finite() {
return Err(AnnotError::Invalid("invalid style".into()));
}
Ok(())
@@ -443,12 +450,6 @@ fn set_redaction_properties(dict: &mut Dict, style: &Style, quads: &[[f64; 8]],
}
}
fn set_non_opaque_style(dict: &mut Dict, opacity: f64) {
if opacity < 1.0 {
dict.set(b"CA".to_vec(), Object::Real(opacity));
}
}
fn add_popup(doc: &mut Document, parent: ObjRef, rect: Option<[f64; 4]>, annots: &mut Vec<Object>) -> Result<(), AnnotError> {
let Some(rect) = rect else { return Ok(()) };
let mut popup = Dict::new();
crates/annot/src/info.rs
+15 -2
View File
@@ -2,7 +2,7 @@ use super::create::finite;
use super::draw::set_appearance;
use super::lookup::{annot_dict, annot_ref, annotation_flags, annots, drawn_upright, is_fill_sign, page_ref, page_refs};
use super::modify::{fillable, touch, unlocked};
use super::types::{FLAG_LOCKED, LineEnding, Meta, NoteIcon, Rgb};
use super::types::{BorderShape, FLAG_LOCKED, LineEnding, Meta, NoteIcon, Rgb};
use super::{AnnotError, Dict, Document, ObjRef, Object, PdfString, appearance};
// ── reading ─────────────────────────────────────────────────────────────────────────────────
@@ -217,6 +217,15 @@ pub struct Props {
pub fill: Option<Rgb>,
/// Text boxes, typewriter text and callouts: the `/DA` font size in points.
pub font_size: Option<f64>,
/// Text boxes, typewriter text and callouts: the border's own colour (`None`: the text's).
pub border_color: Option<Rgb>,
/// Text boxes and typewriter text: the border's shape (callouts have none to choose).
pub border_shape: Option<BorderShape>,
/// How opaque the fill is drawn, for a comment with a fill control (`opacity` is then its
/// line's or border's).
pub fill_opacity: Option<f64>,
/// How opaque a text box's, typewriter text's or callout's text is drawn.
pub text_opacity: Option<f64>,
}
/// The current properties of the comment at `(page, index)`.
@@ -239,7 +248,7 @@ pub fn props(doc: &Document, page: usize, index: usize) -> Option<Props> {
author: text_value(doc, d, b"T").unwrap_or_default(),
subject: text_value(doc, d, b"Subj").unwrap_or_default(),
color,
opacity: d.get(b"CA").and_then(|o| o.as_f64()).unwrap_or(1.0),
opacity: appearance::part_opacity(d).line,
width,
icon: (subtype == "Text").then(|| d.name(b"Name").and_then(|n| NoteIcon::from_name(&String::from_utf8_lossy(n))).unwrap_or(NoteIcon::Note)),
modified: text_value(doc, d, b"M"),
@@ -248,6 +257,10 @@ pub fn props(doc: &Document, page: usize, index: usize) -> Option<Props> {
fillable: fillable(&subtype),
fill: fill_of(d),
font_size: (subtype == "FreeText").then(|| appearance::parse_da(d).1),
border_color: (subtype == "FreeText").then(|| appearance::da_border(d)).flatten(),
border_shape: (subtype == "FreeText" && !d.contains(b"CL")).then(|| BorderShape::of(d)),
fill_opacity: fillable(&subtype).then(|| appearance::part_opacity(d).fill),
text_opacity: (subtype == "FreeText").then(|| appearance::part_opacity(d).text),
subtype,
endings: line_endings_of(d),
})
crates/annot/src/lib.rs
+15 -4
View File
@@ -22,7 +22,9 @@ mod create;
mod draw;
mod info;
mod lookup;
mod model3d;
mod modify;
mod threed;
mod types;
#[cfg(test)]
@@ -36,14 +38,17 @@ use lookup::{annots, page_ref};
pub use lookup::{
has_visible_fill_sign, is_fill_sign, isolate_annotation, isolate_annotations, page_refs, picture_rotation, set_annots, signature_image,
};
pub use model3d::{Model3d, Model3dKind, model_3d, model_3d_annotations};
pub(crate) use modify::n;
pub use modify::{
add_reply, add_text_replacement, deletable, delete_annotation, erase_ink, fillable, goes_with, lift_annotation, move_annotation, set_contents,
set_fill, set_font_size, set_locked, set_marked, set_points, set_rect, set_review_state, set_style, with_replies_and_popups,
add_reply, add_text_replacement, deletable, delete_annotation, erase_ink, fillable, goes_with, lift_annotation, move_annotation,
set_border_color, set_border_shape, set_contents, set_fill, set_font_size, set_locked, set_marked, set_part_opacity, set_points, set_rect,
set_review_state, set_style, with_replies_and_popups,
};
pub use threed::{Activation as Activation3d, MAX_MODEL_BYTES, ModelFormat, NewModel3d, PosterPicture, add_3d_model, check_model};
pub use types::{
AttachIcon, FillMark, LineEnding, Markup, Meta, NOTE_SIZE, NewAnnotation, NoteIcon, OverlayFont, OverlayLook, ReviewState, Rgb, Shape,
StampGroup, StampKind, Style, StyleChange, TextReplacement, rect_quad,
AttachIcon, BorderShape, FillMark, LineEnding, Markup, Meta, NOTE_SIZE, NewAnnotation, NoteIcon, OverlayFont, OverlayLook, ReviewState, Rgb,
Shape, StampGroup, StampKind, Style, StyleChange, TextReplacement, rect_quad,
};
pub mod appearance;
@@ -70,6 +75,12 @@ pub enum AnnotError {
/// with a font that has those characters. The message names the character.
#[error("{0}")]
Undrawable(String),
/// A 3D model file that is neither U3D (starts with `U3D` and a zero byte) nor PRC (starts `PRC`).
#[error("not a 3D model: the file must be U3D (ECMA-363) or PRC (ISO 14739-1)")]
NotAModel,
/// A 3D model larger than `MAX_MODEL_BYTES`.
#[error("the 3D model is {size} bytes; the largest accepted is {max}")]
ModelTooLarge { size: usize, max: usize },
#[error("{0}")]
Cos(#[from] pdfcraft_cos::CosError),
}
crates/annot/src/model3d.rs
+77
View File
@@ -0,0 +1,77 @@
//! Reading the model a 3D annotation shows (ISO 32000-2 §13.6.3 and §13.6.4): its `/3DD` is a
//! 3D stream, or a 3D reference dictionary naming the stream another annotation shares.
use super::lookup::{annots, page_ref};
use super::threed::MAX_MODEL_BYTES;
use super::{AnnotError, Dict, Document, Object};
/// The format of a 3D stream, from its `/Subtype`.
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum Model3dKind {
/// Product Representation Compact (ISO 14739-1).
Prc,
/// Universal 3D (ECMA-363).
U3d,
}
/// A 3D annotation's model: its format and the decoded stream.
#[derive(Debug, Clone, PartialEq)]
pub struct Model3d {
pub kind: Model3dKind,
pub data: Vec<u8>,
}
/// The `/Annots` positions of the 3D annotations on 0-based `page`, in order.
///
/// # Errors
/// [`AnnotError::NoPageTree`] or [`AnnotError::NoSuchPage`].
pub fn model_3d_annotations(doc: &Document, page: usize) -> Result<Vec<usize>, AnnotError> {
let list = annots(doc, page_ref(doc, page)?);
Ok(list.iter().enumerate().filter(|(_, a)| doc.dict(a).is_some_and(|d| d.name(b"Subtype") == Some(b"3D"))).map(|(i, _)| i).collect())
}
/// The model of the 3D annotation at `/Annots` position `index` on 0-based `page`, decoded.
///
/// # Errors
/// [`AnnotError::NoSuchAnnotation`] when there is no annotation there; [`AnnotError::Invalid`]
/// when it is not a 3D annotation, has no readable 3D stream, names a format other than PRC or
/// U3D, or decodes past [`MAX_MODEL_BYTES`].
pub fn model_3d(doc: &Document, page: usize, index: usize) -> Result<Model3d, AnnotError> {
let list = annots(doc, page_ref(doc, page)?);
let entry = list.get(index).ok_or(AnnotError::NoSuchAnnotation { page, index })?;
let annot = doc.dict(entry).ok_or(AnnotError::NoSuchAnnotation { page, index })?;
let at = || format!("annotation {} on page {}", index.saturating_add(1), page.saturating_add(1));
if annot.name(b"Subtype") != Some(b"3D") {
return Err(AnnotError::Invalid(format!("{} is not a 3D annotation", at())));
}
let stream = stream_of(doc, &annot).ok_or_else(|| AnnotError::Invalid(format!("{} has no 3D stream", at())))?;
let Object::Stream(s) = stream.as_ref() else {
return Err(AnnotError::Invalid(format!("{} has no 3D stream", at())));
};
let kind = match s.dict.name(b"Subtype") {
Some(b"PRC") => Model3dKind::Prc,
Some(b"U3D") => Model3dKind::U3d,
other => {
let name = other.map_or_else(|| "none".to_owned(), |n| String::from_utf8_lossy(n).into_owned());
return Err(AnnotError::Invalid(format!("{}'s 3D stream is of an unknown format ({name})", at())));
}
};
let data = doc
.with_direct_parms(s)
.decoded_within(MAX_MODEL_BYTES)
.map_err(|e| AnnotError::Invalid(format!("{}'s 3D stream could not be decoded: {e}", at())))?;
Ok(Model3d { kind, data })
}
/// What an annotation's `/3DD` names: the 3D stream itself, or the `/3D` entry of a 3D reference
/// dictionary. A reference is followed once; the caller refuses anything that is not a stream.
fn stream_of(doc: &Document, annot: &Dict) -> Option<std::sync::Arc<Object>> {
let obj = doc.resolve(annot.get(b"3DD")?);
match obj.as_ref() {
Object::Dict(d) if d.name(b"Type") == Some(b"3DRef") => Some(doc.resolve(d.get(b"3D")?)),
_ => Some(obj),
}
}
#[cfg(test)]
mod tests;
crates/annot/src/model3d/tests.rs
+88
View File
@@ -0,0 +1,88 @@
use std::sync::Arc;
use super::*;
/// A one-page document whose page has `annots` as its `/Annots`, plus `extra` objects from 4 on.
fn doc(annots: &str, extra: &[Vec<u8>]) -> Document {
let mut objs: Vec<Vec<u8>> = vec![
b"<< /Type /Catalog /Pages 2 0 R >>".to_vec(),
b"<< /Type /Pages /Kids [3 0 R] /Count 1 /MediaBox [0 0 612 792] >>".to_vec(),
format!("<< /Type /Page /Parent 2 0 R /Annots {annots} >>").into_bytes(),
];
objs.extend(extra.iter().cloned());
let mut out = b"%PDF-1.7\n".to_vec();
let mut offsets = Vec::new();
for (i, o) in objs.iter().enumerate() {
offsets.push(out.len());
out.extend_from_slice(format!("{} 0 obj\n", i + 1).as_bytes());
out.extend_from_slice(o);
out.extend_from_slice(b"\nendobj\n");
}
let xref = out.len();
out.extend_from_slice(format!("xref\n0 {}\n0000000000 65535 f \n", objs.len() + 1).as_bytes());
for o in offsets {
out.extend_from_slice(format!("{o:010} 00000 n \n").as_bytes());
}
out.extend_from_slice(format!("trailer\n<< /Size {} /Root 1 0 R >>\nstartxref\n{xref}\n%%EOF\n", objs.len() + 1).as_bytes());
Document::open(Arc::new(out)).expect("opens")
}
fn stream(dict: &str, body: &[u8]) -> Vec<u8> {
let mut s = format!("<< {dict} /Length {} >>\nstream\n", body.len()).into_bytes();
s.extend_from_slice(body);
s.extend_from_slice(b"\nendstream");
s
}
fn annot(dd: &str) -> Vec<u8> {
format!("<< /Type /Annot /Subtype /3D /Rect [0 0 100 100] /3DD {dd} >>").into_bytes()
}
#[test]
fn a_plain_3d_stream_is_read_through_its_filters() {
let d = doc("[4 0 R]", &[annot("5 0 R"), stream("/Type /3D /Subtype /PRC /Filter /ASCIIHexDecode", b"505243>")]);
assert_eq!(model_3d(&d, 0, 0), Ok(Model3d { kind: Model3dKind::Prc, data: b"PRC".to_vec() }));
}
#[test]
fn a_3dref_leads_to_the_stream_it_names() {
let d = doc("[4 0 R]", &[annot("<< /Type /3DRef /3D 5 0 R >>"), stream("/Type /3D /Subtype /PRC", b"PRC")]);
assert_eq!(model_3d(&d, 0, 0).map(|m| m.data), Ok(b"PRC".to_vec()));
}
#[test]
fn a_u3d_stream_is_read_and_says_so() {
let d = doc("[4 0 R]", &[annot("5 0 R"), stream("/Type /3D /Subtype /U3D", b"U3D")]);
assert_eq!(model_3d(&d, 0, 0).map(|m| m.kind), Ok(Model3dKind::U3d));
}
#[test]
fn a_3d_annotation_without_a_stream_is_invalid() {
let d = doc("[<< /Type /Annot /Subtype /3D /Rect [0 0 1 1] >>]", &[]);
assert!(matches!(model_3d(&d, 0, 0), Err(AnnotError::Invalid(m)) if m.contains("no 3D stream")));
}
#[test]
fn an_unknown_format_is_invalid() {
let d = doc("[4 0 R]", &[annot("5 0 R"), stream("/Type /3D /Subtype /STL", b"x")]);
assert!(matches!(model_3d(&d, 0, 0), Err(AnnotError::Invalid(m)) if m.contains("unknown format (STL)")));
}
#[test]
fn a_3dref_is_followed_once_and_only_to_a_stream() {
// A reference naming another reference, or one that names itself, is damage: no stream.
let d = doc("[4 0 R]", &[annot("5 0 R"), b"<< /Type /3DRef /3D 6 0 R >>".to_vec(), b"<< /Type /3DRef /3D 5 0 R >>".to_vec()]);
assert!(matches!(model_3d(&d, 0, 0), Err(AnnotError::Invalid(m)) if m.contains("no 3D stream")));
// The stream is under `/3D`; a `/3DD` inside the reference is not the standard's key.
let d = doc("[4 0 R]", &[annot("<< /Type /3DRef /3DD 5 0 R >>"), stream("/Type /3D /Subtype /PRC", b"PRC")]);
assert!(matches!(model_3d(&d, 0, 0), Err(AnnotError::Invalid(m)) if m.contains("no 3D stream")));
}
#[test]
fn another_annotation_is_not_a_3d_one() {
let d = doc("[<< /Type /Annot /Subtype /Square /Rect [0 0 1 1] >> 4 0 R]", &[annot("5 0 R"), stream("/Subtype /PRC", b"PRC")]);
assert!(matches!(model_3d(&d, 0, 0), Err(AnnotError::Invalid(m)) if m.contains("not a 3D annotation")));
assert_eq!(model_3d(&d, 0, 2), Err(AnnotError::NoSuchAnnotation { page: 0, index: 2 }));
assert_eq!(model_3d_annotations(&d, 0), Ok(vec![1]));
assert_eq!(model_3d_annotations(&d, 1), Err(AnnotError::NoSuchPage(1)));
}
crates/annot/src/modify.rs
+264 -22
View File
@@ -5,8 +5,8 @@ use super::draw::{draw_appearance, redraw_or_restore, set_appearance};
use super::info::{line_endings_of, text_value};
use super::lookup::{annot_dict, annot_ref, annotation_flags, annots, page_ref, set_annots};
use super::types::{
FLAG_HIDDEN, FLAG_LOCKED, FLAG_NO_ROTATE, FLAG_NO_ZOOM, FLAG_PRINT, LineEnding, Markup, Meta, NewAnnotation, ReviewState, Rgb, Shape, Style,
StyleChange, TextReplacement,
BorderShape, FLAG_HIDDEN, FLAG_LOCKED, FLAG_NO_ROTATE, FLAG_NO_ZOOM, FLAG_PRINT, LineEnding, Markup, Meta, NewAnnotation, ReviewState, Rgb,
Shape, Style, StyleChange, TextReplacement,
};
use super::{AnnotError, Dict, Document, HashMap, HashSet, ObjRef, Object, PdfString, appearance};
@@ -159,10 +159,13 @@ pub fn with_replies_and_popups(doc: &Document, list: &[Object], doomed: impl Int
}
/// Change a comment's text. Text boxes are redrawn to show it, and their rectangle follows the
/// new text: the wrap width and top edge stay, the height fits the wrapped lines.
/// new text: the wrap width and top edge stay, the height fits the wrapped lines. A circle snug
/// around its text stays snug around the new text, centred where it was.
pub fn set_contents(doc: &mut Document, page: usize, index: usize, text: &str, meta: &Meta) -> Result<(), AnnotError> {
let (_, r) = annot_ref(doc, page, index)?;
let free_text = annot_dict(doc, r).name(b"Subtype") == Some(b"FreeText");
let d = annot_dict(doc, r);
let free_text = d.name(b"Subtype") == Some(b"FreeText");
let snug = free_text && is_snug(&d);
doc.update_dict(r, |d| {
d.set(b"Contents".to_vec(), PdfString::text(text));
// The rich-text version follows the new plain text (text boxes), or would contradict it.
@@ -177,8 +180,16 @@ pub fn set_contents(doc: &mut Document, page: usize, index: usize, text: &str, m
// The box grows and shrinks with the text instead of clipping it. A locked box keeps its
// text editable (as in Acrobat), so the re-fit ignores the lock. New text is drawn even
// where the standard fonts can't draw some of it (marked and reported, see
// `draw_appearance`).
match fitted_box(doc, r, text) {
// `draw_appearance`). A circle snug around its text stays snug around the new text; one
// made larger by hand keeps its size, only growing for more lines.
let fit = if snug {
snug_refit(doc, r)
} else {
let d = annot_dict(doc, r);
let circle = text_box_of(&d).filter(|_| BorderShape::of(&d) == BorderShape::Circle);
fitted_box(doc, r, text).map(|fit| circle.map_or(fit, |old| [fit[0], fit[1].min(old[1]), fit[2], fit[3]]))
};
match fit {
Some(rect) => apply_text_box(doc, r, rect, meta, true)?,
None => draw_appearance(doc, r)?,
}
@@ -197,11 +208,12 @@ fn fitted_box(doc: &Document, r: ObjRef, text: &str) -> Option<[f64; 4]> {
let rect = nums(b"Rect").filter(|v| v.len() == 4 && v.iter().all(|x| x.is_finite()))?;
let rect = [rect[0], rect[1], rect[2], rect[3]];
// A callout re-fits its text box (`/Rect` inset by `/RD`); the leader line keeps its place.
// So does a text box inside an oval or a circle.
let tb = if d.contains(b"CL") {
let rd = nums(b"RD").filter(|v| v.len() == 4 && v.iter().all(|x| *x >= 0.0))?;
[rect[0] + rd[0], rect[1] + rd[1], rect[2] - rd[2], rect[3] - rd[3]]
} else {
rect
text_box_of(&d).unwrap_or(rect)
};
let (x0, top) = (tb[0].min(tb[2]), tb[1].max(tb[3]));
let w = tb[2] - tb[0];
@@ -388,9 +400,134 @@ pub fn set_rect(doc: &mut Document, page: usize, index: usize, rect: [f64; 4], m
if !matches!(subtype.as_str(), "Square" | "Circle" | "FreeText") {
return Err(AnnotError::Invalid(format!("{subtype} comments can't be resized")));
}
// An oval's or a circle's frame is resized: its text box follows inside.
let shape = if subtype == "FreeText" { BorderShape::of(&d) } else { BorderShape::Rectangle };
let rect = match (shape, text_box_of(&d)) {
(BorderShape::Rectangle, _) | (_, None) => rect,
// A circle's text box keeps its proportions and reaches the circle with its longer sides;
// the circle then goes around it again, never closer than its text allows.
(BorderShape::Circle, Some(was)) => {
let (o, w) = (normalize(rect), normalize(was));
let (cx, cy, r) = ((o[0] + o[2]) / 2.0, (o[1] + o[3]) / 2.0, (o[2] - o[0]).min(o[3] - o[1]) / 2.0);
let (wx, wy) = ((w[2] - w[0]) / 2.0, (w[3] - w[1]) / 2.0);
let k = r / wx.max(wy).max(1e-6);
let (ix, iy) = ((wx * k).max(0.5), (wy * k).max(0.5));
[cx - ix, cy - iy, cx + ix, cy + iy]
}
(shape, Some(was)) => shape.inside(rect, border_width_of(&d), was),
};
apply_text_box(doc, r, rect, meta, false)
}
/// A FreeText's text box: `/Rect` inset by `/RD` for a callout or a text box inside an oval or a
/// circle, else `/Rect`. `None` when they can't be read.
pub(crate) fn text_box_of(d: &Dict) -> Option<[f64; 4]> {
let nums = |key: &[u8]| -> Option<Vec<f64>> { d.get(key)?.as_array()?.iter().map(|o| o.as_f64()).collect() };
let rect = normalize(four(&nums(b"Rect")?)?);
if !d.contains(b"CL") && BorderShape::of(d) == BorderShape::Rectangle {
return Some(rect);
}
let rd = appearance::inset(d)?;
let tb = [rect[0] + rd[0], rect[1] + rd[1], rect[2] - rd[2], rect[3] - rd[3]];
(tb[2] - tb[0] >= 1.0 && tb[3] - tb[1] >= 1.0).then_some(tb)
}
/// Give text box or typewriter text `d`, whose text box is `tb`, the border `shape` (`width`
/// wide): `/Rect` around it, `/RD` and `/PCBorderShape`, or for a rectangle `/Rect` the text
/// box itself again.
pub(crate) fn shape_text_box(d: &mut Dict, tb: [f64; 4], shape: BorderShape, width: f64) {
// A circle goes around the text as the box lays it out.
let text = (shape == BorderShape::Circle).then(|| text_extent(d, displayed_sides(d, tb).0 - 2.0 * (2.0 + width.max(0.0))));
let outer = shape.around(tb, width, text);
d.set(b"Rect".to_vec(), num_array(&outer));
if shape == BorderShape::Rectangle {
d.remove(b"PCBorderShape");
d.remove(b"RD");
} else {
d.set(b"PCBorderShape".to_vec(), Object::name(shape.name()));
// A circle touching the text box's sides would put them a rounding error outside it.
let tb = normalize(tb);
let rd = [tb[0] - outer[0], tb[1] - outer[1], outer[2] - tb[2], outer[3] - tb[3]].map(|x| x.max(0.0));
d.set(b"RD".to_vec(), num_array(&rd));
}
}
/// The width of the widest line and the height of the lines of FreeText `d`'s text, as its
/// appearance lays the text out `wrap` wide.
fn text_extent(d: &Dict, wrap: f64) -> [f64; 2] {
let text = d.get(b"Contents").and_then(|o| o.as_string()).map(PdfString::to_text).unwrap_or_default();
let (_, size) = appearance::parse_da(d);
let lines = appearance::wrap(&text, size, wrap.max(1.0));
let widest = lines.iter().map(|l| appearance::text_width(l, size)).fold(0.0, f64::max);
[widest, lines.len().max(1) as f64 * size * 1.2]
}
/// The width and height of FreeText `d`'s text box `tb` as displayed: swapped on a page turned by
/// a quarter, where its text is laid out upright.
fn displayed_sides(d: &Dict, tb: [f64; 4]) -> (f64, f64) {
let tb = normalize(tb);
let (w, h) = (tb[2] - tb[0], tb[3] - tb[1]);
if matches!(appearance::text_turn(d), 90 | 270) { (h, w) } else { (w, h) }
}
/// FreeText `d`'s text box `tb`, with a border `width` wide, made as large as its text wrapped
/// `wrap` wide needs (its widest line and its lines, inside the padding a box sized to its text
/// has) and centred where it was: a circle goes around this, not around a box wider than its text.
fn snug_text_box(d: &Dict, tb: [f64; 4], width: f64, wrap: f64) -> [f64; 4] {
let tb = normalize(tb);
let pad = 2.0 + width.max(0.0);
let [w, h] = text_extent(d, wrap);
let (hx, hy) = (w / 2.0 + pad + 2.0, h / 2.0 + pad + 1.0);
let (hx, hy) = if matches!(appearance::text_turn(d), 90 | 270) { (hy, hx) } else { (hx, hy) };
let (cx, cy) = ((tb[0] + tb[2]) / 2.0, (tb[1] + tb[3]) / 2.0);
[cx - hx, cy - hy, cx + hx, cy + hy]
}
/// Whether FreeText `d` is a circle whose text box is snug around its text ([`snug_text_box`]):
/// it then follows its text as that changes, where one made larger by hand keeps its size.
fn is_snug(d: &Dict) -> bool {
let Some(tb) = text_box_of(d).filter(|_| BorderShape::of(d) == BorderShape::Circle) else { return false };
let width = border_width_of(d);
let s = snug_text_box(d, tb, width, displayed_sides(d, tb).0 - 2.0 * (2.0 + width));
let tb = normalize(tb);
((s[2] - s[0]) - (tb[2] - tb[0])).abs() < 0.5 && ((s[3] - s[1]) - (tb[3] - tb[1])).abs() < 0.5
}
/// The text box of a circle snug around its text for the text and size it has now: its lines up to
/// 300 pt wide, centred where it was.
fn snug_refit(doc: &Document, r: ObjRef) -> Option<[f64; 4]> {
let d = annot_dict(doc, r);
let width = border_width_of(&d);
Some(snug_text_box(&d, text_box_of(&d)?, width, 300.0 - 2.0 * (2.0 + width)))
}
/// Give a text box or typewriter text an oval, a circle or a rectangle for its border, and redraw
/// it. Its text box keeps its place; the oval or circle goes around it. For a circle the text box
/// first shrinks to its text, its lines wrapped as they are.
pub fn set_border_shape(doc: &mut Document, page: usize, index: usize, shape: BorderShape, meta: &Meta) -> Result<(), AnnotError> {
let (_, r) = annot_ref(doc, page, index)?;
unlocked(doc, r)?;
let d = annot_dict(doc, r);
if d.name(b"Subtype") != Some(b"FreeText") || d.contains(b"CL") {
return Err(AnnotError::Invalid("only text boxes and typewriter text have a border shape".into()));
}
let tb = text_box_of(&d).ok_or_else(|| AnnotError::Invalid("the text box can't be read".into()))?;
let tb = match shape {
BorderShape::Circle => {
let width = border_width_of(&d);
snug_text_box(&d, tb, width, displayed_sides(&d, tb).0 - 2.0 * (2.0 + width))
}
_ => tb,
};
refuse_foreign_text(doc, &d)?;
let before = d;
doc.update_dict(r, |d| {
shape_text_box(d, tb, shape, border_width_of(d));
touch(d, meta);
})?;
redraw_or_restore(doc, r, before)
}
/// At most this many points reshape a comment (a drawn polygon has a handful).
const MAX_POINTS: usize = 100_000;
@@ -503,6 +640,8 @@ fn apply_text_box(doc: &mut Document, r: ObjRef, rect: [f64; 4], meta: &Meta, ne
}
None => (rect, None),
};
let (shape, width) = (BorderShape::of(&d), border_width_of(&d));
let free_text = d.name(b"Subtype") == Some(b"FreeText");
doc.update_dict(r, |d| {
d.set(b"Rect".to_vec(), num_array(&outer));
match &cl {
@@ -510,6 +649,8 @@ fn apply_text_box(doc: &mut Document, r: ObjRef, rect: [f64; 4], meta: &Meta, ne
d.set(b"CL".to_vec(), num_array(l));
d.set(b"RD".to_vec(), num_array(&[rect[0] - outer[0], rect[1] - outer[1], outer[2] - rect[2], outer[3] - rect[3]]));
}
// An oval or a circle goes around the text box again.
None if free_text => shape_text_box(d, rect, shape, width),
None => {
d.remove(b"RD");
}
@@ -528,6 +669,8 @@ fn apply_text_box(doc: &mut Document, r: ObjRef, rect: [f64; 4], meta: &Meta, ne
}
}
/// The border width as the appearance draws it: `/BS /W`, else the third number of `/Border`
/// (another app's), else 1.
fn border_width_of(d: &Dict) -> f64 {
appearance::border_width(d)
}
@@ -558,7 +701,8 @@ fn ink_rect_for_width(d: &Dict, old_w: f64, w: f64) -> Option<[f64; 4]> {
Some([inner[0] - base[0] - new_half, inner[1] - base[1] - new_half, inner[2] + base[2] + new_half, inner[3] + base[3] + new_half])
}
/// Change a comment's colour, opacity, line width and/or line endings, and redraw it.
/// Change a comment's colour, opacity (of the whole comment), line width and/or line endings, and
/// redraw it.
pub fn set_style(doc: &mut Document, page: usize, index: usize, change: StyleChange, meta: &Meta) -> Result<(), AnnotError> {
let StyleChange { color, opacity, width, endings } = change;
let (_, r) = annot_ref(doc, page, index)?;
@@ -581,25 +725,34 @@ pub fn set_style(doc: &mut Document, page: usize, index: usize, change: StyleCha
}
refuse_foreign_text(doc, &d)?;
let free_text = subtype == "FreeText";
// Typewriter text has no box of its own: a border goes around its text, so the box grows
// (or shrinks) by the change on every side and the text keeps its place and its lines.
// So does a text box inside an oval or a circle, which goes around it again.
let shape = if free_text { BorderShape::of(&d) } else { BorderShape::Rectangle };
let typewriter = free_text && (d.name(b"IT") == Some(b"FreeTextTypeWriter") || shape != BorderShape::Rectangle) && !d.contains(b"CL");
let regrow = width.filter(|_| typewriter).and_then(|w| {
let by = w.max(0.0) - if d.contains(b"BS") || d.contains(b"Border") { border_width_of(&d) } else { 0.0 };
let rect = grow(text_box_of(&d)?, by);
(by != 0.0 && finite(&rect) && rect[2] - rect[0] >= 1.0 && rect[3] - rect[1] >= 1.0).then_some((rect, w.max(0.0)))
});
let before = d;
doc.update_dict(r, |d| {
if let Some((rect, w)) = regrow {
shape_text_box(d, rect, shape, w);
}
if let Some(c) = color {
if free_text {
let (_, size) = appearance::parse_da(d);
let [r, g, b] = c.map(|x| x.clamp(0.0, 1.0));
d.set(b"DA".to_vec(), PdfString::literal(format!("{} {} {} rg /Helv {} Tf", n(r), n(g), n(b), n(size)).into_bytes()));
d.set(b"DA".to_vec(), free_text_da(c, appearance::da_border(d), size));
rich_text(d);
} else {
d.set(b"C".to_vec(), rgb(c));
}
}
// The whole comment: its fill and text drop opacities of their own.
if let Some(o) = opacity {
let o = o.clamp(0.0, 1.0);
if o < 1.0 {
d.set(b"CA".to_vec(), Object::Real(o));
} else {
d.remove(b"CA");
}
set_part_keys(d, appearance::PartOpacity { line: o, fill: o, text: o });
}
if let Some(w) = width {
let old_w = border_width_of(d);
@@ -629,6 +782,37 @@ pub fn set_style(doc: &mut Document, page: usize, index: usize, change: StyleCha
redraw_or_restore(doc, r, before)
}
/// A FreeText's `/DA`: the text colour, the border's own colour when it has one (the stroking
/// colour, `RG`), Helvetica at `size`.
pub(crate) fn free_text_da(text: Rgb, border: Option<Rgb>, size: f64) -> PdfString {
let [r, g, b] = text.map(|x| x.clamp(0.0, 1.0));
let stroke = border.map(|c| c.map(|x| x.clamp(0.0, 1.0))).map(|[r, g, b]| format!("{} {} {} RG ", n(r), n(g), n(b))).unwrap_or_default();
PdfString::literal(format!("{} {} {} rg {stroke}/Helv {} Tf", n(r), n(g), n(b), n(size)).into_bytes())
}
/// Give a text box, typewriter text or callout a border colour of its own (`None`: the text's
/// again), and redraw it.
pub fn set_border_color(doc: &mut Document, page: usize, index: usize, color: Option<Rgb>, meta: &Meta) -> Result<(), AnnotError> {
let (_, r) = annot_ref(doc, page, index)?;
unlocked(doc, r)?;
let d = annot_dict(doc, r);
if d.name(b"Subtype") != Some(b"FreeText") {
return Err(AnnotError::Invalid("only text boxes, typewriter text and callouts have a border colour".into()));
}
if color.is_some_and(|c| !finite(&c)) {
return Err(AnnotError::Invalid("invalid colour".into()));
}
refuse_foreign_text(doc, &d)?;
let before = d;
doc.update_dict(r, |d| {
let (text, size) = appearance::parse_da(d);
d.set(b"DA".to_vec(), free_text_da(text, color, size));
rich_text(d);
touch(d, meta);
})?;
redraw_or_restore(doc, r, before)
}
/// Change the font size of a text box, typewriter text or callout (`/DA` keeps its colour; the
/// rich-text strings follow) and redraw it. The box grows to hold the wrapped text at the new
/// size and never shrinks, so a box the user drew large stays large. Sizes are 1 to 400 pt.
@@ -643,26 +827,27 @@ pub fn set_font_size(doc: &mut Document, page: usize, index: usize, size: f64, m
return Err(AnnotError::Invalid("font size must be 1\u{2013}400 pt".into()));
}
refuse_foreign_text(doc, &d)?;
let snug = is_snug(&d);
let before = d;
doc.update_dict(r, |d| {
let (col, _) = appearance::parse_da(d);
d.set(b"DA".to_vec(), PdfString::literal(format!("{} {} {} rg /Helv {} Tf", n(col[0]), n(col[1]), n(col[2]), n(size)).into_bytes()));
d.set(b"DA".to_vec(), free_text_da(col, appearance::da_border(d), size));
rich_text(d);
touch(d, meta);
})?;
let text = text_value(doc, &annot_dict(doc, r), b"Contents").unwrap_or_default();
let result = match fitted_box(doc, r, &text) {
Some(fit) => {
let result = match (snug, if snug { snug_refit(doc, r) } else { fitted_box(doc, r, &text) }) {
// A circle snug around its text stays snug around it at the new size.
(true, Some(fit)) => apply_text_box(doc, r, fit, meta, false),
(false, Some(fit)) => {
let cur = annot_dict(doc, r);
let nums: Vec<f64> =
cur.get(b"Rect").and_then(|o| o.as_array()).map(|a| a.iter().filter_map(|x| x.as_f64()).collect()).unwrap_or_default();
// Keep the top and sides; only the bottom edge moves, and only downwards.
match four(&nums).map(normalize) {
match text_box_of(&cur) {
Some(old) if !cur.contains(b"CL") => apply_text_box(doc, r, [fit[0], fit[1].min(old[1]), fit[2], fit[3]], meta, false),
_ => apply_text_box(doc, r, fit, meta, false),
}
}
None => redraw_or_restore(doc, r, before.clone()),
(_, None) => redraw_or_restore(doc, r, before.clone()),
};
if result.is_err() {
doc.set(r, Object::Dict(before));
@@ -707,6 +892,63 @@ pub fn set_fill(doc: &mut Document, page: usize, index: usize, fill: Option<Rgb>
redraw_or_restore(doc, r, before)
}
/// Draw a rectangle's, oval's, polygon's or text box's fill, or a text box's text, at an
/// opacity of its own (0–1), its line or border keeping the comment's (`/CA`), and redraw it.
/// `None` leaves that part as it is drawn now. The fill goes in `/ca` (PDF 2.0's opacity for
/// what is filled), the text in `/PCTextOpacity`, each only where it differs from the one it
/// would otherwise follow.
pub fn set_part_opacity(doc: &mut Document, page: usize, index: usize, fill: Option<f64>, text: Option<f64>, meta: &Meta) -> Result<(), AnnotError> {
let (_, r) = annot_ref(doc, page, index)?;
unlocked(doc, r)?;
let d = annot_dict(doc, r);
let subtype = String::from_utf8_lossy(d.name(b"Subtype").unwrap_or_default()).into_owned();
if fill.is_some() && !fillable(&subtype) {
return Err(AnnotError::Invalid("only rectangles, ovals, polygons and text boxes have a fill of their own".into()));
}
if text.is_some() && subtype != "FreeText" {
return Err(AnnotError::Invalid("only text boxes, typewriter text and callouts have text of their own".into()));
}
if fill.into_iter().chain(text).any(|o| !o.is_finite()) {
return Err(AnnotError::Invalid("invalid opacity".into()));
}
// Check before changing anything: a stale appearance would contradict the new opacity.
if appearance::build(&d).is_none() {
return Err(AnnotError::Unsupported(subtype));
}
refuse_foreign_text(doc, &d)?;
let before = d;
doc.update_dict(r, |d| {
let mut parts = appearance::part_opacity(d);
if let Some(f) = fill {
parts.fill = f.clamp(0.0, 1.0);
}
if let Some(t) = text {
parts.text = t.clamp(0.0, 1.0);
}
set_part_keys(d, parts);
touch(d, meta);
})?;
redraw_or_restore(doc, r, before)
}
/// Write a comment's parts' opacities (see [`appearance::PartOpacity`]): the line's as `/CA`, the
/// fill's as `/ca` and the text's as `/PCTextOpacity`, each only where it differs from the one
/// it would otherwise follow (the fill the line, the text the fill), so a comment drawn at one
/// opacity keeps just `/CA`.
pub(crate) fn set_part_keys(d: &mut Dict, parts: appearance::PartOpacity) {
let appearance::PartOpacity { line, fill, text } = parts;
let mut put = |key: &[u8], value: f64, unless: f64| {
if value == unless {
d.remove(key);
} else {
d.set(key.to_vec(), Object::Real(value));
}
};
put(b"CA", line, 1.0);
put(b"ca", fill, line);
put(b"PCTextOpacity", text, fill);
}
/// Grow a line or polyline's rectangle so a non-`None` ending is not clipped. Idempotent.
fn ensure_ending_room(d: &mut Dict) {
let subtype = d.name(b"Subtype").unwrap_or_default();
crates/annot/src/tests.rs
+508 -3
View File
@@ -2,6 +2,7 @@ use std::sync::Arc;
use pdfcraft_cos::{Document, ObjRef, Object, SaveOptions, write_incremental};
use super::modify::text_box_of;
use super::*;
/// Two pages. Page 1 has an inline Square annotation in a direct /Annots array; page 2 has an
@@ -40,7 +41,7 @@ fn new(page: usize, shape: Shape) -> NewAnnotation {
NewAnnotation {
page,
shape,
style: Style { color: [0.0, 0.4, 1.0], opacity: 1.0, width: 2.0, fill: None },
style: Style { color: [0.0, 0.4, 1.0], opacity: 1.0, width: 2.0, fill: None, ..Style::default() },
contents: "hello".into(),
author: "Ada".into(),
}
@@ -356,6 +357,296 @@ fn restyle_and_resize_redraw_the_appearance() {
assert!(matches!(set_rect(&mut doc, 0, l, [0.0, 0.0, 5.0, 5.0], &meta("")), Err(AnnotError::Invalid(_))));
}
/// Typed text or a text box takes an oval or a circle for its border: the text box keeps its
/// place and the shape goes around it (`/Rect` grows, `/RD` holds the text box, `/PCBorderShape`
/// the shape); back to a rectangle, it is the text box it was. A resized frame takes its text box
/// along, a thicker border leaves it where it is.
#[test]
fn text_boxes_take_an_oval_or_a_circle_around_their_text() {
let close = |a: &[f64], b: &[f64]| a.len() == b.len() && a.iter().zip(b).all(|(x, y)| (x - y).abs() < 0.01);
let nums = |d: &Dict, k: &[u8]| -> Vec<f64> {
d.get(k).and_then(|o| o.as_array()).map(|a| a.iter().filter_map(|x| x.as_f64()).collect()).unwrap_or_default()
};
// Where the text starts: its first line's text matrix.
let text_at = |doc: &Document, d: &Dict| ap_content(doc, d).lines().find(|l| l.ends_with(" Tm") || l.contains(" Tm ")).map(str::to_owned);
let mut doc = fixture();
let mut tw = new(0, Shape::Typewriter { rect: [100.0, 100.0, 200.0, 120.0], font_size: 10.0 });
tw.style.width = 1.0;
let t = add_annotation(&mut doc, &tw, &meta("t")).unwrap();
// In an oval or a circle the text is centred in its text box (left-aligned in a box): one
// line of "hello" at 10 pt in the 100 × 20 text box centred on (150, 110), 3 pt from its top.
let centred = format!("1 0 0 1 {} {} Tm (hello) Tj", n(150.0 - appearance::text_width("hello", 10.0) / 2.0), n(107.0));
assert!(text_at(&doc, &list(&doc, 0)[t]).is_some_and(|l| l.starts_with("1 0 0 1 103 ")), "a box: left-aligned");
set_border_shape(&mut doc, 0, t, BorderShape::Oval, &meta("")).unwrap();
let doc2 = reopen(&doc);
let d = &list(&doc2, 0)[t];
// Around the 100 × 20 text box, centred: √2 its size, two points more for the 1 pt border.
let (ax, ay) = (50.0 * std::f64::consts::SQRT_2 + 2.0, 10.0 * std::f64::consts::SQRT_2 + 2.0);
assert!(close(&rect(d), &[150.0 - ax, 110.0 - ay, 150.0 + ax, 110.0 + ay]), "{:?}", rect(d));
assert!(close(&nums(d, b"RD"), &[ax - 50.0, ay - 10.0, ax - 50.0, ay - 10.0]), "{:?}", nums(d, b"RD"));
assert_eq!(d.name(b"PCBorderShape"), Some(&b"Oval"[..]));
let ap = ap_content(&doc2, d);
assert!(ap.contains(" c\n") && ap.contains("S\n") && !ap.contains("re S"), "an oval, not a box: {ap}");
assert_eq!(text_at(&doc2, d).as_deref(), Some(centred.as_str()), "the text keeps its place");
assert_eq!(props(&doc2, 0, t).unwrap().border_shape, Some(BorderShape::Oval));
// A circle around the text: the text box shrinks to "hello" with its padding, centred where it
// was, and the circle touches its sides (the text a point clear of the border inside).
set_border_shape(&mut doc, 0, t, BorderShape::Circle, &meta("")).unwrap();
let d = &list(&doc, 0)[t];
let hw = appearance::text_width("hello", 10.0) / 2.0;
let r = (hw + 5.0).max(hw.hypot(6.0) + 2.0);
assert!(close(&rect(d), &[150.0 - r, 110.0 - r, 150.0 + r, 110.0 + r]), "{:?}", rect(d));
assert!(close(&nums(d, b"RD"), &[r - hw - 5.0, r - 10.0, r - hw - 5.0, r - 10.0]), "{:?}", nums(d, b"RD"));
assert_eq!(text_at(&doc, d).as_deref(), Some(centred.as_str()));
// A thicker border: the typed text's box grows by the change on every side, so its text keeps
// its place inside the thicker line (as with a rectangle), and the circle goes around that.
set_style(&mut doc, 0, t, StyleChange::new(None, None, Some(3.0), None), &meta("")).unwrap();
let d = &list(&doc, 0)[t];
assert_eq!(text_at(&doc, d).as_deref(), Some(centred.as_str()));
let r3 = (hw + 7.0).max(hw.hypot(6.0) + 4.0);
assert!(close(&rect(d), &[150.0 - r3, 110.0 - r3, 150.0 + r3, 110.0 + r3]), "{:?}", rect(d));
// Its frame resized: the text box follows inside, keeping its proportions, its longer sides
// reaching the circle.
set_rect(&mut doc, 0, t, [150.0 - 2.0 * r3, 110.0 - 2.0 * r3, 150.0 + 2.0 * r3, 110.0 + 2.0 * r3], &meta("")).unwrap();
let d = &list(&doc, 0)[t];
let rd = nums(d, b"RD");
let tb = [rect(d)[0] + rd[0], rect(d)[1] + rd[1], rect(d)[2] - rd[2], rect(d)[3] - rd[3]];
assert!(((tb[2] - tb[0]) / (tb[3] - tb[1]) - (hw + 7.0) / 12.0).abs() < 0.01, "the text box keeps its proportions: {tb:?}");
assert!(close(&rect(d), &[150.0 - 2.0 * r3, 110.0 - 2.0 * r3, 150.0 + 2.0 * r3, 110.0 + 2.0 * r3]), "{:?}", rect(d));
// A rectangle again: the text box it has, nothing around it.
set_border_shape(&mut doc, 0, t, BorderShape::Rectangle, &meta("")).unwrap();
let d = &list(&doc, 0)[t];
assert!(close(&rect(d), &tb), "{:?} vs {tb:?}", rect(d));
assert!(!d.contains(b"RD") && !d.contains(b"PCBorderShape"));
assert!(ap_content(&doc, d).contains("re S"));
// A filled text box made with an oval: the background takes the shape too.
let mut tb = new(0, Shape::TextBox { rect: [300.0, 500.0, 400.0, 540.0], font_size: 12.0 });
(tb.style.fill, tb.style.width, tb.style.border_shape) = (Some([1.0, 1.0, 0.6]), 1.0, BorderShape::Oval);
let b = add_annotation(&mut doc, &tb, &meta("b")).unwrap();
let d = &list(&doc, 0)[b];
assert_eq!(d.name(b"PCBorderShape"), Some(&b"Oval"[..]));
let ap = ap_content(&doc, d);
assert!(ap.contains("1 1 0.6 rg") && ap.contains(" c\nh\nf\n") && !ap.contains(" re f"), "{ap}");
// A callout's border stays a rectangle.
let c = add_annotation(
&mut doc,
&new(
0,
Shape::Callout {
rect: [300.0, 300.0, 400.0, 340.0],
knee: [280.0, 320.0],
point: [250.0, 280.0],
ending: LineEnding::OpenArrow,
font_size: 10.0,
},
),
&meta("c"),
)
.unwrap();
assert!(matches!(set_border_shape(&mut doc, 0, c, BorderShape::Oval, &meta("")), Err(AnnotError::Invalid(_))));
assert_eq!(props(&doc, 0, c).unwrap().border_shape, None);
assert_eq!(BorderShape::parse("CIRCLE"), Some(BorderShape::Circle));
}
/// A circle goes snug around short text, as a number on a drawing is circled: the text box is no
/// wider than its text needs and the circle touches its sides, where one through the box's corners
/// stood well clear. It follows the text and its size as they change; one made larger by hand
/// keeps its size.
#[test]
fn a_circle_goes_snug_around_its_text_and_follows_it() {
let close = |a: (f64, f64), b: (f64, f64)| (a.0 - b.0).abs() < 0.01 && (a.1 - b.1).abs() < 0.01;
let mut doc = fixture();
// "3" at 10 pt with a 1 pt border, in the box Fill & Sign sizes to it: the digit with 2 pt of
// padding and 4 pt to spare across, a line and 6 pt down, the border around that.
let w3 = appearance::text_width("3", 10.0);
let mut typed = new(0, Shape::Typewriter { rect: [100.0, 680.0, 110.0 + w3, 700.0], font_size: 10.0 });
(typed.contents, typed.style.width, typed.style.border_shape) = ("3".into(), 1.0, BorderShape::Circle);
let t = add_annotation(&mut doc, &typed, &meta("t")).unwrap();
let frame = |doc: &Document| {
let r = rect(&list(doc, 0)[t]);
((r[2] - r[0], r[3] - r[1]), ((r[0] + r[2]) / 2.0, (r[1] + r[3]) / 2.0))
};
let centre = (105.0 + w3 / 2.0, 690.0);
// As tall as the box, 20 pt across: through the corners it was 29.
assert!(close(frame(&doc).0, (20.0, 20.0)) && close(frame(&doc).1, centre), "{:?}", frame(&doc));
// The text box Fill & Sign gives a box to type into, 40 pt wide: a circle goes snug around its
// text there too.
let mut wide = typed.clone();
wide.shape = Shape::Typewriter { rect: [300.0, 680.0, 342.0, 700.0], font_size: 10.0 };
wide.style.border_shape = BorderShape::Rectangle;
let b = add_annotation(&mut doc, &wide, &meta("b")).unwrap();
set_border_shape(&mut doc, 0, b, BorderShape::Circle, &meta("")).unwrap();
let r = rect(&list(&doc, 0)[b]);
assert!(close((r[2] - r[0], r[3] - r[1]), (20.0, 20.0)), "{r:?}");
// More text: the circle grows with it, about the same centre; less, and it shrinks back.
set_contents(&mut doc, 0, t, "123", &meta("")).unwrap();
let w = appearance::text_width("123", 10.0) / 2.0;
let d = 2.0 * (w + 5.0).max(10.0).max(w.hypot(6.0) + 2.0);
assert!(close(frame(&doc).0, (d, d)) && close(frame(&doc).1, centre), "{:?}", frame(&doc));
set_contents(&mut doc, 0, t, "3", &meta("")).unwrap();
assert!(close(frame(&doc).0, (20.0, 20.0)) && close(frame(&doc).1, centre), "{:?}", frame(&doc));
// A larger font: a larger circle, still snug.
set_font_size(&mut doc, 0, t, 20.0, &meta("")).unwrap();
let w = appearance::text_width("3", 20.0) / 2.0;
let d = 2.0 * (w + 5.0).max(16.0).max(w.hypot(12.0) + 2.0);
assert!(close(frame(&doc).0, (d, d)) && close(frame(&doc).1, centre), "{:?}", frame(&doc));
set_font_size(&mut doc, 0, t, 10.0, &meta("")).unwrap();
assert!(close(frame(&doc).0, (20.0, 20.0)), "{:?}", frame(&doc));
// Made larger by hand, it keeps its size as its text changes.
set_rect(&mut doc, 0, t, [centre.0 - 20.0, centre.1 - 20.0, centre.0 + 20.0, centre.1 + 20.0], &meta("")).unwrap();
assert!(close(frame(&doc).0, (40.0, 40.0)), "{:?}", frame(&doc));
set_contents(&mut doc, 0, t, "4", &meta("")).unwrap();
assert!(close(frame(&doc).0, (40.0, 40.0)) && close(frame(&doc).1, centre), "{:?}", frame(&doc));
}
/// A circle around a line wider than it is tall touches its text box's sides: `/RD` is 0 there,
/// not a rounding error below it, and a hair below 0 still reads. (A text box that couldn't be
/// read stopped following its text, and its text wrapped in a narrower one.) The circle keeps its
/// centre as its text changes.
#[test]
fn a_circle_touching_its_text_box_keeps_following_its_text() {
let mut doc = fixture();
// Sized to its text: 2 pt of padding and half a point to spare across, a line down.
let (w, h) = (appearance::text_width("Testtext", 12.0) + 4.5, 12.0 * 1.11 + 4.0);
let mut b = new(0, Shape::TextBox { rect: [210.79, 700.0 - h, 210.79 + w, 700.0], font_size: 12.0 });
(b.contents, b.style.width) = ("Testtext".into(), 0.0);
let t = add_annotation(&mut doc, &b, &meta("t")).unwrap();
// Circle, then a 1 pt border, which the circle grows by.
set_border_shape(&mut doc, 0, t, BorderShape::Circle, &meta("")).unwrap();
set_style(&mut doc, 0, t, StyleChange::new(None, None, Some(1.0), None), &meta("")).unwrap();
let lines = |doc: &Document| ap_content(doc, &list(doc, 0)[t]).matches(" Tj").count();
let frame = |doc: &Document| {
let r = rect(&list(doc, 0)[t]);
((r[2] - r[0], r[3] - r[1]), ((r[0] + r[2]) / 2.0, (r[1] + r[3]) / 2.0))
};
let rd = |doc: &Document| -> Vec<f64> { list(doc, 0)[t].get(b"RD").unwrap().as_array().unwrap().iter().map(|o| o.as_f64().unwrap()).collect() };
assert!(rd(&doc).iter().all(|x| *x >= 0.0) && rd(&doc)[0] < 0.01, "touching the sides: {:?}", rd(&doc));
assert_eq!(lines(&doc), 1, "Testtext on one line");
let (wide, centre) = frame(&doc);
set_contents(&mut doc, 0, t, "KRS", &meta("")).unwrap();
let (small, at) = frame(&doc);
assert!(small.0 < wide.0 - 10.0 && (small.0 - small.1).abs() < 0.01, "a smaller circle: {small:?} from {wide:?}");
assert!((at.0 - centre.0).abs() < 0.01 && (at.1 - centre.1).abs() < 0.01, "about its centre: {at:?} {centre:?}");
assert_eq!(lines(&doc), 1);
// A hair below 0, as rounding left it: still read, so the circle follows its text again.
let (_, r) = annot_ref(&mut doc, 0, t).unwrap();
let mut was = rd(&doc);
was[0] = -2.8e-14;
doc.update_dict(r, |d| d.set(b"RD".to_vec(), num_array(&was))).unwrap();
assert!(text_box_of(&list(&doc, 0)[t]).is_some());
set_contents(&mut doc, 0, t, "Testtext", &meta("")).unwrap();
assert!((frame(&doc).0.0 - wide.0).abs() < 0.01 && lines(&doc) == 1, "{:?}", frame(&doc));
assert!(rd(&doc).iter().all(|x| *x >= 0.0), "{:?}", rd(&doc));
}
/// On a page turned a quarter, typed text is laid out as the page is shown (`/PCTextRotate`), and a
/// circle made with it goes around the text as shown: as large as around the same text on an
/// upright page. (Made before the turn was set, its text wrapped at the box's displayed height and
/// the circle went around several lines.)
#[test]
fn a_circle_made_on_a_turned_page_is_sized_as_on_an_upright_one() {
let w = appearance::text_width("12345", 10.0);
let typed = |rect: [f64; 4]| {
let mut t = new(0, Shape::Typewriter { rect, font_size: 10.0 });
(t.contents, t.style.width, t.style.border_shape) = ("12345".into(), 1.0, BorderShape::Circle);
t
};
let size = |doc: &Document, i: usize| {
let r = rect(&list(doc, 0)[i]);
(r[2] - r[0], r[3] - r[1])
};
let mut upright = fixture();
let u = add_annotation(&mut upright, &typed([100.0, 680.0, 110.0 + w, 700.0]), &meta("u")).unwrap();
for turn in [90, 270] {
let mut doc = fixture();
let page = page_refs(&doc).unwrap()[0];
doc.update_dict(page, |d| d.set(b"Rotate".to_vec(), Object::Int(turn))).unwrap();
// The same box as shown: its sides swapped in user space.
let t = add_annotation(&mut doc, &typed([100.0, 680.0, 120.0, 690.0 + w]), &meta("t")).unwrap();
assert_eq!(list(&doc, 0)[t].get(b"PCTextRotate").and_then(Object::as_int), Some(turn));
let (turned, straight) = (size(&doc, t), size(&upright, u));
assert!((turned.0 - straight.0).abs() < 0.01 && (turned.1 - straight.1).abs() < 0.01, "{turn}°: {turned:?}, upright {straight:?}");
}
}
/// A rectangle's fill, and a text box's border, background and text, each drawn at an opacity of
/// its own: the line's in `/CA`, the fill's in `/ca` (PDF 2.0), the text's in `/PCTextOpacity`, and
/// in the appearance `/GS0` (strokes at the line's, fills at the fill's) and `/GS1` for the text.
#[test]
fn parts_are_drawn_at_opacities_of_their_own() {
let num = |d: &Dict, k: &[u8]| d.get(k).and_then(|o| o.as_f64());
let gstate = |doc: &Document, d: &Dict, name: &[u8]| -> Option<(f64, f64)> {
let n = d.get(b"AP").map(|a| doc.resolve(a)).and_then(|a| a.as_dict().and_then(|a| a.reference(b"N")))?;
let obj = doc.get(n);
let Object::Stream(s) = &*obj else { return None };
let res = s.dict.get(b"Resources").map(|r| doc.resolve(r))?;
let gs = res.as_dict()?.get(b"ExtGState").map(|g| doc.resolve(g))?;
let gs = gs.as_dict()?.get(name).map(|g| doc.resolve(g))?;
let gs = gs.as_dict()?;
Some((gs.get(b"CA")?.as_f64()?, gs.get(b"ca")?.as_f64()?))
};
let mut doc = fixture();
let r = add_annotation(&mut doc, &new(0, Shape::Rectangle { rect: [10.0, 10.0, 110.0, 60.0] }), &meta("r")).unwrap();
set_fill(&mut doc, 0, r, Some([1.0, 1.0, 0.0]), &meta("")).unwrap();
set_part_opacity(&mut doc, 0, r, Some(0.3), None, &meta("")).unwrap();
let doc2 = reopen(&doc);
let d = &list(&doc2, 0)[r];
assert_eq!((num(d, b"CA"), num(d, b"ca")), (None, Some(0.3)), "an opaque line, a fill of its own");
assert_eq!(gstate(&doc2, d, b"GS0"), Some((1.0, 0.3)));
let ap = ap_content(&doc2, d);
assert!(ap.starts_with("/GS0 gs") && ap.contains("\nB\n"), "filled and stroked in one go: {ap}");
let p = props(&doc2, 0, r).unwrap();
assert_eq!((p.opacity, p.fill_opacity, p.text_opacity), (1.0, Some(0.3), None));
// The whole comment's opacity (StyleAnnotation, the quick bar): the fill goes along again.
set_style(&mut doc, 0, r, StyleChange::new(None, Some(0.6), None, None), &meta("")).unwrap();
let d = &list(&doc, 0)[r];
assert_eq!((num(d, b"CA"), num(d, b"ca")), (Some(0.6), None));
assert_eq!(props(&doc, 0, r).unwrap().fill_opacity, Some(0.6));
// A text box with its border opaque, its background at 20 % and its text opaque, as made.
let mut tb = new(0, Shape::TextBox { rect: [200.0, 600.0, 400.0, 650.0], font_size: 12.0 });
tb.style.fill = Some([1.0, 1.0, 0.6]);
tb.style.width = 1.0;
(tb.style.fill_opacity, tb.style.text_opacity) = (Some(0.2), Some(1.0));
let t = add_annotation(&mut doc, &tb, &meta("t")).unwrap();
let doc2 = reopen(&doc);
let d = &list(&doc2, 0)[t];
assert_eq!((num(d, b"CA"), num(d, b"ca"), num(d, b"PCTextOpacity")), (None, Some(0.2), Some(1.0)));
assert_eq!((gstate(&doc2, d, b"GS0"), gstate(&doc2, d, b"GS1")), (Some((1.0, 0.2)), Some((1.0, 1.0))));
let ap = ap_content(&doc2, d);
let (gs1, text) = (ap.find("/GS1 gs\n").expect("the text's own state, an operator of its own"), ap.find("BT").expect("text"));
assert!(ap.find(" re f").is_some_and(|fill| fill < gs1) && gs1 < text, "background at 20 %, then the text at its own: {ap}");
// A new fill opacity leaves the text as it is drawn.
set_part_opacity(&mut doc, 0, t, Some(0.5), None, &meta("")).unwrap();
let d = &list(&doc, 0)[t];
assert_eq!((num(d, b"ca"), num(d, b"PCTextOpacity")), (Some(0.5), Some(1.0)));
let p = props(&doc, 0, t).unwrap();
assert_eq!((p.opacity, p.fill_opacity, p.text_opacity), (1.0, Some(0.5), Some(1.0)));
// All at one opacity again: only /CA stays.
set_part_opacity(&mut doc, 0, t, Some(1.0), Some(1.0), &meta("")).unwrap();
let d = &list(&doc, 0)[t];
assert_eq!((num(d, b"CA"), num(d, b"ca"), num(d, b"PCTextOpacity")), (None, None, None));
// Only the parts a comment has, and finite values; nothing changes when refused.
let l = add_annotation(
&mut doc,
&new(0, Shape::Line { from: [0.0, 0.0], to: [10.0, 10.0], start: LineEnding::None, end: LineEnding::None }),
&meta("l"),
)
.unwrap();
assert!(matches!(set_part_opacity(&mut doc, 0, l, Some(0.5), None, &meta("")), Err(AnnotError::Invalid(_))), "a line has no fill");
assert!(matches!(set_part_opacity(&mut doc, 0, r, None, Some(0.5), &meta("")), Err(AnnotError::Invalid(_))), "a rectangle has no text");
let before = list(&doc, 0)[r].clone();
assert!(matches!(set_part_opacity(&mut doc, 0, r, Some(f64::NAN), None, &meta("")), Err(AnnotError::Invalid(_))));
assert_eq!(list(&doc, 0)[r], before);
// Another app's comment with /CA alone is drawn at it throughout, as before.
assert_eq!(
appearance::part_opacity(&Dict::from_iter([(b"CA".to_vec(), Object::Real(0.4))])),
appearance::PartOpacity { line: 0.4, fill: 0.4, text: 0.4 }
);
}
#[test]
fn shapes_can_be_filled_and_unfilled() {
let mut doc = fixture();
@@ -591,6 +882,85 @@ fn fill_and_sign_items_are_drawn() {
assert!(set_style(&mut doc, 1, 0, StyleChange::new(Some([0.0; 3]), None, None, None), &meta("")).is_err());
}
/// A border chosen for typewriter text goes around it in the text's colour. Changing it grows or
/// shrinks the box on every side, so the text stays where it was typed; a text box keeps its box.
#[test]
fn a_typewriter_border_goes_around_its_text_which_stays_put() {
let mut doc = fixture();
let shape = Shape::Typewriter { rect: [100.0, 700.0, 200.0, 716.0], font_size: 11.0 };
let style = Style { width: 1.0, ..Style::default_for(&shape) };
let t = add_annotation(&mut doc, &NewAnnotation { page: 1, shape, style, contents: "Ada".into(), author: "Ada".into() }, &meta("x")).unwrap();
// The text starts 2 pt inside the border: 100 + 1 + 2 across, 716 - 1 - 2 - 0.9 em down.
let at = "1 0 0 1 103 703.1 Tm";
let d = &list(&doc, 1)[t];
let ap = ap_content(&doc, d);
assert!(ap.contains("1 w\n") && ap.contains(" re S"), "a 1 pt border: {ap}");
assert!(ap.contains(at), "{ap}");
for (width, expected) in [(2.0, [99.0, 699.0, 201.0, 717.0]), (0.0, [101.0, 701.0, 199.0, 715.0])] {
set_style(&mut doc, 1, t, StyleChange::new(None, None, Some(width), None), &meta("")).unwrap();
let d = &list(&doc, 1)[t];
assert_eq!(rect(d), expected, "{width} pt border");
let ap = ap_content(&doc, d);
assert!(ap.contains(at), "{width} pt: the text keeps its place: {ap}");
assert_eq!(ap.contains(" re S"), width > 0.0, "{width} pt: {ap}");
}
let shape = Shape::TextBox { rect: [100.0, 600.0, 200.0, 650.0], font_size: 11.0 };
let b = add_annotation(&mut doc, &new(1, shape), &meta("b")).unwrap();
set_style(&mut doc, 1, b, StyleChange::new(None, None, Some(3.0), None), &meta("")).unwrap();
assert_eq!(rect(&list(&doc, 1)[b]), [100.0, 600.0, 200.0, 650.0], "a text box keeps the box it was drawn with");
// Another app's typed text with its width in /Border (no /BS): 4 pt, so 2 pt shrinks the box
// by 2 on every side, as it is drawn.
let shape = Shape::Typewriter { rect: [100.0, 500.0, 200.0, 530.0], font_size: 11.0 };
let o = add_annotation(
&mut doc,
&NewAnnotation { page: 1, style: Style::default_for(&shape), shape: shape.clone(), contents: "Bo".into(), author: "Ada".into() },
&meta("o"),
)
.unwrap();
let (_, r) = annot_ref(&mut doc, 1, o).unwrap();
doc.update_dict(r, |d| {
d.remove(b"BS");
d.set(b"Border".to_vec(), num_array(&[0.0, 0.0, 4.0]));
})
.unwrap();
set_style(&mut doc, 1, o, StyleChange::new(None, None, Some(2.0), None), &meta("")).unwrap();
assert_eq!(rect(&list(&doc, 1)[o]), [102.0, 502.0, 198.0, 528.0], "the imported 4 pt border is read from /Border");
// A border colour that isn't a colour is refused before anything is written.
let mut bad = NewAnnotation { page: 1, style: Style::default_for(&shape), shape, contents: "x".into(), author: "Ada".into() };
bad.style.border_color = Some([f64::NAN, 0.0, 0.0]);
let before = list(&doc, 1).len();
assert!(matches!(add_annotation(&mut doc, &bad, &meta("n")), Err(AnnotError::Invalid(_))));
assert_eq!(list(&doc, 1).len(), before);
}
/// Text keeps its colour and the border gets its own: the stroking colour in `/DA`, which a new
/// text colour or font size keeps, and which `None` gives back to the text's.
#[test]
fn a_text_border_can_have_its_own_colour() {
let mut doc = fixture();
let shape = Shape::Typewriter { rect: [100.0, 700.0, 200.0, 716.0], font_size: 11.0 };
let style = Style { color: [0.0, 0.5, 0.0], width: 1.0, border_color: Some([1.0, 0.0, 0.0]), ..Style::default_for(&shape) };
let t = add_annotation(&mut doc, &NewAnnotation { page: 1, shape, style, contents: "Ada".into(), author: "Ada".into() }, &meta("x")).unwrap();
let da = |doc: &Document| text(&list(doc, 1)[t], b"DA");
assert_eq!(da(&doc), "0 0.5 0 rg 1 0 0 RG /Helv 11 Tf");
let ap = ap_content(&doc, &list(&doc, 1)[t]);
assert!(ap.contains("1 0 0 RG\n1 w"), "the border in red: {ap}");
assert!(ap.contains("0 0.5 0 rg"), "the text in green: {ap}");
assert_eq!(props(&doc, 1, t).unwrap().border_color, Some([1.0, 0.0, 0.0]));
set_style(&mut doc, 1, t, StyleChange::new(Some([0.0, 0.0, 1.0]), None, None, None), &meta("")).unwrap();
assert_eq!(da(&doc), "0 0 1 rg 1 0 0 RG /Helv 11 Tf", "a new text colour keeps the border's");
set_font_size(&mut doc, 1, t, 14.0, &meta("")).unwrap();
assert_eq!(da(&doc), "0 0 1 rg 1 0 0 RG /Helv 14 Tf", "so does a new size");
set_border_color(&mut doc, 1, t, Some([0.5, 0.5, 0.5]), &meta("")).unwrap();
assert!(ap_content(&doc, &list(&doc, 1)[t]).contains("0.5 0.5 0.5 RG\n1 w"));
set_border_color(&mut doc, 1, t, None, &meta("")).unwrap();
assert_eq!(da(&doc), "0 0 1 rg /Helv 14 Tf");
assert!(ap_content(&doc, &list(&doc, 1)[t]).contains("0 0 1 RG\n1 w"), "the text's colour again");
assert_eq!(props(&doc, 1, t).unwrap().border_color, None);
let m = add_annotation(&mut doc, &new(0, Shape::Rectangle { rect: [10.0, 10.0, 50.0, 50.0] }), &meta("m")).unwrap();
assert!(set_border_color(&mut doc, 0, m, Some([1.0, 0.0, 0.0]), &meta("")).is_err(), "a rectangle's colour is its border");
}
#[test]
fn author_subject_and_note_icon_change() {
let mut doc = fixture();
@@ -873,8 +1243,8 @@ fn cloud_ys(path: &str) -> Vec<f64> {
fn replace_text_groups_a_strikeout_and_a_caret() {
let mut doc = fixture();
let quads = [[100.0, 720.0, 200.0, 720.0, 100.0, 706.0, 200.0, 706.0]];
let red = Style { color: [0.89, 0.13, 0.13], opacity: 1.0, width: 1.0, fill: None };
let blue = Style { color: [0.0, 0.47, 0.84], opacity: 1.0, width: 1.0, fill: None };
let red = Style { color: [0.89, 0.13, 0.13], opacity: 1.0, width: 1.0, fill: None, ..Style::default() };
let blue = Style { color: [0.0, 0.47, 0.84], opacity: 1.0, width: 1.0, fill: None, ..Style::default() };
let i = add_text_replacement(
&mut doc,
TextReplacement {
@@ -2251,6 +2621,141 @@ fn freehand_highlight_is_compatible_ink_with_marker_blending() {
}
}
/// A 3D annotation is PDF 1.6 and a PRC model 1.7 (ISO 32000-2 §13.6): an older document's
/// catalog says so, and a newer one is left as it is.
#[test]
fn a_3d_model_raises_an_older_document_to_its_version() {
let catalog_version = |doc: &Document| {
let root = doc.root().unwrap();
doc.get(root).as_dict().unwrap().name(b"Version").map(<[u8]>::to_vec)
};
let add = |doc: &mut Document, model: &[u8]| {
let new = NewModel3d { page: 0, rect: [100.0, 100.0, 300.0, 300.0], model, activation: Activation3d::default(), poster: None, contents: "" };
add_3d_model(doc, &new, &meta("m")).unwrap();
};
let mut doc = fixture();
add(&mut doc, b"U3D\0rest");
assert_eq!(catalog_version(&reopen(&doc)), None, "a 1.7 file needs nothing for U3D");
for (model, version) in [(&b"U3D\0rest"[..], &b"1.6"[..]), (b"PRC rest", b"1.7")] {
let mut doc = fixture();
doc.set_version("1.4");
add(&mut doc, model);
assert_eq!(catalog_version(&reopen(&doc)).as_deref(), Some(version));
}
let mut doc = fixture();
doc.set_version("2.0");
add(&mut doc, b"PRC rest");
assert_eq!(catalog_version(&reopen(&doc)), None, "a 2.0 file takes PRC as it is");
}
/// Before 2.0, PRC is an Adobe extension to PDF 1.7 (§7.12): a PRC model declares it, keeping
/// other extensions and a later ADBE level; U3D and a 2.0 file need nothing.
#[test]
fn a_prc_model_declares_the_adobe_extension_before_pdf_2() {
let adbe = |doc: &Document| {
let root = doc.root().unwrap();
let ext = doc.get(root).as_dict().unwrap().get(b"Extensions").map(|e| doc.resolve(e).as_dict().unwrap().clone());
ext.map(|e| {
let adbe = e.get(b"ADBE").map(|a| a.as_dict().unwrap().clone());
(e.get(b"XYZ").is_some(), adbe.map(|a| (a.name(b"BaseVersion").map(<[u8]>::to_vec), a.get(b"ExtensionLevel").and_then(Object::as_int))))
})
};
let add = |doc: &mut Document, model: &[u8]| {
let new = NewModel3d { page: 0, rect: [100.0, 100.0, 300.0, 300.0], model, activation: Activation3d::default(), poster: None, contents: "" };
add_3d_model(doc, &new, &meta("m")).unwrap();
};
let mut doc = fixture();
add(&mut doc, b"PRC rest");
assert_eq!(adbe(&reopen(&doc)), Some((false, Some((Some(b"1.7".to_vec()), Some(1))))));
let mut doc = fixture();
add(&mut doc, b"U3Drest");
assert_eq!(adbe(&reopen(&doc)), None, "U3D is in ISO 32000-1");
let mut doc = fixture();
doc.set_version("2.0");
add(&mut doc, b"PRC rest");
assert_eq!(adbe(&reopen(&doc)), None, "PRC is in ISO 32000-2");
let mut doc = fixture();
let root = doc.root().unwrap();
let mut level3 = Dict::new();
level3.set(b"BaseVersion".to_vec(), Object::name("1.7"));
level3.set(b"ExtensionLevel".to_vec(), Object::Int(3));
let mut ext = Dict::new();
ext.set(b"ADBE".to_vec(), Object::Dict(level3));
ext.set(b"XYZ".to_vec(), Object::Dict(Dict::new()));
doc.update_dict(root, |c| c.set(b"Extensions".to_vec(), Object::Dict(ext.clone()))).unwrap();
add(&mut doc, b"PRC rest");
assert_eq!(adbe(&reopen(&doc)), Some((true, Some((Some(b"1.7".to_vec()), Some(3))))), "a later level is kept");
ext.remove(b"ADBE");
let mut doc = fixture();
doc.update_dict(root, |c| c.set(b"Extensions".to_vec(), Object::Dict(ext))).unwrap();
add(&mut doc, b"PRC rest");
assert_eq!(adbe(&reopen(&doc)), Some((true, Some((Some(b"1.7".to_vec()), Some(1))))), "other extensions are kept");
let mut doc = fixture();
let mut ext = Dict::new();
ext.set(b"ADBE".to_vec(), Object::Array(vec![Object::Dict(Dict::new())]));
let ext = doc.add(Object::Dict(ext));
doc.update_dict(root, |c| c.set(b"Extensions".to_vec(), Object::Ref(ext))).unwrap();
add(&mut doc, b"PRC rest");
let doc = reopen(&doc);
let ext = doc.resolve(doc.get(doc.root().unwrap()).as_dict().unwrap().get(b"Extensions").unwrap());
assert!(ext.as_dict().unwrap().get(b"ADBE").unwrap().as_array().is_some(), "an array of ADBE extensions is left alone");
}
#[test]
fn a_picture_is_placed_by_its_box_and_a_form_needs_a_positive_size() {
let at = [10.0, 20.0, 110.0, 70.0];
assert_eq!(appearance::picture_place(at, true, None).as_deref(), Some("100 0 0 50 10 20 cm"), "an image fills its unit square");
assert_eq!(appearance::picture_place(at, false, Some((200.0, 25.0))).as_deref(), Some("0.5 0 0 2 10 20 cm"));
for size in [None, Some((0.0, 25.0)), Some((200.0, -1.0))] {
assert_eq!(appearance::picture_place(at, false, size), None, "{size:?}");
}
}
/// The poster's box must lie in the rectangle, and a form poster needs a size to scale by.
#[test]
fn a_3d_poster_outside_its_rectangle_or_without_a_size_is_refused() {
let picture = |at: [f64; 4], size: (f64, f64)| PosterPicture { xobject: ObjRef { num: 6, generation: 0 }, image: false, size, at };
for (poster, why) in [
(picture([50.0, 100.0, 250.0, 300.0], (10.0, 10.0)), "starts left of the rectangle"),
(picture([100.0, 100.0, 300.0, 301.0], (10.0, 10.0)), "ends above it"),
(picture([100.0, 100.0, 300.0, 300.0], (0.0, 10.0)), "has no width"),
(picture([100.0 - 2e-7, 150.0, 100.0 - 1e-7, 200.0], (10.0, 10.0)), "lies outside within the rounding slack"),
] {
let mut doc = fixture();
let new = NewModel3d {
page: 0,
rect: [100.0, 100.0, 300.0, 300.0],
model: b"PRC rest",
activation: Activation3d::default(),
poster: Some(poster),
contents: "",
};
assert!(matches!(add_3d_model(&mut doc, &new, &meta("m")), Err(AnnotError::Invalid(_))), "{why}");
assert_eq!(list(&doc, 0).len(), 1, "{why}: nothing added");
}
}
/// A poster box fitted to the rectangle in floating point ends an ulp outside it: a 7×3 picture
/// centred in [0 100 3.6 110.8] spans -2.2e-16 to 3.6000000000000005. It is taken, inside the
/// rectangle.
#[test]
fn a_3d_poster_fitted_to_its_rectangle_with_rounding_is_taken() {
let mut doc = fixture();
let at = [-2.220446049250313e-16, 104.62857142857143, 3.6000000000000005, 106.17142857142858];
let poster = PosterPicture { xobject: ObjRef { num: 6, generation: 0 }, image: false, size: (7.0, 3.0), at };
let new = NewModel3d {
page: 0,
rect: [0.0, 100.0, 3.6, 110.8],
model: b"PRC rest",
activation: Activation3d::default(),
poster: Some(poster),
contents: "",
};
assert_eq!(add_3d_model(&mut doc, &new, &meta("m")), Ok(1));
}
#[test]
fn indirect_typewriter_names_keep_outer_whitespace_in_summaries() {
let mut doc = fixture();
crates/annot/src/threed.rs
+292
View File
@@ -0,0 +1,292 @@
//! 3D annotations (ISO 32000-2 §13.6.2): a U3D or PRC model shown in a rectangle on a page,
//! with a poster (`/AP /N`) for printing, for readers without 3D, and until it is activated.
use super::create::{finite, normalize, num_array};
use super::lookup::{annots, page_ref, set_annots};
use super::types::{FLAG_PRINT, Meta};
use super::{AnnotError, Dict, Document, ObjRef, Object, PdfString, Stream, appearance, n};
/// The largest model accepted, in bytes: well past real U3D/PRC files, and bounded because the
/// model is held and compressed in memory.
pub const MAX_MODEL_BYTES: usize = 256 << 20;
/// What the poster says when there is no picture.
const PLACEHOLDER_LABEL: &str = "3D model";
/// A 3D stream's data format (§13.6.3): ECMA-363 U3D or ISO 14739-1 PRC.
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum ModelFormat {
U3d,
Prc,
}
impl ModelFormat {
/// The format a model file's first bytes declare: U3D files start `U3D\0` (ECMA-363 file
/// header block type), PRC files start `PRC`.
///
/// # Errors
///
/// [`AnnotError::NotAModel`] for anything else.
pub fn detect(bytes: &[u8]) -> Result<Self, AnnotError> {
if bytes.starts_with(b"U3D\0") {
Ok(Self::U3d)
} else if bytes.starts_with(b"PRC") {
Ok(Self::Prc)
} else {
Err(AnnotError::NotAModel)
}
}
/// The 3D stream's `/Subtype`.
pub fn subtype(self) -> &'static str {
match self {
Self::U3d => "U3D",
Self::Prc => "PRC",
}
}
/// The PDF version a document holding it needs (§13.6): a 3D annotation is 1.6, a PRC stream
/// 1.7 under Adobe's extension, declared by [`add_3d_model`] (PDF 2.0 takes PRC as standard).
pub fn version(self) -> &'static str {
match self {
Self::U3d => "1.6",
Self::Prc => "1.7",
}
}
}
/// When a reader turns the poster into the live model (§13.6.2, the activation dictionary's
/// `/A`). The standard's default is [`Self::Click`].
#[derive(Clone, Copy, Debug, Default, PartialEq, Eq)]
pub enum Activation {
/// `/XA`: when the user clicks it; it prints as its poster.
#[default]
Click,
/// `/PO`: when the page opens.
PageOpen,
/// `/PV`: when the page becomes visible.
PageVisible,
}
impl Activation {
/// Every activation, with the word a tool argument names it by.
pub const ALL: [(Self, &'static str); 3] = [(Self::Click, "click"), (Self::PageOpen, "page_open"), (Self::PageVisible, "page_visible")];
/// The `/A` name.
pub fn name(self) -> &'static str {
match self {
Self::Click => "XA",
Self::PageOpen => "PO",
Self::PageVisible => "PV",
}
}
}
/// A picture already in the document (an image XObject, or a form XObject whose `/Matrix` maps it
/// to `size` points) drawn into `at`, a box inside the annotation's rectangle.
#[derive(Clone, Copy, Debug, PartialEq)]
pub struct PosterPicture {
pub xobject: ObjRef,
pub image: bool,
pub size: (f64, f64),
pub at: [f64; 4],
}
/// A 3D model to add on `page` (0-based) in `rect` (user space).
#[derive(Clone, Debug, PartialEq)]
pub struct NewModel3d<'a> {
pub page: usize,
pub rect: [f64; 4],
pub model: &'a [u8],
pub activation: Activation,
/// The poster's picture; `None` draws a labelled frame.
pub poster: Option<PosterPicture>,
/// `/Contents`: the annotation's alternate description (§12.5.2).
pub contents: &'a str,
}
/// Check a model before anything is added: its size and its format.
///
/// # Errors
///
/// [`AnnotError::ModelTooLarge`] past [`MAX_MODEL_BYTES`], else as [`ModelFormat::detect`].
pub fn check_model(model: &[u8]) -> Result<ModelFormat, AnnotError> {
check_model_size(model, MAX_MODEL_BYTES)?;
ModelFormat::detect(model)
}
fn check_model_size(model: &[u8], max: usize) -> Result<(), AnnotError> {
if model.len() > max {
return Err(AnnotError::ModelTooLarge { size: model.len(), max });
}
Ok(())
}
/// Add a 3D annotation: the model in a Flate-compressed 3D stream referenced from `/3DD` (so the
/// annotation has its own instance, §13.6.2), `/3DA` stating only `/A`, and the poster as `/AP /N`.
/// No view (`/3DV`, `/VA`) is written: readers open on the model's own default view. A document
/// older than [`ModelFormat::version`] is raised to it. Returns its index in the page's `/Annots`.
///
/// # Errors
///
/// As [`check_model`]; [`AnnotError::NoSuchPage`]; [`AnnotError::Invalid`] for a rectangle with no
/// area or a poster box outside it.
pub fn add_3d_model(doc: &mut Document, new: &NewModel3d<'_>, meta: &Meta) -> Result<usize, AnnotError> {
let format = check_model(new.model)?;
let page = page_ref(doc, new.page)?;
let rect = normalize(new.rect);
if !finite(&new.rect) || rect[2] - rect[0] < 1.0 || rect[3] - rect[1] < 1.0 {
return Err(AnnotError::Invalid("invalid rectangle for the 3D model (each side at least 1 pt)".into()));
}
let poster = match &new.poster {
Some(p) => picture_poster(rect, p)?,
None => placeholder_poster(rect),
};
let mut stream = Dict::new();
stream.set(b"Type".to_vec(), Object::name("3D"));
stream.set(b"Subtype".to_vec(), Object::name(format.subtype()));
let stream = doc.add(Object::Stream(Stream::flate(stream, new.model)));
let poster = doc.add(Object::Stream(poster));
let d = annotation_dict(new, rect, page, stream, poster, meta);
let r = doc.add(Object::Dict(d));
let mut list = annots(doc, page);
let index = list.len();
list.push(Object::Ref(r));
set_annots(doc, page, list)?;
at_least_version(doc, format.version())?;
if format == ModelFormat::Prc && !at_least(doc, "2.0") {
declare_prc_extension(doc)?;
}
Ok(index)
}
/// Whether the header or the catalog `/Version` says `min` or later (§7.7.2).
fn at_least(doc: &Document, min: &str) -> bool {
let reached = |v: &[u8]| v.len() >= 3 && v >= min.as_bytes();
let catalog = doc.root().and_then(|root| doc.get(root).as_dict().and_then(|c| c.get(b"Version").cloned()));
let catalog = catalog.and_then(|v| doc.resolve(&v).as_name().map(<[u8]>::to_vec));
reached(doc.version().as_bytes()) || catalog.as_deref().is_some_and(reached)
}
/// PRC came before PDF 2.0 as an Adobe extension to 1.7: such a file declares
/// `/Extensions /ADBE` with ExtensionLevel 1 or later (§7.12). Other extensions and an ADBE entry
/// that already has a level are kept; one that is not a dictionary (an array of them) is left alone.
fn declare_prc_extension(doc: &mut Document) -> Result<(), AnnotError> {
let Some(root) = doc.root() else { return Ok(()) };
let catalog = doc.get(root);
// `None` when present but not a dictionary; an empty one when absent.
let dict_or_new = |o: Option<&Object>| match o.map(|o| doc.resolve(o)) {
None => Some(Dict::new()),
Some(o) => o.as_dict().cloned(),
};
let Some(mut extensions) = dict_or_new(catalog.as_dict().and_then(|c| c.get(b"Extensions"))) else { return Ok(()) };
let Some(mut adbe) = dict_or_new(extensions.get(b"ADBE")) else { return Ok(()) };
if adbe.get(b"ExtensionLevel").map(|l| doc.resolve(l)).and_then(|l| l.as_int()).is_some_and(|l| l >= 1) {
return Ok(());
}
adbe.set(b"BaseVersion".to_vec(), Object::name("1.7"));
adbe.set(b"ExtensionLevel".to_vec(), Object::Int(1));
extensions.set(b"ADBE".to_vec(), Object::Dict(adbe));
doc.update_dict(root, |c| c.set(b"Extensions".to_vec(), Object::Dict(extensions)))?;
Ok(())
}
/// A document whose header and catalog `/Version` both say less than `min` gets `/Version min` in
/// its catalog (§7.7.2: the later of the two applies).
fn at_least_version(doc: &mut Document, min: &str) -> Result<(), AnnotError> {
let Some(root) = doc.root() else { return Ok(()) };
if !at_least(doc, min) {
doc.update_dict(root, |c| c.set(b"Version".to_vec(), Object::name(min)))?;
}
Ok(())
}
fn annotation_dict(new: &NewModel3d<'_>, rect: [f64; 4], page: ObjRef, stream: ObjRef, poster: ObjRef, meta: &Meta) -> Dict {
let mut d = Dict::new();
d.set(b"Type".to_vec(), Object::name("Annot"));
d.set(b"Subtype".to_vec(), Object::name("3D"));
d.set(b"Rect".to_vec(), num_array(&rect));
d.set(b"P".to_vec(), Object::Ref(page));
d.set(b"F".to_vec(), Object::Int(FLAG_PRINT));
if !new.contents.is_empty() {
d.set(b"Contents".to_vec(), PdfString::text(new.contents));
}
if !meta.id.is_empty() {
d.set(b"NM".to_vec(), PdfString::text(&meta.id));
}
if let Some(date) = &meta.date {
d.set(b"M".to_vec(), PdfString::literal(date.as_bytes().to_vec()));
}
d.set(b"3DD".to_vec(), Object::Ref(stream));
let mut activation = Dict::new();
activation.set(b"A".to_vec(), Object::name(new.activation.name()));
d.set(b"3DA".to_vec(), Object::Dict(activation));
let mut ap = Dict::new();
ap.set(b"N".to_vec(), Object::Ref(poster));
d.set(b"AP".to_vec(), Object::Dict(ap));
d
}
/// The poster drawing `p` into its box, which must lie within `rect`.
fn picture_poster(rect: [f64; 4], p: &PosterPicture) -> Result<Stream, AnnotError> {
let at = normalize(p.at);
// A box fitted to `rect` in floating point can end a few ulps outside it: allow that, then clamp.
let slack = rect.iter().fold(1.0_f64, |m, v| m.max(v.abs())) * 1e-9;
let near = finite(&p.at) && at[0] >= rect[0] - slack && at[1] >= rect[1] - slack && at[2] <= rect[2] + slack && at[3] <= rect[3] + slack;
let at = [at[0].max(rect[0]), at[1].max(rect[1]), at[2].min(rect[2]), at[3].min(rect[3])];
if !(near && at[2] > at[0] && at[3] > at[1]) {
return Err(AnnotError::Invalid("the poster must fit inside the 3D model's rectangle".into()));
}
let place = appearance::picture_place(at, p.image, Some(p.size)).ok_or_else(|| AnnotError::Invalid("the poster picture has no size".into()))?;
let mut xobjects = Dict::new();
xobjects.set(b"Pic".to_vec(), Object::Ref(p.xobject));
let mut resources = Dict::new();
resources.set(b"XObject".to_vec(), Object::Dict(xobjects));
Ok(appearance::form(rect, format!("q {place} /Pic Do Q\n").as_bytes(), resources))
}
/// A light frame with "3D model" centred in it, shrunk to fit (left out when even 2 pt won't).
fn placeholder_poster(rect: [f64; 4]) -> Stream {
let [x0, y0, x1, y1] = rect;
let (w, h) = (x1 - x0, y1 - y0);
let mut c =
format!("0.94 g {} {} {} {} re f\n0.45 G 1 w {} {} {} {} re S\n", n(x0), n(y0), n(w), n(h), n(x0 + 0.5), n(y0 + 0.5), n(w - 1.0), n(h - 1.0));
let mut resources = Dict::new();
let unit = appearance::text_width(PLACEHOLDER_LABEL, 1.0);
let size = if unit > 0.0 { (w * 0.8 / unit).min(h * 0.4).min(14.0) } else { 0.0 };
if size >= 2.0 {
let (tx, ty) = (x0 + (w - unit * size) / 2.0, y0 + (h - size * 0.7) / 2.0);
c.push_str(&format!("BT 0.3 g /Helv {} Tf {} {} Td ({PLACEHOLDER_LABEL}) Tj ET\n", n(size), n(tx), n(ty)));
resources.set(b"Font".to_vec(), appearance::helvetica_resource());
}
appearance::form(rect, c.as_bytes(), resources)
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn model_size_is_capped() {
assert_eq!(check_model_size(&[0; 8], 8), Ok(()));
assert_eq!(check_model_size(&[0; 9], 8), Err(AnnotError::ModelTooLarge { size: 9, max: 8 }));
}
#[test]
fn the_format_is_read_from_the_header() {
assert_eq!(ModelFormat::detect(b"U3D\0rest"), Ok(ModelFormat::U3d));
assert_eq!(ModelFormat::detect(b"PRC8800"), Ok(ModelFormat::Prc));
for junk in [&b""[..], b"U3D", b"U3DX", b"PR", b"%PDF-1.7", b"prc"] {
assert_eq!(ModelFormat::detect(junk), Err(AnnotError::NotAModel), "{junk:?}");
}
}
#[test]
fn the_placeholder_label_shrinks_to_fit_and_is_dropped_when_it_cannot() {
let wide = String::from_utf8(placeholder_poster([0.0, 0.0, 300.0, 200.0]).decoded().unwrap().to_vec()).unwrap();
assert!(wide.contains("/Helv 14 Tf") && wide.contains("(3D model) Tj"), "{wide}");
let tiny = placeholder_poster([0.0, 0.0, 4.0, 4.0]);
assert!(!String::from_utf8(tiny.decoded().unwrap().to_vec()).unwrap().contains("Tj"));
assert!(tiny.dict.get(b"Resources").and_then(Object::as_dict).is_none_or(|r| r.get(b"Font").is_none()));
}
}
crates/annot/src/types.rs
+132 -5
View File
@@ -1,5 +1,5 @@
use super::create::normalize;
use super::{ObjRef, appearance};
use super::{Dict, ObjRef, appearance};
pub type Rgb = [f64; 3];
@@ -124,6 +124,28 @@ impl FillMark {
let (w, h) = if matches!(rotation.rem_euclid(360), 90 | 270) { (h, w) } else { (w, h) };
[at[0] - w / 2.0, at[1] - h / 2.0, at[0] + w / 2.0, at[1] + h / 2.0]
}
/// The mark whose `/Name` this is.
pub fn from_name(n: &[u8]) -> Option<Self> {
[FillMark::Check, FillMark::Cross, FillMark::Dot, FillMark::Line].into_iter().find(|m| m.name().as_bytes() == n)
}
/// The pen strokes the mark is drawn with in `rect` (user space): what its appearance strokes,
/// and what a copy of it shows while it is being placed. The dot is a filled circle (radius
/// 0.3 × the shorter side, centred) and has none.
pub fn strokes(self, rect: [f64; 4]) -> Vec<Vec<[f64; 2]>> {
let [x0, y0, x1, y1] = rect;
let (w, h) = (x1 - x0, y1 - y0);
match self {
FillMark::Check => vec![vec![[x0 + w * 0.15, y0 + h * 0.5], [x0 + w * 0.4, y0 + h * 0.2], [x0 + w * 0.88, y0 + h * 0.85]]],
FillMark::Cross => vec![
vec![[x0 + w * 0.18, y0 + h * 0.18], [x1 - w * 0.18, y1 - h * 0.18]],
vec![[x0 + w * 0.18, y1 - h * 0.18], [x1 - w * 0.18, y0 + h * 0.18]],
],
FillMark::Dot => Vec::new(),
FillMark::Line => vec![vec![[x0, y0 + h / 2.0], [x1, y0 + h / 2.0]]],
}
}
}
/// The stamps of Acrobat's stamp palette (drawn in PdfCraft's own style).
@@ -566,6 +588,16 @@ pub struct Style {
pub width: f64,
/// Interior fill for rectangles and ovals (`/IC`), background for text boxes.
pub fill: Option<Rgb>,
/// A text box's, typewriter text's or callout's border colour where it differs from the
/// text's (`None`: the text colour), the stroking colour in its `/DA`.
pub border_color: Option<Rgb>,
/// A text box's or typewriter text's border shape (a callout's is always a rectangle).
pub border_shape: BorderShape,
/// The fill's own opacity, 0–1 (`/ca`), where it differs from `opacity` (which is then the
/// line's or border's); `None`: as `opacity`.
pub fill_opacity: Option<f64>,
/// A text box's text's own opacity, 0–1 (`/PCTextOpacity`); `None`: as the fill's.
pub text_opacity: Option<f64>,
}
/// The style properties changed by one annotation edit. `None` leaves a property unchanged.
@@ -597,7 +629,102 @@ pub struct TextReplacement<'a> {
impl Default for Style {
fn default() -> Self {
Self { color: [1.0, 0.82, 0.0], opacity: 1.0, width: 1.0, fill: None }
Self {
color: [1.0, 0.82, 0.0],
opacity: 1.0,
width: 1.0,
fill: None,
fill_opacity: None,
text_opacity: None,
border_color: None,
border_shape: BorderShape::Rectangle,
}
}
}
/// The shape of a text box's or typewriter text's border (PdfCraft's `/PCBorderShape`, absent for
/// a rectangle). An oval or a circle goes around the text box, which keeps its place: `/Rect`
/// grows to hold it, centred on the text box, and `/RD` says where the text box is inside
/// (§12.5.6.6), as for a callout. Other apps that redraw the comment draw a rectangle.
#[derive(Clone, Copy, Debug, Default, PartialEq, Eq)]
pub enum BorderShape {
#[default]
Rectangle,
Oval,
Circle,
}
impl BorderShape {
pub const ALL: [BorderShape; 3] = [BorderShape::Rectangle, BorderShape::Oval, BorderShape::Circle];
pub fn name(self) -> &'static str {
match self {
BorderShape::Rectangle => "Rectangle",
BorderShape::Oval => "Oval",
BorderShape::Circle => "Circle",
}
}
/// `rectangle`, `oval` or `circle`, in any case.
pub fn parse(s: &str) -> Option<Self> {
Self::ALL.into_iter().find(|b| b.name().eq_ignore_ascii_case(s))
}
/// The border shape of FreeText `d`: always a rectangle for a callout.
pub fn of(d: &Dict) -> Self {
if d.contains(b"CL") {
return BorderShape::Rectangle;
}
match d.name(b"PCBorderShape") {
Some(b"Oval") => BorderShape::Oval,
Some(b"Circle") => BorderShape::Circle,
_ => BorderShape::Rectangle,
}
}
/// The rectangle this shape, with a border `width` wide, takes around text box `tb`, centred on
/// it: an oval through the text box's corners (`√2` its size), a point clear of the border; a
/// circle around the text's lines (`text`: the widest line's width and the lines' height, when
/// known), a point clear of the border and touching the text box's longer sides at least
/// (its corners, padding only, may lie outside), else through the box's corners; the text box
/// itself for a rectangle.
pub fn around(self, tb: [f64; 4], width: f64, text: Option<[f64; 2]>) -> [f64; 4] {
let tb = normalize(tb);
let (cx, cy, hx, hy) = ((tb[0] + tb[2]) / 2.0, (tb[1] + tb[3]) / 2.0, (tb[2] - tb[0]) / 2.0, (tb[3] - tb[1]) / 2.0);
let gap = width.max(0.0) + 1.0;
let (ax, ay) = match self {
BorderShape::Rectangle => return tb,
BorderShape::Oval => (hx * std::f64::consts::SQRT_2 + gap, hy * std::f64::consts::SQRT_2 + gap),
BorderShape::Circle => {
let r = match text {
Some([w, h]) => ((w / 2.0).hypot(h / 2.0) + gap).max(hx).max(hy),
None => hx.hypot(hy) + gap,
};
(r, r)
}
};
[cx - ax, cy - ay, cx + ax, cy + ay]
}
/// The text box inside `outer` for this shape when none is recorded: the inverse of
/// [`Self::around`] for an oval, a box through a circle's corners for a circle, keeping the
/// proportions of `was` (the one it had).
pub(crate) fn inside(self, outer: [f64; 4], width: f64, was: [f64; 4]) -> [f64; 4] {
let o = normalize(outer);
let (cx, cy, hx, hy) = ((o[0] + o[2]) / 2.0, (o[1] + o[3]) / 2.0, (o[2] - o[0]) / 2.0, (o[3] - o[1]) / 2.0);
let gap = width.max(0.0) + 1.0;
let (ix, iy) = match self {
BorderShape::Rectangle => return o,
BorderShape::Oval => ((hx - gap) / std::f64::consts::SQRT_2, (hy - gap) / std::f64::consts::SQRT_2),
BorderShape::Circle => {
let w = normalize(was);
let (wx, wy) = ((w[2] - w[0]) / 2.0, (w[3] - w[1]) / 2.0);
let k = (hx.min(hy) - gap) / wx.hypot(wy).max(1e-6);
(wx * k, wy * k)
}
};
let (ix, iy) = (ix.max(0.5), iy.max(0.5));
[cx - ix, cy - iy, cx + ix, cy + iy]
}
}
@@ -615,7 +742,7 @@ impl Style {
}
Shape::Callout { .. } => ([0.0, 0.0, 0.0], 1.0),
Shape::Caret { .. } | Shape::Attachment { .. } => ([0.0, 0.47, 0.84], 1.0),
Shape::FreehandHighlight { .. } => return Self { color: [1.0, 0.94, 0.0], opacity: 0.35, width: 14.0, fill: None },
Shape::FreehandHighlight { .. } => return Self { color: [1.0, 0.94, 0.0], opacity: 0.35, width: 14.0, ..Self::default() },
Shape::Ink { .. } => ([0.0, 0.4, 0.87], 2.0),
Shape::TextBox { .. } | Shape::Typewriter { .. } => ([0.0, 0.0, 0.0], 0.0),
Shape::Mark { .. } => ([0.0, 0.0, 0.0], 1.5),
@@ -624,9 +751,9 @@ impl Style {
Shape::CustomStamp { .. } => ([0.0, 0.0, 0.0], 0.0),
Shape::TypedSignature { .. } => ([0.0, 0.0, 0.0], 0.0),
// Red outline while marked; a black box once applied.
Shape::Redact { .. } => return Self { color: [0.89, 0.13, 0.13], opacity: 1.0, width: 1.0, fill: Some([0.0, 0.0, 0.0]) },
Shape::Redact { .. } => return Self { color: [0.89, 0.13, 0.13], fill: Some([0.0, 0.0, 0.0]), ..Self::default() },
};
Self { color, opacity: 1.0, width, fill: None }
Self { color, width, ..Self::default() }
}
}
crates/automation/Cargo.toml
+2
View File
@@ -14,6 +14,8 @@ pdfcraft-cos = { workspace = true }
pdfcraft-render = { workspace = true }
pdfcraft-organize = { workspace = true }
pdfcraft-platform = { workspace = true }
pdfcraft-prc = { workspace = true }
pdfcraft-prc-render = { workspace = true }
serde_json = { workspace = true }
sha2 = "0.11"
tempfile = "3.27"
crates/automation/README.md
+9
View File
@@ -250,6 +250,15 @@ choices and values over 4096 UTF-8 bytes return an error.
Example: `form_set_props doc=1 field=agree export_value=Accepted widget=1`.
## Rendering a 3D model
`model_render` draws a page's 3D model (a 3D annotation's PRC stream, directly or through a 3D
reference) to a PNG, without the viewer. `model` picks the page's Nth 3D model (default 1);
`width` and `height` are 1 to 4096 (default 512); `direction` is the view direction from the eye
(default `[-1, 1, -1]`, with Z up unless the view looks along Z); `perspective` defaults to true.
The model is framed to fill the image and shaded by a headlight on white. U3D models are refused
with an error, and the annotation's own views, lighting and render modes are not applied yet.
## Viewer navigation history
A running PdfCraft instance launched with `--control FILE` exposes its per-tab history through
crates/automation/src/bookmark_io.rs
+209
View File
@@ -0,0 +1,209 @@
//! Bookmark trees in and out of documents: `bookmark_export`, `bookmark_import`, `bookmark_copy`.
//!
//! The portable form is JSON: `{"bookmarks": [node, …]}`, where a node is `title`, `page`
//! (1-based, or `null` for one that goes nowhere), `view` (`{"type", "values"}`: a destination
//! type and its values in PDF user space, as the file stores them), `color`, `bold`, `italic`,
//! `open` and `children`. What export writes, import takes back unchanged. A file also carries
//! `"truncated": true` when the export left out bookmarks nested past the listing limit.
use crate::{Args, Automation, Result, ToolError, failed};
use pdfcraft_engine::{DocId, Edit, OutlineNode};
use serde_json::{Value, json};
/// Deepest nesting a bookmark tree may have (the listing reaches 33 levels).
const MAX_DEPTH: usize = 33;
/// The keys a node may have: the ones [`node_json`] writes.
const NODE_KEYS: [&str; 8] = ["title", "page", "view", "color", "bold", "italic", "open", "children"];
impl Automation {
pub(crate) fn bookmark_export(&self, a: &Args) -> Result<Value> {
let id = self.doc(a)?.id;
let path = a.opt_path("path")?.unwrap_or_default();
let tree = self.outline_of(id, &path)?;
let mut out = json!({ "bookmarks": tree.nodes.iter().map(node_json).collect::<Vec<_>>(), "truncated": tree.truncated });
if let Some(file) = a.opt_str("out")? {
let target = self.resolve(file, true)?;
let mut saved = json!({ "bookmarks": out["bookmarks"] });
if tree.truncated {
saved["truncated"] = json!(true);
}
let text = serde_json::to_vec_pretty(&saved).map_err(failed)?;
crate::write_atomic(&target, &text)?;
out["path"] = json!(target);
}
Ok(out)
}
pub(crate) fn bookmark_import(&mut self, a: &Args) -> Result<Value> {
let (nodes, truncated) = match (a.get("bookmarks"), a.opt_str("file")?) {
(Some(list), None) => (nodes_of(list)?, false),
(None, Some(file)) => {
let path = self.resolve(file, false)?;
let value: Value = serde_json::from_slice(&crate::read_input(&path)?).map_err(|e| failed(format!("not a bookmark file: {e}")))?;
let list = value.get("bookmarks").ok_or_else(|| failed("not a bookmark file: it has no \"bookmarks\" list"))?;
(nodes_of(list)?, value.get("truncated") == Some(&Value::Bool(true)))
}
_ => return Err(ToolError::InvalidArgs("give either bookmarks or file".into())),
};
let mut out = self.write_tree(a, nodes)?;
out["truncated"] = json!(truncated);
Ok(out)
}
pub(crate) fn bookmark_copy(&mut self, a: &Args) -> Result<Value> {
let from = a.int("from")?;
let from = DocId(u64::try_from(from).map_err(|_| ToolError::InvalidArgs("from must be positive".into()))?);
let path = a.opt_path("path")?.unwrap_or_default();
let tree = self.outline_of(from, &path)?;
let mut out = self.write_tree(a, tree.nodes)?;
out["truncated"] = json!(tree.truncated);
Ok(out)
}
fn outline_of(&self, id: DocId, path: &[usize]) -> Result<pdfcraft_engine::OutlineTree> {
match self.session.outline_tree(id, path) {
// The engine names the path 0-based; the caller gave it 1-based.
Some(tree) => tree.map_err(|e| match e {
pdfcraft_organize::OutlineError::NoSuchBookmark(_) => failed(format!("there is no bookmark at {:?}", one_based(path))),
e => failed(e),
}),
None => Err(failed(format!("document {} is not open, or its bookmarks can't be read for editing", id.0))),
}
}
/// Write `nodes` into `doc` under `parent` at `position` (or in place of the outline with
/// `replace`), and say which of them went nowhere because the document lacks their page.
fn write_tree(&mut self, a: &Args, bookmarks: Vec<OutlineNode>) -> Result<Value> {
let parent = a.opt_path("parent")?.unwrap_or_default();
let index = a.opt_int("position")?.map_or(usize::MAX, |p| usize::try_from(p.max(1) - 1).unwrap_or(usize::MAX));
let replace = a.opt_bool("replace")?.unwrap_or(false);
let count = self.doc(a)?.info.pages.len();
let unmapped = pdfcraft_engine::outline_unmapped(&bookmarks, count);
let mut out = self.apply(a, Edit::ImportBookmarks { parent, index, bookmarks, replace })?;
out["unmapped"] = json!(unmapped.iter().map(|p| one_based(p)).collect::<Vec<_>>());
Ok(out)
}
}
fn one_based(path: &[usize]) -> Vec<usize> {
path.iter().map(|i| i.saturating_add(1)).collect()
}
/// A node as JSON, leaving out what is at its default (no view, black, regular, no children).
fn node_json(n: &OutlineNode) -> Value {
let mut v = json!({ "title": n.title, "page": n.page.map(|p| p.saturating_add(1)), "open": n.open });
if let Some((kind, values)) = &n.view {
v["view"] = json!({ "type": String::from_utf8_lossy(kind), "values": values });
}
if let Some(c) = n.color {
v["color"] = json!(c);
}
if n.bold {
v["bold"] = json!(true);
}
if n.italic {
v["italic"] = json!(true);
}
if !n.children.is_empty() {
v["children"] = json!(n.children.iter().map(node_json).collect::<Vec<_>>());
}
v
}
fn nodes_of(list: &Value) -> Result<Vec<OutlineNode>> {
let mut count = 0;
list_of(list, &[], &mut count)
}
/// The nodes of one level; `at` is the 1-based path of their parent, for messages.
fn list_of(list: &Value, at: &[usize], count: &mut usize) -> Result<Vec<OutlineNode>> {
let items = list.as_array().ok_or_else(|| bad(at, "children must be a list of bookmarks"))?;
if at.len() >= MAX_DEPTH && !items.is_empty() {
return Err(bad(at, &format!("bookmarks may nest at most {MAX_DEPTH} levels deep")));
}
*count = count.saturating_add(items.len());
if *count > pdfcraft_engine::MAX_IMPORT_BOOKMARKS {
return Err(ToolError::InvalidArgs(format!("at most {} bookmarks can be imported at once", pdfcraft_engine::MAX_IMPORT_BOOKMARKS)));
}
let mut out = Vec::with_capacity(items.len());
for (i, item) in items.iter().enumerate() {
let mut path = at.to_vec();
path.push(i.saturating_add(1));
out.push(node_of(item, &path, count)?);
}
Ok(out)
}
/// One node and, through [`list_of`], its children (recursion bounded by `MAX_DEPTH`).
fn node_of(v: &Value, at: &[usize], count: &mut usize) -> Result<OutlineNode> {
let o = v.as_object().ok_or_else(|| bad(at, "must be an object"))?;
if let Some(k) = o.keys().find(|k| !NODE_KEYS.contains(&k.as_str())) {
return Err(bad(at, &format!("unknown key {k:?} (a bookmark has {})", NODE_KEYS.join(", "))));
}
let field = |k: &str| o.get(k).filter(|x| !x.is_null());
// An empty title is kept: files have them, and export writes them as "".
let title = field("title").and_then(Value::as_str).ok_or_else(|| bad(at, "needs a title"))?.to_string();
let page = match field("page") {
None => None,
Some(p) => Some(p.as_u64().filter(|p| *p >= 1).and_then(|p| usize::try_from(p - 1).ok()).ok_or_else(|| bad(at, "page must be 1 or more"))?),
};
let flag = |k: &str| field(k).map_or(Ok(None), |x| x.as_bool().map(Some).ok_or_else(|| bad(at, &format!("{k} must be true or false"))));
let color = match field("color") {
None => None,
Some(c) => {
let rgb: Option<Vec<f32>> = c.as_array().and_then(|c| c.iter().map(|x| x.as_f64().map(|x| x as f32)).collect());
match rgb.as_deref() {
Some(&[r, g, b]) if [r, g, b].iter().all(|c| (0.0..=1.0).contains(c)) => Some([r, g, b]),
_ => return Err(bad(at, "color must be [r, g, b], each from 0 to 1")),
}
}
};
let children = match field("children") {
None => Vec::new(),
Some(list) => list_of(list, at, count)?,
};
Ok(OutlineNode {
title,
page,
view: field("view").map(|x| view_of(x, at)).transpose()?,
color,
italic: flag("italic")?.unwrap_or(false),
bold: flag("bold")?.unwrap_or(false),
open: flag("open")?.unwrap_or(true),
children,
})
}
/// `{"type": "XYZ", "values": [left, top, zoom]}`: a destination type with exactly its values.
fn view_of(v: &Value, at: &[usize]) -> Result<(Vec<u8>, Vec<Option<f64>>)> {
let kind = v.get("type").and_then(Value::as_str).ok_or_else(|| bad(at, "view needs a type such as XYZ or Fit"))?;
let values = match v.get("values").filter(|x| !x.is_null()) {
None => Vec::new(),
Some(list) => {
let list = list.as_array().filter(|l| l.len() <= 4).ok_or_else(|| bad(at, "view values must be a list of at most 4 numbers or nulls"))?;
list.iter()
.map(|x| match x {
Value::Null => Ok(None),
x => x.as_f64().map(Some).ok_or_else(|| bad(at, "view values must be numbers or null")),
})
.collect::<Result<_>>()?
}
};
let arity = pdfcraft_engine::outline_view_arity(kind.as_bytes()).ok_or_else(|| bad(at, VIEW_TYPES))?;
if values.len() != arity {
return Err(bad(at, &format!("a {kind} view takes {arity} values (numbers, or null to keep the reader's)")));
}
if kind == "FitR" && values.iter().any(Option::is_none) {
return Err(bad(at, "a FitR view needs all four values: left, bottom, right, top"));
}
Ok((kind.as_bytes().to_vec(), values))
}
const VIEW_TYPES: &str = "view type must be XYZ (3 values: left, top, zoom), Fit, FitB (none), FitH, FitV, FitBH, FitBV (1) or FitR (4)";
fn bad(at: &[usize], what: &str) -> ToolError {
if at.is_empty() {
return ToolError::InvalidArgs(format!("bookmarks: {what}"));
}
ToolError::InvalidArgs(format!("bookmark {at:?}: {what}"))
}
crates/automation/src/comments.rs
+153 -8
View File
@@ -5,7 +5,9 @@
use std::collections::HashMap;
use pdfcraft_engine::{Edit, LineEnding, Markup, NewAnnotation, NoteIcon, ReviewState, Rgb, Shape, StampGroup, StampKind, Style, SummarySort};
use pdfcraft_engine::{
BorderShape, Edit, LineEnding, Markup, NewAnnotation, NoteIcon, ReviewState, Rgb, Shape, StampGroup, StampKind, Style, SummarySort,
};
use pdfcraft_render::{Annotation, PageInfo};
use serde_json::{Value, json};
@@ -75,7 +77,7 @@ fn to_user(p: &PageInfo, x: f64, y: f64) -> [f64; 2] {
[ux as f64, uy as f64]
}
fn rect_to_user(p: &PageInfo, r: [f64; 4]) -> [f64; 4] {
pub(crate) fn rect_to_user(p: &PageInfo, r: [f64; 4]) -> [f64; 4] {
let (a, b) = (to_user(p, r[0], r[1]), to_user(p, r[2], r[3]));
[a[0].min(b[0]), a[1].min(b[1]), a[0].max(b[0]), a[1].max(b[1])]
}
@@ -190,6 +192,25 @@ impl Automation {
"rect": crate::view_rect(&doc.info, c.page, c.shown_rect(doc.info.pages.get(c.page).map_or(0, |p| p.rotation))),
"color": c.color.map(hex),
});
// A text box's or typed text's oval or circle (a rectangle is not reported).
if c.subtype == "FreeText"
&& let Some(shape) = doc.comment_props(c.page, c.index).and_then(|p| p.border_shape).filter(|s| *s != BorderShape::Rectangle)
{
v["border_shape"] = json!(shape.name().to_ascii_lowercase());
}
// How opaque it is drawn: as a whole, or its fill and text each at their own.
// Not for replies, which are drawn with the comment they answer.
if !c.is_reply()
&& let Some(p) = doc.comment_props(c.page, c.index)
{
v["opacity"] = json!(p.opacity);
if let Some(f) = p.fill_opacity {
v["fill_opacity"] = json!(f);
}
if let Some(t) = p.text_opacity {
v["text_opacity"] = json!(t);
}
}
// A line's ends, or the corners of connected lines, a polygon or a cloud, as
// comment_edit's `points` takes them.
// Unchanged on a page the viewer could not read, as `rect` is.
@@ -253,6 +274,12 @@ impl Automation {
if endings.is_some() && !matches!(kind, "line" | "arrow" | "polyline" | "callout") {
return Err(ToolError::InvalidArgs("endings apply to a line, arrow, polyline or callout".into()));
}
// Shift while drawing: each end or corner held level or plumb with the one before.
let ortho = a.opt_bool("ortho")?.unwrap_or(false);
if ortho && !matches!(kind, "line" | "arrow" | "polygon" | "cloud" | "polyline") {
return Err(ToolError::InvalidArgs("ortho applies to a line, arrow, polygon, cloud or polyline".into()));
}
let held = |last: [f64; 2], p: [f64; 2]| if ortho { pdfcraft_engine::measure::snap::ortho(last, p) } else { p };
let markup = match kind {
"highlight" => Some(Markup::Highlight),
"underline" => Some(Markup::Underline),
@@ -325,7 +352,8 @@ impl Automation {
Shape::TextBox { rect: rect_to_user(&info, a.need::<4>("rect", "a text box")?), font_size }
}
"line" | "arrow" => {
let (f, t) = (a.need::<2>("from", "a line")?, a.need::<2>("to", "a line")?);
let f = a.need::<2>("from", "a line")?;
let t = held(f, a.need::<2>("to", "a line")?);
let [start, end] = ending_pair(kind, endings.as_deref())?;
Shape::Line { from: to_user(&info, f[0], f[1]), to: to_user(&info, t[0], t[1]), start, end }
}
@@ -352,10 +380,15 @@ impl Automation {
"polygon" | "cloud" | "polyline" => {
let wrong = || ToolError::InvalidArgs(format!("{kind} needs `points`: an array of [x, y] points"));
let pts = a.get("points").and_then(|p| p.as_array()).ok_or_else(wrong)?;
let mut last = None;
let vertices = pts
.iter()
.map(|p| match p.as_array().map(|p| p.iter().map(|v| v.as_f64()).collect::<Vec<_>>()).as_deref() {
Some([Some(x), Some(y)]) => Ok(to_user(&info, *x, *y)),
Some([Some(x), Some(y)]) => {
let p = last.map_or([*x, *y], |l| held(l, [*x, *y]));
last = Some(p);
Ok(to_user(&info, p[0], p[1]))
}
_ => Err(wrong()),
})
.collect::<Result<Vec<_>>>()?;
@@ -425,9 +458,28 @@ impl Automation {
if let Some(o) = a.opt_num("opacity")? {
style.opacity = o.clamp(0.0, 1.0);
}
// The fill's and a text box's text's own opacities (the line keeps `opacity`).
let (fill_opacity, text_opacity) = (a.opt_num("fill_opacity")?, a.opt_num("text_opacity")?);
let texts = matches!(shape, Shape::TextBox { .. } | Shape::Typewriter { .. } | Shape::Callout { .. });
if fill_opacity.is_some() && !(texts || matches!(shape, Shape::Rectangle { .. } | Shape::Oval { .. } | Shape::Polygon { .. })) {
return Err(ToolError::InvalidArgs(
"fill_opacity applies to a rectangle, oval, polygon, cloud, text box, typewriter text or callout".into(),
));
}
if text_opacity.is_some() && !texts {
return Err(ToolError::InvalidArgs("text_opacity applies to a text box, typewriter text or callout".into()));
}
style.fill_opacity = fill_opacity.map(|o| o.clamp(0.0, 1.0));
style.text_opacity = text_opacity.map(|o| o.clamp(0.0, 1.0));
if let Some(w) = a.opt_num("width")? {
style.width = w.max(0.0);
}
if let Some(b) = border_shape(a)? {
if !matches!(shape, Shape::TextBox { .. } | Shape::Typewriter { .. }) {
return Err(ToolError::InvalidArgs("border_shape is for a text box or typewriter text".into()));
}
style.border_shape = b;
}
let contents = a.opt_str("contents")?.unwrap_or_default().to_string();
let author = a.opt_str("author")?.unwrap_or(DEFAULT_AUTHOR).to_string();
// Replace Text: the struck-out text and a grouped caret holding `contents`.
@@ -488,6 +540,7 @@ impl Automation {
}
pub(crate) fn fill_sign_add(&mut self, a: &Args) -> Result<Value> {
let border_shape = border_shape(a)?;
use pdfcraft_engine::FillMark;
let page = self.page(a)?;
let info = self.doc(a)?.info.pages[page].clone();
@@ -515,6 +568,25 @@ impl Automation {
if !(4.0..=300.0).contains(&size) {
return Err(ToolError::InvalidArgs("font_size must be 4–300 pt".into()));
}
let requested_border = a.opt_num("border")?;
if requested_border.is_some() && !matches!(kind, "text" | "date") {
return Err(ToolError::InvalidArgs("border is only for type text or date".into()));
}
let border = requested_border.unwrap_or(0.0);
if !(0.0..=12.0).contains(&border) {
return Err(ToolError::InvalidArgs("border must be 0–12 pt".into()));
}
let border_color = a.opt_str("border_color")?.map(parse_color).transpose()?;
if border_color.is_some() && border <= 0.0 {
return Err(ToolError::InvalidArgs("border_color needs a border".into()));
}
if border_shape.is_some() && !matches!(kind, "text" | "date") {
return Err(ToolError::InvalidArgs("border_shape is only for type text or date".into()));
}
// Without a border an oval or a circle would grow the box around nothing drawn.
if border_shape.is_some_and(|s| s != BorderShape::Rectangle) && border <= 0.0 {
return Err(ToolError::InvalidArgs("border_shape needs a border".into()));
}
if let Some(path) = a.opt_str("path")? {
if !matches!(kind, "signature" | "initials") || a.opt_str("text")?.is_some() {
return Err(ToolError::InvalidArgs("path is only for an image signature or initials; pass either path or text".into()));
@@ -538,13 +610,16 @@ impl Automation {
let edit = image.edit(page, &info, at, kind == "initials", &author).ok_or_else(|| ToolError::InvalidArgs("at must be finite".into()))?;
return self.apply(a, edit);
}
// A circle goes around the text, so its box is no wider than the text needs.
let min_width = if border_shape == Some(BorderShape::Circle) { 0.0 } else { 20.0 };
let text_at = |t: &str| {
let longest = t.lines().map(|line| pdfcraft_engine::annot_text::text_width(line, size)).fold(0.0, f64::max);
let w = (longest + 8.0).clamp(20.0, 600.0);
let w = (longest + 8.0).clamp(min_width, 600.0);
let rows = pdfcraft_engine::annot_text::wrap(t, size, w - 4.0).len().max(1);
let h = rows as f64 * size * 1.2 + 6.0;
let rect = pdfcraft_engine::text_box_at(i64::from(info.rotation), at, w, h).unwrap_or([at[0], at[1] - h, at[0] + w, at[1]]);
Shape::Typewriter { rect, font_size: size }
let [x0, y0, x1, y1] = pdfcraft_engine::text_box_at(i64::from(info.rotation), at, w, h).unwrap_or([at[0], at[1] - h, at[0] + w, at[1]]);
// A border goes around the box, so the text sits where it would without one.
Shape::Typewriter { rect: [x0 - border, y0 - border, x1 + border, y1 + border], font_size: size }
};
let (shape, contents) = match kind {
"text" => {
@@ -590,6 +665,11 @@ impl Automation {
if let Some(color) = requested_color {
style.color = color;
}
if matches!(shape, Shape::Typewriter { .. }) {
style.width = border;
style.border_color = border_color;
style.border_shape = border_shape.unwrap_or_default();
}
self.apply(a, Edit::AddAnnotation(NewAnnotation { page, shape, style, contents, author }))
}
@@ -656,6 +736,21 @@ impl Automation {
let fill = if f.eq_ignore_ascii_case("none") { None } else { Some(parse_color(f)?) };
edits.push(Edit::FillAnnotation { page, index, fill });
}
if let Some(c) = a.opt_str("border_color")? {
let color = if c.eq_ignore_ascii_case("text") { None } else { Some(parse_color(c)?) };
edits.push(Edit::BorderColorAnnotation { page, index, color });
}
if let Some(shape) = border_shape(a)? {
edits.push(Edit::BorderShapeAnnotation { page, index, shape });
}
// After `opacity`, which sets the whole comment's.
let (fill_opacity, text_opacity) = (a.opt_num("fill_opacity")?, a.opt_num("text_opacity")?);
if fill_opacity.into_iter().chain(text_opacity).any(|o| !(0.0..=1.0).contains(&o)) {
return Err(ToolError::InvalidArgs("fill_opacity and text_opacity are 0–1".into()));
}
if fill_opacity.is_some() || text_opacity.is_some() {
edits.push(Edit::PartOpacity { page, index, fill: fill_opacity, text: text_opacity });
}
if let Some(size) = a.opt_num("font_size")? {
if !(1.0..=400.0).contains(&size) {
return Err(ToolError::InvalidArgs("font_size must be 1–400 pt".into()));
@@ -689,7 +784,7 @@ impl Automation {
}
if edits.is_empty() {
return Err(ToolError::InvalidArgs(
"nothing to change: pass contents, color, opacity, width, fill, font_size, rect, move or points".into(),
"nothing to change: pass contents, color, opacity, fill_opacity, text_opacity, width, fill, border_color, border_shape, font_size, rect, move or points".into(),
));
}
let edit = if edits.len() == 1 { edits.remove(0) } else { Edit::Batch { label: "Edit comment".into(), edits } };
@@ -700,6 +795,49 @@ impl Automation {
let (page, index) = self.comment_target(a)?;
self.apply(a, Edit::DeleteAnnotation { page, index })
}
/// The XFDF of the listed comments and their replies (`comments`: 1-based [page, index] pairs).
pub(crate) fn comments_copy(&self, a: &Args) -> Result<Value> {
let id = self.doc(a)?.id;
let all = self.comments(a)?;
let shape = || ToolError::InvalidArgs("comments: [[page, index], …] (1-based, as comment_list reports)".into());
let list = a.get("comments").and_then(Value::as_array).ok_or_else(shape)?;
let mut which = Vec::with_capacity(list.len());
for item in list {
let pair = item.as_array().filter(|p| p.len() == 2).and_then(|p| Some((p.first()?.as_u64()?, p.get(1)?.as_u64()?)));
let Some((page, index)) = pair.filter(|(p, i)| *p >= 1 && *i >= 1) else { return Err(shape()) };
let (page, index) = (usize::try_from(page - 1).map_err(|_| shape())?, usize::try_from(index - 1).map_err(|_| shape())?);
if !all.iter().any(|c| c.page == page && c.index == index) {
return Err(failed(format!("there is no comment {} on page {} (see comment_list)", index + 1, page + 1)));
}
which.push((page, index));
}
let xfdf = self.session.export_comments(id, &which).map_err(failed)?;
Ok(json!({ "xfdf": xfdf, "comments": which.len() }))
}
/// Add the comments of `xfdf` (from `comments_copy`) to `page` as new comments.
pub(crate) fn comments_paste(&mut self, a: &Args) -> Result<Value> {
let page = self.page(a)?;
let xfdf = a.str("xfdf")?.to_owned();
if !pdfcraft_engine::is_xfdf_comments(&xfdf) {
return Err(ToolError::InvalidArgs("xfdf: the text comments_copy returned (XFDF with <annots>)".into()));
}
let offset = match a.nums::<2>("at")? {
Some([x, y]) => {
let info = self.doc(a)?.info.pages.get(page).cloned().ok_or_else(|| failed(format!("no page {}", page + 1)))?;
let [cx, cy] = to_user(&info, x, y);
let b =
pdfcraft_engine::comments_bounds(&xfdf).ok_or_else(|| ToolError::InvalidArgs("xfdf: no comment with a rect to place".into()))?;
Some([cx - (b[0] + b[2]) / 2.0, cy - (b[1] + b[3]) / 2.0])
}
None => None,
};
let before = self.comments(a)?.len();
let mut out = self.apply(a, Edit::PasteComments { xfdf, page, offset })?;
out["comments"] = json!(self.comments(a)?.len().saturating_sub(before));
Ok(out)
}
}
impl Automation {
@@ -744,6 +882,13 @@ impl Automation {
}
}
/// `border_shape`: `rectangle`, `oval` or `circle`.
fn border_shape(a: &Args) -> Result<Option<BorderShape>> {
a.opt_str("border_shape")?
.map(|s| BorderShape::parse(s).ok_or_else(|| ToolError::InvalidArgs(format!("unknown border_shape {s:?}: rectangle, oval or circle"))))
.transpose()
}
#[cfg(test)]
mod tests {
use super::*;
crates/automation/src/dispatch.rs
+40
View File
@@ -69,6 +69,7 @@ impl Automation {
self.apply(&a, edit)?
}
"page_render" => return self.page_render(&a),
"model_render" => return self.model_render(&a),
"comment_image_preview" => return self.comment_image_preview(&a),
"text_extract" => self.text_extract(&a)?,
"text_word_count" => self.text_word_count(&a)?,
@@ -136,6 +137,7 @@ impl Automation {
out["detected"] = json!(given.is_none());
out
}
"page_enhance_scan" => self.page_enhance_scan(&a)?,
"page_delete" => {
let pages = self.pages(&a, "pages")?;
self.apply(&a, Edit::DeletePages { pages })?
@@ -196,6 +198,9 @@ impl Automation {
out
}
"bookmark_list" => self.bookmark_listing(&a)?,
"bookmark_export" => self.bookmark_export(&a)?,
"bookmark_import" => self.bookmark_import(&a)?,
"bookmark_copy" => self.bookmark_copy(&a)?,
"bookmark_add" => {
let page = self.page(&a)?;
let parent = a.opt_path("parent")?.unwrap_or_default();
@@ -769,6 +774,7 @@ impl Automation {
"comment_list" => self.comment_list(&a)?,
"comment_add" => self.comment_add(&a)?,
"stamp_custom" => self.stamp_custom(&a)?,
"model_add" => self.model_add(&a)?,
"comment_reply" => self.comment_reply(&a)?,
"comment_set_status" => self.comment_set_status(&a)?,
"sign_list" => self.sign_list(&a)?,
@@ -861,6 +867,8 @@ impl Automation {
"comments_summarize" => self.comments_summarize(&a)?,
"comment_edit" => self.comment_edit(&a)?,
"comment_delete" => self.comment_delete(&a)?,
"comments_copy" => self.comments_copy(&a)?,
"comments_paste" => self.comments_paste(&a)?,
other => return Err(ToolError::UnknownTool(other.into())),
};
Ok(vec![Content::Json(out)])
@@ -914,3 +922,35 @@ impl Automation {
Ok(out)
}
}
impl Automation {
/// `page_enhance_scan`: Enhance ▸ Scanned document on the chosen pages (all by default).
fn page_enhance_scan(&mut self, a: &Args) -> Result<Value> {
let pages = match a.opt_ints("pages")? {
Some(_) => self.pages(a, "pages")?,
None => Vec::new(),
};
let level = match a.opt_int("level")? {
None => pdfcraft_engine::scan_clean::ScanCleanSettings::default().level,
Some(l @ 0..=100) => l as u8,
Some(l) => return Err(ToolError::InvalidArgs(format!("level must be between 0 and 100, not {l}"))),
};
let mut settings = pdfcraft_engine::scan_clean::ScanCleanSettings::at(level);
if let Some(on) = a.opt_bool("edge_shadows")? {
settings.edge_shadows = on;
}
if let Some(on) = a.opt_bool("despeckle")? {
settings.despeckle = on;
}
let id = self.doc(a)?.id;
let report = self.session.enhance_scanned_pages(id, &pages, settings).map_err(failed)?;
let numbered = |p: &[usize]| p.iter().map(|p| p + 1).collect::<Vec<_>>();
let skipped: Vec<Value> = report.skipped.iter().map(|k| json!({ "pages": numbered(&k.pages), "name": k.name, "reason": k.reason })).collect();
let mut out = summary(self.doc(a)?);
out["pages_cleaned"] = json!(numbered(&report.pages));
out["images"] = json!(report.images);
out["not_scanned"] = json!(numbered(&report.not_scanned));
out["skipped"] = json!(skipped);
Ok(out)
}
}
crates/automation/src/file_access.rs
+12 -2
View File
@@ -394,8 +394,18 @@ mod tests {
// Helpers that are only called with a resolved path. `doc_batch` (batch.rs) resolves its
// inputs and out_dir, walks only inside them (skipping links) and resolves every
// discovered file again before `process` reads it or `publish` writes it.
const RESOLVED_BY_CALLER: &[&str] =
&["write_atomic", "write_atomic_with", "read_input", "read_input_bytes", "read_review_sidecar", "process", "linked", "walk", "publish"];
const RESOLVED_BY_CALLER: &[&str] = &[
"write_atomic",
"write_atomic_with",
"read_input",
"read_input_bytes",
"read_model",
"read_review_sidecar",
"process",
"linked",
"walk",
"publish",
];
let mut dirs = vec![Path::new(env!("CARGO_MANIFEST_DIR")).join("src")];
let mut files = vec![];
while let Some(dir) = dirs.pop() {
crates/automation/src/lib.rs
+3
View File
@@ -27,6 +27,8 @@ mod links;
#[cfg(feature = "mcp")]
pub mod mcp;
mod measure;
mod model3d;
mod model_render;
mod printing;
mod reading;
mod redact;
@@ -50,6 +52,7 @@ use file_access::foreign_share;
use text_cache::TextCache;
mod assemble;
mod bookmark_io;
mod dispatch;
mod documents;
mod error;
crates/automation/src/mcp.rs
+1 -1
View File
@@ -557,7 +557,7 @@ fn tool_json(t: &crate::ToolDef) -> Value {
"title": t.title,
"description": t.description,
"inputSchema": t.input_schema,
"annotations": { "title": t.title, "readOnlyHint": t.read_only, "destructiveHint": t.destructive, "idempotentHint": t.read_only || matches!(t.name, "doc_save" | "doc_optimize" | "doc_export_data" | "measure_export" | "doc_export_images" | "doc_export_office" | "doc_export_all_images" | "doc_export_text" | "image_save"), "openWorldHint": false },
"annotations": { "title": t.title, "readOnlyHint": t.read_only, "destructiveHint": t.destructive, "idempotentHint": t.read_only || matches!(t.name, "doc_save" | "doc_optimize" | "doc_export_data" | "measure_export" | "doc_export_images" | "doc_export_office" | "doc_export_all_images" | "doc_export_text" | "image_save" | "bookmark_export"), "openWorldHint": false },
})
}
crates/automation/src/model3d.rs
+114
View File
@@ -0,0 +1,114 @@
//! 3D model tools (Acrobat: Rich Media ▸ Add 3D, AGENTS.md §3).
use std::path::Path;
use pdfcraft_engine::{Activation3d, Edit, MAX_3D_MODEL_BYTES, MarkFile, Model3dFormat};
use serde_json::Value;
use crate::comments::rect_to_user;
use crate::{Args, Automation, Result, ToolError, failed};
/// The largest side of a 3D model's box, in points: the largest page PDF allows (ISO 32000-2 Annex C).
const MAX_MODEL_SIDE: f64 = 14_400.0;
impl Automation {
/// A U3D or PRC model in `rect` on `page`, its poster the picture at `poster` (a PDF page or an
/// image) fitted inside the box, or a labelled frame.
pub(crate) fn model_add(&mut self, a: &Args) -> Result<Value> {
let page = self.page(a)?;
let info = self.doc(a)?.info.pages.get(page).cloned().ok_or_else(|| ToolError::InvalidArgs(format!("page {} does not exist", page + 1)))?;
let r = a.nums::<4>("rect")?.ok_or_else(|| ToolError::InvalidArgs("missing argument rect".into()))?;
let (w, h) = ((r[2] - r[0]).abs(), (r[3] - r[1]).abs());
if !((1.0..=MAX_MODEL_SIDE).contains(&w) && (1.0..=MAX_MODEL_SIDE).contains(&h)) {
return Err(ToolError::InvalidArgs(format!(
"rect must be [x0, y0, x1, y1] with each side at least 1 pt and at most {MAX_MODEL_SIDE} pt"
)));
}
let activation = match a.opt_str("activation")? {
None => Activation3d::default(),
Some(word) => Activation3d::ALL
.iter()
.find(|(_, w)| *w == word)
.map(|(act, _)| *act)
.ok_or_else(|| ToolError::InvalidArgs(format!("activation must be one of: {}", Activation3d::ALL.map(|(_, w)| w).join(", "))))?,
};
let path = self.resolve(a.str("path")?, false)?;
let model = read_model(&path, MAX_3D_MODEL_BYTES)?;
let file = path.file_name().map(|n| n.to_string_lossy().into_owned()).unwrap_or_default();
let name = a.opt_str("name")?.map(str::to_owned).unwrap_or_else(|| file.rsplit_once('.').map_or(file.clone(), |(s, _)| s.to_owned()));
let poster = match a.opt_str("poster")? {
None => None,
Some(p) => {
let path = self.resolve(p, false)?;
let bytes = crate::read_input(&path)?;
let name = path.file_name().map(|n| n.to_string_lossy().into_owned()).unwrap_or_default();
let page = usize::try_from(a.opt_int("poster_page")?.unwrap_or(1).max(1)).unwrap_or(1).saturating_sub(1);
Some(MarkFile { name, bytes: std::sync::Arc::new(bytes), page })
}
};
let edit = Edit::AddModel3d { page, rect: rect_to_user(&info, r), model: std::sync::Arc::new(model), poster, activation, name };
self.apply(a, edit)
}
}
/// Read a resolved model file of at most `max` bytes. Its size (from the file system) and its first
/// bytes are checked before the rest is read, so a large file that is not a model is never loaded.
fn read_model(path: &Path, max: usize) -> Result<Vec<u8>> {
use std::io::Read as _;
let io = |e: std::io::Error| failed(format!("{}: {e}", path.display()));
let too_large = |size: u64| failed(format!("the 3D model is {size} bytes; the largest accepted is {max}"));
let cap = u64::try_from(max).unwrap_or(u64::MAX);
let mut file = std::fs::File::open(path).map_err(io)?;
if let Ok(meta) = file.metadata()
&& meta.is_file()
&& meta.len() > cap
{
return Err(too_large(meta.len()));
}
let mut bytes = Vec::new();
(&mut file).take(4).read_to_end(&mut bytes).map_err(io)?;
Model3dFormat::detect(&bytes).map_err(failed)?;
let rest = cap.saturating_add(1).saturating_sub(bytes.len() as u64);
file.take(rest).read_to_end(&mut bytes).map_err(io)?;
if bytes.len() > max {
return Err(failed(format!("the 3D model is more than {max} bytes, the largest accepted")));
}
Ok(bytes)
}
#[cfg(test)]
mod tests {
use super::*;
fn file(dir: &Path, name: &str, bytes: &[u8]) -> std::path::PathBuf {
let path = dir.join(name);
std::fs::write(&path, bytes).unwrap();
path
}
#[test]
fn a_model_is_read_whole_when_it_fits() {
let dir = tempfile::tempdir().unwrap();
let mut prc = b"PRC".to_vec();
prc.extend_from_slice(&[7; 40]);
assert_eq!(read_model(&file(dir.path(), "m.prc", &prc), 64).unwrap(), prc);
assert_eq!(read_model(&file(dir.path(), "m.u3d", b"U3D\0x"), 6).unwrap(), b"U3D\0x");
}
#[test]
fn a_model_past_the_cap_is_refused_by_its_size_before_it_is_read() {
let dir = tempfile::tempdir().unwrap();
let path = file(dir.path(), "big.prc", &[b'P', b'R', b'C', 0, 0, 0, 0, 0, 0, 0]);
let err = read_model(&path, 4).unwrap_err().to_string();
assert!(err.contains("is 10 bytes") && err.contains("largest accepted is 4"), "the file's own size, from the file system: {err}");
}
#[test]
fn a_file_that_is_not_a_model_is_refused_by_its_header() {
let dir = tempfile::tempdir().unwrap();
for junk in [&b"%PDF-1.7 a page"[..], b"PR", b"", b"U3DX and more"] {
let err = read_model(&file(dir.path(), "junk.bin", junk), 64).unwrap_err().to_string();
assert!(err.contains("not a 3D model"), "{junk:?}: {err}");
}
}
}
crates/automation/src/model_render.rs
+104
View File
@@ -0,0 +1,104 @@
//! `model_render`: a page's 3D model (ISO 32000-2 §13.6) drawn to a PNG, without a viewer.
use pdfcraft_engine::model3d::Model3dKind;
use pdfcraft_prc::PrcFile;
use pdfcraft_prc_render::{Camera, RenderOptions, assemble, render};
use super::Automation;
use super::dispatch::Content;
use super::error::{Result, ToolError, failed};
use super::json::Args;
use super::pages::encode_png;
/// Default and largest side of the image, in pixels.
pub(crate) const DEFAULT_SIDE: u32 = 512;
pub(crate) const MAX_SIDE: u32 = 4096;
/// The default view: from the front, right and above, as an isometric view looks.
const DEFAULT_DIRECTION: [f64; 3] = [-1.0, 1.0, -1.0];
impl Automation {
pub(crate) fn model_render(&mut self, a: &Args) -> Result<Vec<Content>> {
let page = self.page(a)?;
let width = side(a, "width")?;
let height = side(a, "height")?;
let direction = direction(a)?;
let perspective = a.opt_bool("perspective")?.unwrap_or(true);
let doc = self.doc(a)?;
let found = doc.model_3d_annotations(page).map_err(failed)?;
let index = match a.opt_int("model")? {
None => *found.first().ok_or_else(|| failed(format!("page {} has no 3D model", page.saturating_add(1))))?,
Some(n) => {
let pos = usize::try_from(n).ok().and_then(|n| n.checked_sub(1)).ok_or_else(|| Args::wrong("model", "a positive integer"))?;
*found
.get(pos)
.ok_or_else(|| ToolError::InvalidArgs(format!("page {} has {} 3D models, not {n}", page.saturating_add(1), found.len())))?
}
};
let model = doc.model_3d(page, index).map_err(failed)?;
if model.kind == Model3dKind::U3d {
return Err(failed("this 3D model is in U3D, which can't be drawn yet; only PRC models can"));
}
let prc = PrcFile::parse(&model.data).map_err(|e| failed(format!("the 3D model could not be read: {e}")))?;
let meshes = assemble(&prc).map_err(failed)?;
let aspect = f64::from(width) / f64::from(height);
let camera = Camera::fit_meshes(&meshes, direction, up_for(direction), perspective, aspect).map_err(failed)?;
let image = render(&meshes, &camera, &RenderOptions { width, height, ..RenderOptions::default() }).map_err(failed)?;
let data = encode_png(image.width, image.height, &premultiplied(image.rgba))?;
Ok(vec![Content::Png { data, width: image.width, height: image.height }])
}
}
fn side(a: &Args, key: &str) -> Result<u32> {
match a.opt_int(key)? {
None => Ok(DEFAULT_SIDE),
Some(n) => u32::try_from(n).ok().filter(|n| (1..=MAX_SIDE).contains(n)).ok_or_else(|| Args::wrong(key, &format!("between 1 and {MAX_SIDE}"))),
}
}
/// `direction`: three finite numbers, not all zero, from the eye towards the model.
fn direction(a: &Args) -> Result<[f64; 3]> {
let Some(v) = a.get("direction") else { return Ok(DEFAULT_DIRECTION) };
let wrong = || Args::wrong("direction", "three numbers [x, y, z], not all zero");
let items = v.as_array().filter(|x| x.len() == 3).ok_or_else(wrong)?;
let mut out = [0.0; 3];
for (o, x) in out.iter_mut().zip(items) {
*o = x.as_f64().filter(|f| f.is_finite()).ok_or_else(wrong)?;
}
if out.iter().all(|c| *c == 0.0) {
return Err(wrong());
}
Ok(out)
}
/// Z is up, as CAD models are drawn, unless the view looks along it.
fn up_for([x, y, z]: [f64; 3]) -> [f64; 3] {
if x.hypot(y) <= z.abs() * 1e-6 { [0.0, 1.0, 0.0] } else { [0.0, 0.0, 1.0] }
}
/// Straight RGBA to the premultiplied RGBA `encode_png` takes.
fn premultiplied(mut rgba: Vec<u8>) -> Vec<u8> {
for [r, g, b, alpha] in rgba.as_chunks_mut::<4>().0 {
let a = u16::from(*alpha);
for c in [r, g, b] {
*c = u8::try_from((u16::from(*c) * a + 127) / 255).unwrap_or(u8::MAX);
}
}
rgba
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn premultiplying_scales_colour_by_alpha() {
assert_eq!(premultiplied(vec![200, 100, 0, 255, 200, 100, 0, 128, 9, 9, 9, 0]), vec![200, 100, 0, 255, 100, 50, 0, 128, 0, 0, 0, 0]);
}
#[test]
fn a_view_along_z_keeps_y_up() {
assert_eq!(up_for([0.0, 0.0, -1.0]), [0.0, 1.0, 0.0]);
assert_eq!(up_for(DEFAULT_DIRECTION), [0.0, 0.0, 1.0]);
}
}
crates/automation/src/tools.rs
+112 -5
View File
@@ -181,6 +181,13 @@ pub fn tools() -> Vec<ToolDef> {
"hide_text_block": { "type": "integer", "minimum": 1, "description": "Render an inline editor background without this paragraph (number from text_paragraphs, resolved as text_edit does: refused once an edit merged that paragraph into a neighbour). Does not modify the document." },
"hide_added_text": { "type": "integer", "minimum": 1, "description": "Render a retyping background without this added text item (added index from object_list). Mutually exclusive with hide_text_block. Does not modify the document." }
}), &["doc", "page"])),
t("model_render", "Render a 3D model", "Draw a page's 3D model (a 3D annotation's PRC stream) to a PNG, framed to fill the image, with headlight shading on a white background. model is its 1-based position among the page's 3D models (default 1). direction is the view direction [x, y, z] from the eye towards the model (default [-1, 1, -1]); Z is up unless the view looks along Z. perspective defaults to true. U3D models can't be drawn yet and give an error. Does not modify the document.")
.ro()
.with(schema(json!({ "doc": doc(), "page": { "type": "integer", "minimum": 1 }, "model": { "type": "integer", "minimum": 1 },
"width": { "type": "integer", "minimum": 1, "maximum": 4096 }, "height": { "type": "integer", "minimum": 1, "maximum": 4096 },
"direction": { "type": "array", "items": { "type": "number" }, "minItems": 3, "maxItems": 3 },
"perspective": { "type": "boolean" }
}), &["doc", "page"])),
t("object_list", "List editable objects", "List existing paragraphs, Image/Form artwork and added content once, with source kind/index references, displayed rectangles and document generation. Indexes change when content changes; use generation in object_move to reject stale references.")
.ro().with(schema(json!({"doc":doc(),"page":{"type":"integer","minimum":1}}), &["doc","page"])),
t("object_move", "Move objects together", "Atomically translate 1–1000 mixed references from object_list by [dx, dy] in displayed points (right/down). Preserves glyph codes, fonts, image data and relative spacing; success is one undo step. Text used as a clipping path and vertical text are refused. A bad/stale reference refuses the entire move.")
@@ -233,6 +240,18 @@ pub fn tools() -> Vec<ToolDef> {
}),
&["doc", "page"],
)),
t("page_enhance_scan", "Enhance scanned file", "Clean up scanned pages in place (Scan & OCR ▸ Enhance ▸ Scanned document). A page is a scan when its pictures cover at least half of it; the pictures covering at least a tenth of it are cleaned and written back into the same image objects at the same size and colour space (a JPEG stays a JPEG at quality 90, other grey/RGB images are stored losslessly, a 1-bit image stays 1-bit), so the page's invisible OCR text, comments, links and other content stay as they were, and an image drawn on several pages is cleaned once. level (0–100, default 50) is background removal: the paper's illumination is evened out so a grey, yellowed or unevenly lit background turns white, then the contrast is raised so faint strokes darken; colour ink stays colour. edge_shadows (default: on unless level is 0) lightens shadows along the scan's edges (book gutters, a copier's dark border). despeckle (default: from level 60) turns isolated dark specks white; it is the only change a 1-bit scan takes. Returns pages_cleaned (1-based), images (rewritten), not_scanned (pages that are no scan) and skipped ([{pages, name, reason}]: CMYK, indexed, JPEG 2000, JBIG2 or CCITT images, a scan drawn inside grouped Form artwork, broken data, or nothing to change). One undoable step; none when nothing changed.")
.cmd("ocr.enhance")
.with(schema(
json!({
"doc": doc(),
"pages": pages("to clean up (default: all)"),
"level": { "type": "integer", "minimum": 0, "maximum": 100 },
"edge_shadows": { "type": "boolean" },
"despeckle": { "type": "boolean" },
}),
&["doc"],
)),
t("page_delete", "Delete pages", "Delete pages. Undoable until saved.")
.destructive()
.cmd("page.delete")
@@ -323,6 +342,50 @@ pub fn tools() -> Vec<ToolDef> {
}),
&["doc", "title", "page"],
)),
t(
"bookmark_export",
"Export bookmarks",
"The bookmark tree, or the subtree at path, as a list of nodes bookmark_import takes back unchanged. A node has title, page (1-based, null when it goes nowhere), open, and when set view ({type, values}: a destination type from ISO 32000-2 §12.3.2.2 Table 149 and its values in PDF user space, as the file stores them), color ([r, g, b], 0-1), bold, italic and children. truncated is true when bookmarks nested deeper than 33 levels were left out. With out, the tree is also written there as a JSON file {\"bookmarks\": [...]}, with \"truncated\": true when that is so; out is overwritten.",
)
.destructive()
.with(schema(
json!({ "doc": doc(), "path": path("The bookmark whose subtree to export (omit for the whole outline)"), "out": path_arg() }),
&["doc"],
)),
t(
"bookmark_import",
"Import bookmarks",
"Add a tree of bookmarks, as bookmark_export gives it (bookmarks) or as a JSON file it wrote (file), under a parent bookmark (or at the top level) at a position, or with replace in place of the whole outline. A bookmark whose page this document lacks is kept without a destination and listed in unmapped (1-based paths into the given tree). A view takes exactly its type's values (XYZ 3, FitR 4, FitH/FitV/FitBH/FitBV 1, Fit/FitB none). Give exactly one of bookmarks and file. A title may be empty but not missing; a key a node does not have is refused. Every node is checked before anything is written: at most 100000 bookmarks, nested at most 33 levels counting the parent's. truncated is true when the file says its export left bookmarks out. Undoable.",
)
.destructive()
.with(schema(
json!({
"doc": doc(),
"bookmarks": { "type": "array", "items": { "type": "object" }, "minItems": 1, "description": "Nodes as bookmark_export returns them." },
"file": path_arg(),
"parent": path("Parent bookmark (omit for the top level)"),
"position": { "type": "integer", "minimum": 1, "description": "1-based position among the parent's children (default: last)." },
"replace": { "type": "boolean", "description": "Replace the document's bookmarks instead of adding to them (default false; leave parent out)." }
}),
&["doc"],
)),
t(
"bookmark_copy",
"Copy bookmarks from another document",
"Copy the bookmarks of document from (or the subtree at path) into doc, by page number: under a parent bookmark (or at the top level) at a position, or with replace in place of doc's outline. A bookmark whose page doc lacks is kept without a destination and listed in unmapped (1-based paths into the copied tree). truncated is true when bookmarks nested deeper than 33 levels were left out of the copy. Undoable.",
)
.destructive()
.with(schema(
json!({
"doc": doc(),
"from": { "type": "integer", "minimum": 1, "description": "Id of the document to copy from." },
"path": path("The bookmark whose subtree to copy (omit for the whole outline)"),
"parent": path("Parent bookmark in doc (omit for the top level)"),
"position": { "type": "integer", "minimum": 1, "description": "1-based position among the parent's children (default: last)." },
"replace": { "type": "boolean", "description": "Replace doc's bookmarks instead of adding to them (default false; leave parent out)." }
}),
&["doc", "from"],
)),
t("bookmark_rename", "Rename a bookmark", "Change a bookmark's title. Undoable.")
.with(schema(json!({ "doc": doc(), "path": path("The bookmark"), "title": { "type": "string", "minLength": 1 } }), &["doc", "path", "title"])),
t("bookmark_delete", "Delete a bookmark", "Delete a bookmark and the bookmarks under it. Undoable.")
@@ -469,6 +532,20 @@ pub fn tools() -> Vec<ToolDef> {
}),
&["doc", "page", "path"],
)),
t("model_add", "Add a 3D model", "Add a 3D model (Acrobat: Rich Media ▸ Add 3D): the U3D or PRC file at path, recognised by its first bytes, as a 3D annotation in rect [x0, y0, x1, y1] (points from the top-left of the displayed page; each side 1–14400 pt; at most 256 MiB). Its poster, what prints and what readers show until the model is activated, is the picture at poster (a PDF page, poster_page default 1, or an image) fitted inside rect keeping its shape and centred; without poster, a frame labelled \"3D model\". activation: click (default), page_open or page_visible. name is its description (default: the file name). No views are added: readers open the model on its own default view. Undoable.")
.with(schema(
json!({
"doc": doc(),
"page": { "type": "integer", "minimum": 1 },
"path": { "type": "string" },
"rect": { "type": "array", "items": { "type": "number" }, "minItems": 4, "maxItems": 4 },
"poster": { "type": "string" },
"poster_page": { "type": "integer", "minimum": 1 },
"activation": { "type": "string", "enum": ["click", "page_open", "page_visible"] },
"name": { "type": "string" },
}),
&["doc", "page", "path", "rect"],
)),
t("form_set_image", "Set an image field's picture", "Show an image file (PNG, JPEG, TIFF, GIF or BMP) in an image field or button, scaled to fit and centred (what clicking an image field does). On an XFA form the picture is kept in its data too (base64, as Designer saves it) and the form shows it from there. Undoable.")
.with(schema(json!({ "doc": doc(), "field": { "type": "string" }, "path": { "type": "string" } }), &["doc", "field", "path"])),
t(
@@ -781,9 +858,13 @@ pub fn tools() -> Vec<ToolDef> {
"icon": { "type": "string", "description": "note: Comment, Note, Help, Insert, Key, NewParagraph, Paragraph (the standard names), Check, Circle, Cross, CrossHairs, Star, RightArrow, UpArrow; attachment: PushPin, Paperclip, Graph, Tag." },
"color": color(),
"fill": color(),
"opacity": { "type": "number", "minimum": 0, "maximum": 1 },
"opacity": { "type": "number", "minimum": 0, "maximum": 1, "description": "The whole comment's opacity (its line or border's, where its fill or text has its own)." },
"fill_opacity": { "type": "number", "minimum": 0, "maximum": 1, "description": "The fill's own opacity: rectangle, oval, polygon, cloud, text box, typewriter text or callout." },
"text_opacity": { "type": "number", "minimum": 0, "maximum": 1, "description": "The text's own opacity: text box, typewriter text or callout." },
"width": { "type": "number", "minimum": 0, "description": "Line width in points." },
"border_shape": { "type": "string", "enum": ["rectangle", "oval", "circle"], "description": "A text box's or typewriter text's border shape: an oval or a circle goes around the text box." },
"endings": { "type": "array", "items": { "type": "string", "enum": ["None", "Square", "Circle", "Diamond", "OpenArrow", "ClosedArrow", "Butt", "ROpenArrow", "RClosedArrow", "Slash"] }, "minItems": 1, "maxItems": 2, "description": "Line endings. Two names (start, end) for a line, arrow or polyline; one name for a callout." },
"ortho": { "type": "boolean", "description": "line/arrow/polygon/cloud/polyline: hold each end or corner level or plumb with the one before, whichever is nearer (Shift while drawing)." },
"font_size": { "type": "number", "exclusiveMinimum": 0 },
}),
&["doc", "page", "type"],
@@ -914,13 +995,17 @@ pub fn tools() -> Vec<ToolDef> {
json!({ "doc": doc(), "sort": { "type": "string", "enum": ["page", "author", "date", "type"] }, "out": save_out, "open": open }),
&["doc"],
)),
t("comment_edit", "Edit a comment", "Change a comment's text, colour, opacity, line width, font size (text box/typewriter/callout), fill (rectangle/oval/polygon/text box/callout; \"none\" removes it), rectangle (rectangle/oval/text box/stamp), position (`move` [dx, dy] in points) or shape (`points`: a line's two ends, or the corners of connected lines, a polygon or a cloud). Stamps keep their original appearance when resized. One undo step.").with(schema(
t("comment_edit", "Edit a comment", "Change a comment's text, colour, opacity (the whole comment's, or its fill's and text's own), line width, font size (text box/typewriter/callout), fill (rectangle/oval/polygon/text box/callout; \"none\" removes it), border colour (`border_color`, text box/typewriter/callout; \"text\" gives it the text's colour again), rectangle (rectangle/oval/text box/stamp), position (`move` [dx, dy] in points) or shape (`points`: a line's two ends, or the corners of connected lines, a polygon or a cloud). Stamps keep their original appearance when resized. One undo step.").with(schema(
comment_ref(json!({
"contents": { "type": "string" },
"color": color(),
"opacity": { "type": "number", "minimum": 0, "maximum": 1 },
"opacity": { "type": "number", "minimum": 0, "maximum": 1, "description": "The whole comment's opacity: its fill and text drop opacities of their own." },
"fill_opacity": { "type": "number", "minimum": 0, "maximum": 1, "description": "The fill's own opacity (rectangle, oval, polygon, cloud, text box, typewriter text, callout); the line or border keeps the comment's." },
"text_opacity": { "type": "number", "minimum": 0, "maximum": 1, "description": "The text's own opacity (text box, typewriter text, callout)." },
"width": { "type": "number", "minimum": 0 },
"fill": { "type": "string", "description": "Interior colour of a rectangle, oval, polygon, text box or callout: #RRGGBB, a colour name, or \"none\" for no fill." },
"border_color": { "type": "string", "description": "A text box's, typewriter text's or callout's border colour: #RRGGBB, a colour name, or \"text\" for the text's colour." },
"border_shape": { "type": "string", "enum": ["rectangle", "oval", "circle"], "description": "A text box's or typewriter text's border shape: an oval or a circle goes around the text, which keeps its place." },
"font_size": { "type": "number", "minimum": 1, "maximum": 400, "description": "Font size in points of a text box, typewriter text or callout; the box grows to hold the text." },
"rect": { "type": "array", "items": { "type": "number" }, "minItems": 4, "maxItems": 4 },
"move": point(),
@@ -931,6 +1016,25 @@ pub fn tools() -> Vec<ToolDef> {
t("comment_image_preview", "Preview image signature layers", "Return a PNG of the page without the selected image signature/initials (layer=background, default), or its embedded image with alpha (layer=image). Includes the displayed rectangle, document rotation, image_rotation (how far clockwise the image is shown turned) and annotation opacity. Cache these layers for live placement/resizing; commit once with comment_edit. Does not change the document or undo history.")
.ro().with(schema(comment_ref(json!({ "layer": { "type": "string", "enum": ["background", "image"], "default": "background" }, "dpi": { "type": "number", "minimum": 1, "maximum": 600, "default": 96 } })), &["doc"])),
t("comment_delete", "Delete a comment", "Delete a comment with its pop-up and replies, or one reply (its id, or page and index, from comment_list) with its pop-up and the replies to it; the comment and its other replies stay. Undoable.").destructive().with(schema(comment_ref(json!({})), &["doc"])),
t(
"comments_copy",
"Copy comments",
"The XFDF of these comments, each with its replies: what Copy puts on the clipboard. Paste it into this or another document with comments_paste. `comments`: [[page, index], …], 1-based as comment_list reports them.",
)
.ro()
.with(schema(
json!({
"doc": doc(),
"comments": { "type": "array", "items": { "type": "array", "items": { "type": "integer", "minimum": 1 }, "minItems": 2, "maxItems": 2 }, "minItems": 1 }
}),
&["doc", "comments"],
)),
t(
"comments_paste",
"Paste comments",
"Add the comments of an XFDF text (from comments_copy) to `page` as new comments: copies under names of their own, so nothing in the document is replaced and pasting twice gives two sets; replies keep their copied parent. `at` [x, y] (points, origin at the top-left of the displayed page, y down, as comment_add takes them) puts the centre of the copies' bounding box there; without it they keep their positions. Undoable.",
)
.with(schema(json!({ "doc": doc(), "page": { "type": "integer", "minimum": 1 }, "xfdf": { "type": "string" }, "at": point() }), &["doc", "page", "xfdf"])),
t(
"doc_protect",
"Protect with passwords",
@@ -1306,7 +1410,7 @@ pub fn tools() -> Vec<ToolDef> {
t(
"fill_sign_add",
"Fill & Sign: type text or place a mark",
"Fill in a form that has no fields, as Acrobat's Fill & Sign does: type text (`text`, default 10 pt; `font_size` 4–300 pt; optional `color` as #RRGGBB for text/date, default black), place a check, cross, dot or line, today's date (in Preferences ▸ Date format, see fill_sign_date_format, or `format`; `font_size` also applies), or a signature or initials (either `text` drawn in a script font, or `path` to a local PNG/JPEG image up to 4 MiB and 32 megapixels, scaled down to 4 megapixels, preserving transparency; with `scan: true`, `path` is instead a photo of a signature on paper, within the same limits: the signature is found, and the paper, shadows and table around it are removed, with `sensitivity` 0–100, default 50, picking up fainter ink when higher). `at` [x, y] is in points from the top-left of the page: text's top-left, mark's centre, signature's left edge centred vertically. Image signatures fit within 150 pt wide and 32 pt tall (24 pt for initials). Creates movable, undoable annotations. Typed text the standard fonts can't draw is saved and drawn as `?` (`warnings` in the result says so); a typed signature neither the script font nor its fallback can draw is refused (pass `path` instead).",
"Fill in a form that has no fields, as Acrobat's Fill & Sign does: type text (`text`, default 10 pt; `font_size` 4–300 pt; optional `color` as #RRGGBB for text/date, default black; `border` 0–12 pt around it, default none, in its colour or `border_color`), place a check, cross, dot or line, today's date (in Preferences ▸ Date format, see fill_sign_date_format, or `format`; `font_size` also applies), or a signature or initials (either `text` drawn in a script font, or `path` to a local PNG/JPEG image up to 4 MiB and 32 megapixels, scaled down to 4 megapixels, preserving transparency; with `scan: true`, `path` is instead a photo of a signature on paper, within the same limits: the signature is found, and the paper, shadows and table around it are removed, with `sensitivity` 0–100, default 50, picking up fainter ink when higher). `at` [x, y] is in points from the top-left of the page: text's top-left, mark's centre, signature's left edge centred vertically. Image signatures fit within 150 pt wide and 32 pt tall (24 pt for initials). Creates movable, undoable annotations. Typed text the standard fonts can't draw is saved and drawn as `?` (`warnings` in the result says so); a typed signature neither the script font nor its fallback can draw is refused (pass `path` instead).",
)
.cmd("sign.fill.text")
.with(schema(
@@ -1318,6 +1422,9 @@ pub fn tools() -> Vec<ToolDef> {
"text": { "type": "string", "minLength": 1 },
"font_size": { "type": "number", "minimum": 4, "maximum": 300, "description": "For type text or date (default 10 pt)." },
"color": color(),
"border": { "type": "number", "minimum": 0, "maximum": 12, "description": "For type text or date: a border around it in points, in the text's colour (default 0, none)." },
"border_color": { "type": "string", "description": "With border: its own colour, #RRGGBB or a colour name." },
"border_shape": { "type": "string", "enum": ["rectangle", "oval", "circle"], "description": "With border, for type text or date: the border's shape (default rectangle); an oval or a circle goes around the text." },
"path": { "type": "string", "description": "PNG or JPEG for signature/initials; pass either path or text." },
"scan": { "type": "boolean", "description": "With path: scan a photo of a signature on paper (default false)." },
"sensitivity": { "type": "integer", "minimum": 0, "maximum": 100, "description": "With scan: how faint ink may be, 0–100 (default 50)." },
@@ -1459,7 +1566,7 @@ pub fn tools() -> Vec<ToolDef> {
.with(schema(json!({ "doc": doc() }), &["doc"])),
t("doc_open_revision", "Open a revision", "Open saved revision `revision` (1 = the oldest) of a document as a new, unsaved document, to see the file as it was then.")
.with(schema(json!({ "doc": doc(), "revision": { "type": "integer", "minimum": 1 } }), &["doc", "revision"])),
t("doc_optimize", "Optimize PDF", "Write an optimized copy to `path` (Acrobat's PDF Optimizer). preset: standard (default; Reduce File Size's choices) or mobile (images to 96 ppi above 144, JPEG high quality); the options below change it. compatibility: retain (default; keeps the document's version, or 1.5 when object streams are written), 1.4, 1.5, 1.6, 1.7 or 2.0. merge_fragments (default true) first puts a page picture drawn as many stacked image strips (as some PDF printers write it) back together into one image, so it can be resampled (merged_images, fragments_merged). color / gray: { downsample, ppi, above_ppi, compression: jpeg|zip|retain, quality 1–100 } (defaults: downsample to 150 ppi above 225, JPEG 60). mono: { downsample, ppi, above_ppi, compression: ccitt_g4|jbig2|retain } for black-and-white images and image masks (default: averaged down to 300 ppi above 450 ppi, counted in images_resampled; stored losslessly as CCITT Group 4 or JBIG2 (or 1-bit Flate, once resampled or packed from more bits) when that is smaller, Group 3 fax images, JBIG2 images, black-and-white RGB and indexed images using only palette entries 0 and 1 among them, counted in images_recompressed; retain resamples only, a CCITT image staying CCITT, a JBIG2 image JBIG2 and any other becoming Flate). Images past what one run may decode stay as they are, counted in images_refused; streams it only stores with Flate past that stay LZW, uncounted. Images are measured where pages draw them and replaced only when smaller. discard_*: thumbnails (default true), alternate_images (true), tags, print_settings; flate_unencoded (true); object_streams (default true; always off for 1.4); lzw_to_flate (true) decodes LZW streams and stores them with Flate instead, leaving any it cannot decode unchanged (lzw_streams counts the streams converted); remove_invalid_links and remove_unreferenced_dests (true); images and forms the pages list but never draw are always dropped (unused_xobjects). discard: Remove Hidden Information categories (metadata, attachments, comments, form-fields, hidden-text, hidden-layers, bookmarks, links-actions-scripts, private-data). Signed documents are refused. The open document is unchanged.")
t("doc_optimize", "Optimize PDF", "Write an optimized copy to `path` (Acrobat's PDF Optimizer). preset: standard (default; Reduce File Size's choices) or mobile (images to 96 ppi above 144, JPEG high quality); the options below change it. compatibility: retain (default; keeps the document's version, or 1.5 when object streams are written), 1.4, 1.5, 1.6, 1.7 or 2.0. merge_fragments (default true) first puts a page picture drawn as many stacked image strips (as some PDF printers write it) back together into one image, so it can be resampled (merged_images, fragments_merged). color / gray: { downsample, ppi, above_ppi, compression: jpeg|zip|retain, quality 1–100 } (defaults: downsample to 150 ppi above 225, JPEG 60). mono: { downsample, ppi, above_ppi, compression: ccitt_g4|jbig2|retain } for black-and-white images and image masks (default: averaged down to 300 ppi above 450 ppi, counted in images_resampled; stored losslessly as CCITT Group 4 or JBIG2 (or 1-bit Flate, once resampled or packed from more bits) when that is smaller, up to 16 JBIG2 images sharing one symbol dictionary (/JBIG2Globals) where that is smaller still (image masks and DeviceGray images named directly, of at most 64 Mi pixels, which redaction can store again without it), Group 3 fax images, JBIG2 images, black-and-white RGB and indexed images using only palette entries 0 and 1 among them, counted in images_recompressed; retain resamples only, a CCITT image staying CCITT, a JBIG2 image JBIG2 and any other becoming Flate). Images past what one run may decode stay as they are, counted in images_refused; streams it only stores with Flate past that stay LZW, uncounted. Images are measured where pages draw them and replaced only when smaller. discard_*: thumbnails (default true), alternate_images (true), tags, print_settings; flate_unencoded (true); object_streams (default true; always off for 1.4); lzw_to_flate (true) decodes LZW streams and stores them with Flate instead, leaving any it cannot decode unchanged (lzw_streams counts the streams converted); remove_invalid_links and remove_unreferenced_dests (true); images and forms the pages list but never draw are always dropped (unused_xobjects). discard: Remove Hidden Information categories (metadata, attachments, comments, form-fields, hidden-text, hidden-layers, bookmarks, links-actions-scripts, private-data). Signed documents are refused. The open document is unchanged.")
.cmd("optimize.advanced")
.with(schema(
json!({
crates/automation/tests/it/automation.rs
+470
View File
@@ -2135,6 +2135,104 @@ fn numbering_pages_through_tools() {
assert!(matches!(a.call("page_number", &json!({ "doc": doc, "from": 1, "to": 1, "style": "klingon" })), Err(ToolError::InvalidArgs(_))));
}
/// Typed text and text boxes take an oval or a circle for their border through the tools;
/// comment_list reports it (a rectangle is not reported). Only typed text, dates and text boxes
/// have one, and only the three shapes.
#[test]
fn text_borders_take_an_oval_or_a_circle_through_the_tools() {
let dir = workdir("border-shape");
let mut a = auto(&dir);
let doc = ok(&mut a, "doc_open", json!({ "path": "a.pdf" }))["doc"].as_u64().unwrap();
ok(
&mut a,
"fill_sign_add",
json!({ "doc": doc, "page": 1, "type": "text", "at": [20, 40], "text": "12", "border": 1, "border_shape": "circle" }),
);
ok(
&mut a,
"comment_add",
json!({ "doc": doc, "page": 1, "type": "textbox", "rect": [10, 200, 190, 240], "contents": "Hello", "width": 1, "border_shape": "Oval" }),
);
let list = |a: &mut Automation| ok(a, "comment_list", json!({ "doc": doc, "page": 1 }))["comments"].as_array().unwrap().clone();
let shapes: Vec<Value> = list(&mut a).iter().map(|c| c["border_shape"].clone()).collect();
assert_eq!(shapes, [json!("circle"), json!("oval")]);
let typed = list(&mut a)[0].clone();
// Snug around "12": the box no wider than the text needs, the circle as wide as that box.
let r: Vec<f64> = typed["rect"].as_array().unwrap().iter().map(|v| v.as_f64().unwrap()).collect();
let w = pdfcraft_engine::annot_text::text_width("12", 10.0) + 10.0;
assert!((r[2] - r[0] - w).abs() < 0.05 && (r[3] - r[1] - w).abs() < 0.05, "{r:?}");
ok(&mut a, "comment_edit", json!({ "doc": doc, "page": typed["page"], "index": typed["index"], "border_shape": "oval" }));
assert_eq!(list(&mut a)[0]["border_shape"], "oval");
ok(&mut a, "comment_edit", json!({ "doc": doc, "page": typed["page"], "index": typed["index"], "border_shape": "rectangle" }));
assert!(list(&mut a)[0].get("border_shape").is_none(), "a rectangle is not reported");
for (tool, args) in [
("fill_sign_add", json!({ "doc": doc, "page": 1, "type": "check", "at": [60, 60], "border_shape": "oval" })),
("fill_sign_add", json!({ "doc": doc, "page": 1, "type": "text", "text": "7", "at": [60, 60], "border_shape": "circle" })),
("comment_add", json!({ "doc": doc, "page": 1, "type": "rectangle", "rect": [80, 10, 120, 40], "border_shape": "oval" })),
("comment_edit", json!({ "doc": doc, "page": typed["page"], "index": typed["index"], "border_shape": "star" })),
] {
assert!(matches!(a.call(tool, &args), Err(ToolError::InvalidArgs(_))), "{args}");
}
let callout = ok(
&mut a,
"comment_add",
json!({ "doc": doc, "page": 1, "type": "callout", "rect": [110, 100, 190, 140], "to": [40, 60], "contents": "Look" }),
);
assert!(matches!(
a.call("comment_edit", &json!({ "doc": doc, "page": 1, "index": callout["comment"]["index"], "border_shape": "oval" })),
Err(ToolError::Failed(_))
));
}
/// A rectangle's fill and a text box's text each take an opacity of their own through the tools
/// (the line or border keeps the comment's `opacity`); `opacity` sets the whole comment's again.
#[test]
fn comment_parts_take_opacities_of_their_own_through_the_tools() {
let dir = workdir("part-opacity");
let mut a = auto(&dir);
let doc = ok(&mut a, "doc_open", json!({ "path": "a.pdf" }))["doc"].as_u64().unwrap();
ok(
&mut a,
"comment_add",
json!({ "doc": doc, "page": 1, "type": "rectangle", "rect": [10, 10, 60, 40], "fill": "yellow", "fill_opacity": 0.25 }),
);
ok(
&mut a,
"comment_add",
json!({ "doc": doc, "page": 1, "type": "textbox", "rect": [10, 200, 190, 240], "contents": "Hello", "fill": "#FFFF99", "width": 1, "fill_opacity": 0.2, "text_opacity": 1 }),
);
let find = |a: &mut Automation, kind: &str| {
let list = ok(a, "comment_list", json!({ "doc": doc, "page": 1 }));
list["comments"].as_array().unwrap().iter().find(|c| c["type"] == kind).unwrap().clone()
};
let rect = find(&mut a, "Square");
assert_eq!((rect["opacity"].as_f64(), rect["fill_opacity"].as_f64(), rect.get("text_opacity")), (Some(1.0), Some(0.25), None));
let tb = find(&mut a, "FreeText");
assert_eq!((tb["opacity"].as_f64(), tb["fill_opacity"].as_f64(), tb["text_opacity"].as_f64()), (Some(1.0), Some(0.2), Some(1.0)));
// The fill changes, the text stays; an undo step.
ok(&mut a, "comment_edit", json!({ "doc": doc, "page": tb["page"], "index": tb["index"], "fill_opacity": 0.5 }));
let tb = find(&mut a, "FreeText");
assert_eq!((tb["fill_opacity"].as_f64(), tb["text_opacity"].as_f64()), (Some(0.5), Some(1.0)));
assert_eq!(ok(&mut a, "edit_undo", json!({ "doc": doc }))["undone"], "Change comment properties");
ok(&mut a, "edit_redo", json!({ "doc": doc }));
// The whole comment's opacity: everything at it again.
ok(&mut a, "comment_edit", json!({ "doc": doc, "page": rect["page"], "index": rect["index"], "opacity": 0.6 }));
let rect = find(&mut a, "Square");
assert_eq!((rect["opacity"].as_f64(), rect["fill_opacity"].as_f64()), (Some(0.6), Some(0.6)));
// Only the parts a comment has, at 0–1.
for (tool, args) in [
("comment_add", json!({ "doc": doc, "page": 1, "type": "line", "from": [0, 0], "to": [50, 50], "fill_opacity": 0.5 })),
("comment_add", json!({ "doc": doc, "page": 1, "type": "rectangle", "rect": [80, 10, 120, 40], "text_opacity": 0.5 })),
("comment_edit", json!({ "doc": doc, "page": rect["page"], "index": rect["index"], "fill_opacity": 2 })),
] {
assert!(matches!(a.call(tool, &args), Err(ToolError::InvalidArgs(_))), "{args}");
}
assert!(matches!(
a.call("comment_edit", &json!({ "doc": doc, "page": rect["page"], "index": rect["index"], "text_opacity": 0.5 })),
Err(ToolError::Failed(_))
));
}
#[test]
fn comments_through_tools() {
let dir = workdir("comments");
@@ -3741,6 +3839,62 @@ fn enhance_camera_image_through_tools() {
));
}
#[test]
fn enhance_scanned_file_cleans_the_scan_in_place_through_tools() {
let dir = workdir("enhance-scan");
// A 300 × 400 grey scan: dull paper, a dark gutter at the left, text bars and a speck.
let img = image::GrayImage::from_fn(300, 400, |x, y| {
let light = 0.55 + 0.45 * (x as f32 / 60.0).min(1.0);
let ink = if ((60..360).contains(&y) && (y - 60) % 40 < 6 && (60..240).contains(&x)) || (x, y) == (150, 25) { 0.12 } else { 0.78 };
image::Luma([(ink * light * 255.0) as u8])
});
img.save(dir.join("scan.jpg")).unwrap();
let mut a = auto(&dir);
let scan = ok(&mut a, "doc_create", json!({ "from": "images", "paths": ["scan.jpg"] }))["doc"].as_u64().unwrap();
ok(&mut a, "image_save", json!({ "doc": scan, "page": 1, "image": 1, "path": "before.jpg" }));
let r = ok(&mut a, "page_enhance_scan", json!({ "doc": scan, "level": 100 }));
assert_eq!(r["undo"], "Enhance scanned file");
assert_eq!((r["pages_cleaned"].clone(), r["images"].clone(), r["not_scanned"].clone()), (json!([1]), json!(1), json!([])));
assert_eq!(r["skipped"], json!([]));
// In place: still one picture of the same size, still a JPEG, now with white paper.
let images = ok(&mut a, "page_images", json!({ "doc": scan, "page": 1 }));
assert_eq!((images["count"].clone(), images["images"][0]["pixels"].clone()), (json!(1), json!([300, 400])));
let saved = ok(&mut a, "image_save", json!({ "doc": scan, "page": 1, "image": 1, "path": "after" }));
assert_eq!(saved["format"], "jpg");
let (before, after) = (image::open(dir.join("before.jpg")).unwrap().to_luma8(), image::open(dir.join("after.jpg")).unwrap().to_luma8());
for (x, y) in [(150, 40), (3, 40), (3, 380), (150, 25)] {
assert!(
before.get_pixel(x, y)[0] < 200 && after.get_pixel(x, y)[0] >= 230,
"({x}, {y}): {:?} → {:?}",
before.get_pixel(x, y),
after.get_pixel(x, y)
);
}
assert!(after.get_pixel(150, 62)[0] <= 60, "a stroke stays dark");
// Undo puts the scan back as it was.
ok(&mut a, "edit_undo", json!({ "doc": scan }));
ok(&mut a, "image_save", json!({ "doc": scan, "page": 1, "image": 1, "path": "undone.jpg" }));
assert_eq!(std::fs::read(dir.join("undone.jpg")).unwrap(), std::fs::read(dir.join("before.jpg")).unwrap());
// Switches: despeckle alone leaves everything but the speck.
let r = ok(&mut a, "page_enhance_scan", json!({ "doc": scan, "pages": [1], "level": 0, "despeckle": true, "edge_shadows": false }));
assert_eq!(r["images"], 1);
// A document of text pages has no scans: nothing changes, and it says so.
let text = ok(&mut a, "doc_open", json!({ "path": "a.pdf" }))["doc"].as_u64().unwrap();
let r = ok(&mut a, "page_enhance_scan", json!({ "doc": text }));
assert_eq!((r["images"].clone(), r["not_scanned"].clone(), r["undo"].clone()), (json!(0), json!([1, 2, 3]), Value::Null));
// Bad arguments are refused before anything changes.
for bad in [
json!({ "doc": scan, "level": 101 }),
json!({ "doc": scan, "pages": [9] }),
json!({ "doc": scan, "despeckle": "yes" }),
json!({ "doc": scan, "pages": [] }),
] {
assert!(matches!(a.call("page_enhance_scan", &bad), Err(ToolError::InvalidArgs(_))), "{bad}");
}
}
#[test]
fn enhance_camera_image_finds_the_edges_by_itself() {
let dir = workdir("enhance-camera-auto");
@@ -3969,6 +4123,48 @@ fn fill_sign_text_and_date_colours_are_saved_and_validated() {
}
}
#[test]
fn fill_sign_text_border_is_saved_and_validated() {
let dir = workdir("fill-sign-text-border");
let mut a = auto(&dir);
let doc = ok(&mut a, "doc_open", json!({ "path": "a.pdf" }))["doc"].as_u64().unwrap();
ok(&mut a, "fill_sign_add", json!({ "doc": doc, "page": 1, "type": "text", "at": [20, 40], "text": "Plain" }));
ok(&mut a, "fill_sign_add", json!({ "doc": doc, "page": 1, "type": "text", "at": [20, 80], "text": "Plain", "border": 1.5 }));
let comments = ok(&mut a, "comment_list", json!({ "doc": doc }));
let width = |k: usize| {
let rect = comments["comments"][k]["rect"].as_array().unwrap();
rect[2].as_f64().unwrap() - rect[0].as_f64().unwrap()
};
assert!((width(1) - width(0) - 3.0).abs() < 1e-6, "1.5 pt on each side: {comments}");
ok(&mut a, "doc_save", json!({ "doc": doc, "path": "framed.pdf" }));
let saved = std::fs::read(dir.join("framed.pdf")).unwrap();
let pdf = pdfcraft_cos::Document::open(std::sync::Arc::new(saved)).unwrap();
let page = pdfcraft_model::pages(&pdf).remove(0);
let annotations = pdf.resolve(page.dict.get(b"Annots").unwrap());
let annotation = pdf.resolve(annotations.as_array().unwrap().get(1).unwrap());
let bs = annotation.as_dict().unwrap().get(b"BS").unwrap().as_dict().unwrap().get(b"W").unwrap().as_f64().unwrap();
assert_eq!(bs, 1.5, "saved annotation has the requested border");
// Its own colour, then the text's again through comment_edit.
ok(
&mut a,
"fill_sign_add",
json!({ "doc": doc, "page": 1, "type": "text", "at": [20, 160], "text": "Red", "border": 1, "border_color": "#ff0000" }),
);
let list = ok(&mut a, "comment_list", json!({ "doc": doc, "page": 1 }));
let red = list["comments"].as_array().unwrap().iter().find(|c| c["contents"] == "Red").unwrap()["index"].clone();
ok(&mut a, "comment_edit", json!({ "doc": doc, "page": 1, "index": red, "border_color": "text" }));
assert!(
a.call("fill_sign_add", &json!({ "doc": doc, "page": 1, "type": "text", "at": [20, 200], "text": "No", "border_color": "red" })).is_err(),
"no border"
);
for (kind, border) in [("text", -1.0), ("text", 13.0), ("check", 1.0)] {
assert!(matches!(
a.call("fill_sign_add", &json!({ "doc": doc, "page": 1, "type": kind, "at": [20, 120], "text": "No", "border": border })),
Err(ToolError::InvalidArgs(_))
));
}
}
#[test]
fn image_signatures_through_tools_preserve_transparency_and_survive_save() {
let dir = workdir("image-signatures");
@@ -6437,6 +6633,44 @@ fn drawing_comments_through_tools() {
assert!(matches!(png[0], Content::Png { .. }));
}
/// comment_add `ortho`, Shift's hold while drawing: a line's end is held level or plumb with its
/// start, and each corner of connected lines or a polygon with the corner before.
#[test]
fn lines_and_corners_are_held_level_or_plumb_through_the_tool() {
let dir = workdir("ortho");
let mut a = auto(&dir);
let doc = ok(&mut a, "doc_open", json!({ "path": "a.pdf" }))["doc"].as_u64().unwrap();
for args in [
json!({ "type": "line", "from": [20, 40], "to": [120, 52], "ortho": true }),
json!({ "type": "arrow", "from": [180, 40], "to": [170, 140], "ortho": true }),
json!({ "type": "polyline", "points": [[20, 200], [90, 210], [95, 260], [130, 250]], "ortho": true }),
json!({ "type": "line", "from": [20, 280], "to": [120, 290] }),
] {
let mut args = args;
args.as_object_mut().unwrap().extend([("doc".to_string(), json!(doc)), ("page".to_string(), json!(1))]);
ok(&mut a, "comment_add", args);
}
let list = ok(&mut a, "comment_list", json!({ "doc": doc, "page": 1 }));
// In the order added: the list isn't sorted by index.
let mut listed = list["comments"].as_array().unwrap().clone();
listed.sort_by_key(|c| c["index"].as_u64());
let points: Vec<Vec<[f64; 2]>> =
listed.iter().map(|c| c["points"].as_array().unwrap().iter().map(|p| [p[0].as_f64().unwrap(), p[1].as_f64().unwrap()]).collect()).collect();
let want: [&[[f64; 2]]; 4] = [
&[[20.0, 40.0], [120.0, 40.0]],
&[[180.0, 40.0], [180.0, 140.0]],
&[[20.0, 200.0], [90.0, 200.0], [90.0, 260.0], [130.0, 260.0]],
&[[20.0, 280.0], [120.0, 290.0]],
];
assert_eq!(points.len(), want.len(), "{points:?}");
for (got, want) in points.iter().zip(want) {
assert_eq!(got.len(), want.len(), "{got:?}");
assert!(got.iter().zip(want).all(|(g, w)| (g[0] - w[0]).abs() < 0.01 && (g[1] - w[1]).abs() < 0.01), "{got:?} for {want:?}");
}
let refused = a.call("comment_add", &json!({ "doc": doc, "page": 1, "type": "rectangle", "rect": [10, 10, 50, 50], "ortho": true }));
assert!(matches!(refused, Err(ToolError::InvalidArgs(_))), "ortho is for lines and corners: {refused:?}");
}
#[test]
fn signing_with_an_image_typed_or_drawn_graphic_through_tools() {
let dir = workdir("signing-graphic");
@@ -10194,6 +10428,31 @@ fn changing_a_lists_options_through_tools_keeps_its_saved_selection_on_the_same_
assert_eq!(saved_selection_indices(&dir.join("letters.pdf"), "letters"), Some(vec![0, 1]), "B and C are now items 0 and 1");
}
/// Clearing a field's length limit through `form_set_props` (`"max_length": 0`) leaves it with
/// none, even when its parent field has one it would otherwise inherit.
#[test]
fn clearing_a_kids_length_limit_through_tools_does_not_bring_back_its_parents() {
let objs = [
"<< /Type /Catalog /Pages 2 0 R /AcroForm 4 0 R >>",
"<< /Type /Pages /Kids [3 0 R] /Count 1 /MediaBox [0 0 400 300] >>",
"<< /Type /Page /Parent 2 0 R /Annots [6 0 R] >>",
"<< /Fields [5 0 R] /DA (/Helv 12 Tf 0 g) /DR << /Font << /Helv 7 0 R >> >> >>",
"<< /T (address) /FT /Tx /MaxLen 3 /Kids [6 0 R] >>",
"<< /Type /Annot /Subtype /Widget /T (city) /Parent 5 0 R /MaxLen 8 /Rect [10 250 150 268] /P 3 0 R >>",
"<< /Type /Font /Subtype /Type1 /BaseFont /Helvetica /Encoding /WinAnsiEncoding >>",
];
let dir = workdir("kid-max-len");
std::fs::write(dir.join("form.pdf"), pdf_of(&objs)).unwrap();
let mut a = auto(&dir);
let doc = ok(&mut a, "doc_open", json!({ "path": "form.pdf" }))["doc"].as_u64().unwrap();
ok(&mut a, "form_set_props", json!({ "doc": doc, "field": "address.city", "max_length": 0 }));
ok(&mut a, "form_fill", json!({ "doc": doc, "values": { "address.city": "Absalom" } }));
let fields = ok(&mut a, "form_fields", json!({ "doc": doc }));
let city = fields["fields"].as_array().unwrap().iter().find(|f| f["name"] == "address.city").unwrap().clone();
assert!(city.get("max_length").is_none(), "no limit, not the parent's 3: {city}");
assert_eq!(city["value"], "Absalom", "the value is not cut to the parent's 3 characters");
}
/// Clearing a field's default through `form_set_props` (`"default": ""`) leaves it with none,
/// even when its parent field has one it would otherwise inherit.
#[test]
@@ -10642,3 +10901,214 @@ fn multiline_fill_sign_engine_round_trip_preserves_lines_and_appearance() {
assert_eq!(rendered, re.call("page_render", &json!({"doc":reopened,"page":1,"dpi":72})).unwrap(), "saved appearances are identical");
std::fs::remove_dir_all(dir).unwrap();
}
/// Export, then import into another document and copy between open documents: the tree comes
/// back unchanged, a page the target lacks is reported, and the import undoes in one step.
#[test]
fn bookmarks_export_import_and_copy_between_documents() {
let dir = workdir("bookmark-io");
let mut a = auto(&dir);
let src = ok(&mut a, "doc_open", json!({ "path": "a.pdf" }))["doc"].as_u64().unwrap();
ok(&mut a, "bookmark_add", json!({ "doc": src, "title": "Intro", "page": 1 }));
ok(&mut a, "bookmark_add", json!({ "doc": src, "title": "Detail", "page": 3, "parent": [1] }));
ok(&mut a, "bookmark_add", json!({ "doc": src, "title": "End", "page": 2 }));
ok(&mut a, "bookmark_set_style", json!({ "doc": src, "path": [1], "color": [1, 0, 0], "bold": true }));
let exported = ok(&mut a, "bookmark_export", json!({ "doc": src, "out": "marks.json" }));
assert_eq!(exported["truncated"], false);
let tree = exported["bookmarks"].clone();
assert_eq!((tree[0]["title"].clone(), tree[0]["color"].clone(), tree[0]["bold"].clone()), (json!("Intro"), json!([1.0, 0.0, 0.0]), json!(true)));
assert_eq!((tree[0]["children"][0]["title"].clone(), tree[0]["children"][0]["page"].clone()), (json!("Detail"), json!(3)));
assert_eq!(tree[0]["view"]["type"], "XYZ");
let subtree = ok(&mut a, "bookmark_export", json!({ "doc": src, "path": [1, 1] }))["bookmarks"].clone();
assert_eq!(subtree, json!([tree[0]["children"][0]]));
// b.pdf has two pages: Detail goes to the third, so it arrives without a destination.
let dst = ok(&mut a, "doc_open", json!({ "path": "b.pdf" }))["doc"].as_u64().unwrap();
let imported = ok(&mut a, "bookmark_import", json!({ "doc": dst, "file": "marks.json" }));
assert_eq!(imported["unmapped"], json!([[1, 1]]));
let back = ok(&mut a, "bookmark_export", json!({ "doc": dst }))["bookmarks"].clone();
assert_eq!(back[0]["children"][0]["page"], json!(null));
assert_eq!((back[0]["title"].clone(), back[1].clone()), (tree[0]["title"].clone(), tree[1].clone()));
ok(&mut a, "edit_undo", json!({ "doc": dst }));
assert_eq!(ok(&mut a, "bookmark_list", json!({ "doc": dst }))["bookmarks"], json!([]));
// Into a document with every page, the tree is unchanged, wherever it is copied from.
let same = ok(&mut a, "doc_open", json!({ "path": "a.pdf" }))["doc"].as_u64().unwrap();
ok(&mut a, "bookmark_add", json!({ "doc": same, "title": "Old", "page": 1 }));
let copied = ok(&mut a, "bookmark_copy", json!({ "doc": same, "from": src, "replace": true }));
assert_eq!(copied["unmapped"], json!([]));
assert_eq!(ok(&mut a, "bookmark_export", json!({ "doc": same }))["bookmarks"], tree);
ok(&mut a, "bookmark_import", json!({ "doc": same, "bookmarks": subtree, "parent": [2], "position": 1 }));
assert_eq!(ok(&mut a, "bookmark_list", json!({ "doc": same }))["bookmarks"][1]["children"][0]["title"], "Detail");
}
/// A one-page document whose outline has an untitled item, a /FitR destination with three of
/// its four numbers, and a chain nested `deep` levels below them.
fn odd_outline(deep: usize) -> Vec<u8> {
let first_deep = 7;
let mut objs = vec![
"<< /Type /Catalog /Pages 2 0 R /Outlines 4 0 R >>".to_string(),
"<< /Type /Pages /Kids [3 0 R] /Count 1 /MediaBox [0 0 200 300] >>".into(),
"<< /Type /Page /Parent 2 0 R >>".into(),
"<< /Type /Outlines /First 5 0 R /Last 6 0 R >>".into(),
"<< /Title () /Parent 4 0 R /Next 6 0 R /Dest [3 0 R /Fit] >>".into(),
format!(
"<< /Title (Short) /Parent 4 0 R /Prev 5 0 R /Dest [3 0 R /FitR 0 0 10]{} >>",
if deep > 0 { format!(" /First {first_deep} 0 R /Last {first_deep} 0 R") } else { String::new() }
),
];
for i in 0..deep {
let (me, parent) = (first_deep + i, if i == 0 { 6 } else { first_deep + i - 1 });
let child = if i + 1 < deep { format!(" /First {0} 0 R /Last {0} 0 R", me + 1) } else { String::new() };
objs.push(format!("<< /Title (Level) /Parent {parent} 0 R /Dest [3 0 R /Fit]{child} >>"));
}
pdf_of(&objs.iter().map(String::as_str).collect::<Vec<_>>())
}
/// What a real file holds and the JSON form can't say exactly still copies and round-trips: an
/// empty title stays empty, a short /FitR keeps its page and drops the view, and bookmarks past
/// the listing depth are reported as left out at every step.
#[test]
fn odd_bookmarks_round_trip_and_a_cut_tree_says_so() {
let dir = workdir("bookmark-io-odd");
std::fs::write(dir.join("odd.pdf"), odd_outline(0)).unwrap();
std::fs::write(dir.join("deep.pdf"), odd_outline(40)).unwrap();
let mut a = auto(&dir);
let src = ok(&mut a, "doc_open", json!({ "path": "odd.pdf" }))["doc"].as_u64().unwrap();
let tree = ok(&mut a, "bookmark_export", json!({ "doc": src, "out": "odd.json" }))["bookmarks"].clone();
assert_eq!((tree[0]["title"].clone(), tree[1]["page"].clone(), tree[1].get("view")), (json!(""), json!(1), None));
let dst = ok(&mut a, "doc_open", json!({ "path": "b.pdf" }))["doc"].as_u64().unwrap();
assert_eq!(ok(&mut a, "bookmark_import", json!({ "doc": dst, "file": "odd.json", "replace": true }))["truncated"], false);
// A destination with no view is written as XYZ with nulls: the same page, the reader's view.
let back = ok(&mut a, "bookmark_export", json!({ "doc": dst }))["bookmarks"].clone();
assert_eq!((back[0].clone(), back[1]["title"].clone(), back[1]["page"].clone()), (tree[0].clone(), json!("Short"), json!(1)));
assert_eq!(back[1]["view"], json!({ "type": "XYZ", "values": [null, null, null] }));
let copied = ok(&mut a, "bookmark_copy", json!({ "doc": dst, "from": src, "replace": true }));
assert_eq!((copied["unmapped"].clone(), copied["truncated"].clone()), (json!([]), json!(false)));
match a.call("bookmark_export", &json!({ "doc": src, "path": [3] })) {
Err(ToolError::Failed(m)) => assert!(m.contains("no bookmark at [3]"), "{m}"),
other => panic!("expected no bookmark, got {other:?}"),
}
let deep = ok(&mut a, "doc_open", json!({ "path": "deep.pdf" }))["doc"].as_u64().unwrap();
assert_eq!(ok(&mut a, "bookmark_export", json!({ "doc": deep, "out": "deep.json" }))["truncated"], true);
let saved: Value = serde_json::from_slice(&std::fs::read(dir.join("deep.json")).unwrap()).unwrap();
assert_eq!(saved["truncated"], true);
assert_eq!(ok(&mut a, "bookmark_import", json!({ "doc": dst, "file": "deep.json", "replace": true }))["truncated"], true);
assert_eq!(ok(&mut a, "bookmark_copy", json!({ "doc": dst, "from": deep, "replace": true }))["truncated"], true);
let export = tools().into_iter().find(|t| t.name == "bookmark_export").unwrap();
assert!(export.destructive && !export.read_only, "bookmark_export overwrites out");
std::fs::remove_dir_all(dir).unwrap();
}
#[test]
fn a_misspelt_bookmark_key_is_refused_not_dropped() {
let dir = workdir("bookmark-io-key");
let mut a = auto(&dir);
let doc = ok(&mut a, "doc_open", json!({ "path": "a.pdf" }))["doc"].as_u64().unwrap();
let tree = json!([{ "title": "Parent", "children": [{ "title": "Red", "page": 1, "colour": [1, 0, 0] }] }]);
match a.call("bookmark_import", &json!({ "doc": doc, "bookmarks": tree })) {
Err(ToolError::InvalidArgs(m)) => assert!(m.contains("[1, 1]") && m.contains("\"colour\""), "{m}"),
other => panic!("expected InvalidArgs, got {other:?}"),
}
assert_eq!(ok(&mut a, "bookmark_list", json!({ "doc": doc }))["bookmarks"], json!([]));
std::fs::remove_dir_all(dir).unwrap();
}
#[test]
fn a_bad_bookmark_tree_is_refused_whole_and_names_the_node() {
let dir = workdir("bookmark-io-bad");
let mut a = auto(&dir);
let doc = ok(&mut a, "doc_open", json!({ "path": "a.pdf" }))["doc"].as_u64().unwrap();
let refused = |a: &mut Automation, bookmarks: Value| match a.call("bookmark_import", &json!({ "doc": doc, "bookmarks": bookmarks })) {
Err(ToolError::InvalidArgs(m)) => m,
other => panic!("expected InvalidArgs, got {other:?}"),
};
let fine = json!({ "title": "Fine", "page": 1 });
let m = refused(&mut a, json!([fine, { "title": "Parent", "children": [{ "page": 1 }] }]));
assert!(m.contains("[2, 1]") && m.contains("title"), "{m}");
let m = refused(&mut a, json!([{ "title": 5 }]));
assert!(m.contains("[1]") && m.contains("title"), "{m}");
let m = refused(&mut a, json!([fine, { "title": "Bad", "page": 0 }]));
assert!(m.contains("[2]") && m.contains("page"), "{m}");
let m = refused(&mut a, json!([{ "title": "Bad", "color": [2, 0, 0] }]));
assert!(m.contains("color"), "{m}");
let m = refused(&mut a, json!([{ "title": "Bad", "view": { "type": "XYZ", "values": [1, 2, 3, 4, 5] } }]));
assert!(m.contains("view"), "{m}");
let mut deep = json!({ "title": "Leaf" });
for _ in 0..40 {
deep = json!({ "title": "Level", "children": [deep] });
}
assert!(refused(&mut a, json!([deep])).contains("33 levels"));
let m = refused(&mut a, json!([fine, { "title": "Z", "page": 1, "view": { "type": "Zoom" } }]));
assert!(m.contains("[2]") && m.contains("XYZ (3 values"), "{m}");
let m = refused(&mut a, json!([fine, { "title": "Z", "page": 1, "view": { "type": "FitH", "values": [1, 2] } }]));
assert!(m.contains("[2]") && m.contains("FitH view takes 1"), "{m}");
let m = refused(&mut a, json!([{ "title": "R", "view": { "type": "FitR", "values": [0, null, 1, 1] } }]));
assert!(m.contains("all four"), "{m}");
assert!(matches!(a.call("bookmark_import", &json!({ "doc": doc })), Err(ToolError::InvalidArgs(_))));
assert!(matches!(a.call("bookmark_import", &json!({ "doc": doc, "bookmarks": [fine], "file": "x.json" })), Err(ToolError::InvalidArgs(_))));
// A real bookmark file outside the root is refused for being outside, not for being missing.
let outside = dir.with_extension("outside.json");
std::fs::write(&outside, br#"{"bookmarks": [{"title": "Outside", "page": 1}]}"#).unwrap();
let name = format!("../{}", outside.file_name().unwrap().to_string_lossy());
match a.call("bookmark_import", &json!({ "doc": doc, "file": name })) {
Err(ToolError::Failed(m)) => assert!(m.contains("outside the allowed directory"), "{m}"),
other => panic!("expected the root to refuse it, got {other:?}"),
}
let _ = std::fs::remove_file(&outside);
let out = dir.with_extension("out.json");
let name = format!("../{}", out.file_name().unwrap().to_string_lossy());
match a.call("bookmark_export", &json!({ "doc": doc, "out": name })) {
Err(ToolError::Failed(m)) => assert!(m.contains("outside the allowed directory"), "{m}"),
other => panic!("expected the root to refuse it, got {other:?}"),
}
assert!(!out.exists());
assert_eq!(ok(&mut a, "bookmark_list", json!({ "doc": doc }))["bookmarks"], json!([]));
}
/// Copy and paste comments between documents through the tools: the copies carry their replies,
/// land on the asked-for page under their own names (a second paste gives a second set), and a
/// plain text is refused.
#[test]
fn comments_copy_and_paste_between_documents_through_tools() {
let dir = workdir("comments-clipboard");
let mut a = auto(&dir);
let src = ok(&mut a, "doc_open", json!({ "path": "a.pdf" }))["doc"].as_u64().unwrap();
ok(&mut a, "comment_add", json!({ "doc": src, "page": 2, "type": "note", "at": [30, 30], "contents": "Please review", "author": "Ada" }));
ok(&mut a, "comment_add", json!({ "doc": src, "page": 2, "type": "rectangle", "rect": [20, 20, 80, 60], "contents": "Box" }));
ok(&mut a, "comment_reply", json!({ "doc": src, "page": 2, "index": 1, "text": "Done", "author": "Bob" }));
// Positions as comment_list reports them (a note's pop-up takes a position of its own).
let listed = ok(&mut a, "comment_list", json!({ "doc": src, "page": 2 }));
let pairs: Vec<Value> = listed["comments"].as_array().unwrap().iter().map(|c| json!([c["page"], c["index"]])).collect();
assert_eq!(pairs.len(), 2, "{listed}");
let copied = ok(&mut a, "comments_copy", json!({ "doc": src, "comments": pairs }));
assert_eq!(copied["comments"], 2);
let xfdf = copied["xfdf"].as_str().unwrap().to_owned();
assert!(xfdf.contains("inreplyto=") && xfdf.contains("<square"), "{xfdf}");
let dst = ok(&mut a, "doc_open", json!({ "path": "b.pdf" }))["doc"].as_u64().unwrap();
let r = ok(&mut a, "comments_paste", json!({ "doc": dst, "page": 1, "xfdf": xfdf }));
assert_eq!(r["comments"], 3, "two comments and the reply: {r}");
assert_eq!(r["undo"], "Paste comments");
let list = ok(&mut a, "comment_list", json!({ "doc": dst }));
let comments = list["comments"].as_array().unwrap();
assert_eq!(comments.len(), 2, "the reply is in its thread, not a comment of its own: {list}");
assert!(comments.iter().all(|c| c["page"] == 1), "{list}");
assert_eq!(comments.iter().filter(|c| c["contents"] == "Please review").count(), 1);
// The same again: a second set, not a replacement of the first.
ok(&mut a, "comments_paste", json!({ "doc": dst, "page": 1, "xfdf": copied["xfdf"] }));
assert_eq!(ok(&mut a, "comment_list", json!({ "doc": dst }))["comments"].as_array().unwrap().len(), 4);
// Placed: the copies' box is centred at `at` (top-left page coordinates, like comment_add).
let r = ok(&mut a, "comments_paste", json!({ "doc": dst, "page": 2, "xfdf": copied["xfdf"], "at": [100, 150] }));
assert_eq!(r["comments"], 3);
let placed = ok(&mut a, "comment_list", json!({ "doc": dst, "page": 2 }));
let rects: Vec<Vec<f64>> =
placed["comments"].as_array().unwrap().iter().map(|c| c["rect"].as_array().unwrap().iter().map(|v| v.as_f64().unwrap()).collect()).collect();
let (x0, y0, x1, y1) =
rects.iter().fold((f64::MAX, f64::MAX, f64::MIN, f64::MIN), |(a, b, c, d), r| (a.min(r[0]), b.min(r[1]), c.max(r[2]), d.max(r[3])));
assert!(((x0 + x1) / 2.0 - 100.0).abs() < 1.0 && ((y0 + y1) / 2.0 - 150.0).abs() < 1.0, "centred at [100, 150]: {rects:?}");
assert!(matches!(a.call("comments_paste", &json!({ "doc": dst, "page": 1, "xfdf": "hello" })), Err(ToolError::InvalidArgs(_))));
assert!(a.call("comments_copy", &json!({ "doc": src, "comments": [[9, 9]] })).is_err(), "no such comment");
assert!(matches!(a.call("comments_copy", &json!({ "doc": src, "comments": [[2]] })), Err(ToolError::InvalidArgs(_))));
assert_eq!(ok(&mut a, "comment_list", json!({ "doc": src }))["comments"].as_array().unwrap().len(), 2, "the source is untouched");
}
crates/automation/tests/it/automation_roots.rs
+4
View File
@@ -145,6 +145,8 @@ fn tools_that_read_a_path_are_held_to_the_read_root() {
ok(&mut server, "doc_open", json!({"path": s(&base.join("in/a.pdf"))}));
let (secret, pic, gone) = (s(&base.join("outside/secret.pdf")), s(&base.join("outside/pic.png")), s(&base.join("outside/gone.bin")));
let out = s(&base.join("out/result.pdf"));
std::fs::write(base.join("in/m.prc"), b"PRC model").unwrap();
let model = s(&base.join("in/m.prc"));
for (tool, args) in [
("page_insert_file", json!({"doc": 1, "path": secret, "at": 1})),
("page_replace", json!({"doc": 1, "pages": [1], "path": secret})),
@@ -153,6 +155,8 @@ fn tools_that_read_a_path_are_held_to_the_read_root() {
("doc_create", json!({"from": "images", "paths": [pic]})),
("stamp_custom", json!({"doc": 1, "page": 1, "path": secret, "at": [10, 10]})),
("stamp_custom", json!({"doc": 1, "page": 1, "path": secret, "rect": [10, 10, 493, 300]})),
("model_add", json!({"doc": 1, "page": 1, "path": secret, "rect": [10, 10, 100, 100]})),
("model_add", json!({"doc": 1, "page": 1, "path": model, "rect": [10, 10, 100, 100], "poster": pic})),
("page_add_image", json!({"doc": 1, "page": 1, "path": pic})),
("image_edit", json!({"doc": 1, "page": 1, "image": 1, "action": "replace", "path": pic})),
("comment_add", json!({"doc": 1, "page": 1, "type": "attachment", "at": [10, 10], "path": secret})),
crates/automation/tests/it/main.rs
+2
View File
@@ -8,6 +8,8 @@ mod batch;
mod button_caption;
mod form_bulk;
mod hostile_args;
mod model_3d;
mod model_render;
mod optimize_content_boundaries;
mod redact_after_stray_q_form;
mod redact_text_after_q;
crates/automation/tests/it/model_3d.rs
+120
View File
@@ -0,0 +1,120 @@
//! `model_add` embeds a U3D or PRC model as a 3D annotation (ISO 32000-2 §13.6.2) with a poster,
//! end to end: the tool, the saved file, undo, and the refusals.
use pdfcraft_automation::{Automation, Content, ToolError};
use pdfcraft_cos::{Dict, Document, Object, Stream};
use serde_json::{Value, json};
fn tool(a: &mut Automation, name: &str, args: Value) -> Value {
match a.call(name, &args).unwrap().remove(0) {
Content::Json(value) => value,
other => panic!("expected JSON: {other:?}"),
}
}
fn png(w: u32, h: u32) -> Vec<u8> {
let mut out = Vec::new();
let mut encoder = png::Encoder::new(&mut out, w, h);
encoder.set_color(png::ColorType::Rgb);
encoder.set_depth(png::BitDepth::Eight);
encoder.write_header().unwrap().write_image_data(&vec![90; (w * h * 3) as usize]).unwrap();
out
}
fn stream(doc: &Document, obj: &Object) -> Stream {
match doc.resolve(obj).as_ref() {
Object::Stream(s) => s.clone(),
other => panic!("expected a stream: {other:?}"),
}
}
/// The 3D annotations on the first page of the saved file.
fn saved_3d(dir: &std::path::Path) -> (Document, Vec<Dict>) {
let doc = Document::open(std::sync::Arc::new(std::fs::read(dir.join("out.pdf")).unwrap())).unwrap();
let page = &pdfcraft_model::pages(&doc)[0];
let list = page.dict.get(b"Annots").map(|a| doc.resolve(a).as_array().cloned().unwrap()).unwrap_or_default();
let found = list.iter().map(|o| doc.resolve(o).as_dict().unwrap().clone()).filter(|d| d.name(b"Subtype") == Some(&b"3D"[..])).collect();
(doc, found)
}
#[test]
fn model_add_embeds_the_model_with_a_fitted_poster() {
let dir = tempfile::tempdir().unwrap();
let mut prc = b"PRC".to_vec();
prc.extend_from_slice(&[7; 200]);
std::fs::write(dir.path().join("bracket.prc"), &prc).unwrap();
std::fs::write(dir.path().join("poster.png"), png(80, 40)).unwrap();
let mut a = Automation::new().with_root(dir.path()).unwrap();
let doc = tool(&mut a, "doc_create", json!({ "from": "blank", "width": 600, "height": 800 }))["doc"].as_u64().unwrap();
// A 300 × 300 box from the top-left of the page: user space [100, 400, 400, 700].
let call =
json!({ "doc": doc, "page": 1, "path": "bracket.prc", "rect": [100, 100, 400, 400], "poster": "poster.png", "activation": "page_open" });
tool(&mut a, "model_add", call);
tool(&mut a, "doc_save", json!({ "doc": doc, "path": "out.pdf" }));
let (saved, found) = saved_3d(dir.path());
let [annot] = found.as_slice() else { panic!("one 3D annotation: {found:?}") };
let rect: Vec<f64> = annot.get(b"Rect").unwrap().as_array().unwrap().iter().filter_map(Object::as_f64).collect();
assert_eq!(rect, [100.0, 400.0, 400.0, 700.0]);
assert_eq!(annot.get(b"Contents").and_then(Object::as_string).map(|s| s.to_text()).as_deref(), Some("bracket"));
let model = stream(&saved, annot.get(b"3DD").unwrap());
assert_eq!(model.dict.name(b"Type"), Some(&b"3D"[..]));
assert_eq!(model.dict.name(b"Subtype"), Some(&b"PRC"[..]));
assert_eq!(model.dict.name(b"Filter"), Some(&b"FlateDecode"[..]));
assert_eq!(model.decoded().unwrap().to_vec(), prc, "the model is stored unchanged");
let activation = saved.resolve(annot.get(b"3DA").unwrap()).as_dict().unwrap().clone();
assert_eq!(activation.name(b"A"), Some(&b"PO"[..]));
// The 80 × 40 picture fitted in the 300 × 300 box: 300 × 150, centred top to bottom.
let ap = saved.resolve(annot.get(b"AP").unwrap()).as_dict().unwrap().clone();
let poster = stream(&saved, ap.get(b"N").unwrap());
let content = String::from_utf8(poster.decoded().unwrap().to_vec()).unwrap();
assert!(content.contains("300 0 0 150 100 475 cm /Pic Do"), "{content}");
assert_eq!(tool(&mut a, "edit_undo", json!({ "doc": doc }))["undone"], "Add 3D model");
tool(&mut a, "doc_save", json!({ "doc": doc, "path": "out.pdf" }));
assert!(saved_3d(dir.path()).1.is_empty(), "undo removes it");
}
#[test]
fn model_add_without_a_poster_draws_a_labelled_frame_and_activates_on_click() {
let dir = tempfile::tempdir().unwrap();
let mut u3d = b"U3D\0".to_vec();
u3d.extend_from_slice(&[0; 60]);
std::fs::write(dir.path().join("part.u3d"), &u3d).unwrap();
let mut a = Automation::new().with_root(dir.path()).unwrap();
let doc = tool(&mut a, "doc_create", json!({ "from": "blank", "width": 600, "height": 800 }))["doc"].as_u64().unwrap();
tool(&mut a, "model_add", json!({ "doc": doc, "page": 1, "path": "part.u3d", "rect": [50, 50, 250, 200], "name": "Part" }));
tool(&mut a, "doc_save", json!({ "doc": doc, "path": "out.pdf" }));
let (saved, found) = saved_3d(dir.path());
let [annot] = found.as_slice() else { panic!("one 3D annotation: {found:?}") };
assert_eq!(stream(&saved, annot.get(b"3DD").unwrap()).dict.name(b"Subtype"), Some(&b"U3D"[..]));
let activation = saved.resolve(annot.get(b"3DA").unwrap()).as_dict().unwrap().clone();
assert_eq!(activation.name(b"A"), Some(&b"XA"[..]), "the standard's default, stated");
let ap = saved.resolve(annot.get(b"AP").unwrap()).as_dict().unwrap().clone();
let content = String::from_utf8(stream(&saved, ap.get(b"N").unwrap()).decoded().unwrap().to_vec()).unwrap();
assert!(content.contains("re S") && content.contains("(3D model) Tj"), "{content}");
}
#[test]
fn model_add_refuses_what_is_not_a_model_and_adds_nothing() {
let dir = tempfile::tempdir().unwrap();
std::fs::write(dir.path().join("notes.txt"), b"PR not a model").unwrap();
std::fs::write(dir.path().join("part.prc"), b"PRC....").unwrap();
let mut a = Automation::new().with_root(dir.path()).unwrap();
let doc = tool(&mut a, "doc_create", json!({ "from": "blank", "width": 600, "height": 800 }))["doc"].as_u64().unwrap();
let not_a_model = a.call("model_add", &json!({ "doc": doc, "page": 1, "path": "notes.txt", "rect": [10, 10, 100, 100] }));
assert!(matches!(&not_a_model, Err(ToolError::Failed(m)) if m.contains("U3D") && m.contains("PRC")), "{not_a_model:?}");
for (args, why) in [
(json!({ "rect": [10, 10, 10, 100] }), "zero width"),
(json!({ "rect": [0, 0, 20000, 100] }), "wider than a PDF page can be"),
(json!({}), "no rect"),
(json!({ "rect": [10, 10, 100, 100], "activation": "now" }), "an unknown activation"),
] {
let mut call = json!({ "doc": doc, "page": 1, "path": "part.prc" });
call.as_object_mut().unwrap().extend(args.as_object().unwrap().clone());
assert!(matches!(a.call("model_add", &call), Err(ToolError::InvalidArgs(_))), "{why}");
}
tool(&mut a, "doc_save", json!({ "doc": doc, "path": "out.pdf" }));
assert!(saved_3d(dir.path()).1.is_empty());
}
crates/automation/tests/it/model_render.rs
+112
View File
@@ -0,0 +1,112 @@
//! `model_render` draws a page's 3D model headlessly: a synthetic PDF holding a synthetic PRC.
use std::path::PathBuf;
use pdfcraft_automation::{Automation, Content};
use serde_json::{Value, json};
/// One triangle at z = 0 facing +z, with corners (-1,-1), (1,-1) and (0,1), as a PRC file of one
/// file structure with no product tree, in hexadecimal for the stream's `/ASCIIHexDecode`.
const TRIANGLE_PRC_HEX: &str = "505243c91f0000c91f00000100000002000000030000000400000000000000000000000000000000000000010000000500000006000000070000000800000000000000060000006b0000009a000000a3000000ab000000cd000000d5000000dd000000e500000000000000505243c91f0000c91f0000050000000600000007000000080000000000000000000000000000000000000000000000789c63000000010001789c030000000001789c9b71c087758da2422b43084fe3140e038713464cd759381a162800006ed70828789c030000000001789c030000000001789c030000000001";
/// A one-page PDF whose page holds a 3D annotation over a stream of `/Subtype /{subtype}`.
fn pdf(subtype: &str, hex: &str) -> Vec<u8> {
let stream = format!("<< /Type /3D /Subtype /{subtype} /Filter /ASCIIHexDecode /Length {} >>\nstream\n{hex}>\nendstream", hex.len() + 1);
let objs: Vec<String> = vec![
"<< /Type /Catalog /Pages 2 0 R >>".into(),
"<< /Type /Pages /Kids [3 0 R] /Count 1 /MediaBox [0 0 200 200] >>".into(),
"<< /Type /Page /Parent 2 0 R /Annots [<< /Type /Annot /Subtype /Square /Rect [0 0 10 10] >> 4 0 R] >>".into(),
"<< /Type /Annot /Subtype /3D /Rect [20 20 180 180] /3DD 5 0 R >>".into(),
stream,
];
let mut out = b"%PDF-1.7\n".to_vec();
let mut offsets = Vec::new();
for (i, o) in objs.iter().enumerate() {
offsets.push(out.len());
out.extend_from_slice(format!("{} 0 obj\n{o}\nendobj\n", i + 1).as_bytes());
}
let xref = out.len();
out.extend_from_slice(format!("xref\n0 {}\n0000000000 65535 f \n", objs.len() + 1).as_bytes());
for o in offsets {
out.extend_from_slice(format!("{o:010} 00000 n \n").as_bytes());
}
out.extend_from_slice(format!("trailer\n<< /Size {} /Root 1 0 R >>\nstartxref\n{xref}\n%%EOF\n", objs.len() + 1).as_bytes());
out
}
/// A fresh folder with `prc.pdf` and `u3d.pdf` in `in`, `outside.pdf` beside it, and an
/// automation rooted at `in`.
fn setup(test: &str) -> (PathBuf, Automation) {
let dir = std::env::temp_dir().join(format!("pdfcraft-model3d-{test}-{}", std::process::id()));
let _ = std::fs::remove_dir_all(&dir);
std::fs::create_dir_all(dir.join("in")).unwrap();
std::fs::write(dir.join("in/prc.pdf"), pdf("PRC", TRIANGLE_PRC_HEX)).unwrap();
std::fs::write(dir.join("in/u3d.pdf"), pdf("U3D", "553344")).unwrap();
std::fs::write(dir.join("outside.pdf"), pdf("PRC", TRIANGLE_PRC_HEX)).unwrap();
let a = Automation::new().with_root(dir.join("in")).unwrap();
(dir, a)
}
fn open(a: &mut Automation, path: &str) -> u64 {
match a.call("doc_open", &json!({ "path": path })).unwrap().remove(0) {
Content::Json(v) => v["doc"].as_u64().unwrap(),
other => panic!("expected JSON: {other:?}"),
}
}
/// The RGB of pixel (x, y) of a PNG.
fn rgb(data: &[u8], x: usize, y: usize) -> [u8; 3] {
let mut reader = png::Decoder::new(std::io::Cursor::new(data)).read_info().unwrap();
let mut pixels = vec![0; reader.output_buffer_size().unwrap()];
let info = reader.next_frame(&mut pixels).unwrap();
let channels = info.line_size / info.width as usize;
let at = y * info.line_size + x * channels;
[pixels[at], pixels[at + 1], pixels[at + 2]]
}
#[test]
fn a_prc_model_is_drawn_over_the_background() {
let (_dir, mut a) = setup("prc");
let doc = open(&mut a, "prc.pdf");
let args = json!({ "doc": doc, "page": 1, "width": 64, "height": 48, "direction": [0, 0, -1], "perspective": false });
let out = a.call("model_render", &args).unwrap();
let Content::Png { data, width, height } = &out[0] else { panic!("expected a PNG") };
assert_eq!((*width, *height), (64, 48));
assert_eq!(rgb(data, 0, 0), [255, 255, 255], "the corner is background");
assert_eq!(rgb(data, 32, 24), [190, 192, 200], "the centre is the triangle, lit head on");
// In perspective the headlight meets the face at a slight angle, so only not-background holds.
let out = a.call("model_render", &json!({ "doc": doc, "page": 1, "width": 64, "height": 48, "direction": [0, 0, -1] })).unwrap();
let Content::Png { data, .. } = &out[0] else { panic!("expected a PNG") };
assert_ne!(rgb(data, 32, 24), [255, 255, 255], "the centre is drawn in perspective too");
// The same model named explicitly, and the default isometric view, both draw.
assert!(a.call("model_render", &json!({ "doc": doc, "page": 1, "model": 1 })).is_ok());
}
#[test]
fn a_u3d_model_or_a_missing_one_is_a_clear_error() {
let (_dir, mut a) = setup("u3d");
let doc = open(&mut a, "u3d.pdf");
let err = a.call("model_render", &json!({ "doc": doc, "page": 1 })).unwrap_err().to_string();
assert!(err.contains("U3D"), "{err}");
let doc = open(&mut a, "prc.pdf");
let err = a.call("model_render", &json!({ "doc": doc, "page": 1, "model": 2 })).unwrap_err().to_string();
assert!(err.contains("has 1 3D models, not 2"), "{err}");
for bad in [json!({ "direction": [0, 0, 0] }), json!({ "direction": [1, 2] }), json!({ "width": 0 }), json!({ "height": 5000 })] {
let mut args = json!({ "doc": doc, "page": 1 });
args.as_object_mut().unwrap().extend(bad.as_object().unwrap().clone());
assert!(a.call("model_render", &args).is_err(), "{bad}");
}
}
#[test]
fn a_model_outside_the_root_is_never_opened_to_be_drawn() {
let (dir, mut a) = setup("root");
let outside = dir.join("outside.pdf");
for path in [Value::from("../outside.pdf"), Value::from(outside.to_string_lossy().into_owned())] {
assert!(a.call("doc_open", &json!({ "path": path })).is_err(), "{path} lies outside the root");
}
// The tool takes no path of its own: only a document opened inside the root reaches it.
assert!(a.call("model_render", &json!({ "doc": 1, "page": 1 })).is_err());
let doc = open(&mut a, "prc.pdf");
assert!(a.call("model_render", &json!({ "doc": doc, "page": 1 })).is_ok());
}
crates/cos/README.md
+19 -5
View File
@@ -53,7 +53,18 @@ another. Edits never touch the original bytes.
- **Every cross-reference form:** classic tables with their quirky subsections, cross-reference
streams, hybrid-reference files (`/XRefStm`), and object streams.
- **Repair.** A file whose xref is wrong or missing is reconstructed by scanning for `obj`
headers. Repairs are *recorded*, never silent, so the UI can say the file was damaged.
headers. Repairs are *recorded*, never silent, so the UI can say the file was damaged. A live
in-file reference may match either its xref generation or a same-number object header at
that indexed offset; a missing, free, or neither-matching slot still resolves to null.
Header probes inspect at most 128 bytes at the indexed and header-relative offsets, without
parsing object bodies. On open, disagreements in retained xref revisions are recorded once
per object, up to 256 objects plus an omitted-details summary. This diagnostic scan does
not expand compact xref ranges; the winning xref entry remains authoritative for free
slots. Compressed objects require generation zero. Explicit `set` overlays accept only the
supplied generation; `set_existing` and `update_dict` retain both accepted generations of a
damaged live slot, including through snapshots, and require full rewrite. Headers beyond
this bounded probe are not additionally recovered. Materialized primary and hybrid rows are
diagnosed before folding, so a newer row does not hide a retained disagreement.
- **Tolerant, never fatal.** The lexer works on a byte slice and a position; it allocates nothing
for skipped data, bounds-checks every index, and caps nesting depth. Malformed input yields
`CosError`, never a panic — this crate is the first thing a hostile file meets. Exhausted xref entries are skipped with a repair log; exhausted /Size hints allocate unused numbers without overflowing or replacing existing objects.
@@ -66,14 +77,17 @@ another. Edits never touch the original bytes.
## Writing
`write_incremental` appends only the edited objects, a cross-reference section **in the same
style as the file's last one**, and a trailer with `/Prev`. The original bytes are never modified,
which is what keeps a signature over an earlier revision valid.
`write_incremental` ordinarily appends only the edited objects, a cross-reference section **in
the same style as the file's last one**, and a trailer with `/Prev`, preserving the original
bytes and signatures over earlier revisions. It delegates to `write_full` when reconstruction,
encryption changes or `full_save_required()` demand a rewrite. Repair-aware edits to a live
xref/header generation disagreement use this fallback rather than append a contradictory row.
`write_full` writes only what the trailer can reach, renumbered densely — the Save As and
garbage-collecting path. By default (`SaveOptions::object_streams`) it packs non-stream objects
into compressed object streams with a cross-reference stream (PDF 1.5+); turning that off writes a
classic table any reader accepts.
classic table any reader accepts. Both forms map accepted xref/header aliases to one output
object, retaining shared identity such as page-tree parent links; unrelated generations stay null.
Dictionaries keep their key order and streams keep their *encoded* bytes, so rewritten objects stay
recognisable in a diff and byte-stable when re-serialized.
crates/cos/src/budget.rs
+6
View File
@@ -25,6 +25,12 @@ pub const MAX_PAGE_DECODED: usize = 256 << 20;
/// Form XObjects one page's content walk may enter, each entry counted again.
pub const MAX_PAGE_FORMS: usize = 50_000;
/// The most pixels an image redaction decodes, to clear the ones under a mark or to store again
/// without the `/JBIG2Globals` it shares with a redacted image (`pdfcraft_redact`). Optimize has
/// JBIG2 images share a dictionary only within it (`pdfcraft_optimize`), so that redaction can
/// store each of them again: one it couldn't would refuse the redaction.
pub const MAX_REDACTED_IMAGE_PIXELS: u64 = 64 << 20;
/// The steps a sample of a JPEG costs: a few bytes can hold a page of one grey, and decoding,
/// resampling and storing it took about 20 to 45 ns a sample.
const JPEG_WORK: u64 = 2;
crates/cos/src/document.rs
+5
View File
@@ -19,6 +19,7 @@ mod reader;
mod state;
mod xref;
pub(crate) use objects::Held;
#[cfg(test)]
pub(crate) use reader::find;
#[cfg(test)]
@@ -62,6 +63,10 @@ pub struct Document {
security: Option<Arc<pdfcraft_crypt::SecurityHandler>>,
/// Object number of the `/Encrypt` dictionary (never encrypted itself).
encrypt_num: Option<u32>,
/// What that dictionary and `/ID` refer to, which the key comes from, and so is as stored.
encrypt_members: Arc<HashSet<u32>>,
/// Object numbers a scan of the file didn't find: the file doesn't change, so it never will.
unfound: Arc<Mutex<HashSet<u32>>>,
/// Encryption was added, changed or removed since opening: only a full save can apply it.
encryption_changed: bool,
/// Set by edits that must not leave earlier revisions in the file (redaction).
crates/cos/src/document/objects.rs
+194 -29
View File
@@ -1,6 +1,8 @@
use std::collections::HashSet;
use super::reader::find;
use super::state::{LENGTHS, LoadGuard, ObjStm, ObjectReader, ReadState, Slot, SourceIdentity, held_bytes, objstm_member};
use super::xref::{Revision, XrefEntry};
use super::xref::{Revision, XrefEntry, XrefIndex};
use super::{
Arc, CosError, Cow, Dict, Document, Lexer, MAX_DECODED, MAX_JBIG2_GLOBALS, ObjRef, Object, Stream, is_whitespace, parse_indirect_shared,
};
@@ -129,8 +131,8 @@ impl Document {
pub fn xref_entry_is_exact(&self, num: u32) -> bool {
match self.xref_entry(num) {
None | Some(XrefEntry::Free { .. }) => true,
Some(XrefEntry::InFile { offset, .. }) => {
let at = |off: usize| matches!(parse_indirect_shared(&self.data, off, &|_| None, &self.endstreams), Ok((id, _)) if id.num == num);
Some(XrefEntry::InFile { offset, generation }) => {
let at = |off: usize| matches!(parse_indirect_shared(&self.data, off, &|_| None, &self.endstreams), Ok((id, _)) if id == ObjRef::new(num, generation));
// Offsets may count from a header that does not start the file.
usize::try_from(offset).is_ok_and(|off| at(off) || (self.header_offset != 0 && at(off.saturating_add(self.header_offset))))
}
@@ -151,19 +153,107 @@ impl Document {
// ── object access ───────────────────────────────────────────────────────────────────────
/// Fetch an object by reference. Missing objects are `Null` (ISO 32000-2 §7.3.10).
/// Fetch an object by reference. Undefined references are `Null` (ISO 32000-2 §7.3.10).
/// A live file slot also accepts its actual header generation when a damaged xref row
/// disagrees; the disagreement is recorded in [`Self::repair_log`]. Free slots never do.
pub fn get(&self, r: ObjRef) -> Arc<Object> {
self.try_get(r.num).unwrap_or_else(|_| Arc::new(Object::Null))
self.try_get_reference(r).unwrap_or_else(|_| Arc::new(Object::Null))
}
/// Follow a reference if `o` is one; otherwise return `o` itself.
pub fn resolve(&self, o: &Object) -> Arc<Object> {
match o {
Object::Ref(r) => self.get(*r),
other => Arc::new(other.clone()),
pub(super) fn reference_matches(&self, r: ObjRef) -> bool {
match self.overlay.get(&r.num) {
Some(Slot::Set(generation, _)) => r.generation == *generation,
Some(Slot::RepairedSet(generation, alias, _)) => r.generation == *generation || r.generation == *alias,
Some(Slot::Freed(_)) => false,
None => match self.entries.lookup(r.num, &self.data) {
Some(XrefEntry::InFile { offset, generation }) => {
r.generation == generation || self.indexed_header_reference(r.num, offset) == Some(r)
}
Some(XrefEntry::InStream { .. }) => r.generation == 0,
None | Some(XrefEntry::Free { .. }) => false,
},
}
}
pub(super) fn try_get_reference(&self, r: ObjRef) -> Result<Arc<Object>, CosError> {
if !self.reference_matches(r) {
return Ok(Arc::new(Object::Null));
}
self.try_get(r.num)
}
/// A repair probe, not a whole-file search: only the indexed header (or the same offset
/// relative to a shifted PDF header) can establish this additional generation. A small
/// window bounds hostile whitespace/tokens; ordinary xref-matching reads are unchanged.
pub(super) fn indexed_header_reference(&self, num: u32, offset: u64) -> Option<ObjRef> {
let at = |off: usize| {
let bytes = self.data.get(off..off.saturating_add(128).min(self.data.len()))?;
let mut lexer = Lexer::new(bytes, 0);
lexer.skip_ws();
let number = std::str::from_utf8(lexer.token()).ok()?.parse::<u32>().ok()?;
lexer.skip_ws();
let generation = std::str::from_utf8(lexer.token()).ok()?.parse::<u16>().ok()?;
(number == num && lexer.eat_keyword(b"obj")).then_some(ObjRef::new(number, generation))
};
let off = usize::try_from(offset).ok()?;
at(off).or_else(|| if self.header_offset == 0 { None } else { off.checked_add(self.header_offset).and_then(at) })
}
/// Add at most one bounded diagnostic per object. False stops diagnostic probing only;
/// parsers must still read the remaining xref rows needed to build their index.
pub(super) fn inspect_xref_generation_repair(&self, num: u32, entry: XrefEntry, recorded: &mut HashSet<u32>, log: &mut Vec<String>) -> bool {
let XrefEntry::InFile { offset, generation } = entry else { return true };
if recorded.contains(&num) {
return true;
}
let Some(header) = self.indexed_header_reference(num, offset) else { return true };
if header.generation == generation {
return true;
}
if recorded.len() == 256 {
log.push("additional xref/header generation disagreements omitted after 256 objects".into());
return false;
}
recorded.insert(num);
log.push(format!("object {num}: xref generation {generation} disagrees with object header generation {}", header.generation));
true
}
/// Diagnose once at open without expanding compact ranges or decoding an xref twice.
/// Materialized readers retain bounded notes while they read accepted sections, before
/// newer rows hide older ones. Discarded parse attempts/rebuilt indexes carry no notes.
pub(super) fn note_xref_generation_repairs(&self, log: &mut Vec<String>) {
match &*self.entries {
XrefIndex::Materialized(_, notes) => log.extend(notes.iter().cloned()),
XrefIndex::Compact { revisions, .. } => {
let mut recorded = HashSet::new();
for revision in revisions {
for section in std::iter::once(&revision.primary).chain(revision.hybrid.iter()) {
for num in section.numbers() {
if let Some(entry) = section.lookup(num, &self.data)
&& !self.inspect_xref_generation_repair(num, entry, &mut recorded, log)
{
return;
}
}
}
}
}
}
}
/// Follow a reference if `o` is one; otherwise return `o` itself. An object can be nothing but
/// a reference to another, which is followed too, up to [`MAX_REFERENCE_HOPS`] of them.
pub fn resolve(&self, o: &Object) -> Arc<Object> {
let Object::Ref(r) = o else { return Arc::new(o.clone()) };
let mut held = self.get(*r);
for _ in 0..MAX_REFERENCE_HOPS {
let Object::Ref(r) = &*held else { break };
held = self.get(*r);
}
held
}
/// `s` with the `/DecodeParms` its filters use made direct, so [`Stream::decoded`] reads them:
/// an indirect dictionary resolved, and of an array (itself perhaps indirect), the entries at
/// the filters' places (`/Crypt` counts, ISO 32000-2 §7.4.1) resolved, the rest dropped. Of
@@ -218,7 +308,7 @@ impl Document {
}
/// `o`, or the object it refers to, without copying either.
fn held<'o>(&self, o: &'o Object) -> Held<'o> {
pub(crate) fn held<'o>(&self, o: &'o Object) -> Held<'o> {
match o {
Object::Ref(_) => Held::Resolved(self.resolve(o)),
o => Held::Direct(o),
@@ -384,7 +474,7 @@ impl Document {
pub fn try_get(&self, num: u32) -> Result<Arc<Object>, CosError> {
match self.overlay.get(&num) {
Some(Slot::Set(_, o)) => return Ok(o.clone()),
Some(Slot::Set(_, o) | Slot::RepairedSet(_, _, o)) => return Ok(o.clone()),
Some(Slot::Freed(_)) => return Ok(Arc::new(Object::Null)),
None => {}
}
@@ -395,6 +485,7 @@ impl Document {
// cycles such as `6 0 obj << /Length 6 0 R >>`. Bound the nesting on every path.
let _guard =
LoadGuard::enter(num).ok_or_else(|| CosError::Syntax { offset: 0, detail: format!("reference cycle while loading object {num}") })?;
let _scope = crate::security::FollowScope::enter();
let obj = self.load(num, 0)?;
let obj = Arc::new(obj);
// The editing cache is bounded like the scan reader's: re-parsing a file-backed view
@@ -514,11 +605,11 @@ impl Document {
None | Some(XrefEntry::Free { .. }) => Ok(Object::Null),
Some(XrefEntry::InFile { offset, .. }) => {
let resolve = |r: ObjRef| {
if let Some(known) = LENGTHS.with(|m| m.borrow().get(&r.num).copied()) {
if let Some(known) = LENGTHS.with(|m| m.borrow().get(&r).copied()) {
return known;
}
let length = self.try_get(r.num).ok().and_then(|o| o.as_int());
LENGTHS.with(|m| m.borrow_mut().insert(r.num, length));
let length = self.try_get_reference(r).ok().and_then(|o| o.as_int());
LENGTHS.with(|m| m.borrow_mut().insert(r, length));
length
};
let off = offset as usize;
@@ -562,7 +653,11 @@ impl Document {
let Object::Stream(s) = &*loaded else { return Err(CosError::MissingObject(num)) };
let n = s.dict.int(b"N").unwrap_or(0).clamp(0, 1_000_000) as usize;
let first = usize::try_from(s.dict.int(b"First").unwrap_or(0).max(0)).ok();
let data = s.decoded_within(self.stream_limit)?;
// Its parameters can be given by reference, and can sit in another object stream. A scan of
// the file's object streams comes here without a read around it.
let _scope = crate::security::FollowScope::enter();
let readable = crate::security::following(Cow::Borrowed(s), || self.with_direct_parms(s));
let data = readable.decoded_within(self.stream_limit)?;
let mut lx = Lexer::new(&data, 0);
let mut index = Vec::with_capacity(n);
for _ in 0..n {
@@ -577,8 +672,14 @@ impl Document {
Ok(stm)
}
/// Last-resort lookup: scan the file for `num G obj`.
/// Last-resort lookup: scan the file for `num G obj`. What isn't there is remembered: each miss
/// reads the whole file, and a file can refer to such an object from everywhere.
fn scan_for(&self, num: u32) -> Option<(ObjRef, Object)> {
if self.unfound.lock().is_ok_and(|unfound| unfound.contains(&num)) {
return None;
}
#[cfg(test)]
super::tests::FILE_SCANS.with(|scans| scans.set(scans.get() + 1));
let needle = format!("{num} ");
let data = &self.data;
let mut found = None;
@@ -594,6 +695,11 @@ impl Document {
found = Some((id, o)); // keep the last (newest) definition
}
}
if found.is_none()
&& let Ok(mut unfound) = self.unfound.lock()
{
unfound.insert(num);
}
found
}
@@ -620,6 +726,38 @@ impl Document {
}
}
/// Replace an accepted existing reference without losing a damaged source slot's aliases.
/// Unlike [`Self::set`], this preserves both xref/header generations when they disagree.
/// Such an edit forces a full rewrite so the output has consistent references, rather than
/// appending a new contradictory xref row. Missing, free and stale references are errors.
pub fn set_existing(&mut self, r: ObjRef, obj: impl Into<Object>) -> Result<(), CosError> {
if !self.reference_matches(r) {
return Err(CosError::MissingObject(r.num));
}
let alias = self.alternate_generation(r);
if let Some(alias) = alias {
let obj = self.adopt(obj.into());
self.overlay.insert(r.num, Slot::RepairedSet(r.generation, alias, Arc::new(obj)));
} else {
self.set(r, obj);
}
Ok(())
}
/// Other accepted generation for a known-live reference to a damaged slot.
pub(super) fn alternate_generation(&self, r: ObjRef) -> Option<u16> {
match self.overlay.get(&r.num) {
Some(Slot::RepairedSet(generation, alias, _)) => Some(if r.generation == *generation { *alias } else { *generation }),
Some(_) => None,
None => match self.entries.lookup(r.num, &self.data) {
Some(XrefEntry::InFile { offset, generation }) => self.indexed_header_reference(r.num, offset).and_then(|header| {
(header.generation != generation).then_some(if r.generation == generation { header.generation } else { generation })
}),
_ => None,
},
}
}
/// Add a new indirect object and return its reference.
pub fn add(&mut self, obj: impl Into<Object>) -> ObjRef {
let r = ObjRef::new(self.allocate_num(), 0);
@@ -659,8 +797,9 @@ impl Document {
///
/// Used when the working snapshot was restored from an earlier revision and the next
/// incremental save must still append to this file (the last saved revision). Objects
/// that already match are left untouched. Reads go through private readers so neither
/// document's editing cache is filled.
/// that already match are left untouched, except members of an object stream of this file
/// whose stream number `current` frees or reuses: those are written again on their own.
/// Reads go through private readers so neither document's editing cache is filled.
pub fn overlay_changes_from(&mut self, current: &Document) -> Result<(), CosError> {
let saved_nums = self.object_numbers();
let current_nums = current.object_numbers();
@@ -678,12 +817,29 @@ impl Document {
let generation = current.generation(num);
let same = saved_nums.binary_search(&num).is_ok()
&& self.generation(num) == generation
&& saved_reader.try_get(ObjRef::new(num, 0))? == current_reader.try_get(ObjRef::new(num, 0))?;
&& self.alternate_generation(ObjRef::new(num, generation)) == current.alternate_generation(ObjRef::new(num, generation))
&& saved_reader.try_get(ObjRef::new(num, generation))? == current_reader.try_get(ObjRef::new(num, generation))?;
if !same {
changed.push(num);
}
}
}
// An object stream of this file that is freed or replaced takes its members with it, so
// a member that is otherwise unchanged is written again on its own.
let gone = |n: u32| freed.binary_search(&n).is_ok() || changed.binary_search(&n).is_ok();
let orphaned: Vec<u32> = current_nums
.iter()
.copied()
.filter(|&num| changed.binary_search(&num).is_err())
.filter(|&num| {
let packed_in = match self.xref_entry(num) {
Some(XrefEntry::InStream { stream, .. }) => Some(stream),
_ => None,
};
packed_in.is_some_and(gone)
})
.collect();
changed.extend(orphaned);
for num in freed {
let generation = self.generation(num);
self.free(ObjRef::new(num, generation));
@@ -691,8 +847,14 @@ impl Document {
let mut current_reader = current.object_reader();
for num in changed {
let generation = current.generation(num);
let object = current_reader.try_get(ObjRef::new(num, 0))?;
self.set(ObjRef::new(num, generation), (*object).clone());
let object = current_reader.try_get(ObjRef::new(num, generation))?;
if let Some(alias) = current.alternate_generation(ObjRef::new(num, generation)) {
let object = self.adopt((*object).clone());
self.overlay.insert(num, Slot::RepairedSet(generation, alias, Arc::new(object)));
self.next_num = self.next_num.max(num.saturating_add(1));
} else {
self.set(ObjRef::new(num, generation), (*object).clone());
}
}
*self.trailer_mut() = current.trailer().clone();
Ok(())
@@ -703,14 +865,13 @@ impl Document {
let mut obj = (*self.get(r)).clone();
let d = obj.as_dict_mut().ok_or(CosError::NotADictionary(r))?;
f(d);
self.set(r, obj);
Ok(())
self.set_existing(r, obj)
}
/// Current generation of an object number.
pub fn generation(&self, num: u32) -> u16 {
match self.overlay.get(&num) {
Some(Slot::Set(g, _)) => *g,
Some(Slot::Set(g, _) | Slot::RepairedSet(g, _, _)) => *g,
Some(Slot::Freed(g)) => *g,
None => match self.entries.lookup(num, &self.data) {
Some(XrefEntry::InFile { generation, .. }) => generation,
@@ -722,7 +883,7 @@ impl Document {
pub(crate) fn overlay_entries(&self) -> impl Iterator<Item = (u32, u16, Option<Arc<Object>>)> + '_ {
self.overlay.iter().map(|(n, s)| match s {
Slot::Set(g, o) => (*n, *g, Some(o.clone())),
Slot::Set(g, o) | Slot::RepairedSet(g, _, o) => (*n, *g, Some(o.clone())),
Slot::Freed(g) => (*n, *g, None),
})
}
@@ -745,7 +906,7 @@ impl Document {
if source.peek().is_some_and(|n| *n == num) {
source.next();
}
if matches!(slot, Slot::Set(..)) {
if matches!(slot, Slot::Set(..) | Slot::RepairedSet(..)) {
numbers.push(num);
}
}
@@ -763,7 +924,7 @@ impl Document {
pub(super) fn root_is_catalog(&self) -> bool {
let Some(r) = self.root() else { return false };
matches!(self.try_get(r.num).as_deref(), Ok(Object::Dict(d)) if d.name(b"Type") == Some(b"Catalog") || d.contains(b"Pages"))
matches!(self.try_get_reference(r).as_deref(), Ok(Object::Dict(d)) if d.name(b"Type") == Some(b"Catalog") || d.contains(b"Pages"))
}
}
@@ -786,9 +947,13 @@ const DECODE_PARMS_KEYS: [&[u8]; 13] = [
b"ColorTransform",
];
/// How many references in a row [`Document::resolve`] follows: enough for any file, and it ends a
/// cycle of them.
const MAX_REFERENCE_HOPS: usize = 32;
/// An object read in place, or the object a reference names as the document holds it: either way
/// nothing is copied ([`Document::held`]).
enum Held<'o> {
pub(crate) enum Held<'o> {
Direct(&'o Object),
Resolved(Arc<Object>),
}
crates/cos/src/document/open.rs
+59 -9
View File
@@ -1,7 +1,11 @@
use super::reader::find;
use super::state::Credentials;
use super::state::{Credentials, Slot};
use super::xref::XrefIndex;
use super::{Arc, BTreeMap, CosError, Dict, Document, ObjRef, Object};
use super::{Arc, BTreeMap, CosError, Dict, Document, HashSet, ObjRef, Object};
/// More objects than `/Encrypt` and `/ID` refer to in any file, so that one which refers to the page
/// tree, say, doesn't have a whole document read to open it.
const MAX_ENCRYPT_MEMBERS: usize = 64;
impl Document {
/// A new document with an empty page tree (for split / extract / combine). It has no
@@ -27,6 +31,8 @@ impl Document {
version: "1.7".into(),
security: None,
encrypt_num: None,
encrypt_members: Arc::default(),
unfound: Arc::default(),
encryption_changed: false,
full_save: false,
out_security: None,
@@ -111,6 +117,8 @@ impl Document {
version,
security: None,
encrypt_num: None,
encrypt_members: Arc::default(),
unfound: Arc::default(),
encryption_changed: false,
full_save: false,
out_security: None,
@@ -148,6 +156,9 @@ impl Document {
}
doc.reconstruct(&mut log)?;
}
// Header generations are plaintext, including in encrypted files. Inspect only the
// bounded headers, not object bodies, and retain the compact xref representation.
doc.note_xref_generation_repairs(&mut log);
doc.repair_log = Arc::new(log);
// Reconstruction may have found a different trailer: (re)authenticate against it.
if (!ok || !authenticated)
@@ -185,6 +196,8 @@ impl Document {
};
let params = crate::security::encrypt_dict(self, &dict);
let id0 = self.first_id();
// Read here, while nothing is decrypted yet.
let members = self.encryption_members(enc);
let handler = match credentials {
Credentials::Password(password) => Arc::new(pdfcraft_crypt::SecurityHandler::open(params, &id0, *password).map_err(|e| match e {
pdfcraft_crypt::CryptError::WrongPassword if password.is_none() => CosError::NeedsPassword,
@@ -197,6 +210,7 @@ impl Document {
};
self.security = Some(handler);
self.encrypt_num = num;
self.encrypt_members = Arc::new(members);
// Objects read while locating the catalog were read before decryption was possible.
self.cache = Arc::default();
self.objstms = Arc::default();
@@ -205,10 +219,42 @@ impl Document {
/// The first file identifier, part of the encryption key (empty when there is none).
fn first_id(&self) -> Vec<u8> {
match self.trailer.get(b"ID") {
Some(Object::Array(a)) => a.first().and_then(|s| s.as_string()).map(|s| s.bytes.clone()).unwrap_or_default(),
_ => Vec::new(),
let id = self.trailer.get(b"ID").map(|id| self.resolve(id));
let first = id.as_deref().and_then(Object::as_array).and_then(|a| a.first()).map(|s| self.resolve(s));
first.as_deref().and_then(Object::as_string).map(|s| s.bytes.clone()).unwrap_or_default()
}
/// The objects `/Encrypt`, given as `enc`, and `/ID` refer to, the ones they refer to, and so
/// on, up to [`MAX_ENCRYPT_MEMBERS`]. Their strings aren't encrypted (ISO 32000-2 §7.6.2): the
/// key is made from them, so they have to read as stored, and stay so. A number with no object
/// behind it, and a stream, which those dictionaries never hold, isn't one: an object added
/// under that number later, or a page's content a hostile `/Encrypt` names, is encrypted.
fn encryption_members(&self, enc: &Object) -> HashSet<u32> {
let mut pending = Vec::new();
crate::writer::collect_refs(enc, &mut |r| pending.push(r));
if let Some(id) = self.trailer.get(b"ID") {
crate::writer::collect_refs(id, &mut |r| pending.push(r));
}
let mut members = HashSet::new();
while let Some(r) = pending.pop() {
if members.len() >= MAX_ENCRYPT_MEMBERS || members.contains(&r.num) {
continue;
}
let held = self.get(r);
if matches!(&*held, Object::Null | Object::Stream(_)) {
continue;
}
members.insert(r.num);
crate::writer::collect_refs(&held, &mut |x| pending.push(x));
}
members
}
/// Whether object `num` is written as it is held, being one of [`Document::encryption_members`],
/// unedited, and the save writing the same `/Encrypt`. What was edited or added is encrypted
/// whatever its number.
pub(crate) fn stored_as_is(&self, num: u32) -> bool {
!self.encryption_changed && !self.is_edited(num) && self.encrypt_members.contains(&num)
}
/// Lift the restrictions of a document opened with its user password by entering the
@@ -245,8 +291,9 @@ impl Document {
}
/// The next save must rewrite the whole file (see [`Document::require_full_save`]).
/// Repair-aware edits retaining two accepted generations cannot use one incremental slot.
pub fn full_save_required(&self) -> bool {
self.full_save
self.full_save || self.overlay.values().any(|slot| matches!(slot, Slot::RepairedSet(..)))
}
/// Security was set or removed since the document was opened (the next save applies it).
@@ -270,8 +317,9 @@ impl Document {
/// always a full rewrite. Returns the handler (authenticated as owner).
pub fn set_encryption(&mut self, params: &pdfcraft_crypt::NewEncryption) -> Result<(), CosError> {
// The file identifier is part of the key; make sure it exists and keep it.
let id0 = match self.trailer.get(b"ID") {
Some(Object::Array(a)) if a.len() == 2 => a[0].as_string().map(|s| s.bytes.clone()).unwrap_or_default(),
let id = self.trailer.get(b"ID").map(|id| self.resolve(id));
let id0 = match id.as_deref() {
Some(Object::Array(a)) if a.len() == 2 => self.first_id(),
_ => {
let id = pdfcraft_crypt::file_id(&params.seed);
let s = Object::String(crate::PdfString { bytes: id.clone(), hex: true });
@@ -338,7 +386,9 @@ impl Document {
/// Decrypt a freshly parsed indirect object if the document is encrypted.
pub(super) fn decrypted(&self, id: ObjRef, o: Object) -> Object {
match &self.security {
Some(h) if Some(id.num) != self.encrypt_num => crate::security::transform(h, &o, id.num, id.generation, crate::security::Pass::Decrypt),
Some(h) if Some(id.num) != self.encrypt_num && !self.encrypt_members.contains(&id.num) => {
crate::security::transform(self, h, &o, id.num, id.generation, crate::security::Pass::Decrypt)
}
_ => o,
}
}
crates/cos/src/document/reader.rs
+23 -8
View File
@@ -17,8 +17,8 @@ impl Document {
return Ok((index, trailer, revisions, Vec::new()));
}
}
let (entries, trailer, revisions, oddities) = self.read_xref_chain_materialized(log)?;
Ok((XrefIndex::Materialized(Arc::new(entries)), trailer, revisions, oddities))
let (entries, trailer, revisions, oddities, generation_repairs) = self.read_xref_chain_materialized(log)?;
Ok((XrefIndex::Materialized(Arc::new(entries), Arc::new(generation_repairs)), trailer, revisions, oddities))
}
fn read_compact_xref_chain(&self, log: &mut Vec<String>) -> Result<Option<(XrefIndex, Dict, Vec<Revision>)>, CosError> {
@@ -323,7 +323,7 @@ impl Document {
pub(super) fn read_xref_chain_materialized(
&self,
log: &mut Vec<String>,
) -> Result<(BTreeMap<u32, XrefEntry>, Dict, Vec<Revision>, Vec<(u64, u32)>), CosError> {
) -> Result<(BTreeMap<u32, XrefEntry>, Dict, Vec<Revision>, Vec<(u64, u32)>, Vec<String>), CosError> {
let data = &self.data;
let tail_start = data.len().saturating_sub(4096);
let sx = rfind(data, b"startxref", tail_start).ok_or(CosError::Syntax { offset: data.len(), detail: "no startxref".into() })?;
@@ -343,6 +343,9 @@ impl Document {
// At most MAX_XREF_OBJECTS entries are kept in all; a stream past that is read again.
let mut hybrids: HashMap<i64, Option<BTreeMap<u32, XrefEntry>>> = HashMap::new();
let (mut hybrid_rows, mut noted) = (0usize, HashSet::new());
let mut generation_repairs = Vec::new();
let mut recorded = HashSet::new();
let mut diagnose = true;
let mut next = Some(first);
while let Some(off) = next.take() {
if !visited.insert(off) || visited.len() > 10_000 {
@@ -371,6 +374,12 @@ impl Document {
}
};
revisions.push(Revision { xref_offset: off, is_stream, header_relative });
for (&num, &entry) in &sect {
if !diagnose {
break;
}
diagnose = self.inspect_xref_generation_repair(num, entry, &mut recorded, &mut generation_repairs);
}
// Hybrid files (§7.5.8.4): the /XRefStm entries supplement the table and replace the
// free entries it gives objects only a cross-reference stream can locate.
let hybrid = sect_trailer.int(b"XRefStm").filter(|x| *x >= 0);
@@ -390,6 +399,12 @@ impl Document {
let mut stm = BTreeMap::new();
let read = match self.read_section(x as usize, &mut stm, &mut odd, log) {
Ok(_) => {
for (&num, &entry) in &stm {
if !diagnose {
break;
}
diagnose = self.inspect_xref_generation_repair(num, entry, &mut recorded, &mut generation_repairs);
}
for (&num, &e) in &stm {
match sect.get(&num) {
None | Some(XrefEntry::Free { .. }) => {
@@ -427,7 +442,7 @@ impl Document {
for k in [&b"Prev"[..], b"XRefStm", b"Type", b"W", b"Index", b"Filter", b"DecodeParms", b"Length"] {
trailer.remove(k);
}
Ok((entries, trailer, revisions, oddities))
Ok((entries, trailer, revisions, oddities, generation_repairs))
}
/// Read one section at `off`, adding entries not already known (newer sections win). `odd`
@@ -606,7 +621,7 @@ impl Document {
}
if found > 0 {
log.push(format!("the cross-reference data leaves out {found} objects the file contains; located them by their object headers"));
self.entries = Arc::new(XrefIndex::Materialized(Arc::new(filled)));
self.entries = Arc::new(XrefIndex::Materialized(Arc::new(filled), Arc::default()));
}
self.located = Arc::new(located);
}
@@ -631,7 +646,7 @@ impl Document {
if entries.is_empty() {
return Err(CosError::Syntax { offset: 0, detail: "no objects found".into() });
}
self.entries = Arc::new(XrefIndex::Materialized(Arc::new(entries.clone())));
self.entries = Arc::new(XrefIndex::Materialized(Arc::new(entries.clone()), Arc::default()));
self.cache.lock().map_err(|_| CosError::Poisoned)?.clear();
// Objects inside object streams.
for num in objstms {
@@ -648,7 +663,7 @@ impl Document {
}
}
let object_count = entries.len();
self.entries = Arc::new(XrefIndex::Materialized(Arc::new(entries)));
self.entries = Arc::new(XrefIndex::Materialized(Arc::new(entries), Arc::default()));
self.cache.lock().map_err(|_| CosError::Poisoned)?.clear();
// Trailer: the last `trailer` dictionary or cross-reference stream dictionary with a
// usable /Root (last in the file, whichever kind), else the catalog itself. One whose
@@ -658,7 +673,7 @@ impl Document {
// A file whose only trailers are cross-reference streams (and whose `startxref` is gone
// or points nowhere) keeps its /Info, /ID and /Encrypt this way.
let rank = |at: usize, d: &Dict| {
let root = self.try_get(d.reference(b"Root")?.num).ok()?;
let root = self.try_get_reference(d.reference(b"Root")?).ok()?;
let Object::Dict(c) = &*root else { return None };
Some((c.name(b"Type") == Some(b"Catalog") || c.contains(b"Pages"), at))
};
crates/cos/src/document/state.rs
+6 -2
View File
@@ -7,7 +7,7 @@ thread_local! {
/// Indirect `/Length` values resolved during the current outermost load: a damaged file
/// retries a parse several ways, and each retry would resolve the same lengths again
/// (exponential in the chain of streams whose lengths point at further streams).
pub(super) static LENGTHS: std::cell::RefCell<HashMap<u32, Option<i64>>> = std::cell::RefCell::new(HashMap::new());
pub(super) static LENGTHS: std::cell::RefCell<HashMap<ObjRef, Option<i64>>> = std::cell::RefCell::new(HashMap::new());
}
/// Marks an object as being loaded on this thread. Loading can re-enter (indirect stream
@@ -65,6 +65,9 @@ pub(super) fn objstm_member(num: i64, offset: i64, first: Option<usize>, len: us
#[derive(Clone, Debug, PartialEq)]
pub(super) enum Slot {
Set(u16, Arc<Object>),
/// Both generations accepted from a damaged live source slot survive a repair-aware edit.
/// A full rewrite is required: a new xref slot cannot encode two generations.
RepairedSet(u16, u16, Arc<Object>),
Freed(u16),
}
@@ -103,6 +106,7 @@ impl ReadState {
a == b
&& match (x, y) {
(Slot::Set(g, o), Slot::Set(h, p)) => g == h && Arc::ptr_eq(o, p),
(Slot::RepairedSet(g, a, o), Slot::RepairedSet(h, b, p)) => g == h && a == b && Arc::ptr_eq(o, p),
(Slot::Freed(g), Slot::Freed(h)) => g == h,
_ => false,
}
@@ -133,7 +137,7 @@ impl ObjectReader {
}
pub(crate) fn try_get(&mut self, reference: ObjRef) -> Result<Arc<Object>, CosError> {
let object = self.document.try_get(reference.num);
let object = self.document.try_get_reference(reference);
// A missed eviction here only costs a re-parse, so never fail the read on a poisoned
// lock. The editing-cache evictors propagate instead; they guard user edits.
if let Ok(mut streams) = self.document.objstms.lock() {
crates/cos/src/document/tests.rs
+182 -11
View File
@@ -1,5 +1,147 @@
use super::*;
std::thread_local! {
/// How many times `Document::scan_for` read the file on this thread.
pub(super) static FILE_SCANS: std::cell::Cell<usize> = const { std::cell::Cell::new(0) };
}
/// A scan that doesn't find an object isn't made again: each reads the whole file, and a file can
/// refer to such an object from every object it has.
#[test]
fn what_a_scan_of_the_file_did_not_find_is_remembered() {
let body = b"%PDF-1.7\n1 0 obj << /Type /Catalog /Pages 2 0 R >> endobj\n2 0 obj << /Type /Pages /Kids [] /Count 0 >> endobj\n";
let mut out = body.to_vec();
let xref = out.len();
out.extend_from_slice(b"xref\n0 4\n0000000000 65535 f \n");
out.extend_from_slice(format!("{:010} 00000 n \n", 9).as_bytes());
out.extend_from_slice(format!("{:010} 00000 n \n", body.iter().position(|b| *b == b'\n').unwrap() + 49).as_bytes());
out.extend_from_slice(b"0000000001 00000 n \n");
out.extend_from_slice(format!("trailer << /Size 4 /Root 1 0 R >>\nstartxref\n{xref}\n%%EOF\n").as_bytes());
let doc = Document::open(Arc::new(out)).unwrap();
FILE_SCANS.with(|scans| scans.set(0));
for _ in 0..5 {
assert!(doc.try_get(3).is_err());
}
assert_eq!(FILE_SCANS.with(std::cell::Cell::get), 1);
let clone = doc.clone();
assert!(clone.try_get(3).is_err());
assert_eq!(FILE_SCANS.with(std::cell::Cell::get), 1, "a clone has the same file");
}
/// An object can be nothing but a reference to another. A chain of them is followed, and one that
/// goes round ends in what isn't a dictionary or a name, without hanging.
#[test]
fn references_to_references_are_followed() {
let mut doc = Document::new_empty();
let name = doc.add(Object::name("Target"));
let one = doc.add(Object::Ref(name));
let two = doc.add(Object::Ref(one));
assert_eq!(doc.resolve(&Object::Ref(two)).as_name(), Some(b"Target".as_slice()));
let a = doc.add(Object::Null);
let b = doc.add(Object::Ref(a));
doc.set(a, Object::Ref(b));
assert_eq!(doc.resolve(&Object::Ref(a)).as_name(), None);
}
/// An object stream's `/DecodeParms` can be given by reference: without them a predictor isn't
/// undone, so its members were lost.
#[test]
fn an_object_stream_with_parameters_by_reference_is_read() {
let members = b"3 0 << /Hidden (yes) >>";
// PNG Sub, one row.
let mut rows = vec![1u8];
let mut before = 0u8;
for b in members {
rows.push(b.wrapping_sub(before));
before = *b;
}
let coded = pdfcraft_filters::encode_flate(&rows);
let mut out = b"%PDF-1.5\n".to_vec();
let mut offs = [0usize; 7];
let mut obj = |out: &mut Vec<u8>, n: usize, body: &[u8]| {
offs[n] = out.len();
out.extend_from_slice(format!("{n} 0 obj\n").as_bytes());
out.extend_from_slice(body);
out.extend_from_slice(b"\nendobj\n");
};
obj(&mut out, 1, b"<< /Type /Catalog /Pages 2 0 R >>");
obj(&mut out, 2, b"<< /Type /Pages /Kids [] /Count 0 /Extra 3 0 R >>");
let mut stm = format!("<< /Type /ObjStm /N 1 /First 4 /Filter /FlateDecode /DecodeParms 6 0 R /Length {} >>\nstream\n", coded.len()).into_bytes();
stm.extend_from_slice(&coded);
stm.extend_from_slice(b"\nendstream");
obj(&mut out, 4, &stm);
let data = [2u8, 0, 4, 0];
let mut xs = format!("<< /Type /XRef /Size 7 /W [1 2 1] /Index [3 1] /Length {} >>\nstream\n", data.len()).into_bytes();
xs.extend_from_slice(&data);
xs.extend_from_slice(b"\nendstream");
obj(&mut out, 5, &xs);
obj(&mut out, 6, format!("<< /Predictor 11 /Columns {} >>", members.len()).as_bytes());
let table = out.len();
out.extend_from_slice(b"xref\n0 7\n0000000000 65535 f \n");
for (n, off) in offs.iter().enumerate().skip(1) {
let line = if n == 3 { "0000000000 65535 f \n".to_string() } else { format!("{off:010} 00000 n \n") };
out.extend_from_slice(line.as_bytes());
}
out.extend_from_slice(format!("trailer << /Size 7 /Root 1 0 R /XRefStm {} >>\nstartxref\n{table}\n%%EOF\n", offs[5]).as_bytes());
let doc = Document::open(Arc::new(out)).unwrap();
let hidden = doc.get(ObjRef::new(3, 0));
assert_eq!(hidden.as_dict().and_then(|d| d.get(b"Hidden")).and_then(Object::as_string).map(|s| s.bytes.clone()), Some(b"yes".to_vec()));
}
/// Each entry of an `/Encrypt` dictionary, and of a crypt filter in it, reads the same by reference.
/// Read as absent, `/StmF` or `/CFM` was `/Identity`, so a whole file's streams stayed ciphertext.
#[test]
fn encrypt_dictionary_entries_read_the_same_by_reference() {
let string = |s: &str| Object::String(crate::PdfString::text(s));
let mut filter = Dict::new();
filter.set(b"CFM".to_vec(), Object::name("AESV2"));
filter.set(b"Length".to_vec(), Object::Int(5));
let cf = |filter: Dict| {
let mut cf = Dict::new();
cf.set(b"StdCF".to_vec(), Object::Dict(filter));
cf.set(b"Other".to_vec(), Object::Dict(Dict::new()));
Object::Dict(cf)
};
let mut d = Dict::new();
let entries = [
("Filter", Object::name("Standard")),
("V", Object::Int(4)),
("R", Object::Int(4)),
("Length", Object::Int(128)),
("P", Object::Int(-44)),
("O", string("owner")),
("U", string("user")),
("EncryptMetadata", Object::Bool(false)),
("CF", cf(filter.clone())),
("StmF", Object::name("StdCF")),
("StrF", Object::name("Other")),
("EFF", Object::name("StdCF")),
];
for (key, value) in &entries {
d.set(key.as_bytes().to_vec(), value.clone());
}
let mut doc = Document::new_empty();
let direct = crate::security::encrypt_dict(&doc, &d);
assert_eq!((direct.stm_f.as_slice(), direct.str_f.as_slice(), direct.ef_f.as_slice()), (&b"StdCF"[..], &b"Other"[..], &b"StdCF"[..]));
assert_eq!(direct.length_bits, 40, "the filter's own length, in bytes");
assert!(!direct.encrypt_metadata);
assert_eq!(direct.crypt_filters.iter().filter(|(_, m)| *m == pdfcraft_crypt::Method::Aes128).count(), 1);
for (key, value) in &entries {
let mut one = d.clone();
let r = doc.add(value.clone());
one.set(key.as_bytes().to_vec(), Object::Ref(r));
assert_eq!(crate::security::encrypt_dict(&doc, &one), direct, "/{key}");
}
for key in ["CFM", "Length"] {
let mut by_reference = filter.clone();
let r = doc.add(filter.get(key.as_bytes()).unwrap().clone());
by_reference.set(key.as_bytes().to_vec(), Object::Ref(r));
let mut one = d.clone();
one.set(b"CF".to_vec(), cf(by_reference));
assert_eq!(crate::security::encrypt_dict(&doc, &one), direct, "/{key} in the filter");
}
}
/// /DecodeParms given by reference, whole, as array entries or as the numbers inside, are read
/// once made direct: here a Flate predictor, which the stream alone doesn't see. Only the entries the filters use are
/// resolved, and nothing is for a chain longer than a decoder takes.
@@ -457,7 +599,7 @@ pub(crate) fn build(bodies: &[&str], trailer: &str) -> Vec<u8> {
fn assert_compact_matches_materialized(doc: &Document) {
assert!(matches!(doc.entries.as_ref(), XrefIndex::Compact { .. }), "valid xrefs should not retain a per-entry map");
let (entries, _, revisions, _) = doc.read_xref_chain_materialized(&mut Vec::new()).unwrap();
let (entries, _, revisions, _, _) = doc.read_xref_chain_materialized(&mut Vec::new()).unwrap();
assert_eq!(revisions.as_slice(), doc.revisions());
let max = entries.keys().next_back().copied().unwrap_or(0);
assert_eq!(doc.entries.max_num(), max);
@@ -545,7 +687,7 @@ fn malformed_size_uses_the_materialized_compatibility_path() {
let at = bytes.windows(marker.len()).position(|window| window == marker).unwrap();
bytes[at..at + marker.len()].copy_from_slice(b"/Size 2");
let doc = Document::open(Arc::new(bytes)).unwrap();
assert!(matches!(doc.entries.as_ref(), XrefIndex::Materialized(_)));
assert!(matches!(doc.entries.as_ref(), XrefIndex::Materialized(..)));
assert_eq!(doc.next_num, 4);
assert_eq!(doc.object_numbers(), vec![1, 2, 3]);
}
@@ -1039,15 +1181,18 @@ fn reads_classic_revision_and_edits_are_copy_on_write() {
fn object_numbers_merge_source_and_overlay_in_order() {
let mut doc = Document::new_empty();
doc.overlay.clear();
doc.entries = Arc::new(XrefIndex::Materialized(Arc::new(BTreeMap::from([
(0, XrefEntry::Free { next_generation: 65535 }),
(2, XrefEntry::InFile { offset: 10, generation: 7 }),
(4, XrefEntry::Free { next_generation: 9 }),
(6, XrefEntry::InStream { stream: 80, index: 1 }),
(8, XrefEntry::InFile { offset: 20, generation: 5 }),
(10, XrefEntry::InFile { offset: 30, generation: 3 }),
(u32::MAX, XrefEntry::InFile { offset: 40, generation: 1 }),
]))));
doc.entries = Arc::new(XrefIndex::Materialized(
Arc::new(BTreeMap::from([
(0, XrefEntry::Free { next_generation: 65535 }),
(2, XrefEntry::InFile { offset: 10, generation: 7 }),
(4, XrefEntry::Free { next_generation: 9 }),
(6, XrefEntry::InStream { stream: 80, index: 1 }),
(8, XrefEntry::InFile { offset: 20, generation: 5 }),
(10, XrefEntry::InFile { offset: 30, generation: 3 }),
(u32::MAX, XrefEntry::InFile { offset: 40, generation: 1 }),
])),
Arc::default(),
));
doc.set(ObjRef::new(1, 2), Object::Int(1)); // before the first live source slot
doc.set(ObjRef::new(2, 8), Object::Int(2)); // override, without duplication
doc.set(ObjRef::new(4, 9), Object::Int(4)); // reuse a source free slot
@@ -1704,6 +1849,32 @@ fn a_stream_whose_length_is_itself_and_never_ends_loads_quickly() {
rx.recv_timeout(std::time::Duration::from_secs(20)).expect("loading must not retry exponentially");
}
#[test]
fn a_restored_revision_keeps_what_the_saved_file_held_in_object_streams() {
// The earlier revision: an empty document read back from a file whose catalog and pages
// sit in an object stream.
let earlier = Document::open(Arc::new(crate::write_full(&Document::new_empty(), &crate::SaveOptions::default()).unwrap())).unwrap();
// The saved revision adds an object the catalog links to, so a full save numbers its objects
// anew; an undo then returns to the earlier graph.
let mut later = earlier.clone();
let extra = later.add(Object::Int(7));
let root = later.root().unwrap();
later.update_dict(root, |c| c.set(b"Extra".to_vec(), Object::Ref(extra))).unwrap();
let saved = crate::write_full(&later, &crate::SaveOptions::default()).unwrap();
let mut base = Document::open(Arc::new(saved)).unwrap();
assert!(
base.object_numbers().iter().any(|&n| matches!(base.xref_entry(n), Some(XrefEntry::InStream { .. }))),
"the saved file holds objects in a stream"
);
base.overlay_changes_from(&earlier).unwrap();
let file = crate::write_incremental(&base, &crate::SaveOptions::default()).unwrap();
let reread = Document::open(Arc::new(file)).unwrap();
let catalog = reread.get(reread.root().unwrap()).as_dict().unwrap().clone();
assert!(catalog.get(b"Extra").is_none(), "the restored catalog");
let pages = reread.resolve(catalog.get(b"Pages").unwrap());
assert_eq!(pages.as_dict().and_then(|d| d.name(b"Type")), Some(&b"Pages"[..]), "the pages survive: {pages:?}");
}
/// The cache keeps the sum of what its objects hold as they come and go, so neither reader has
/// to measure the cache again on every read (#5873).
#[test]
crates/cos/src/document/xref.rs
+16 -18
View File
@@ -41,7 +41,7 @@ pub(super) struct XrefRevisionSection {
#[derive(Clone, Debug)]
pub(super) enum XrefIndex {
Compact { revisions: Vec<XrefRevisionSection>, max_num: u32, has_entries: bool },
Materialized(Arc<BTreeMap<u32, XrefEntry>>),
Materialized(Arc<BTreeMap<u32, XrefEntry>>, Arc<Vec<String>>),
}
pub(super) const MAX_COMPACT_XREF_ROWS: usize = 2_000_000;
@@ -149,8 +149,8 @@ impl XrefSection {
}
}
pub(super) fn extend_numbers(&self, numbers: &mut Vec<u32>) {
for range in &self.ranges {
pub(super) fn numbers(&self) -> impl Iterator<Item = u32> + '_ {
self.ranges.iter().flat_map(|range| {
let available = match &self.data {
XrefSectionData::Table => range.count as usize,
XrefSectionData::Stream { bytes, widths } => {
@@ -158,14 +158,12 @@ impl XrefSection {
bytes.len().saturating_sub(range.offset).checked_div(row_width).unwrap_or(0)
}
};
for i in 0..available.min(range.count as usize) {
if let Ok(i) = u32::try_from(i)
&& let Some(num) = range.start.checked_add(i)
{
numbers.push(num);
}
}
}
(0..available.min(range.count as usize)).filter_map(move |i| u32::try_from(i).ok().and_then(|i| range.start.checked_add(i)))
})
}
pub(super) fn extend_numbers(&self, numbers: &mut Vec<u32>) {
numbers.extend(self.numbers());
}
}
@@ -189,20 +187,20 @@ impl XrefRevisionSection {
impl XrefIndex {
pub(super) fn empty() -> Self {
Self::Materialized(Arc::new(BTreeMap::new()))
Self::Materialized(Arc::new(BTreeMap::new()), Arc::default())
}
pub(super) fn lookup(&self, num: u32, file: &[u8]) -> Option<XrefEntry> {
match self {
Self::Compact { revisions, .. } => revisions.iter().find_map(|revision| revision.lookup(num, file)),
Self::Materialized(entries) => entries.get(&num).copied(),
Self::Materialized(entries, _) => entries.get(&num).copied(),
}
}
pub(super) fn is_empty(&self) -> bool {
match self {
Self::Compact { has_entries, .. } => !has_entries,
Self::Materialized(entries) => entries.is_empty(),
Self::Materialized(entries, _) => entries.is_empty(),
}
}
@@ -211,14 +209,14 @@ impl XrefIndex {
pub(super) fn remove_exhausted(&mut self) -> bool {
match self {
Self::Compact { .. } => false,
Self::Materialized(entries) => Arc::make_mut(entries).remove(&u32::MAX).is_some(),
Self::Materialized(entries, _) => Arc::make_mut(entries).remove(&u32::MAX).is_some(),
}
}
/// Every entry, newest section first, as the materialized reader keeps them.
pub(super) fn to_map(&self, file: &[u8]) -> BTreeMap<u32, XrefEntry> {
match self {
Self::Materialized(entries) => entries.as_ref().clone(),
Self::Materialized(entries, _) => entries.as_ref().clone(),
Self::Compact { revisions, .. } => {
let mut numbers = Vec::new();
for revision in revisions {
@@ -232,13 +230,13 @@ impl XrefIndex {
pub(super) fn max_num(&self) -> u32 {
match self {
Self::Compact { max_num, .. } => *max_num,
Self::Materialized(entries) => entries.keys().next_back().copied().unwrap_or(0),
Self::Materialized(entries, _) => entries.keys().next_back().copied().unwrap_or(0),
}
}
pub(super) fn object_numbers(&self, file: &[u8]) -> Vec<u32> {
match self {
Self::Materialized(entries) => {
Self::Materialized(entries, _) => {
entries.iter().filter_map(|(&num, entry)| (!matches!(entry, XrefEntry::Free { .. })).then_some(num)).collect()
}
Self::Compact { revisions, .. } => {
crates/cos/src/lib.rs
+1 -1
View File
@@ -22,7 +22,7 @@ mod parser;
mod security;
mod writer;
pub use budget::{DecodeBudget, ImageBudget, MAX_PAGE_DECODED, MAX_PAGE_FORMS};
pub use budget::{DecodeBudget, ImageBudget, MAX_PAGE_DECODED, MAX_PAGE_FORMS, MAX_REDACTED_IMAGE_PIXELS};
pub use bytes::Bytes;
pub use document::{Document, ReadState, Revision, SourceIdentity, XrefEntry};
pub use object::{Dict, MAX_DECODED, MAX_JBIG2_GLOBALS, Name, ObjRef, Object, PdfString, Stream};
crates/cos/src/security.rs
+214 -80
View File
@@ -7,10 +7,69 @@
//! false. Streams naming a crypt filter (`/Crypt` in `/Filter`) use that filter; embedded files
//! use `/EFF`.
use pdfcraft_crypt::{EncryptDict, Method, SecurityHandler, StreamKind};
use std::cell::Cell;
use pdfcraft_crypt::{EncryptDict, Method, SecurityHandler, StreamKind};
use pdfcraft_filters::MAX_CHAIN;
use crate::document::Held;
use crate::{Dict, Document, Object, PdfString, Stream};
/// How many references one read or write of an object may follow, with all that it loads on the
/// way. How an object is treated can be given by reference, and what that loads can do the same: a
/// file could make one read a tree of them, each object read again once the cache has let it go.
/// An object takes a few of these, and two dozen if every part of it is given by reference; past
/// this they read as absent. The parameters of an object stream, read together, count as one.
const MAX_FOLLOWED: u32 = 32;
std::thread_local! {
/// How many [`FollowScope`]s this thread is inside.
static SCOPES: Cell<u32> = const { Cell::new(0) };
/// The references followed in the outermost of them.
static FOLLOWED: Cell<u32> = const { Cell::new(0) };
/// The references refused in the outermost of them, past [`MAX_FOLLOWED`].
static REFUSED: Cell<u32> = const { Cell::new(0) };
}
/// One read or write of an object, with everything it loads on the way: what [`MAX_FOLLOWED`] counts.
pub(crate) struct FollowScope;
impl FollowScope {
pub(crate) fn enter() -> Self {
if SCOPES.replace(SCOPES.get().saturating_add(1)) == 0 {
FOLLOWED.set(0);
REFUSED.set(0);
}
FollowScope
}
}
impl Drop for FollowScope {
fn drop(&mut self) {
SCOPES.set(SCOPES.get().saturating_sub(1));
}
}
/// What `f` gives, counted against [`MAX_FOLLOWED`], or `spent` past it. Called inside a
/// [`FollowScope`], which is what starts the count again.
pub(crate) fn following<T>(spent: T, f: impl FnOnce() -> T) -> T {
if FOLLOWED.get() >= MAX_FOLLOWED {
REFUSED.set(REFUSED.get().saturating_add(1));
return spent;
}
FOLLOWED.set(FOLLOWED.get() + 1);
f()
}
/// `o`, or the object it refers to, for what decides how a stream or dictionary is treated. Past
/// [`MAX_FOLLOWED`] a reference reads as absent.
fn followed<'o>(doc: &Document, o: &'o Object) -> Held<'o> {
match o {
Object::Ref(_) => following(Held::Direct(&Object::Null), || doc.held(o)),
o => Held::Direct(o),
}
}
fn bytes_of(doc: &Document, d: &Dict, key: &[u8]) -> Vec<u8> {
d.get(key).map(|v| doc.resolve(v)).and_then(|v| v.as_string().map(|s| s.bytes.clone())).unwrap_or_default()
}
@@ -18,12 +77,12 @@ fn bytes_of(doc: &Document, d: &Dict, key: &[u8]) -> Vec<u8> {
/// Read the handler parameters from an `/Encrypt` dictionary.
pub(crate) fn encrypt_dict(doc: &Document, d: &Dict) -> EncryptDict {
let int = |k: &[u8], default: i64| d.get(k).map(|v| doc.resolve(v)).and_then(|v| v.as_int()).unwrap_or(default);
let name = |k: &[u8]| d.get(k).and_then(|v| v.as_name()).map(|n| n.to_vec()).unwrap_or_else(|| b"Identity".to_vec());
let name = |d: &Dict, k: &[u8]| d.get(k).and_then(|v| doc.resolve(v).as_name().map(<[u8]>::to_vec));
let mut crypt_filters = Vec::new();
if let Some(cf) = d.get(b"CF").map(|c| doc.resolve(c)).and_then(|c| c.as_dict().cloned()) {
for (k, v) in cf.iter() {
let Some(f) = doc.resolve(v).as_dict().cloned() else { continue };
let method = match f.name(b"CFM") {
let method = match name(&f, b"CFM").as_deref() {
Some(b"V2") => Method::Rc4,
Some(b"AESV2") => Method::Aes128,
Some(b"AESV3") => Method::Aes256,
@@ -40,7 +99,8 @@ pub(crate) fn encrypt_dict(doc: &Document, d: &Dict) -> EncryptDict {
.get(b"CF")
.map(|c| doc.resolve(c))
.and_then(|c| c.as_dict().and_then(|c| c.get(b"StdCF").map(|f| doc.resolve(f))))
.and_then(|f| f.as_dict().and_then(|f| f.int(b"Length")))
.and_then(|f| f.as_dict().and_then(|f| f.get(b"Length").map(|n| doc.resolve(n))))
.and_then(|n| n.as_int())
.filter(|n| *n > 0);
let length_bits = match (v, cf_len) {
(5.., _) => 256,
@@ -50,7 +110,7 @@ pub(crate) fn encrypt_dict(doc: &Document, d: &Dict) -> EncryptDict {
_ => int(b"Length", 40),
};
EncryptDict {
filter: d.name(b"Filter").map(|n| n.to_vec()).unwrap_or_default(),
filter: name(d, b"Filter").unwrap_or_default(),
v,
r: int(b"R", 0),
length_bits,
@@ -60,28 +120,46 @@ pub(crate) fn encrypt_dict(doc: &Document, d: &Dict) -> EncryptDict {
ue: bytes_of(doc, d, b"UE"),
perms: bytes_of(doc, d, b"Perms"),
p: int(b"P", -1) as i32,
encrypt_metadata: !matches!(d.get(b"EncryptMetadata"), Some(Object::Bool(false))),
encrypt_metadata: !matches!(d.get(b"EncryptMetadata").map(|v| doc.resolve(v)).as_deref(), Some(Object::Bool(false))),
crypt_filters,
stm_f: name(b"StmF"),
str_f: name(b"StrF"),
ef_f: d.get(b"EFF").and_then(|v| v.as_name()).map(|n| n.to_vec()).unwrap_or_default(),
stm_f: name(d, b"StmF").unwrap_or_else(|| b"Identity".to_vec()),
str_f: name(d, b"StrF").unwrap_or_else(|| b"Identity".to_vec()),
ef_f: name(d, b"EFF").unwrap_or_default(),
}
}
/// The crypt filter a stream names in its `/Filter` chain (`/Crypt` + `/DecodeParms /Name`).
fn named_crypt_filter(dict: &Dict) -> Option<Vec<u8>> {
let names: Vec<&[u8]> = match dict.get(b"Filter") {
Some(Object::Name(n)) => vec![n],
Some(Object::Array(a)) => a.iter().filter_map(Object::as_name).collect(),
/// Whether `f`, an entry of a `/Filter` chain or its only filter, is `/Crypt`.
fn is_crypt(doc: &Document, f: &Object) -> bool {
followed(doc, f).as_name() == Some(b"Crypt")
}
/// The crypt filter a stream names in its `/Filter` chain (`/Crypt` + `/DecodeParms /Name`), each
/// given directly or by reference, in an array or not. A chain longer than any decoder takes names
/// none: nothing can read that stream anyway, and every stream of a file can share one long array.
/// A `/Name` that reads as absent means `/Identity`, unless a reference on the way to it was refused
/// (see [`MAX_FOLLOWED`]): that says nothing about the stream, so it names none, and the document's
/// own filter is used.
fn named_crypt_filter(doc: &Document, dict: &Dict) -> Option<Vec<u8>> {
let refused = REFUSED.get();
let filter = followed(doc, dict.get(b"Filter")?);
let i = match &*filter {
Object::Array(a) if a.len() <= MAX_CHAIN + 1 => a.iter().position(|f| is_crypt(doc, f))?,
f if is_crypt(doc, f) => 0,
_ => return None,
};
let i = names.iter().position(|n| *n == b"Crypt")?;
let parms = match dict.get(b"DecodeParms") {
Some(Object::Dict(d)) if i == 0 => Some(d.clone()),
Some(Object::Array(a)) => a.get(i).and_then(|p| p.as_dict().cloned()),
let parms = dict.get(b"DecodeParms").map(|p| followed(doc, p));
let entry = match parms.as_deref() {
Some(Object::Array(a)) => a.get(i).map(|p| followed(doc, p)),
Some(p) if i == 0 => Some(Held::Direct(p)),
_ => None,
};
Some(parms.and_then(|p| p.name(b"Name").map(|n| n.to_vec())).unwrap_or_else(|| b"Identity".to_vec()))
let name = entry.as_deref().and_then(Object::as_dict).and_then(|p| p.get(b"Name")).map(|n| followed(doc, n));
match name.as_deref().and_then(Object::as_name) {
Some(name) => Some(name.to_vec()),
// /Identity here would read ciphertext as data, and write the stream in the clear.
None if REFUSED.get() > refused => None,
None => Some(b"Identity".to_vec()),
}
}
/// A signature dictionary (field signature, document timestamp or usage-rights signature),
@@ -91,42 +169,67 @@ fn named_crypt_filter(dict: &Dict) -> Option<Vec<u8>> {
/// `/Contents` alongside an array `/ByteRange` (`/Type` is optional). A signature dictionary
/// has no `/Subtype`; a dictionary with one (an annotation, whose `/Contents` is the comment
/// text) or with another `/Type` is never exempt, whatever else it carries.
fn is_signature_dict(d: &Dict) -> bool {
if d.get(b"Subtype").is_some() {
/// With `doc` its entries are read through references; a dictionary nested in another object
/// has none (see [`transform`]).
fn is_signature_dict(doc: Option<&Document>, d: &Dict) -> bool {
// All this decides is what happens to such a /Contents.
if !matches!(d.get(b"Contents"), Some(Object::String(_))) {
return false;
}
match d.get(b"Type") {
Some(_) => matches!(d.name(b"Type"), Some(b"Sig" | b"DocTimeStamp")),
None => matches!(d.get(b"Contents"), Some(Object::String(_))) && matches!(d.get(b"ByteRange"), Some(Object::Array(_))),
// A null entry, or one that refers to nothing, is an absent one (ISO 32000-2 §7.3.7, §7.3.10).
let entry = |key: &[u8]| {
let value = d.get(key)?;
let held = doc.map_or(Held::Direct(value), |doc| followed(doc, value));
(!matches!(*held, Object::Null)).then_some(held)
};
if entry(b"Subtype").is_some() {
return false;
}
match entry(b"Type") {
Some(t) => matches!(t.as_name(), Some(b"Sig" | b"DocTimeStamp")),
None => matches!(entry(b"ByteRange").as_deref(), Some(Object::Array(_))),
}
}
/// Remove the `/Crypt` entry from a stream's filter chain, with its decode parameters.
fn drop_crypt_filter(dict: &mut Dict) {
let i = match dict.get(b"Filter") {
Some(Object::Array(a)) => a.iter().position(|n| n.as_name() == Some(b"Crypt")),
Some(Object::Name(n)) if n.as_slice() == b"Crypt" => Some(0),
_ => None,
/// Longer than any filter's name. A name can be of any length, and a chain every stream shares can
/// hold a long one.
const MAX_FILTER_NAME: usize = 127;
/// The dictionary of `source` without the `/Crypt` entries of its filter chain and their decode
/// parameters. The two come out direct when given by reference, which every stream of a file can
/// share, so of what was shared only small values are copied: filter names, and the parameters
/// decoders read ([`Document::with_direct_parms`]). The chain is short, as [`named_crypt_filter`]
/// found a filter in it.
fn drop_crypt_filter(doc: &Document, source: &Stream) -> Dict {
let entry = |f: &Object| match f {
f if is_crypt(doc, f) => Object::name("Crypt"),
Object::Name(n) if n.len() <= MAX_FILTER_NAME => f.clone(),
Object::Ref(_) => f.clone(),
_ => Object::Null,
};
let Some(i) = i else { return };
let mut filters = match dict.get(b"Filter") {
Some(Object::Array(a)) => a.clone(),
_ => Vec::new(),
let filter = source.dict.get(b"Filter").map(|f| followed(doc, f));
let chain: Vec<Object> = match filter.as_deref() {
Some(Object::Array(a)) => a.iter().map(entry).collect(),
Some(f) => vec![entry(f)],
None => Vec::new(),
};
if i < filters.len() {
filters.remove(i);
}
let crypt: Vec<bool> = chain.iter().map(|f| f.as_name() == Some(b"Crypt")).collect();
let mut chained = Stream::from_raw(source.dict.clone(), Vec::new());
chained.dict.set(b"Filter".to_vec(), Object::Array(chain.clone()));
// A lone parameter dictionary belongs to the first filter (as `named_crypt_filter` reads it).
match dict.get(b"DecodeParms").cloned() {
Some(Object::Array(mut parms)) if i < parms.len() => {
parms.remove(i);
dict.set(b"DecodeParms".to_vec(), Object::Array(parms));
}
Some(Object::Dict(_)) if i == 0 => {
let parms = match doc.with_direct_parms(&chained).dict.get(b"DecodeParms") {
Some(Object::Array(p)) => Some(Object::Array(p.iter().zip(&crypt).filter(|(_, c)| !**c).map(|(p, _)| p.clone()).collect())),
Some(Object::Dict(_)) if crypt.first() == Some(&true) => None,
_ => source.dict.get(b"DecodeParms").cloned(),
};
let mut dict = source.dict.clone();
match parms {
Some(p) => dict.set(b"DecodeParms".to_vec(), p),
None => {
dict.remove(b"DecodeParms");
}
_ => {}
}
let filters: Vec<Object> = chain.into_iter().zip(&crypt).filter(|(_, c)| !**c).map(|(f, _)| f).collect();
if filters.is_empty() {
dict.remove(b"Filter");
if matches!(dict.get(b"DecodeParms"), Some(Object::Array(p)) if p.is_empty()) {
@@ -135,6 +238,7 @@ fn drop_crypt_filter(dict: &mut Dict) {
} else {
dict.set(b"Filter".to_vec(), Object::Array(filters));
}
dict
}
/// What [`transform`] does to an object.
@@ -152,51 +256,81 @@ pub(crate) enum Pass {
/// Decrypt or encrypt every string and stream in an indirect object, except a signature
/// dictionary's `/Contents` (see [`is_signature_dict`]).
pub(crate) fn transform(h: &SecurityHandler, o: &Object, num: u32, generation: u16, pass: Pass) -> Object {
///
/// `o` is as `doc` holds it: what a stream's dictionary gives by reference is read from `doc`, so
/// its references have to be the document's, not a save's renumbered ones.
pub(crate) fn transform(doc: &Document, h: &SecurityHandler, o: &Object, num: u32, generation: u16, pass: Pass) -> Object {
let _scope = FollowScope::enter();
let decrypt = pass == Pass::Decrypt;
let strings = |s: &PdfString| {
let bytes = if decrypt { h.decrypt_string(num, generation, &s.bytes) } else { h.encrypt_string(num, generation, &s.bytes) };
PdfString { bytes, hex: s.hex || !decrypt }
};
walk(o, &strings, h, num, generation, pass)
// References are followed here only, for the object itself, never for what is nested in it.
// `walk` goes deep, and a load from down there adds that depth again for each object loaded,
// which overflowed the stack on a 10 KB file. A stream can't be nested, and a dictionary that
// is reads as it says directly.
match o {
Object::Stream(s) => stream(doc, s, &strings, h, num, generation, pass),
Object::Dict(d) if is_signature_dict(Some(doc), d) => signature_dict(d, &strings),
o => walk(o, &strings),
}
}
fn walk(o: &Object, strings: &dyn Fn(&PdfString) -> PdfString, h: &SecurityHandler, num: u32, generation: u16, pass: Pass) -> Object {
let decrypt = pass == Pass::Decrypt;
/// `d`, a signature dictionary, with the strings of everything but its `/Contents` transformed.
fn signature_dict(d: &Dict, strings: &dyn Fn(&PdfString) -> PdfString) -> Object {
Object::Dict(
d.iter()
.map(|(k, v)| match (k.as_slice(), v) {
(b"Contents", Object::String(_)) => (k.clone(), v.clone()),
_ => (k.clone(), walk(v, strings)),
})
.collect(),
)
}
fn walk(o: &Object, strings: &dyn Fn(&PdfString) -> PdfString) -> Object {
match o {
Object::String(s) => Object::String(strings(s)),
Object::Array(a) => Object::Array(a.iter().map(|x| walk(x, strings, h, num, generation, pass)).collect()),
Object::Dict(d) if is_signature_dict(d) => Object::Dict(
d.iter()
.map(|(k, v)| match (k.as_slice(), v) {
(b"Contents", Object::String(_)) => (k.clone(), v.clone()),
_ => (k.clone(), walk(v, strings, h, num, generation, pass)),
})
.collect(),
),
Object::Dict(d) => Object::Dict(d.iter().map(|(k, v)| (k.clone(), walk(v, strings, h, num, generation, pass))).collect()),
Object::Stream(s) => {
let mut dict: Dict = s.dict.iter().map(|(k, v)| (k.clone(), walk(v, strings, h, num, generation, pass))).collect();
let ty = s.dict.name(b"Type");
// Cross-reference streams are never encrypted (§7.5.8), but only reading may trust the
// label: a stream a producer calls /XRef stays plaintext if writing believes it, and the
// writer builds its own cross-reference streams outside this transform.
if (decrypt && ty == Some(b"XRef")) || (ty == Some(b"Metadata") && !h.encrypts_metadata()) {
return Object::Stream(Stream { dict, raw: s.raw.clone() });
}
let mut named = named_crypt_filter(&s.dict);
if pass == Pass::EncryptReplacing && named.as_deref().is_some_and(|n| !h.dict().crypt_filters.iter().any(|(f, _)| f == n)) {
drop_crypt_filter(&mut dict);
named = None;
}
let kind = match (&named, ty) {
(Some(n), _) => StreamKind::Named(n),
(None, Some(b"EmbeddedFile")) => StreamKind::EmbeddedFile,
_ => StreamKind::Normal,
};
let raw = if decrypt { h.decrypt_stream(num, generation, &s.raw, kind) } else { h.encrypt_stream(num, generation, &s.raw, kind) };
Object::Stream(Stream { dict, raw: raw.into() })
}
Object::Array(a) => Object::Array(a.iter().map(|x| walk(x, strings)).collect()),
Object::Dict(d) if is_signature_dict(None, d) => signature_dict(d, strings),
Object::Dict(d) => Object::Dict(d.iter().map(|(k, v)| (k.clone(), walk(v, strings))).collect()),
other => other.clone(),
}
}
fn stream(
doc: &Document,
s: &Stream,
strings: &dyn Fn(&PdfString) -> PdfString,
h: &SecurityHandler,
num: u32,
generation: u16,
pass: Pass,
) -> Object {
let decrypt = pass == Pass::Decrypt;
let walked = |d: &Dict| -> Dict { d.iter().map(|(k, v)| (k.clone(), walk(v, strings))).collect() };
let ty = s.dict.get(b"Type").map(|t| followed(doc, t));
let ty = ty.as_deref().and_then(Object::as_name);
// Cross-reference streams are never encrypted (§7.5.8), but only reading may trust the
// label: a stream a producer calls /XRef stays plaintext if writing believes it, and the
// writer builds its own cross-reference streams outside this transform.
if (decrypt && ty == Some(b"XRef")) || (ty == Some(b"Metadata") && !h.encrypts_metadata()) {
return Object::Stream(Stream { dict: walked(&s.dict), raw: s.raw.clone() });
}
let mut named = named_crypt_filter(doc, &s.dict);
// Before the walk, which then encrypts what is in the dictionary this leaves.
let foreign = pass == Pass::EncryptReplacing && named.as_deref().is_some_and(|n| !h.dict().crypt_filters.iter().any(|(f, _)| f == n));
let dropped = foreign.then(|| drop_crypt_filter(doc, s));
if dropped.is_some() {
named = None;
}
let dict = walked(dropped.as_ref().unwrap_or(&s.dict));
let kind = match (&named, ty) {
(Some(n), _) => StreamKind::Named(n),
(None, Some(b"EmbeddedFile")) => StreamKind::EmbeddedFile,
_ => StreamKind::Normal,
};
let raw = if decrypt { h.decrypt_stream(num, generation, &s.raw, kind) } else { h.encrypt_stream(num, generation, &s.raw, kind) };
Object::Stream(Stream { dict, raw: raw.into() })
}
crates/cos/src/writer.rs
+61 -15
View File
@@ -133,11 +133,12 @@ fn write_stream(s: &Stream, out: &mut Vec<u8>) {
/// The object as written: encrypted with its (output) number when the document is encrypted,
/// except the `/Encrypt` dictionary itself (`source_num` is its number in the source document).
/// `o` has the document's numbers in its references ([`crate::security::transform`]).
fn prepared<'a>(doc: &Document, source_num: u32, num: u32, generation: u16, o: &'a Object) -> Cow<'a, Object> {
match doc.output_security() {
(Some(h), encrypt_num) if Some(source_num) != encrypt_num => {
(Some(h), encrypt_num) if Some(source_num) != encrypt_num && !doc.stored_as_is(source_num) => {
let pass = if doc.encryption_changed() { crate::security::Pass::EncryptReplacing } else { crate::security::Pass::Encrypt };
Cow::Owned(crate::security::transform(h, o, num, generation, pass))
Cow::Owned(crate::security::transform(doc, h, o, num, generation, pass))
}
_ => Cow::Borrowed(o),
}
@@ -287,6 +288,10 @@ pub fn write_full(doc: &Document, opts: &SaveOptions) -> Result<Vec<u8>, CosErro
queue.push_back(r);
}
}
// A direct /Encrypt can refer to objects, /CF or /StmF, say.
if let Some(encrypt @ Object::Dict(_)) = trailer_in.get(b"Encrypt") {
visit_refs(encrypt, &mut queue);
}
// Unknown trailer entries are document data. Objects they alone reference must be written
// too, or a full rewrite would keep the key and drop its value.
for (key, value) in trailer_in.iter() {
@@ -302,8 +307,17 @@ pub fn write_full(doc: &Document, opts: &SaveOptions) -> Result<Vec<u8>, CosErro
if matches!(*o, Object::Null) {
continue; // dangling reference: written as null by renumbering below
}
map.insert(r, output_number(order.len())?);
order.push((r, matches!(*o, Object::Stream(_))));
// Accepted xref/header aliases denote one slot, including page-tree cycles.
// Check for null first: an unrelated or freed generation must remain dangling.
let canonical = ObjRef::new(r.num, reader.document.generation(r.num));
if let Some(&num) = map.get(&canonical) {
map.insert(r, num);
continue;
}
let num = output_number(order.len())?;
map.insert(r, num);
map.insert(canonical, num);
order.push((canonical, matches!(*o, Object::Stream(_))));
visit_refs(&o, &mut queue);
}
let version = reader.document.version().chars().take(3).collect::<String>();
@@ -314,10 +328,15 @@ pub fn write_full(doc: &Document, opts: &SaveOptions) -> Result<Vec<u8>, CosErro
out.extend_from_slice(&[b'%', 0xE2, 0xE3, 0xCF, 0xD3, b'\n']);
// The /Encrypt dictionary and everything it references (e.g. indirect /CF crypt-filter
// dictionaries) are needed before anything can be decrypted: never inside object streams.
let mut encryption_objects: std::collections::HashSet<ObjRef> = std::collections::HashSet::new();
let mut pending: Vec<ObjRef> = trailer_in.reference(b"Encrypt").into_iter().collect();
let mut encryption_objects: std::collections::HashSet<u32> = std::collections::HashSet::new();
let mut pending: Vec<ObjRef> = Vec::new();
if let Some(encrypt) = trailer_in.get(b"Encrypt") {
collect_refs(encrypt, &mut |r| pending.push(r));
}
while let Some(r) = pending.pop() {
if encryption_objects.insert(r) {
if let Some(&num) = map.get(&r)
&& encryption_objects.insert(num)
{
collect_refs(&reader.get(r), &mut |x| pending.push(x));
}
}
@@ -325,20 +344,22 @@ pub fn write_full(doc: &Document, opts: &SaveOptions) -> Result<Vec<u8>, CosErro
// damaged and could not be repaired", before it even asks for the password), while other
// readers accept it. qpdf keeps the catalog out of object streams too. Verified with
// Acrobat DC on macOS (#774).
let root_outside = if reader.document.output_security().0.is_some() { trailer_in.reference(b"Root") } else { None };
let root_outside =
if reader.document.output_security().0.is_some() { trailer_in.reference(b"Root").and_then(|r| map.get(&r)).copied() } else { None };
let mut rows: BTreeMap<u32, Row> = BTreeMap::new();
// Keep only references, not renamed copies of the entire object graph. Standalone
// objects still precede object streams, preserving the established output order.
let mut packed: Vec<(u32, ObjRef)> = Vec::new();
for (i, (r, is_stream)) in order.iter().enumerate() {
let num = output_number(i)?;
let in_stream = opts.object_streams && !is_stream && !encryption_objects.contains(r) && root_outside != Some(*r);
let in_stream = opts.object_streams && !is_stream && !encryption_objects.contains(&num) && root_outside != Some(num);
if in_stream {
packed.push((num, *r));
} else {
let o = renumber(&*reread(&mut reader, *r, *is_stream)?, &map);
let o = reread(&mut reader, *r, *is_stream)?;
let o = renumber(&prepared(&reader.document, r.num, num, 0, &o), &map);
rows.insert(num, Row::InFile(out.len() as u64, 0));
write_indirect(num, 0, &prepared(&reader.document, r.num, num, 0, &o), &mut out);
write_indirect(num, 0, &o, &mut out);
}
}
// Object streams take the numbers after the last object. Their contents are not encrypted
@@ -368,11 +389,11 @@ pub fn write_full(doc: &Document, opts: &SaveOptions) -> Result<Vec<u8>, CosErro
trailer.set(key.to_vec(), Object::Ref(ObjRef::new(*r, 0)));
}
}
if let Some(Object::Dict(d)) = trailer_in.get(b"Encrypt") {
trailer.set(b"Encrypt".to_vec(), Object::Dict(d.clone()));
if let Some(encrypt @ Object::Dict(_)) = trailer_in.get(b"Encrypt") {
trailer.set(b"Encrypt".to_vec(), renumber(encrypt, &map));
}
if let Some(id) = trailer_in.get(b"ID") {
trailer.set(b"ID".to_vec(), id.clone());
trailer.set(b"ID".to_vec(), direct_id(&reader.document, id));
}
for (key, value) in trailer_in.iter() {
if !trailer_key_rewritten(key) {
@@ -442,6 +463,22 @@ impl ObjectStreamBatch {
}
}
/// `/ID` for the trailer, direct, as it has to be with `/Encrypt`: nothing else in the file holds
/// what it refers to now. Given by reference it is two entries at most, however many references a
/// file gave, so that one large string isn't copied for each. An entry that isn't a string stays in
/// its place as an empty one, which is how the key read it.
fn direct_id(doc: &Document, id: &Object) -> Object {
if matches!(id, Object::Array(a) if !a.iter().any(|s| matches!(s, Object::Ref(_)))) {
return id.clone();
}
let entry = |s: &Object| match &*doc.resolve(s) {
s @ Object::String(_) => s.clone(),
_ => Object::String(crate::PdfString { bytes: Vec::new(), hex: true }),
};
let held = doc.resolve(id);
Object::Array(held.as_array().into_iter().flatten().take(2).map(entry).collect())
}
/// Trailer keys a full rewrite generates itself. Everything else is copied, with references
/// renumbered, so a save does not drop metadata the reader accepted.
fn trailer_key_rewritten(key: &[u8]) -> bool {
@@ -486,7 +523,7 @@ pub(crate) fn unlinked_additions(doc: &Document) -> HashSet<u32> {
pending.into_keys().collect()
}
fn collect_refs(o: &Object, f: &mut impl FnMut(ObjRef)) {
pub(crate) fn collect_refs(o: &Object, f: &mut impl FnMut(ObjRef)) {
match o {
Object::Ref(r) => f(*r),
Object::Array(a) => a.iter().for_each(|x| collect_refs(x, f)),
@@ -663,6 +700,15 @@ fn civil_from_days(days: i64) -> (i64, u32, u32) {
#[cfg(test)]
mod tests {
/// `/ID` by reference comes out direct, each entry where it was: the key is made from the first.
#[test]
fn a_file_identifier_keeps_its_entries_in_their_places() {
let doc = Document::new_empty();
let string = |bytes: &[u8]| Object::String(crate::PdfString { bytes: bytes.to_vec(), hex: true });
let nowhere = Object::Ref(ObjRef::new(9_999, 0));
assert_eq!(super::direct_id(&doc, &Object::Array(vec![nowhere, string(b"ab")])), Object::Array(vec![string(b""), string(b"ab")]));
}
use proptest::prelude::*;
use super::*;
crates/cos/tests/it/encryption.rs
+823
View File
@@ -494,3 +494,826 @@ fn a_stream_labelled_xref_is_encrypted_when_saved() {
let bytes = write_full(&doc, &SaveOptions { object_streams: false, ..SaveOptions::default() }).unwrap();
assert!(!bytes.windows(SECRET.len()).any(|w| w == SECRET.as_bytes()), "the labelled stream was written in the clear");
}
/// The content stream's reference.
fn contents_of(doc: &Document) -> ObjRef {
let catalog = doc.get(doc.root().unwrap()).as_dict().cloned().unwrap();
let pages = doc.get(catalog.reference(b"Pages").unwrap()).as_dict().cloned().unwrap();
let page = doc.resolve(&pages.get(b"Kids").unwrap().as_array().unwrap()[0]).as_dict().cloned().unwrap();
page.reference(b"Contents").unwrap()
}
/// The content stream's data as loaded, whatever its filters say.
fn raw_content(doc: &Document) -> String {
match &*doc.get(contents_of(doc)) {
Object::Stream(s) => String::from_utf8_lossy(&s.raw).trim().to_string(),
_ => panic!("no content stream"),
}
}
/// Gives the content stream the `/Filter` and `/DecodeParms` of `selector`.
fn select(doc: &mut Document, (filter, parms): (Object, Object)) {
let contents = contents_of(doc);
let Object::Stream(mut s) = doc.get(contents).as_ref().clone() else { panic!("no content stream") };
s.dict.set(b"Filter".to_vec(), filter);
s.dict.set(b"DecodeParms".to_vec(), parms);
doc.set(contents, Object::Stream(s));
}
/// `/Crypt` naming `name`, in the `form`th spelling: all direct, then each part by reference.
/// `None` past the last.
fn selector(doc: &mut Document, form: usize, name: &str) -> Option<(Object, Object)> {
let crypt = Object::name("Crypt");
let parms = |n: Object| {
let mut d = pdfcraft_cos::Dict::new();
d.set(b"Name".to_vec(), n);
Object::Dict(d)
};
let named = parms(Object::name(name));
Some(match form {
0 => (crypt, named),
1 => (crypt, Object::Ref(doc.add(named))),
2 => (Object::Array(vec![crypt]), Object::Array(vec![Object::Ref(doc.add(named))])),
3 => (Object::Array(vec![crypt]), Object::Ref(doc.add(Object::Array(vec![named])))),
4 => (crypt, parms(Object::Ref(doc.add(Object::name(name))))),
5 => (Object::Ref(doc.add(crypt)), named),
6 => (Object::Ref(doc.add(Object::Array(vec![crypt]))), Object::Array(vec![named])),
7 => (Object::Array(vec![Object::Ref(doc.add(crypt))]), Object::Array(vec![named])),
// Behind a filter given by reference: places in the two arrays pair up all the same.
8 => (Object::Array(vec![Object::Ref(doc.add(Object::name("ASCIIHexDecode"))), crypt]), Object::Array(vec![Object::Null, named])),
// The two entries in different forms, which is how a lone filter is usually spelled.
9 => (Object::Array(vec![crypt]), named),
10 => (crypt, Object::Array(vec![named])),
// The filter twice, which names the first's.
11 => (Object::Array(vec![crypt.clone(), crypt]), Object::Array(vec![named.clone(), named])),
// A reference to a reference.
12 => {
let name = doc.add(Object::name(name));
let hop = doc.add(Object::Ref(name));
(crypt, parms(Object::Ref(hop)))
}
_ => return None,
})
}
fn holds(bytes: &[u8], text: &str) -> bool {
String::from_utf8_lossy(bytes).contains(text)
}
/// An object `/Encrypt` names is written as stored only while it is as stored. A file whose
/// `/Encrypt` names its `/Info` reads that dictionary as stored; once the title is edited, both
/// saves encrypt it rather than writing the new title in the clear.
#[test]
fn an_edited_object_that_encrypt_names_is_encrypted() {
let mut doc = Document::open(Arc::new(fixture())).unwrap();
doc.set_encryption(&NewEncryption {
algorithm: Algorithm::Aes128,
user_password: "",
owner_password: "owner",
permissions: -1,
encrypt_metadata: true,
seed: [9; 32],
})
.unwrap();
let info = doc.trailer().reference(b"Info").unwrap();
let mut dict = doc.get(doc.trailer().reference(b"Encrypt").unwrap()).as_dict().cloned().unwrap();
dict.set(b"Junk".to_vec(), Object::Ref(info));
doc.trailer_mut().set(b"Encrypt".to_vec(), Object::Dict(dict));
let bytes = write_full(&doc, &SaveOptions::default()).unwrap();
let mut doc = Document::open(Arc::new(bytes)).unwrap();
let info = doc.trailer().reference(b"Info").unwrap();
doc.update_dict(info, |d| d.set(b"Title".to_vec(), PdfString::text("Merger plan"))).unwrap();
for (save, saved) in [("incremental", write_incremental(&doc, &SaveOptions::default())), ("full", write_full(&doc, &SaveOptions::default()))] {
assert!(!holds(&saved.unwrap(), "Merger plan"), "{save}: the edited title was written in the clear");
}
}
/// A stream can select its crypt filter by reference at every step: the chain, an entry of it, the
/// parameters, an entry of those, the name (ISO 32000-2 §7.3.10, §7.4.10). Read as absent, the
/// filter was `/Identity`, so such a stream was saved in the clear and, in a file from elsewhere,
/// its ciphertext read as data. A full save renumbers: what it follows has the document's numbers.
#[test]
fn a_crypt_filter_selected_by_reference_is_used() {
for alg in [Algorithm::Aes128, Algorithm::Aes256] {
for form in 0.. {
let mut doc = Document::open(Arc::new(protect(alg, "", "owner", -1))).unwrap();
let Some(selected) = selector(&mut doc, form, "StdCF") else { break };
select(&mut doc, selected);
let saved = write_incremental(&doc, &SaveOptions::default()).unwrap();
assert!(!holds(&saved, "Secret page"), "{alg:?} form {form}: saved encrypted");
let back = Document::open(Arc::new(saved)).unwrap();
assert_eq!(raw_content(&back), SECRET, "{alg:?} form {form}: read back");
let full = write_full(&back, &SaveOptions::default()).unwrap();
assert!(!holds(&full, "Secret page"), "{alg:?} form {form}: renumbered, encrypted");
assert_eq!(raw_content(&Document::open(Arc::new(full)).unwrap()), SECRET, "{alg:?} form {form}: renumbered");
}
}
}
/// A reference to a stream that takes `follows` references to read cold, this one included: its
/// nine filters are all read, the last leading to the next stream.
fn spends(doc: &mut Document, follows: usize) -> Object {
let left = follows - 1;
let mut filters: Vec<Object> = (0..left.min(8)).map(|i| Object::Ref(doc.add(Object::Int(i as i64)))).collect();
if left > 8 {
filters.push(spends(doc, left - 8));
}
let mut dict = pdfcraft_cos::Dict::new();
dict.set(b"Filter".to_vec(), Object::Array(filters));
Object::Ref(doc.add(Object::Stream(pdfcraft_cos::Stream::from_raw(dict, Vec::new()))))
}
/// Gives the content stream a `/Type` that takes `follows` references to read, ahead of its filter.
fn spend_first(doc: &mut Document, follows: usize) {
let spent = spends(doc, follows);
let contents = contents_of(doc);
let Object::Stream(mut s) = doc.get(contents).as_ref().clone() else { panic!("no content stream") };
s.dict.set(b"Type".to_vec(), spent);
doc.set(contents, Object::Stream(s));
}
/// Past what one read may follow a reference reads as absent, and a crypt filter with no `/Name` is
/// `/Identity`. Together those would read this stream's ciphertext as data, so what can't be told
/// is the document's own filter, as a `/Filter` that can't be read is.
#[test]
fn a_crypt_filter_past_what_a_read_may_follow_is_the_documents_own() {
let mut doc = Document::open(Arc::new(protect(Algorithm::Aes128, "", "owner", -1))).unwrap();
let selected = selector(&mut doc, 1, "StdCF").unwrap();
select(&mut doc, selected);
spend_first(&mut doc, 40);
let saved = write_incremental(&doc, &SaveOptions::default()).unwrap();
assert!(!holds(&saved, "Secret page"), "saved encrypted");
let back = Document::open(Arc::new(saved)).unwrap();
assert_eq!(raw_content(&back), SECRET);
let full = write_full(&back, &SaveOptions::default()).unwrap();
assert!(!holds(&full, "Secret page"), "renumbered, encrypted");
assert_eq!(raw_content(&Document::open(Arc::new(full)).unwrap()), SECRET, "renumbered");
}
/// A read that takes all it may follow has refused nothing: a crypt filter with no `/Name` after it
/// is still `/Identity`, and the stream is in the clear.
#[test]
fn a_crypt_filter_with_no_name_after_all_a_read_may_follow_is_identity() {
let mut doc = Document::open(Arc::new(protect(Algorithm::Aes128, "", "owner", -1))).unwrap();
select(&mut doc, (Object::name("Crypt"), Object::Dict(pdfcraft_cos::Dict::new())));
spend_first(&mut doc, 32);
let saved = write_incremental(&doc, &SaveOptions::default()).unwrap();
assert!(holds(&saved, "Secret page"), "written as the stream asks");
assert_eq!(raw_content(&Document::open(Arc::new(saved)).unwrap()), SECRET);
}
/// `/Identity` selected by reference keeps a stream in the clear in a document that encrypts the
/// rest, in every spelling, on both saves.
#[test]
fn an_identity_crypt_filter_selected_by_reference_leaves_a_stream_alone() {
for form in 0.. {
let mut doc = Document::open(Arc::new(protect(Algorithm::Aes128, "", "owner", -1))).unwrap();
let Some(selected) = selector(&mut doc, form, "Identity") else { break };
select(&mut doc, selected);
let saved = write_incremental(&doc, &SaveOptions::default()).unwrap();
assert!(holds(&saved, "Secret page"), "form {form}: written as the stream asks");
let back = Document::open(Arc::new(saved)).unwrap();
assert_eq!(raw_content(&back), SECRET, "form {form}");
let full = write_full(&back, &SaveOptions::default()).unwrap();
assert!(holds(&full, "Secret page"), "form {form}: renumbered");
assert_eq!(raw_content(&Document::open(Arc::new(full)).unwrap()), SECRET, "form {form}: renumbered");
}
}
/// New protection drops a crypt filter it doesn't define, whichever way the stream selected it. The
/// chain comes out direct. Parameters in the stream's own dictionary stay, encrypted with it; of an
/// array in an object of its own, which every stream can share, only what decoders read is copied.
#[test]
fn protecting_drops_a_crypt_filter_selected_by_reference() {
let protected = |doc: &mut Document| {
doc.set_encryption(&NewEncryption {
algorithm: Algorithm::Aes256,
user_password: "",
owner_password: "owner",
permissions: -1,
encrypt_metadata: true,
seed: [7; 32],
})
.unwrap();
write_full(doc, &SaveOptions::default()).unwrap()
};
for form in 0.. {
let mut doc = Document::open(Arc::new(fixture())).unwrap();
let Some(selected) = selector(&mut doc, form, "MyCF") else { break };
select(&mut doc, selected);
let bytes = protected(&mut doc);
assert!(!holds(&bytes, "Secret page"), "form {form}: encrypted");
assert!(!holds(&bytes, "MyCF"), "form {form}: no selector left in the stream");
let back = Document::open(Arc::new(bytes.clone())).unwrap();
assert_eq!(raw_content(&back), SECRET, "form {form}");
let contents = back.get(contents_of(&back));
let filter = contents.as_dict().unwrap().get(b"Filter");
match form {
8 => assert!(
matches!(filter, Some(Object::Array(a)) if matches!(a.as_slice(), [Object::Ref(_)])),
"form 8: the other filter stays: {filter:?}"
),
_ => assert_eq!(filter, None, "form {form}"),
}
assert!(hayro_syntax::Pdf::new(bytes).is_ok(), "form {form}");
}
for shared in [false, true] {
let mut doc = Document::open(Arc::new(fixture())).unwrap();
let mut named = pdfcraft_cos::Dict::new();
named.set(b"Name".to_vec(), Object::name("MyCF"));
let mut other = pdfcraft_cos::Dict::new();
other.set(b"Note".to_vec(), Object::String(PdfString::text("Private note")));
other.set(b"Predictor".to_vec(), Object::Int(1));
let parms = Object::Array(vec![Object::Dict(named), Object::Dict(other)]);
let parms = if shared { Object::Ref(doc.add(parms)) } else { parms };
select(&mut doc, (Object::Array(vec![Object::name("Crypt"), Object::name("ASCIIHexDecode")]), parms));
let bytes = protected(&mut doc);
assert!(!holds(&bytes, "Private note"), "shared: {shared}");
let back = Document::open(Arc::new(bytes)).unwrap();
let contents = back.get(contents_of(&back));
let kept = contents.as_dict().unwrap().get(b"DecodeParms").unwrap().as_array().unwrap();
assert_eq!(kept.len(), 1, "one filter left, one entry");
let kept = kept[0].as_dict().unwrap();
assert_eq!(kept.get(b"Predictor"), Some(&Object::Int(1)), "shared: {shared}");
assert_eq!(kept.get(b"Note").is_some(), !shared, "shared: {shared}");
}
}
/// Every stream of a file can share one large object as its `/DecodeParms` or `/Filter`, which
/// costs nothing until new protection drops a crypt filter from each. Copied whole, a 1 MB file
/// came out at 1 GB.
#[test]
fn protecting_copies_no_large_object_into_every_stream() {
let pad = Object::Array(vec![Object::Int(1); 100_000]);
let mut doc = Document::open(Arc::new(fixture())).unwrap();
let mut large = pdfcraft_cos::Dict::new();
large.set(b"Pad".to_vec(), pad.clone());
let parms = doc.add(Object::Array(vec![Object::Null, Object::Dict(large)]));
let filter = doc.add(Object::Array(vec![Object::name("Crypt"), Object::Name(vec![b'A'; 100_000]), pad]));
let mut streams = Vec::new();
for _ in 0..200 {
let mut by_parms = pdfcraft_cos::Dict::new();
by_parms.set(b"Filter".to_vec(), Object::Array(vec![Object::name("Crypt"), Object::name("ASCIIHexDecode")]));
by_parms.set(b"DecodeParms".to_vec(), Object::Ref(parms));
streams.push(Object::Ref(doc.add(Object::Stream(pdfcraft_cos::Stream::from_raw(by_parms, b"41>".to_vec())))));
let mut by_filter = pdfcraft_cos::Dict::new();
by_filter.set(b"Filter".to_vec(), Object::Ref(filter));
streams.push(Object::Ref(doc.add(Object::Stream(pdfcraft_cos::Stream::from_raw(by_filter, b"41>".to_vec())))));
}
doc.update_dict(doc.root().unwrap(), |c| c.set(b"Test".to_vec(), Object::Array(streams))).unwrap();
doc.set_encryption(&NewEncryption {
algorithm: Algorithm::Aes256,
user_password: "",
owner_password: "owner",
permissions: -1,
encrypt_metadata: true,
seed: [7; 32],
})
.unwrap();
let bytes = write_full(&doc, &SaveOptions { object_streams: false, ..SaveOptions::default() }).unwrap();
assert!(bytes.len() < 1 << 20, "{} bytes", bytes.len());
}
/// A file can name one large string thousands of times in `/ID`, which holds two. An encrypted file
/// keeps its `/ID`, so each copy would be in the file a save writes.
#[test]
fn a_file_identifier_of_many_references_is_two_strings() {
let mut doc = Document::open(Arc::new(fixture())).unwrap();
doc.set_encryption(&NewEncryption {
algorithm: Algorithm::Aes128,
user_password: "",
owner_password: "owner",
permissions: -1,
encrypt_metadata: true,
seed: [9; 32],
})
.unwrap();
let large = doc.add(Object::String(PdfString { bytes: vec![0xAB; 100_000], hex: true }));
doc.trailer_mut().set(b"ID".to_vec(), Object::Array(vec![Object::Ref(large); 500]));
let bytes = write_full(&doc, &SaveOptions { object_streams: false, ..SaveOptions::default() }).unwrap();
assert!(bytes.len() < 1 << 20, "{} bytes", bytes.len());
}
/// What a file can do with those references: point a stream's selector at the stream itself, or at
/// nothing. They read as absent, so a stream with `/Crypt` is `/Identity` (ISO 32000-2 §7.4.10),
/// and one that names no filter follows the document's. Loading ends either way.
#[test]
fn a_crypt_filter_reference_to_itself_or_nothing_reads_as_absent() {
let named = |n: Object| {
let mut d = pdfcraft_cos::Dict::new();
d.set(b"Name".to_vec(), n);
Object::Dict(d)
};
for (what, encrypted) in [
("/Filter to itself", true),
("/Filter nowhere", true),
("/DecodeParms to itself", false),
("/DecodeParms nowhere", false),
("/Name to itself", false),
("/Name nowhere", false),
] {
let mut doc = Document::open(Arc::new(protect(Algorithm::Aes128, "", "owner", -1))).unwrap();
let me = Object::Ref(contents_of(&doc));
let nowhere = Object::Ref(ObjRef::new(9_999, 0));
let crypt = Object::name("Crypt");
select(
&mut doc,
match what {
"/Filter to itself" => (me, Object::Null),
"/Filter nowhere" => (nowhere, Object::Null),
"/DecodeParms to itself" => (crypt, me),
"/DecodeParms nowhere" => (crypt, nowhere),
"/Name to itself" => (crypt, named(me)),
_ => (crypt, named(nowhere)),
},
);
let saved = write_incremental(&doc, &SaveOptions::default()).unwrap();
assert_eq!(!holds(&saved, "Secret page"), encrypted, "{what}");
assert_eq!(raw_content(&Document::open(Arc::new(saved)).unwrap()), SECRET, "{what}");
}
}
/// A lone parameter dictionary belongs to the first filter, so `/Crypt` behind another filter has
/// none of its own and is `/Identity` (ISO 32000-2 §7.4.1), whatever that dictionary says.
#[test]
fn a_lone_parameter_dictionary_is_the_first_filters() {
let mut doc = Document::open(Arc::new(protect(Algorithm::Aes128, "", "owner", -1))).unwrap();
let mut parms = pdfcraft_cos::Dict::new();
parms.set(b"Name".to_vec(), Object::name("StdCF"));
let first = Object::Array(vec![Object::name("Crypt"), Object::name("ASCIIHexDecode")]);
let second = Object::Array(vec![Object::name("ASCIIHexDecode"), Object::name("Crypt")]);
select(&mut doc, (first, Object::Dict(parms.clone())));
assert!(!holds(&write_incremental(&doc, &SaveOptions::default()).unwrap(), "Secret page"), "first: it names StdCF");
select(&mut doc, (second, Object::Dict(parms)));
assert!(holds(&write_incremental(&doc, &SaveOptions::default()).unwrap(), "Secret page"), "second: it names nothing");
}
/// A lone parameter dictionary is the first filter's, so dropping a `/Crypt` behind that filter
/// leaves it, and dropping one in front of it takes it away.
#[test]
fn protecting_keeps_the_parameters_of_the_filter_before_a_crypt_filter() {
let protected = |filters: Vec<Object>| {
let mut doc = Document::open(Arc::new(fixture())).unwrap();
let mut parms = pdfcraft_cos::Dict::new();
parms.set(b"Name".to_vec(), Object::name("MyCF"));
select(&mut doc, (Object::Array(filters), Object::Dict(parms)));
doc.set_encryption(&NewEncryption {
algorithm: Algorithm::Aes256,
user_password: "",
owner_password: "owner",
permissions: -1,
encrypt_metadata: true,
seed: [7; 32],
})
.unwrap();
let back = Document::open(Arc::new(write_full(&doc, &SaveOptions::default()).unwrap())).unwrap();
back.get(contents_of(&back)).as_dict().unwrap().contains(b"DecodeParms")
};
let (hex, crypt) = (Object::name("ASCIIHexDecode"), Object::name("Crypt"));
assert!(protected(vec![hex.clone(), crypt.clone()]), "the first filter's, which stays");
assert!(!protected(vec![crypt, hex]), "/Crypt's, which goes");
}
/// Dropping `/Crypt` takes its entry out of `/DecodeParms`, and an array too short to have one is
/// left as it is, given directly or by reference.
#[test]
fn protecting_leaves_a_parameter_array_too_short_for_the_crypt_filter() {
for by_reference in [false, true] {
let mut doc = Document::open(Arc::new(fixture())).unwrap();
let short = Object::Array(vec![Object::Null]);
let parms = if by_reference { Object::Ref(doc.add(short)) } else { short };
select(&mut doc, (Object::Array(vec![Object::name("ASCIIHexDecode"), Object::name("Crypt")]), parms));
doc.set_encryption(&NewEncryption {
algorithm: Algorithm::Aes256,
user_password: "",
owner_password: "owner",
permissions: -1,
encrypt_metadata: true,
seed: [7; 32],
})
.unwrap();
let bytes = write_full(&doc, &SaveOptions::default()).unwrap();
assert!(!holds(&bytes, "Secret page"), "by reference: {by_reference}");
let back = Document::open(Arc::new(bytes)).unwrap();
assert_eq!(raw_content(&back), SECRET, "by reference: {by_reference}");
let contents = back.get(contents_of(&back));
let filter = contents.as_dict().unwrap().get(b"Filter").unwrap().as_array().unwrap();
assert_eq!(filter.len(), 1, "only /Crypt goes");
assert_eq!(filter[0].as_name(), Some(b"ASCIIHexDecode".as_slice()), "the other filter stays as it was");
}
}
/// A chain of up to [`pdfcraft_filters::MAX_CHAIN`] filters and `/Crypt` is read; a longer one
/// names no crypt filter, so its entries aren't followed for every stream sharing it.
#[test]
fn a_chain_longer_than_any_decoder_takes_names_no_crypt_filter() {
let chain = |others: usize| {
let mut doc = Document::open(Arc::new(protect(Algorithm::Aes128, "", "owner", -1))).unwrap();
let mut filters = vec![Object::name("ASCIIHexDecode"); others];
filters.push(Object::name("Crypt"));
let array = doc.add(Object::Array(filters));
select(&mut doc, (Object::Ref(array), Object::Null));
holds(&write_incremental(&doc, &SaveOptions::default()).unwrap(), "Secret page")
};
assert!(chain(pdfcraft_filters::MAX_CHAIN), "the longest chain read: /Identity, in the clear");
assert!(!chain(pdfcraft_filters::MAX_CHAIN + 1), "one more: the document's filter");
}
/// A stream's `/Type` by reference picks its handling like one given directly: metadata left
/// readable under `/EncryptMetadata false`, an embedded file under `/EFF`.
#[test]
fn a_stream_type_given_by_reference_picks_its_crypt_filter() {
let mut doc = Document::open(Arc::new(fixture())).unwrap();
let catalog = doc.get(doc.root().unwrap()).as_dict().cloned().unwrap();
let meta = catalog.reference(b"Metadata").unwrap();
let ty = doc.add(Object::name("Metadata"));
let Object::Stream(mut s) = doc.get(meta).as_ref().clone() else { panic!("no metadata stream") };
s.dict.set(b"Type".to_vec(), Object::Ref(ty));
doc.set(meta, Object::Stream(s));
doc.set_encryption(&NewEncryption {
algorithm: Algorithm::Aes256,
user_password: "",
owner_password: "o",
permissions: -1,
encrypt_metadata: false,
seed: [5; 32],
})
.unwrap();
let bytes = write_full(&doc, &SaveOptions::default()).unwrap();
assert!(holds(&bytes, "<x:xmpmeta>meta</x:xmpmeta>"), "search engines can read XMP");
let mut doc = Document::open(Arc::new(protect(Algorithm::Aes128, "", "owner", -1))).unwrap();
let encrypt = doc.trailer().reference(b"Encrypt").unwrap();
doc.update_dict(encrypt, |e| e.set(b"EFF".to_vec(), Object::name("Identity"))).unwrap();
let mut doc = Document::open(Arc::new(write_incremental(&doc, &SaveOptions::default()).unwrap())).unwrap();
let contents = contents_of(&doc);
let ty = doc.add(Object::name("EmbeddedFile"));
let Object::Stream(mut s) = doc.get(contents).as_ref().clone() else { panic!("no content stream") };
s.dict.set(b"Type".to_vec(), Object::Ref(ty));
doc.set(contents, Object::Stream(s));
assert!(holds(&write_incremental(&doc, &SaveOptions::default()).unwrap(), "Secret page"), "/EFF /Identity leaves it alone");
}
/// What makes a dictionary a signature's can be given by reference, and what is null or refers to
/// nothing is absent: `/Contents` stays as written in each of these.
#[test]
fn a_signature_dictionary_is_known_in_every_form() {
let mut doc = Document::open(Arc::new(fixture())).unwrap();
let ty = doc.add(Object::name("Sig"));
let range = Object::Array([0, 10, 20, 30].into_iter().map(Object::Int).collect());
let by_reference = doc.add(range.clone());
let nowhere = Object::Ref(ObjRef::new(9_999, 0));
let forms: [(Vec<(&str, Object)>, &str); 5] = [
(vec![("Type", Object::Ref(ty))], "DEAD"),
(vec![("ByteRange", Object::Ref(by_reference))], "BEEF"),
(vec![("Type", Object::Null), ("ByteRange", range.clone())], "CAFE"),
(vec![("Type", nowhere.clone()), ("ByteRange", range.clone())], "F00D"),
(vec![("Subtype", nowhere), ("ByteRange", range)], "D00D"),
];
let mut signatures = Vec::new();
for (entries, contents) in forms {
let mut sig = pdfcraft_cos::Dict::new();
for (key, value) in entries {
sig.set(key.as_bytes().to_vec(), value);
}
let bytes = (0..contents.len()).step_by(2).map(|i| u8::from_str_radix(&contents[i..i + 2], 16).unwrap()).collect();
sig.set(b"Contents".to_vec(), Object::String(PdfString { bytes, hex: true }));
signatures.push(Object::Ref(doc.add(Object::Dict(sig))));
}
// Nested in another object, it follows nothing, but known directly it is one all the same.
let mut nested = pdfcraft_cos::Dict::new();
nested.set(b"Type".to_vec(), Object::name("Sig"));
nested.set(b"Contents".to_vec(), Object::String(PdfString { bytes: vec![0xAB, 0xBA], hex: true }));
let mut field = pdfcraft_cos::Dict::new();
field.set(b"V".to_vec(), Object::Dict(nested));
signatures.push(Object::Ref(doc.add(Object::Dict(field))));
// Neither of what makes one: a string /Contents alone is anybody's, so it's encrypted.
let mut other = pdfcraft_cos::Dict::new();
other.set(b"Contents".to_vec(), Object::String(PdfString { bytes: vec![0xAB, 0xCD], hex: true }));
signatures.push(Object::Ref(doc.add(Object::Dict(other))));
doc.update_dict(doc.root().unwrap(), |c| c.set(b"Test".to_vec(), Object::Array(signatures))).unwrap();
doc.set_encryption(&NewEncryption {
algorithm: Algorithm::Aes256,
user_password: "",
owner_password: "owner",
permissions: -1,
encrypt_metadata: true,
seed: [3; 32],
})
.unwrap();
let bytes = write_full(&doc, &SaveOptions { object_streams: false, ..SaveOptions::default() }).unwrap();
for contents in ["DEAD", "BEEF", "CAFE", "F00D", "D00D", "ABBA"] {
assert!(holds(&bytes, &format!("<{contents}>")), "{contents}: written as is");
}
assert!(!holds(&bytes, "<ABCD>"), "without a /ByteRange it is no signature");
}
/// What a test needs to write an update of a classic file protected with AES-128.
struct Classic {
bytes: Vec<u8>,
root: ObjRef,
pages: ObjRef,
info: ObjRef,
encrypt: ObjRef,
id: Vec<u8>,
dict: pdfcraft_cos::Dict,
}
fn classic_file() -> Classic {
let mut doc = Document::open(Arc::new(fixture())).unwrap();
doc.set_encryption(&NewEncryption {
algorithm: Algorithm::Aes128,
user_password: "",
owner_password: "owner",
permissions: -1,
encrypt_metadata: true,
seed: [9; 32],
})
.unwrap();
let bytes = write_full(&doc, &SaveOptions { object_streams: false, ..SaveOptions::default() }).unwrap();
let doc = Document::open(Arc::new(bytes.clone())).unwrap();
let encrypt = doc.trailer().reference(b"Encrypt").unwrap();
let id = doc.trailer().get(b"ID").unwrap().as_array().unwrap()[0].as_string().unwrap().bytes.clone();
let dict = doc.get(encrypt).as_dict().cloned().unwrap();
let pages = doc.get(doc.root().unwrap()).as_dict().unwrap().reference(b"Pages").unwrap();
Classic { root: doc.root().unwrap(), pages, info: doc.trailer().reference(b"Info").unwrap(), bytes, encrypt, id, dict }
}
fn hex(bytes: &[u8]) -> String {
bytes.iter().map(|b| format!("{b:02X}")).collect()
}
/// `bytes` with an update appended: `objects`, and a trailer of `trailer` entries, in which `@n`
/// stands for the offset of object `n`. An object with nothing in it gets an entry of the
/// cross-reference table that points at nothing.
fn updated(bytes: &[u8], objects: &[(u32, Vec<u8>)], trailer: &str) -> Vec<u8> {
let at = bytes.windows(9).rposition(|w| w == b"startxref").unwrap();
let prev: usize = String::from_utf8_lossy(&bytes[at + 9..]).split_whitespace().next().unwrap().parse().unwrap();
let mut out = bytes.to_vec();
let mut table = String::from("xref\n");
let mut trailer = trailer.to_string();
for (num, body) in objects {
if body.is_empty() {
table.push_str(&format!("{num} 1\n0000000010 00000 n \n"));
continue;
}
trailer = trailer.replace(&format!("@{num}"), &out.len().to_string());
table.push_str(&format!("{num} 1\n{:010} 00000 n \n", out.len()));
out.extend_from_slice(format!("{num} 0 obj\n").as_bytes());
out.extend_from_slice(body);
out.extend_from_slice(b"\nendobj\n");
}
let xref = out.len();
out.extend_from_slice(table.as_bytes());
out.extend_from_slice(format!("trailer\n<< {trailer} /Size 2000 /Prev {prev} >>\nstartxref\n{xref}\n%%EOF\n").as_bytes());
out
}
/// The `/Encrypt` dictionary of `file`, with `/O`, `/StmF` and `/StrF` as given.
fn encrypt_with(file: &Classic, o: &str, filter: &str) -> Vec<u8> {
let string = |key: &[u8]| hex(&file.dict.get(key).unwrap().as_string().unwrap().bytes);
format!(
"<< /Filter /Standard /V 4 /R 4 /Length 128 /P {} /O {o} /U <{}> /CF << /StdCF << /CFM /AESV2 /Length 16 /AuthEvent /DocOpen >> >> /StmF /{filter} /StrF /{filter} >>",
file.dict.int(b"P").unwrap(),
string(b"U")
)
.into_bytes()
}
/// An object stream can have its parameters by reference, and a stream's crypt filter can be
/// selected by a member of it: legal together, though each is rare. What reads the stream first
/// goes through the object stream, and members must come out of it as they do read directly. They
/// didn't, while a read made for a selector wasn't allowed to follow its own: the object stream was
/// kept undecoded, its members lost for good, and a full save wrote nulls without saying so.
#[test]
fn an_object_stream_reached_from_a_selector_keeps_its_members() {
let file = classic_file();
let parms = "[<< /Name /Identity >> null]";
let header = format!("902 0 903 {} ", parms.len() + 1);
let members = format!("{header}{parms} << /Keep (yes) >>");
let (mut rows, mut before) = (vec![1u8], 0u8);
for b in members.bytes() {
rows.push(b.wrapping_sub(before));
before = b;
}
let coded = pdfcraft_cos::Stream::flate(pdfcraft_cos::Dict::new(), &rows).raw;
let mut object_stream =
format!("<< /Type /ObjStm /N 2 /First {} /Filter /FlateDecode /DecodeParms 911 0 R /Length {} >>\nstream\n", header.len(), coded.len())
.into_bytes();
object_stream.extend_from_slice(&coded);
object_stream.extend_from_slice(b"\nendstream");
let data = [2u8, 0x03, 0x8E, 0, 2, 0x03, 0x8E, 1];
let mut xref = format!("<< /Type /XRef /Size 2000 /W [1 2 1] /Index [902 2] /Length {} >>\nstream\n", data.len()).into_bytes();
xref.extend_from_slice(&data);
xref.extend_from_slice(b"\nendstream");
let string = |key: &[u8]| format!("<{}>", hex(&file.dict.get(key).unwrap().as_string().unwrap().bytes));
let objects = [
(file.root.num, format!("<< /Type /Catalog /Pages {} 0 R /A 900 0 R /B 903 0 R >>", file.pages.num).into_bytes()),
(file.encrypt.num, encrypt_with(&file, &string(b"O"), "Identity")),
(900, b"<< /Filter [/Crypt /ASCIIHexDecode] /DecodeParms 902 0 R /Length 1 >>\nstream\n>\nendstream".to_vec()),
(910, object_stream),
(911, format!("<< /Predictor 11 /Columns {} >>", members.len()).into_bytes()),
(912, xref),
];
let trailer = format!("/Root {} 0 R /Encrypt {} 0 R /ID [<{id}> <{id}>] /XRefStm @912", file.root.num, file.encrypt.num, id = hex(&file.id));
let bytes = updated(&file.bytes, &objects, &trailer);
let keep =
|doc: &Document, r: &Object| doc.resolve(r).as_dict().and_then(|d| d.get(b"Keep")).and_then(Object::as_string).map(|s| s.bytes.clone());
let member = Object::Ref(ObjRef::new(903, 0));
let doc = Document::open(Arc::new(bytes.clone())).unwrap();
assert_eq!(keep(&doc, &member).as_deref(), Some(b"yes".as_slice()), "read first");
let doc = Document::open(Arc::new(bytes.clone())).unwrap();
let _ = doc.get(ObjRef::new(900, 0));
assert_eq!(keep(&doc, &member).as_deref(), Some(b"yes".as_slice()), "read after the stream selecting from it");
let saved = write_full(&Document::open(Arc::new(bytes)).unwrap(), &SaveOptions::default()).unwrap();
let back = Document::open(Arc::new(saved)).unwrap();
let catalog = back.get(back.root().unwrap());
assert_eq!(keep(&back, catalog.as_dict().unwrap().get(b"B").unwrap()).as_deref(), Some(b"yes".as_slice()), "after a full save");
}
/// `walk` goes as deep as an object nests, so following a reference from down there adds that depth
/// again for each object it loads: this file, under 10 KB, overflowed a 2 MiB stack and took the
/// process with it. Only the object itself follows references, and it does so in very little stack.
#[test]
fn nested_dictionaries_load_nothing_on_a_small_stack() {
let file = classic_file();
let first = 960u32;
let objects: Vec<(u32, Vec<u8>)> = (first..first + 40)
.map(|num| {
let open = "[".repeat(120);
let close = "]".repeat(120);
(num, format!("{open}<< /Contents () /Type {} 0 R >>{close}", num + 1).into_bytes())
})
.collect();
let trailer = format!("/Root {} 0 R /Encrypt {} 0 R /ID [<{id}> <{id}>]", file.root.num, file.encrypt.num, id = hex(&file.id));
let bytes = updated(&file.bytes, &objects, &trailer);
let reading = std::thread::Builder::new().stack_size(1 << 20).spawn(move || {
let doc = Document::open(Arc::new(bytes)).unwrap();
doc.get(ObjRef::new(first, 0)).as_array().is_some()
});
assert_eq!(reading.unwrap().join().ok(), Some(true));
}
/// The object itself can follow references, a ladder of them as long as loading may nest, and that
/// fits a small stack too.
#[test]
fn a_ladder_of_signature_checks_fits_a_small_stack() {
let file = classic_file();
let first = 960u32;
let objects: Vec<(u32, Vec<u8>)> =
(first..first + 40).map(|num| (num, format!("<< /Contents () /Type {} 0 R >>", num + 1).into_bytes())).collect();
let trailer = format!("/Root {} 0 R /Encrypt {} 0 R /ID [<{id}> <{id}>]", file.root.num, file.encrypt.num, id = hex(&file.id));
let bytes = updated(&file.bytes, &objects, &trailer);
let reading = std::thread::Builder::new().stack_size(1 << 20).spawn(move || {
let doc = Document::open(Arc::new(bytes)).unwrap();
doc.get(ObjRef::new(first, 0)).as_dict().is_some()
});
assert_eq!(reading.unwrap().join().ok(), Some(true));
}
/// The strings of `/Encrypt` are never encrypted (ISO 32000-2 §7.6.2), and can be objects of their
/// own. Read as ordinary strings they were decrypted, and a full save encrypted them again under
/// another number, so the file it wrote no longer opened with its own password.
#[test]
fn encrypt_strings_given_by_reference_stay_as_stored() {
let file = classic_file();
let string = |key: &[u8]| hex(&file.dict.get(key).unwrap().as_string().unwrap().bytes);
let replaced = format!(
"<< /Filter /Standard /V 4 /R 4 /Length 128 /P {} /O 900 0 R /U 901 0 R /CF << /StdCF << /CFM /AESV2 /Length 16 /AuthEvent /DocOpen >> >> /StmF /StdCF /StrF /StdCF >>",
file.dict.int(b"P").unwrap()
);
let objects = [
(file.encrypt.num, replaced.into_bytes()),
(900, format!("<{}>", string(b"O")).into_bytes()),
(901, format!("<{}>", string(b"U")).into_bytes()),
];
let trailer =
format!("/Root {} 0 R /Info {} 0 R /Encrypt {} 0 R /ID [<{id}> <{id}>]", file.root.num, file.info.num, file.encrypt.num, id = hex(&file.id));
let source = Document::open(Arc::new(updated(&file.bytes, &objects, &trailer))).unwrap();
assert_eq!(content(&source), SECRET, "the file reads");
assert_eq!(title(&source).as_deref(), Some("Quarterly numbers"));
for object_streams in [false, true] {
let saved = write_full(&source, &SaveOptions { object_streams, ..SaveOptions::default() }).unwrap();
let back = Document::open(Arc::new(saved)).unwrap();
assert_eq!(content(&back), SECRET, "object streams: {object_streams}");
assert_eq!(title(&back).as_deref(), Some("Quarterly numbers"), "object streams: {object_streams}");
}
}
/// A save keeps what a file is protected with: its `/Encrypt`, which holds the handler and the
/// permissions, its `/ID`, and so its file key, under every algorithm and by either kind of save.
#[test]
fn a_save_keeps_the_handler_the_key_and_the_permissions() {
for alg in ALL {
let doc = Document::open_with_password(Arc::new(protect(alg, "user", "owner", -1 ^ 0b1_0000)), Some("user")).unwrap();
let held = |d: &Document| {
let encrypt = d.resolve(d.trailer().get(b"Encrypt").unwrap()).as_dict().unwrap().clone();
(encrypt, d.trailer().get(b"ID").cloned(), d.security().unwrap().file_key().to_vec(), format!("{:?}", d.permissions()))
};
let before = held(&doc);
let mut edited = doc.clone();
let _ = edited.add(Object::Int(1));
for (save, saved) in [
("incremental", write_incremental(&edited, &SaveOptions::default()).unwrap()),
("full", write_full(&doc, &SaveOptions::default()).unwrap()),
] {
let back = Document::open_with_password(Arc::new(saved), Some("user")).unwrap();
assert!(held(&back) == before, "{alg:?}: a {save} save");
assert_eq!(content(&back), SECRET, "{alg:?}: a {save} save");
}
}
}
/// `/ID` has to be direct in an encrypted file, but one that isn't has its key made from the same
/// bytes. A full save writes it direct, and an incremental one leaves it as it was.
#[test]
fn a_file_identifier_given_by_reference_is_read() {
let file = classic_file();
let id = hex(&file.id);
let reference = |entries: &str, objects: &[(u32, Vec<u8>)]| {
let trailer = format!("/Root {} 0 R /Encrypt {} 0 R /ID {entries}", file.root.num, file.encrypt.num);
updated(&file.bytes, objects, &trailer)
};
let whole = reference("900 0 R", &[(900, format!("[<{id}> <{id}>]").into_bytes())]);
let entries = reference("[900 0 R 901 0 R]", &[(900, format!("<{id}>").into_bytes()), (901, format!("<{id}>").into_bytes())]);
for file in [whole, entries] {
let source = Document::open(Arc::new(file)).unwrap();
assert_eq!(content(&source), SECRET);
let back = Document::open(Arc::new(write_full(&source, &SaveOptions::default()).unwrap())).unwrap();
assert_eq!(content(&back), SECRET);
assert!(matches!(back.trailer().get(b"ID"), Some(Object::Array(_))), "written direct");
let mut edited = source.clone();
let _ = edited.add(Object::Int(1));
let back = Document::open(Arc::new(write_incremental(&edited, &SaveOptions::default()).unwrap())).unwrap();
assert_eq!(content(&back), SECRET, "an incremental save");
assert_eq!(back.trailer().get(b"ID"), edited.trailer().get(b"ID"), "and leaves it as it was");
}
}
/// A direct `/Encrypt` can refer to objects. A full save dropped those it found only there, and kept
/// its references as they were under the old numbers, so nothing in the file read afterwards.
#[test]
fn a_direct_encrypt_dictionary_keeps_what_it_refers_to() {
for object_streams in [false, true] {
let mut doc = Document::open(Arc::new(fixture())).unwrap();
doc.set_encryption(&NewEncryption {
algorithm: Algorithm::Aes128,
user_password: "",
owner_password: "owner",
permissions: -1,
encrypt_metadata: true,
seed: [9; 32],
})
.unwrap();
let mut dict = doc.get(doc.trailer().reference(b"Encrypt").unwrap()).as_dict().cloned().unwrap();
let filter = doc.add(Object::name("StdCF"));
let filters = doc.add(dict.get(b"CF").unwrap().clone());
dict.set(b"StmF".to_vec(), Object::Ref(filter));
dict.set(b"CF".to_vec(), Object::Ref(filters));
doc.trailer_mut().set(b"Encrypt".to_vec(), Object::Dict(dict));
let options = SaveOptions { object_streams, ..SaveOptions::default() };
let saved = write_full(&doc, &options).unwrap();
let back = Document::open(Arc::new(saved)).unwrap();
assert_eq!(content(&back), SECRET, "object streams: {object_streams}");
let again = write_full(&back, &options).unwrap();
assert_eq!(content(&Document::open(Arc::new(again)).unwrap()), SECRET, "object streams: {object_streams}, saved again");
}
}
/// How a stream is treated can be given by reference, and what that loads can be a stream too. A
/// file can make a ladder of those, where each rung reads the rung below, then two trees wider than
/// the object cache, which push that rung out of it, then the rung below again. Each rung would
/// take twice what the one below does, if one read could follow any number of references.
#[test]
fn selectors_that_lead_to_selectors_load_at_once() {
let mut doc = Document::open(Arc::new(protect(Algorithm::Aes128, "", "owner", -1))).unwrap();
let stream_of = |doc: &mut Document, filters: Vec<Object>| {
let mut dict = pdfcraft_cos::Dict::new();
dict.set(b"Filter".to_vec(), Object::Array(filters));
doc.add(Object::Stream(pdfcraft_cos::Stream::from_raw(dict, Vec::new())))
};
let tree = |doc: &mut Document| {
let mut level: Vec<ObjRef> = (0..6_561).map(|i| doc.add(Object::Int(i))).collect();
while level.len() > 1 {
let chunks: Vec<Vec<Object>> = level.chunks(9).map(|c| c.iter().map(|r| Object::Ref(*r)).collect()).collect();
level = chunks.into_iter().map(|c| stream_of(doc, c)).collect();
}
Object::Ref(level[0])
};
let (first, second) = (tree(&mut doc), tree(&mut doc));
let mut top = doc.add(Object::Int(0));
// Twelve rungs of two, and the trees' five levels, stay inside the 32 that loading may nest.
for _ in 0..12 {
let (a, b) = (stream_of(&mut doc, vec![Object::Ref(top)]), stream_of(&mut doc, vec![Object::Ref(top)]));
top = stream_of(&mut doc, vec![Object::Ref(a), first.clone(), second.clone(), Object::Ref(b)]);
}
let saved = write_incremental(&doc, &SaveOptions::default()).unwrap();
let (done, finished) = std::sync::mpsc::channel();
std::thread::spawn(move || {
let doc = Document::open(Arc::new(saved)).unwrap();
let _ = done.send(doc.get(top).as_dict().is_some());
});
// With the budget it takes milliseconds, and without it longer than anyone would wait, so a
// margin this wide can't fail on a busy machine.
assert_eq!(finished.recv_timeout(std::time::Duration::from_secs(120)), Ok(true), "loading must not fan out");
}
crates/cos/tests/it/generations.rs
+475
View File
@@ -0,0 +1,475 @@
//! An indirect reference identifies an object number AND its generation (ISO 32000-2 §7.3.10).
use std::sync::Arc;
use hayro_syntax::object::{Dict as HayroDict, ObjectIdentifier};
use pdfcraft_cos::{Document, ObjRef, Object, SaveOptions, XrefEntry, write_full, write_incremental};
/// Object 3 was reused: only 3 1 R refers to the current value, never 3 0 R.
fn reused_object() -> Arc<Vec<u8>> {
object_file(&[
(1, 0, "<< /Type /Catalog /Pages 2 0 R /Stale 3 0 R /Current 3 1 R >>"),
(2, 0, "<< /Type /Pages /Kids [] /Count 0 >>"),
(3, 1, "42"),
])
}
fn object_file(objects: &[(u32, u16, &str)]) -> Arc<Vec<u8>> {
object_file_with_entry(objects, None)
}
fn object_file_with_entry(objects: &[(u32, u16, &str)], changed: Option<(u32, u16, bool)>) -> Arc<Vec<u8>> {
let mut bytes = b"%PDF-1.7\n".to_vec();
let mut offsets = Vec::new();
for (num, generation, value) in objects {
offsets.push((*num, bytes.len(), *generation));
bytes.extend_from_slice(format!("{num} {generation} obj\n{value}\nendobj\n").as_bytes());
}
let xref = bytes.len();
let size = objects.len() + 1;
bytes.extend_from_slice(format!("xref\n0 {size}\n0000000000 65535 f \n").as_bytes());
for (num, offset, generation) in offsets {
let (generation, free) = changed.filter(|(n, _, _)| *n == num).map_or((generation, false), |(_, g, free)| (g, free));
let (offset, kind) = if free { (0, 'f') } else { (offset, 'n') };
bytes.extend_from_slice(format!("{offset:010} {generation:05} {kind} \n").as_bytes());
}
bytes.extend_from_slice(format!("trailer\n<< /Size {size} /Root 1 0 R >>\nstartxref\n{xref}\n%%EOF\n").as_bytes());
Arc::new(bytes)
}
fn contradictory_entry(free: bool) -> Arc<Vec<u8>> {
object_file_with_entry(
&[(1, 0, "<< /Type /Catalog /Pages 2 0 R /Stale 3 0 R /Current 3 1 R >>"), (2, 0, "<< /Type /Pages /Kids [] /Count 0 >>"), (3, 1, "42")],
Some((3, if free { 2 } else { 0 }, free)),
)
}
#[test]
fn contradictory_xref_generation_recovers_the_matching_header_and_records_repair() {
for warm in [false, true] {
let doc = Document::open(contradictory_entry(false)).unwrap();
if warm {
assert_eq!(*doc.try_get(3).unwrap(), Object::Int(42));
}
assert_eq!(*doc.get(ObjRef::new(3, 1)), Object::Int(42));
assert_eq!(*doc.resolve(&Object::Ref(ObjRef::new(3, 1))), Object::Int(42));
// Preserve the existing xref-matching behavior, but never accept an unrelated generation.
assert_eq!(*doc.get(ObjRef::new(3, 0)), Object::Int(42));
assert_eq!(*doc.get(ObjRef::new(3, 2)), Object::Null);
assert!(!doc.xref_entry_is_exact(3), "a contradictory generation is not an exact xref entry");
assert!(
doc.repair_log()
.iter()
.any(|line| { line.contains("object 3") && line.contains("xref generation 0") && line.contains("header generation 1") }),
"{:?}",
doc.repair_log()
);
let repairs = doc.repair_log().to_vec();
for _ in 0..3 {
assert_eq!(*doc.get(ObjRef::new(3, 1)), Object::Int(42));
}
assert_eq!(doc.repair_log(), repairs);
let mut edited = doc.clone();
edited.free(ObjRef::new(3, 1));
for generation in [0, 1, 2] {
assert_eq!(*edited.get(ObjRef::new(3, generation)), Object::Null);
}
edited.set(ObjRef::new(3, 2), Object::Int(99));
assert_eq!(*edited.get(ObjRef::new(3, 1)), Object::Null);
assert_eq!(*edited.get(ObjRef::new(3, 2)), Object::Int(99));
assert_eq!(*doc.get(ObjRef::new(3, 1)), Object::Int(42));
}
}
/// A damaged row may name generation zero while the retained object and catalog use one.
fn contradictory_dictionary() -> Arc<Vec<u8>> {
object_file_with_entry(
&[(1, 0, "<< /Type /Catalog /Pages 2 0 R /Current 3 1 R >>"), (2, 0, "<< /Type /Pages /Kids [] /Count 0 >>"), (3, 1, "<< /Value 42 >>")],
Some((3, 0, false)),
)
}
fn assert_repaired_dictionary(bytes: Vec<u8>, value: i32) {
let bytes = Arc::new(bytes);
let oracle = hayro_syntax::Pdf::new(bytes.clone()).unwrap();
let catalog = oracle.xref().get::<HayroDict<'_>>(ObjectIdentifier::new(1, 0)).unwrap();
assert_eq!(catalog.get::<HayroDict<'_>>(b"Current").unwrap().get::<i32>(b"Value"), Some(value));
let back = Document::open(bytes).unwrap();
let catalog = back.get(back.root().unwrap());
let current = back.resolve(catalog.as_dict().unwrap().get(b"Current").unwrap());
assert_eq!(current.as_dict().unwrap().int(b"Value"), Some(i64::from(value)));
}
#[test]
fn repaired_header_references_survive_full_saves_and_dictionary_edits() {
let original = contradictory_dictionary();
let doc = Document::open(original.clone()).unwrap();
for object_streams in [false, true] {
assert_repaired_dictionary(write_full(&doc, &SaveOptions { object_streams, ..SaveOptions::default() }).unwrap(), 42);
}
let mut edited = doc.clone();
edited.update_dict(ObjRef::new(3, 1), |dict| dict.set(b"Value".to_vec(), Object::Int(77))).unwrap();
assert_eq!(edited.generation(3), 1);
let incremental = write_incremental(&edited, &SaveOptions::default()).unwrap();
assert!(edited.full_save_required(), "repair-aware dictionary edits normalize on full rewrite");
assert_repaired_dictionary(incremental, 77);
for object_streams in [false, true] {
assert_repaired_dictionary(write_full(&edited, &SaveOptions { object_streams, ..SaveOptions::default() }).unwrap(), 77);
}
let mut restored = doc.clone();
restored.overlay_changes_from(&edited).unwrap();
assert_repaired_dictionary(write_incremental(&restored, &SaveOptions::default()).unwrap(), 77);
}
#[test]
fn free_xref_slots_never_revive_retained_object_headers() {
for warm in [false, true] {
let doc = Document::open(contradictory_entry(true)).unwrap();
assert!(matches!(doc.xref_entry(3), Some(XrefEntry::Free { next_generation: 2 })));
if warm {
assert_eq!(*doc.try_get(3).unwrap(), Object::Null);
}
for generation in [0, 1, 2] {
assert_eq!(*doc.get(ObjRef::new(3, generation)), Object::Null);
}
assert_eq!(*doc.resolve(&Object::Ref(ObjRef::new(3, 1))), Object::Null);
assert!(doc.repair_log().is_empty(), "{:?}", doc.repair_log());
}
// Control: the retained bytes really are a readable object when its slot is live.
let live = Document::open(contradictory_entry(false)).unwrap();
assert_eq!(*live.try_get(3).unwrap(), Object::Int(42));
}
#[test]
fn incremental_generation_identity_is_visible_to_an_independent_reader() {
let original = reused_object();
let mut doc = Document::open(original.clone()).unwrap();
doc.set(ObjRef::new(3, 1), Object::Int(77));
let bytes = write_incremental(&doc, &SaveOptions::default()).unwrap();
assert!(bytes.starts_with(&original));
let oracle = hayro_syntax::Pdf::new(Arc::new(bytes)).unwrap();
let catalog = oracle.xref().get::<HayroDict<'_>>(ObjectIdentifier::new(1, 0)).unwrap();
assert_eq!(catalog.get::<i32>(b"Current"), Some(77));
assert_eq!(catalog.get::<i32>(b"Stale"), None);
}
#[test]
fn generation_repair_diagnostics_are_bounded_without_disabling_recovery() {
let mut objects = vec![(1, 0, "<< /Type /Catalog /Pages 2 0 R >>"), (2, 0, "<< /Type /Pages /Kids [] /Count 0 >>")];
objects.extend((3..303).map(|num| (num, 1, "42")));
let bytes = object_file(&objects);
let bytes = String::from_utf8(bytes.as_ref().clone()).unwrap().replace("00001 n \n", "00000 n \n");
for materialized in [false, true] {
let bytes = if materialized { bytes.replace("/Size 303", "/Size 302") } else { bytes.clone() };
let doc = Document::open(Arc::new(bytes.into_bytes())).unwrap();
assert_eq!(doc.repair_log().len(), 257);
assert!(doc.repair_log().last().unwrap().contains("omitted after 256 objects"));
assert_eq!(*doc.get(ObjRef::new(3, 1)), Object::Int(42));
assert_eq!(*doc.get(ObjRef::new(302, 1)), Object::Int(42));
assert_eq!(*doc.get(ObjRef::new(302, 2)), Object::Null);
assert_eq!(doc.repair_log().len(), 257);
}
}
#[test]
fn references_require_the_current_generation_even_with_a_warm_cache() {
let bytes = reused_object();
let oracle = hayro_syntax::Pdf::new(bytes.clone()).unwrap();
assert_eq!(oracle.xref().get::<i32>(ObjectIdentifier::new(3, 0)), None);
assert_eq!(oracle.xref().get::<i32>(ObjectIdentifier::new(3, 1)), Some(42));
let doc = Document::open(bytes).unwrap();
assert!(doc.repair_log().is_empty());
assert_eq!(*doc.get(ObjRef::new(3, 0)), Object::Null);
assert_eq!(*doc.try_get(3).unwrap(), Object::Int(42));
assert_eq!(*doc.get(ObjRef::new(3, 1)), Object::Int(42));
assert_eq!(*doc.get(ObjRef::new(3, 0)), Object::Null);
assert_eq!(*doc.resolve(&Object::Ref(ObjRef::new(3, 0))), Object::Null);
assert_eq!(*doc.get(ObjRef::new(3, 2)), Object::Null);
}
#[test]
fn overlay_generation_changes_do_not_retarget_old_references_or_clones() {
let mut doc = Document::open(reused_object()).unwrap();
let snapshot = doc.clone();
let old = ObjRef::new(3, 1);
let new = ObjRef::new(3, 2);
doc.free(old);
assert_eq!(*doc.get(old), Object::Null);
doc.set(new, Object::Int(99));
assert_eq!(*doc.get(old), Object::Null);
assert_eq!(*doc.get(new), Object::Int(99));
assert_eq!(*snapshot.get(old), Object::Int(42));
assert_eq!(*snapshot.get(new), Object::Null);
let saved = write_incremental(&doc, &SaveOptions::default()).unwrap();
let back = Document::open(Arc::new(saved)).unwrap();
assert!(back.repair_log().is_empty());
assert_eq!(*back.get(old), Object::Null);
assert_eq!(*back.get(new), Object::Int(99));
}
#[test]
fn restoring_snapshot_preserves_changes_to_nonzero_generation_objects() {
for generation in [1, 2] {
let mut saved = Document::open(reused_object()).unwrap();
let mut current = saved.clone();
let reference = ObjRef::new(3, generation);
current.set(reference, Object::Int(99));
current.update_dict(current.root().unwrap(), |catalog| catalog.set(b"Current".to_vec(), Object::Ref(reference))).unwrap();
saved.overlay_changes_from(&current).unwrap();
assert_eq!(*saved.get(reference), Object::Int(99), "generation {generation}");
let back = Document::open(Arc::new(write_incremental(&saved, &SaveOptions::default()).unwrap())).unwrap();
assert!(back.repair_log().is_empty());
assert_eq!(*back.get(reference), Object::Int(99), "generation {generation}");
let catalog = back.get(back.root().unwrap());
assert_eq!(*back.resolve(catalog.as_dict().unwrap().get(b"Current").unwrap()), Object::Int(99));
}
}
#[test]
fn full_save_does_not_turn_stale_references_into_live_ones() {
let doc = Document::open(reused_object()).unwrap();
for object_streams in [false, true] {
let saved = write_full(&doc, &SaveOptions { object_streams, ..SaveOptions::default() }).unwrap();
let saved = Arc::new(saved);
let oracle = hayro_syntax::Pdf::new(saved.clone()).unwrap();
let catalog = oracle.xref().get::<HayroDict<'_>>(ObjectIdentifier::new(1, 0)).unwrap();
assert_eq!(catalog.get::<i32>(b"Current"), Some(42));
assert_eq!(catalog.get::<i32>(b"Stale"), None);
let back = Document::open(saved).unwrap();
assert!(back.repair_log().is_empty());
let catalog = back.get(back.root().unwrap());
let catalog = catalog.as_dict().unwrap();
assert_eq!(catalog.get(b"Stale"), Some(&Object::Null));
assert_eq!(*back.resolve(catalog.get(b"Current").unwrap()), Object::Int(42));
}
}
#[test]
fn compressed_objects_only_accept_generation_zero() {
let doc = Document::open(reused_object()).unwrap();
let saved = write_full(&doc, &SaveOptions::default()).unwrap();
let back = Document::open(Arc::new(saved)).unwrap();
let catalog = back.get(back.root().unwrap());
let current = catalog.as_dict().unwrap().reference(b"Current").unwrap();
assert_eq!(current.generation, 0);
assert!(matches!(back.xref_entry(current.num), Some(XrefEntry::InStream { .. })));
assert_eq!(*back.get(current), Object::Int(42));
assert_eq!(*back.get(ObjRef::new(current.num, 1)), Object::Null);
}
#[test]
fn indirect_stream_lengths_require_the_referenced_generation() {
for generation in [0, 1] {
// The embedded endstream is data when /Length resolves, but the existing missing-
// length recovery stops there when the reference is stale.
let stream = format!("<< /Length 4 {generation} R >>\nstream\nA\nendstream\nB\nendstream");
let doc = Document::open(object_file(&[
(1, 0, "<< /Type /Catalog /Pages 2 0 R >>"),
(2, 0, "<< /Type /Pages /Kids [] /Count 0 >>"),
(3, 0, &stream),
(4, 1, "13"),
]))
.unwrap();
let object = doc.get(ObjRef::new(3, 0));
let Object::Stream(stream) = &*object else { panic!("expected stream") };
assert_eq!(&stream.raw[..], if generation == 1 { b"A\nendstream\nB".as_slice() } else { b"A".as_slice() });
}
}
/// A newer valid row must not hide damage in a retained older revision.
fn shadowed_generation_disagreement(materialized: bool, shifted: bool) -> Arc<Vec<u8>> {
let original = contradictory_dictionary();
let text = std::str::from_utf8(&original).unwrap();
let previous: usize = text.rsplit("startxref\n").next().unwrap().lines().next().unwrap().parse().unwrap();
let mut bytes = (*original).clone();
let offset = bytes.len();
bytes.extend_from_slice(b"3 1 obj\n<< /Value 77 >>\nendobj\n");
let xref = bytes.len();
let size = if materialized { 3 } else { 4 };
bytes.extend_from_slice(
format!("xref\n3 1\n{offset:010} 00001 n \ntrailer\n<< /Size {size} /Root 1 0 R /Prev {previous} >>\nstartxref\n{xref}\n%%EOF\n").as_bytes(),
);
if shifted {
bytes.splice(..0, b"junk before the header\n".iter().copied());
}
Arc::new(bytes)
}
#[test]
fn retained_generation_disagreements_are_reported_in_materialized_mode_too() {
for materialized in [false, true] {
for shifted in [false, true] {
let doc = Document::open(shadowed_generation_disagreement(materialized, shifted)).unwrap();
assert_eq!(doc.revisions().len(), 2, "the xref chain is retained");
assert!(!doc.reconstructed());
assert_eq!(doc.get(ObjRef::new(3, 1)).as_dict().unwrap().int(b"Value"), Some(77));
assert!(
doc.repair_log().iter().any(|line| line.contains("object 3: xref generation 0 disagrees with object header generation 1")),
"materialized={materialized}, shifted={shifted}: {:?}",
doc.repair_log()
);
}
}
}
#[test]
fn repair_aware_edits_keep_both_aliases_and_full_save_normalizes_them() {
let mut doc = Document::open(contradictory_entry(false)).unwrap();
let snapshot = doc.clone();
doc.set_existing(ObjRef::new(3, 0), Object::Int(77)).unwrap();
for generation in [0, 1] {
assert_eq!(*doc.get(ObjRef::new(3, generation)), Object::Int(77));
assert_eq!(*snapshot.get(ObjRef::new(3, generation)), Object::Int(42));
}
let state = doc.read_state();
assert!(doc.set_existing(ObjRef::new(3, 2), Object::Int(99)).is_err());
assert!(state.matches(&doc));
assert_eq!(*doc.get(ObjRef::new(3, 2)), Object::Null);
let mut restored = snapshot.clone();
restored.overlay_changes_from(&doc).unwrap();
for changed in [&doc, &restored] {
assert!(changed.full_save_required());
for bytes in [
write_incremental(changed, &SaveOptions::default()).unwrap(),
write_full(changed, &SaveOptions { object_streams: false, ..SaveOptions::default() }).unwrap(),
] {
let bytes = Arc::new(bytes);
let oracle = hayro_syntax::Pdf::new(bytes.clone()).unwrap();
let catalog = oracle.xref().get::<HayroDict<'_>>(ObjectIdentifier::new(1, 0)).unwrap();
assert_eq!(catalog.get::<i32>(b"Current"), Some(77));
assert_eq!(catalog.get::<i32>(b"Stale"), Some(77), "both source aliases stay live");
let back = Document::open(bytes).unwrap();
assert!(back.repair_log().is_empty(), "normalized output has no generation repairs");
}
}
// Explicit replacement keeps its narrower, caller-supplied generation contract.
doc.set(ObjRef::new(3, 2), Object::Int(99));
assert_eq!(*doc.get(ObjRef::new(3, 0)), Object::Null);
assert_eq!(*doc.get(ObjRef::new(3, 1)), Object::Null);
assert_eq!(*doc.get(ObjRef::new(3, 2)), Object::Int(99));
}
#[test]
fn hybrid_generation_diagnostics_observe_rows_before_the_table_hides_them() {
let mut bytes = b"%PDF-1.7\n".to_vec();
let mut offsets = Vec::new();
for (num, generation, body) in [(1, 0, "<< /Type /Catalog /Pages 2 0 R >>"), (2, 0, "<< /Type /Pages /Kids [] /Count 0 >>"), (3, 1, "42")] {
offsets.push(bytes.len());
bytes.extend_from_slice(format!("{num} {generation} obj\n{body}\nendobj\n").as_bytes());
}
let hybrid = bytes.len();
offsets.push(hybrid);
bytes.extend_from_slice(b"4 0 obj\n<< /Type /XRef /Size 5 /W [1 4 2] /Index [3 1] /Length 7 >>\nstream\n");
bytes.push(1);
bytes.extend_from_slice(&(offsets[2] as u32).to_be_bytes());
bytes.extend_from_slice(&0u16.to_be_bytes());
bytes.extend_from_slice(b"\nendstream\nendobj\n");
let xref = bytes.len();
bytes.extend_from_slice(b"xref\n0 5\n0000000000 65535 f \n");
for (i, offset) in offsets.into_iter().enumerate() {
let generation = u16::from(i == 2);
bytes.extend_from_slice(format!("{offset:010} {generation:05} n \n").as_bytes());
}
bytes.extend_from_slice(format!("trailer\n<< /Size 5 /Root 1 0 R /XRefStm {hybrid} >>\nstartxref\n{xref}\n%%EOF\n").as_bytes());
let doc = Document::open(Arc::new(bytes)).unwrap();
assert!(!doc.reconstructed());
assert!(doc.xref_oddities().iter().any(|(_, num)| *num == 3), "the conflicting hybrid row is set aside");
assert_eq!(*doc.get(ObjRef::new(3, 1)), Object::Int(42));
assert_eq!(*doc.get(ObjRef::new(3, 0)), Object::Null, "the table still determines accepted references");
assert!(doc.repair_log().iter().any(|line| line.contains("object 3: xref generation 0 disagrees with object header generation 1")));
}
#[test]
fn discarded_xref_repair_notes_do_not_survive_object_header_reconstruction() {
let bytes =
String::from_utf8((*contradictory_dictionary()).clone()).unwrap().replace("/Root 1 0 R", "/Root 99 0 R").replace("/Size 4", "/Size 3");
let doc = Document::open(Arc::new(bytes.into_bytes())).unwrap();
assert!(doc.reconstructed());
assert!(doc.repair_log().iter().all(|line| !line.contains("xref generation")), "{:?}", doc.repair_log());
}
#[test]
fn restoring_equal_values_still_restores_the_accepted_generation_aliases() {
for edited in [false, true] {
for value in [42, 77] {
let mut current = Document::open(contradictory_entry(false)).unwrap();
if edited {
current.set_existing(ObjRef::new(3, 0), Object::Int(42)).unwrap();
}
let mut saved = Document::open(object_file(&[
(1, 0, "<< /Type /Catalog /Pages 2 0 R /Stale 3 0 R /Current 3 1 R >>"),
(2, 0, "<< /Type /Pages /Kids [] /Count 0 >>"),
(3, 0, &value.to_string()),
]))
.unwrap();
assert_eq!(*saved.get(ObjRef::new(3, 1)), Object::Null);
saved.overlay_changes_from(&current).unwrap();
for generation in [0, 1] {
assert_eq!(*saved.get(ObjRef::new(3, generation)), Object::Int(42), "edited={edited}, value={value}");
}
assert!(saved.full_save_required());
}
}
}
#[test]
fn repaired_page_tree_aliases_keep_one_identity_in_full_saves() {
let source = object_file_with_entry(
&[
(1, 0, "<< /Type /Catalog /Pages 2 1 R /Stale 2 2 R >>"),
(2, 1, "<< /Type /Pages /Kids [3 0 R] /Count 1 >>"),
(3, 0, "<< /Type /Page /Parent 2 0 R /MediaBox [0 0 10 10] >>"),
],
Some((2, 0, false)),
);
for edited in [false, true] {
let mut doc = Document::open(source.clone()).unwrap();
if edited {
doc.update_dict(ObjRef::new(2, 0), |dict| dict.set(b"Edited".to_vec(), Object::Bool(true))).unwrap();
}
let mut restored = Document::open(source.clone()).unwrap();
restored.overlay_changes_from(&doc).unwrap();
for current in [&doc, &restored] {
for object_streams in [false, true] {
let opts = SaveOptions { object_streams, ..SaveOptions::default() };
let mut saves = vec![write_full(current, &opts).unwrap()];
if edited {
saves.push(write_incremental(current, &opts).unwrap());
}
for bytes in saves {
let back = Document::open(Arc::new(bytes)).unwrap();
let catalog = back.get(back.root().unwrap());
let catalog = catalog.as_dict().unwrap();
let pages_ref = catalog.reference(b"Pages").unwrap();
let pages = back.get(pages_ref);
let child_ref = pages.as_dict().unwrap().get(b"Kids").unwrap().as_array().unwrap()[0].as_ref().unwrap();
let child = back.get(child_ref);
assert_eq!(child.as_dict().unwrap().reference(b"Parent"), Some(pages_ref), "edited={edited}, packed={object_streams}");
assert!(matches!(catalog.get(b"Stale"), Some(Object::Null)), "unrelated generations stay null");
assert!(back.repair_log().is_empty());
}
}
}
}
}
#[test]
fn restoring_explicit_sets_removes_old_generation_aliases() {
for value in [42, 77] {
let mut current = Document::open(contradictory_entry(false)).unwrap();
current.set(ObjRef::new(3, 0), Object::Int(42));
for repaired_overlay in [false, true] {
let mut saved = Document::open(contradictory_entry(false)).unwrap();
if repaired_overlay {
saved.set_existing(ObjRef::new(3, 0), Object::Int(value)).unwrap();
}
saved.overlay_changes_from(&current).unwrap();
assert_eq!(*saved.get(ObjRef::new(3, 0)), Object::Int(42));
assert_eq!(*saved.get(ObjRef::new(3, 1)), Object::Null, "value={value}, overlay={repaired_overlay}");
}
}
}
crates/cos/tests/it/main.rs
+1
View File
@@ -3,6 +3,7 @@
mod bounded_writer;
mod corpus;
mod encryption;
mod generations;
mod proptests;
mod shared_streams;
mod writer;
crates/cos/tests/it/writer.rs
+26 -5
View File
@@ -358,11 +358,32 @@ fn objects_referenced_by_the_encrypt_dictionary_stay_outside_object_streams() {
let cf = doc.get(enc_ref).as_dict().unwrap().get(b"CF").cloned().unwrap();
let cf_ref = doc.add(cf);
doc.update_dict(enc_ref, |d| d.set(b"CF".to_vec(), Object::Ref(cf_ref))).unwrap();
let packed = write_full(&doc, &SaveOptions::default()).unwrap();
let back = Document::open(Arc::new(packed.clone())).expect("reopens");
assert!(back.repair_log().is_empty(), "{:?}", back.repair_log());
assert_eq!(title(&back), b"Caf\xe9 (draft)");
assert_eq!(hayro_pages(&packed), 4);
// Visit this alias from Root before the encryption walk reaches /CF's other alias.
doc.update_dict(doc.root().unwrap(), |d| d.set(b"CFProbe".to_vec(), Object::Ref(cf_ref))).unwrap();
let classic = write_full(&doc, &SaveOptions { object_streams: false, ..SaveOptions::default() }).unwrap();
let saved = Document::open(Arc::new(classic.clone())).unwrap();
let cf_num = saved.get(saved.root().unwrap()).as_dict().unwrap().reference(b"CFProbe").unwrap().num;
for damaged in [false, true] {
let mut bytes = classic.clone();
if damaged {
// Equal-length edits preserve the xref offsets and its generation-zero row.
for (from, to) in [(format!("\n{cf_num} 0 obj"), format!("\n{cf_num} 1 obj")), (format!("/CF {cf_num} 0 R"), format!("/CF {cf_num} 1 R"))]
{
let start = bytes.windows(from.len()).position(|w| w == from.as_bytes()).unwrap();
bytes[start..start + from.len()].copy_from_slice(to.as_bytes());
}
}
let doc = Document::open(Arc::new(bytes)).unwrap();
let packed = write_full(&doc, &SaveOptions::default()).unwrap();
let back = Document::open(Arc::new(packed.clone())).expect("reopens");
assert!(back.repair_log().is_empty(), "{:?}", back.repair_log());
assert_eq!(title(&back), b"Caf\xe9 (draft)");
assert_eq!(hayro_pages(&packed), 4);
let probe = back.get(back.root().unwrap()).as_dict().unwrap().reference(b"CFProbe").unwrap();
let encrypt = back.get(back.trailer().reference(b"Encrypt").unwrap());
assert_eq!(encrypt.as_dict().unwrap().reference(b"CF"), Some(probe));
assert!(matches!(back.xref_entry(probe.num), Some(pdfcraft_cos::XrefEntry::InFile { .. })), "crypt filters must stay outside object streams");
}
}
#[test]
crates/edit/src/images.rs
+6
View File
@@ -1691,6 +1691,12 @@ pub(crate) fn selectable_images(
.collect())
}
/// Whether an XObject drawn at or after operator `from` of the page's joined operators shows
/// inside `rect` (user space).
pub(crate) fn painted_over(doc: &Document, page: usize, rect: [f64; 4], from: usize) -> Result<bool, EditError> {
Ok(page_images(doc, page)?.iter().any(|i| i.op >= from && i.visible_rect.and_then(|v| intersection(v, rect)).is_some()))
}
fn transform_image(
doc: &mut Document,
page: &pdfcraft_model::Page,
crates/edit/src/lib.rs
+1 -1
View File
@@ -77,7 +77,7 @@ use marks::{matrix_number, n};
pub(crate) use place::undrawable;
#[cfg(test)]
use place::{ISOLATION_MAX_DEPTH, OLD_WRAP_CLOSE, WRAP_CLOSE, WRAP_OPEN, is_wrap_close, page_fonts, stream_bytes, stream_tag, tagged, wrapper_role};
use place::{check, check_on_top, contents, page_list, place_tagged};
use place::{Reach, check, check_on_top, contents, draw_on_top, lift_on_top, page_list, place_tagged, revealed_under_paint};
pub use place::{stamp, undrawable_message};
#[cfg(test)]
crates/edit/src/place.rs
+70
View File
@@ -371,6 +371,76 @@ pub(crate) fn place_tagged(doc: &mut Document, page: &pdfcraft_model::Page, tag:
Ok(mark_ref)
}
/// Draw `ops`, written for the space whose CTM is `ctm`, on top of everything on `page`, tagged
/// `Edited`. A rewritten invisible paragraph goes here when [`revealed_under_paint`]: an OCR
/// layer written before the scan image it lies under would keep the revealed text hidden (#3973).
pub(crate) fn draw_on_top(doc: &mut Document, page: &pdfcraft_model::Page, ctm: [f64; 6], ops: &[pdfcraft_content::Op]) -> Result<(), EditError> {
if ops.is_empty() {
return Ok(());
}
use pdfcraft_content::Op;
let mut all = vec![Op::new("q", vec![]), Op::new("cm", ctm.iter().map(|v| pdfcraft_content::num(*v)).collect())];
all.extend_from_slice(ops);
all.push(Op::new("Q", vec![]));
place_tagged(doc, page, "Edited", pdfcraft_content::serialize_ops(&all), false).map(drop)
}
/// Where a rewritten paragraph can be drawn, in user space ([`revealed_under_paint`]).
pub(crate) struct Reach {
/// The paragraph's old box, where its cover goes.
pub rect: [f64; 4],
/// The move asked for (`BlockStyle::offset`).
pub offset: [f64; 2],
/// The widest line the rewrite may draw, how many lines it has, and how far apart they are.
pub width: f64,
pub rows: usize,
pub pitch: f64,
pub upright: bool,
}
impl Reach {
/// The old box and the moved one, grown to `width` either way (alignment may grow a line left
/// or right), down to `rows` lines and up one `pitch` above its top (a larger size on the first line, however many lines the old box had); turned text grows by the lines' diagonal every way.
/// Unbounded when a number is not finite.
pub(crate) fn bounds(&self) -> [f64; 4] {
const UNBOUNDED: [f64; 4] = [f64::NEG_INFINITY, f64::NEG_INFINITY, f64::INFINITY, f64::INFINITY];
let ([dx, dy], r) = (self.offset, self.rect);
if !r.iter().chain(&self.offset).chain([&self.width, &self.pitch]).all(|v| v.is_finite()) {
return UNBOUNDED;
}
let moved = [r[0] + dx, r[1] + dy, r[2] + dx, r[3] + dy];
let rows = f64::from(u32::try_from(self.rows).unwrap_or(u32::MAX));
let (width, height) = (self.width.abs(), self.pitch.abs() * rows);
let grown = if self.upright {
[moved[0].min(moved[2] - width), moved[1].min(moved[3] - height), moved[2].max(moved[0] + width), moved[3] + self.pitch.abs()]
} else {
let d = width.hypot(height);
[moved[0] - d, moved[1] - d, moved[2] + d, moved[3] + d]
};
let b = [r[0].min(grown[0]), r[1].min(grown[1]), r[2].max(grown[2]), r[3].max(grown[3])];
if b.iter().all(|v| v.is_finite()) { b } else { UNBOUNDED }
}
}
/// Whether a rewritten paragraph that was `invisible` goes on top of page `page`: something after
/// operator `at` of the page's joined `ops` paints over `reach` (user space, [`Reach::bounds`]), as the scan image
/// does over an OCR layer written before it (#3973). XObjects count where they are drawn; an inline
/// image or a shading after it counts anywhere, and so does any XObject when the page's images
/// cannot be listed.
pub(crate) fn revealed_under_paint(doc: &Document, page: usize, invisible: bool, reach: [f64; 4], ops: &[pdfcraft_content::Op], at: usize) -> bool {
let later = ops.get(at..).unwrap_or_default();
if !invisible || later.iter().any(|op| op.is("BI") || op.is("sh")) {
return invisible;
}
later.iter().any(|op| op.is("Do")) && crate::images::painted_over(doc, page, reach, at).unwrap_or(true)
}
/// The operators to draw on top of the page ([`draw_on_top`]): all of `paragraph` when `over`,
/// none otherwise. The page was checked to take them ([`check_on_top`]) before the edit began.
pub(crate) fn lift_on_top(over: bool, paragraph: &mut Vec<pdfcraft_content::Op>) -> Vec<pdfcraft_content::Op> {
if over { std::mem::take(paragraph) } else { Vec::new() }
}
fn too_deep_on_top() -> EditError {
EditError::Invalid("the page's marked content is nested too deeply to add anything on top of it".into())
}
crates/edit/src/tests.rs
+120
View File
@@ -6213,6 +6213,126 @@ fn group_selection_refuses_incomplete_content_and_resource_budget_overruns() {
assert!(editable_objects(&doc, 0).is_err(), "an undecodable stream is refused");
}
/// Invisible text moves on top only when something painted after it covers it: an XObject drawn
/// over its box (an image, or a form holding one), or any inline image or shading. An image
/// clipped away entirely, or drawn elsewhere, does not move it.
#[test]
fn revealed_text_moves_on_top_only_when_painted_over() {
let on_top_of = |doc: Document| {
let block = text::text_blocks(&doc, 0).unwrap().remove(0);
let ops = pdfcraft_content::parse(&page_content_bytes(&doc, 0)).ops;
let at = ops.iter().position(|o| o.is("BT")).unwrap();
revealed_under_paint(&doc, 0, block.invisible, block.rect, &ops, at)
};
let on_top = |content: &str| on_top_of(image_page_with(content));
let scan = "q 200 0 0 100 100 100 cm /Im0 Do Q";
assert!(on_top(&format!("BT /F1 12 Tf 3 Tr 110 150 Td (Under) Tj ET {scan}")), "the scan covers it");
assert!(!on_top(&format!("BT /F1 12 Tf 3 Tr 400 600 Td (Away) Tj ET {scan}")), "an image elsewhere");
assert!(!on_top(&format!("{scan} BT /F1 12 Tf 3 Tr 110 150 Td (Over) Tj ET")), "the scan is drawn first");
assert!(!on_top(&format!("BT /F1 12 Tf 0 Tr 110 150 Td (Shown) Tj ET {scan}")), "visible text stays where it is");
assert!(!on_top("BT /F1 12 Tf 3 Tr 110 150 Td (Path) Tj ET 0 0 600 800 re f"), "a filled path does not move it");
assert!(on_top("BT /F1 12 Tf 3 Tr 400 600 Td (Shade) Tj ET /Sh sh"), "a shading after it");
assert!(on_top("BT /F1 12 Tf 3 Tr 400 600 Td (Inline) Tj ET BI /W 1 /H 1 /CS /G /BPC 8 ID x EI"), "an inline image after it");
let under = "BT /F1 12 Tf 3 Tr 110 150 Td (Under) Tj ET";
let elsewhere = "q 50 0 0 50 400 600 cm /Im0 Do Q";
assert!(!on_top(&format!("{under} q 0 0 10 10 re W n {scan} Q")), "the scan is clipped away entirely");
assert!(on_top(&format!("{under} {elsewhere} {scan}")), "two images, the second over it");
assert!(on_top(&format!("{under} {scan} {elsewhere}")), "two images, the first over it");
assert!(!on_top(&format!("{under} {elsewhere} {elsewhere}")), "two images, neither over it");
let in_form = |content: &str| {
let page = "<< /Type /Page /Parent 2 0 R /MediaBox [0 0 600 800] /Contents 4 0 R /Resources << /Font << /F1 5 0 R >> /XObject << /Fm0 7 0 R >> >> >>";
let image = stream("/Type /XObject /Subtype /Image /Width 1 /Height 1 /ColorSpace /DeviceGray /BitsPerComponent 8", "A");
let form = stream("/Type /XObject /Subtype /Form /BBox [0 0 1 1] /Resources << /XObject << /Im0 6 0 R >> >>", "/Im0 Do");
let objs = [
"<< /Type /Catalog /Pages 2 0 R >>".to_string(),
"<< /Type /Pages /Kids [3 0 R] /Count 1 >>".into(),
page.into(),
stream("", content),
"<< /Type /Font /Subtype /Type1 /BaseFont /Helvetica >>".into(),
image,
form,
];
on_top_of(build(&objs.iter().map(String::as_str).collect::<Vec<_>>()))
};
assert!(in_form(&format!("{under} q 200 0 0 100 100 100 cm /Fm0 Do Q")), "a form holding the scan, over it");
assert!(!in_form(&format!("{under} q 50 0 0 50 400 600 cm /Fm0 Do Q")), "a form holding an image elsewhere");
}
/// Revealed OCR text is checked where the rewrite lands, not only where it was: an image drawn
/// after it over just the moved or grown text still lifts the text on top, on both rewrite paths.
#[test]
fn an_ocr_rewrite_goes_on_top_of_paint_over_where_it_lands() {
let scan = "q 200 0 0 100 100 100 cm /Im0 Do Q";
let moved = text::BlockStyle { offset: Some([-290.0, -450.0]), ..Default::default() };
let sized = text::BlockStyle { size: Some(24.0), ..Default::default() };
let spaced = text::BlockStyle { line_spacing: Some(3.0), ..Default::default() };
let range = text::RangeStyle {
range: 0..2,
family: None,
resource: None,
size: Some(120.0),
color: None,
underline: None,
char_spacing: None,
scale: None,
};
let enlarged = text::BlockStyle { ranges: vec![range], ..Default::default() };
let above = "q 200 0 0 60 100 200 cm /Im0 Do Q";
let cases = [
("BT /F1 12 Tf 3 Tr 400 600 Td (Away) Tj ET", "Moved", &moved),
("BT /F1 12 Tf 3 Tr 400 600 Td (Away) Tj ET", "Moved\nTwice", &moved),
("BT /F1 12 Tf 3 Tr 40 150 Td (Hi) Tj ET", "Hello there, this line now runs well past the scan's left edge", &text::BlockStyle::default()),
("BT /F1 12 Tf 3 Tr 40 150 Td (Hi) Tj ET", "Hello there,\nthis line now runs well past the scan's left edge", &text::BlockStyle::default()),
("BT /F1 12 Tf 3 Tr 110 230 Td (Above) Tj ET", "One\nTwo\nThree\nFour\nFive\nSix", &text::BlockStyle::default()),
];
let low = "q 200 0 0 30 100 300 cm /Im0 Do Q";
let cases = cases.into_iter().map(|(c, e, s)| (c, e, s, scan)).chain([
("BT /F1 12 Tf 3 Tr 110 500 Td (Above) Tj ET", "One\nTwo\nThree\nFour\nFive\nSix", &sized, low),
("BT /F1 12 Tf 3 Tr 110 500 Td (Above) Tj ET", "One\nTwo\nThree\nFour\nFive\nSix", &spaced, low),
("BT /F1 12 Tf 3 Tr 110 150 Td (Hi there) Tj ET", "Hi there", &enlarged, above),
]);
for (content, edit, style, scan) in cases {
let mut doc = image_page_with(&format!("{content} {scan}"));
let before = text::text_blocks(&doc, 0).unwrap().remove(0);
let ops = pdfcraft_content::parse(&page_content_bytes(&doc, 0)).ops;
let at = ops.iter().position(|o| o.is("BT")).unwrap();
assert!(!revealed_under_paint(&doc, 0, true, before.rect, &ops, at), "{edit:?}: the old box alone is clear of the scan");
text::rewrite_block(&mut doc, 0, 0, Some(edit), style).unwrap();
let ops = pdfcraft_content::parse(&page_content_bytes(&doc, 0)).ops;
let scan_at = ops.iter().position(|o| o.is("Do")).unwrap();
let shows = ["Tj", "TJ", "'", "\""];
assert!(ops.iter().skip(scan_at).any(|o| shows.iter().any(|s| o.is(s))), "{edit:?}: the rewrite is drawn over the scan");
}
}
/// [`crate::Reach`] grows the old box to the move, the width either way and the lines below.
#[test]
fn a_rewrites_reach_covers_its_move_width_and_lines() {
let reach = |offset, width, rows, upright| crate::Reach { rect: [10.0, 90.0, 30.0, 100.0], offset, width, rows, pitch: 10.0, upright }.bounds();
assert_eq!(reach([0.0, 0.0], 20.0, 1, true), [10.0, 90.0, 30.0, 110.0]);
assert_eq!(reach([5.0, -50.0], 20.0, 1, true), [10.0, 40.0, 35.0, 100.0]);
assert_eq!(reach([0.0, 0.0], 50.0, 3, true), [-20.0, 70.0, 60.0, 110.0]);
assert_eq!(reach([0.0, 0.0], 30.0, 4, false), [-40.0, 40.0, 80.0, 150.0]);
assert_eq!(reach([f64::NAN, 0.0], 20.0, 1, true)[0], f64::NEG_INFINITY);
let larger = crate::Reach { rect: [10.0, 90.0, 30.0, 100.0], offset: [0.0, 0.0], width: 20.0, rows: 1, pitch: 24.0, upright: true };
assert_eq!(larger.bounds(), [10.0, 76.0, 30.0, 124.0], "a larger size grows the line up as well as down");
}
/// Revealed OCR text may have to go on top of the page; a page whose marked content nests too deep
/// for that is refused before either rewrite path changes anything.
#[test]
fn an_ocr_rewrite_refused_on_top_leaves_the_document_as_it_was() {
let content = format!("{} BT /F1 20 Tf 3 Tr 42 44 Td (John) Tj ET /Im Do {}", "/X BMC ".repeat(10_001), "EMC ".repeat(10_001));
let mut doc = text_page(&content);
assert!(text::text_blocks(&doc, 0).unwrap()[0].invisible);
let before = write_full(&doc, &SaveOptions::default()).unwrap();
for edit in ["Jane", "Jane\nDoe"] {
let error = text::rewrite_block(&mut doc, 0, 0, Some(edit), &text::BlockStyle::default()).unwrap_err();
assert!(error.to_string().contains("nested too deeply"), "{edit:?}: {error}");
assert!(write_full(&doc, &SaveOptions::default()).unwrap() == before, "{edit:?}: a refused rewrite changes nothing");
}
}
#[test]
fn group_move_budget_refusals_keep_the_original_page_selectable() {
let text = text_page(&format!("{} BT /F1 12 Tf 40 400 Td (Target) Tj ET", "n ".repeat(99_995)));
crates/edit/src/text/blocks.rs
+2 -1
View File
@@ -31,7 +31,8 @@ pub struct TextBlock {
/// Byte ranges in `editor_text` and their source line indexes.
pub spans: Vec<(std::ops::Range<usize>, usize)>,
/// Every line is invisible text (rendering mode 3): an OCR layer over a scanned image, whose
/// pixels show the words. A rewrite draws the new text visibly (see [`BlockStyle::cover`]).
/// pixels show the words. A rewrite draws the new text visibly, on top of the page when
/// the scan is drawn after it (see [`BlockStyle::cover`]).
pub invisible: bool,
}
crates/edit/src/text/interp.rs
+62 -11
View File
@@ -18,13 +18,33 @@ pub(super) struct TextState {
/// graphics state the text was painted with that [`TextState::ops`] doesn't restate.
pub(super) stroke: Vec<Op>,
pub(super) line_width: Option<f64>,
ext_gstate: Vec<Vec<u8>>,
ext_gstate: Vec<GsSelection>,
/// The dash pattern and line cap in force (see [`Ts`]): compared, not restated.
pub(super) dash: Dash,
pub(super) line_cap: i64,
}
impl TextState {
/// The state a page's content starts in (ISO 32000-2 §8.4.1, §9.3.1): what is in force on
/// top of the page, after its content wrapped in `q`/`Q`.
pub(crate) fn initial() -> Self {
TextState {
font: None,
char_spacing: 0.0,
word_spacing: 0.0,
scale: 1.0,
leading: 0.0,
rise: 0.0,
fill: Vec::new(),
render_mode: 0,
stroke: Vec::new(),
line_width: None,
ext_gstate: Vec::new(),
dash: (Vec::new(), 0.0),
line_cap: 0,
}
}
pub(super) fn of(ts: &Ts) -> Self {
TextState {
font: ts.font.as_ref().map(|(n, _)| (String::from_utf8_lossy(n).into_owned(), ts.size)),
@@ -54,13 +74,17 @@ impl TextState {
///
/// A `gs` has no inverse, so a selection `current` has and this state doesn't stays in force:
/// run these inside a `q`/`Q` that starts from `current`, one per state drawn.
pub(crate) fn paint_ops(&self, current: &TextState) -> Vec<Op> {
///
/// `ctm` is the CTM where they run. A soft mask is placed by the CTM its `gs` ran at
/// (ISO 32000-2 §11.6.5.2), so one selected at another CTM is selected there and the CTM
/// set back.
pub(crate) fn paint_ops(&self, current: &TextState, ctm: [f64; 6]) -> Vec<Op> {
let n = pdfcraft_content::num;
let mut v = Vec::new();
// The selections both share stay in force; the ones after them are selected again in order.
let shared = self.ext_gstate.iter().zip(&current.ext_gstate).take_while(|(a, b)| a == b).count();
for name in self.ext_gstate.iter().skip(shared) {
v.push(Op::new("gs", vec![Object::Name(name.clone())]));
for sel in self.ext_gstate.iter().skip(shared) {
sel.push_ops(&mut v, ctm);
}
// A reselected ExtGState may set `/LW`, so the width (already the effective one) follows it.
let reselected = !v.is_empty();
@@ -222,7 +246,7 @@ pub(super) struct Ts {
/// The line width a `w` or an ExtGState's `/LW` set (`None`: the initial 1.0).
pub(super) line_width: Option<f64>,
/// The ExtGState resources selected with `gs`, oldest first (see [`select_gs`]).
pub(super) ext_gstate: Vec<Vec<u8>>,
pub(super) ext_gstate: Vec<GsSelection>,
/// The dash pattern a `d` or an ExtGState's `/D` set (initially solid: `[] 0`).
pub(super) dash: Dash,
/// The line cap a `J` or an ExtGState's `/LC` set (initially 0, butt).
@@ -255,12 +279,38 @@ const MAX_GS: usize = 32;
/// Record `gs /name` in `list`. A `gs` sets every parameter its dictionary names to an absolute
/// value, so a later selection of the same dictionary overrides the earlier one completely: the
/// earlier entry goes, which keeps replaying the list equivalent to replaying the stream.
fn select_gs(list: &mut Vec<Vec<u8>>, name: &[u8]) {
list.retain(|n| n != name);
fn select_gs(list: &mut Vec<GsSelection>, name: &[u8], anchor: Option<[f64; 6]>) {
list.retain(|s| s.name != name);
if list.len() >= MAX_GS {
list.remove(0);
}
list.push(name.to_vec());
list.push(GsSelection { name: name.to_vec(), anchor });
}
/// One `gs` selection: the ExtGState's resource name, and the CTM it ran at when the dictionary
/// has a soft mask, which that CTM places (ISO 32000-2 §11.6.5.2).
#[derive(Clone, Debug, PartialEq)]
pub(crate) struct GsSelection {
name: Vec<u8>,
anchor: Option<[f64; 6]>,
}
impl GsSelection {
/// Select this ExtGState from where the CTM is `ctm`: at its anchor, then back to `ctm`. Both
/// moves need an invertible matrix; without one it is selected where it runs.
fn push_ops(&self, v: &mut Vec<Op>, ctm: [f64; 6]) {
let gs = Op::new("gs", vec![Object::Name(self.name.clone())]);
let moves = self.anchor.filter(|a| *a != ctm).and_then(|a| {
let (anchor, here) = (Matrix(a), Matrix(ctm));
Some((anchor.then(&here.invert()?), here.then(&anchor.invert()?)))
});
let Some((to, back)) = moves else {
v.push(gs);
return;
};
let cm = |m: Matrix| Op::new("cm", m.0.iter().map(|x| pdfcraft_content::num(*x)).collect());
v.extend([cm(to), gs, cm(back)]);
}
}
/// The text-showing operators of one stream with their text, font, box and baseline.
@@ -879,14 +929,15 @@ pub(super) fn interpret(
b"Ts" => ts.rise = op.num(0).unwrap_or(0.0),
b"Tr" => ts.render_mode = op.operands.first().and_then(Object::as_int).unwrap_or(-1),
b"gs" => {
if let Some(name) = op.name(0) {
select_gs(&mut ts.ext_gstate, name);
}
let mut held = None;
let params = match op.name(0).and_then(|name| states.get(name)) {
Some(p) => super::dict_in_place(doc, p, &mut held),
None => None,
};
if let Some(name) = op.name(0) {
let masked = params.and_then(|p| p.get(b"SMask")).is_some_and(|m| doc.resolve(m).as_dict().is_some());
select_gs(&mut ts.ext_gstate, name, masked.then_some(ts.ctm.0));
}
if let Some(params) = params {
ts.gs.apply(doc, params);
// `/LW` sets the line width as `w` does (ISO 32000-2 §8.4.5).
crates/edit/src/text/rewrite.rs
+32 -5
View File
@@ -702,11 +702,13 @@ fn block_ops(plan: &BlockPlan<'_>, writer: &LineWriter<'_>, wrapped: &[(String,
let mut ops = vec![Op::new("q", vec![])];
ops.extend(reveal_ops(plan.block, plan.lines, plan.style, plan.emit.ctm)?);
// Moved: a translation inside that state.
let mut ctm = Matrix(plan.emit.ctm);
if let Some((ax, ay)) = plan.movement {
ops.push(Op::new("cm", vec![n(1.0), n(0.0), n(0.0), n(1.0), n(ax), n(ay)]));
ctm = Matrix([1.0, 0.0, 0.0, 1.0, ax, ay]).then(&ctm);
}
// Painted as the paragraph was, including what its text object set before the first show.
ops.extend(plan.state.paint_ops(plan.after));
ops.extend(plan.state.paint_ops(plan.after, ctm.0));
ops.push(Op::new("BT", vec![]));
let mut state = plan.state.clone();
state.word_spacing = 0.0;
@@ -816,6 +818,11 @@ pub(crate) fn rewrite_block_from(
let first = members.first().copied().ok_or_else(|| EditError::Invalid("the paragraph has no source runs".into()))?;
members.iter().try_for_each(|line| line.ensure_editable())?;
let p = page_dict(doc, page)?;
// Revealed OCR text may go on top of the page (see `revealed_under_paint`): a page that can't
// take that is refused here, before either rewrite path changes anything.
if b.invisible {
crate::check_on_top(doc, &p)?;
}
let mut res = p.dict.get(b"Resources").map(|r| doc.resolve(r)).and_then(|r| r.as_dict().cloned()).unwrap_or_default();
let mut fonts_res = res.get(b"Font").map(|f| doc.resolve(f)).and_then(|f| f.as_dict().cloned()).unwrap_or_default();
let dispatch = choose_rewrite_path(doc, &fonts_res, &b, &members, first, text, style)?;
@@ -930,9 +937,28 @@ pub(crate) fn rewrite_block_from(
char_spacing: o_state.char_spacing,
lead,
};
let plan =
BlockPlan { block: &b, lines: &lines, style, emit, after, movement, state: o_state, tm, size, underline: style.underline == Some(true) };
let at = if split { start } else { emit.bt_op };
// Revealed OCR text that something is painted over goes on top of the page, where it starts
// from the initial state.
let widest = wrapped.iter().map(|(l, _, _)| advance.of(l)).fold(width, f64::max);
let reach = crate::Reach { rect: b.rect, offset: [dx, dy], width: widest, rows: wrapped.len(), pitch: lead * k, upright };
let over = crate::revealed_under_paint(doc, page, b.invisible, reach.bounds(), &ops, at);
let initial = TextState::initial();
let drawn_from = if over { &initial } else { after };
let plan = BlockPlan {
block: &b,
lines: &lines,
style,
emit,
after: drawn_from,
movement,
state: o_state,
tm,
size,
underline: style.underline == Some(true),
};
let mut block_ops = block_ops(&plan, &writer, &wrapped, font.show_font.clone())?;
let on_top = crate::lift_on_top(over, &mut block_ops);
let n = pdfcraft_content::num;
block_ops.extend(after.ops().into_iter().filter(|op| !op.is("Tf") || after.font.is_some()));
if split {
@@ -940,7 +966,6 @@ pub(crate) fn rewrite_block_from(
block_ops.push(Op::new("BT", vec![]));
block_ops.push(Op::new("Tm", emit.tlm.iter().map(|v| n(*v)).collect()));
}
let at = if split { start } else { emit.bt_op };
let new_data = splice(&pieces, &ops, |i| {
let insert = if i == at { std::mem::take(&mut block_ops) } else { Vec::new() };
(insert, !dropped.contains(&i))
@@ -957,12 +982,14 @@ pub(crate) fn rewrite_block_from(
d.set(b"Resources".to_vec(), Object::Dict(res));
}
})?;
crate::draw_on_top(doc, &p, emit.ctm, &on_top)?;
Ok(LineEdit { substituted: font.substituted })
}
/// The cover (when [`BlockStyle::cover`] asks for one) and visible rendering that start a
/// rewritten invisible paragraph, drawn in the space whose CTM is `ctm`. Nothing for visible text.
/// The paragraph's `q`/`Q` restores the rendering mode for the invisible text after it.
/// The paragraph's `q`/`Q` restores the rendering mode for the invisible text after it; one
/// painted over where it was is drawn on top of the page instead ([`crate::draw_on_top`]).
pub(crate) fn reveal_ops(block: &TextBlock, lines: &[TextLine], style: &BlockStyle, ctm: [f64; 6]) -> Result<Vec<Op>, EditError> {
if !block.invisible {
return Ok(Vec::new());
crates/edit/src/text_regions/styled.rs
+30 -7
View File
@@ -1,3 +1,4 @@
use super::super::interp::TextState;
use super::grouping::{across, is_centred, row_rect};
use super::{
BlockStyle, CraftFont, Dict, Document, EditError, HashMap, LineEdit, MAX_TYPE3_GLYPHS, Matrix, Metrics, Object, Op, PdfString, Pieces,
@@ -401,6 +402,16 @@ fn paragraph_geometry(doc: &Document, page: &pdfcraft_model::Page, block: &TextB
Ok((dx, dy, style.width.unwrap_or(natural).clamp(1.0, 14_400.0)))
}
/// How far below a row of `size` the next one is written: the source's leading unless a size
/// or line spacing is chosen.
fn row_step(style: &BlockStyle, size: f64, source_leading: f64) -> f64 {
if style.line_spacing.is_none() && style.size.is_none() {
source_leading
} else {
style.line_spacing.unwrap_or(1.5).clamp(0.5, 5.0) * style.size.unwrap_or(size)
}
}
type WrappedStyledRows = (StyledRows, RowRanges, Vec<usize>, Vec<bool>, Vec<usize>);
fn wrap_styled_rows(
@@ -976,6 +987,7 @@ pub(crate) fn rewrite_styled_block(
let o = &lead.origin;
// What is in force where the rewrite is drawn, and what the text after it expects.
let after = if split { &o.state } else { &o.bt_state };
let at = if split { start } else { o.bt_op };
validate_contextual_rows(&rows, &kept)?;
let measured: Vec<_> = rows.iter().map(|row| measure(row)).collect::<Result<_, _>>()?;
let shaped = prepare_shaped_rows(doc, &fonts, &rows, &kept, advance)?;
@@ -989,6 +1001,7 @@ pub(crate) fn rewrite_styled_block(
.ok_or_else(|| EditError::Invalid("the paragraph is drawn in a space it cannot be edited in".into()))?;
let (tx, ty) = back.apply(dx, dy);
block_ops.push(Op::new("cm", vec![n(1.0), n(0.0), n(0.0), n(1.0), n(tx), n(ty)]));
let run_ctm = Matrix([1.0, 0.0, 0.0, 1.0, tx, ty]).then(&Matrix(o.ctm)).0;
// A new row can land one leading above the next paragraph and look like the same
// paragraph. When every row is in this rewrite, a paragraph marker keeps them in one
// scope so the neighbour stays out (#2517). A row left in the source stream cannot share
@@ -1007,6 +1020,19 @@ pub(crate) fn rewrite_styled_block(
})
.rfind(|v| *v > 0.0)
.unwrap_or(first.size * 1.5);
// Revealed OCR text that something is painted over goes on top of the page, where it starts
// from the initial state.
let largest =
block.lines.iter().filter_map(|i| lines.get(*i)).map(|l| l.size).chain(style.ranges.iter().filter_map(|r| r.size)).fold(first.size, f64::max);
// At least one line of the largest glyphs, so a larger range reaches up as far as it draws.
const LINE_EM: f64 = 1.2;
let pitch = source_leading.max(row_step(style, largest, source_leading)).max(largest * LINE_EM);
let widest = measured.iter().copied().fold(width, f64::max);
// This path lays rows out along user-space baselines, so the paragraph is upright here.
let reach = crate::Reach { rect: block.rect, offset: [dx, dy], width: widest, rows: rows.len(), pitch, upright: true };
let over = crate::revealed_under_paint(doc, page, block.invisible, reach.bounds(), &ops, at);
let initial = TextState::initial();
let drawn_from = if over { &initial } else { after };
let mut baseline = first.origin.baseline;
for (row_index, row) in rows.iter().enumerate() {
let anchor = block.rows.get(row_index).and_then(|r| r.first()).and_then(|i| lines.get(*i));
@@ -1073,7 +1099,7 @@ pub(crate) fn rewrite_styled_block(
// object (`gs`, stroke, line width, `Tr`; #3508/#3509) from the state the rewrite is
// drawn in: a `gs` has no inverse, so one run's ExtGState would stay for the next.
block_ops.push(Op::new("q", vec![]));
block_ops.extend(state.paint_ops(after));
block_ops.extend(state.paint_ops(drawn_from, run_ctm));
if let Some((_, shaped, _)) = shaped_runs.get(&(row_index, pos)) {
let resource = glyph_resources.get(&(row_index, pos)).ok_or_else(|| EditError::Invalid("shaped font is missing".into()))?;
let mut properties = Dict::new();
@@ -1164,21 +1190,17 @@ pub(crate) fn rewrite_styled_block(
x += segment_width;
pos = end;
}
baseline -= if style.line_spacing.is_none() && style.size.is_none() {
source_leading
} else {
style.line_spacing.unwrap_or(1.5).clamp(0.5, 5.0) * style.size.unwrap_or(anchor.size)
};
baseline -= row_step(style, anchor.size, source_leading);
}
push_whole_paragraph_marker(&mut block_ops, marked, "EMC");
block_ops.push(Op::new("Q", vec![]));
let on_top = crate::lift_on_top(over, &mut block_ops);
block_ops.extend(after.ops());
if split {
block_ops.insert(0, Op::new("ET", vec![]));
block_ops.push(Op::new("BT", vec![]));
block_ops.push(Op::new("Tm", o.tlm.iter().map(|v| n(*v)).collect()));
}
let at = if split { start } else { o.bt_op };
// A kept source row sits outside `block_ops`. Bound the whole paragraph when a marker is
// safe. The close is not part of `block_ops`: that vector can end inside a text object
// when the rewrite splits one, and a kept row's operators come later in the stream.
@@ -1190,5 +1212,6 @@ pub(crate) fn rewrite_styled_block(
d.set(b"Contents".to_vec(), contents);
d.set(b"Resources".to_vec(), Object::Dict(res));
})?;
crate::draw_on_top(doc, &p, o.ctm, &on_top)?;
Ok(LineEdit { substituted })
}
crates/engine/src/camera_clean.rs
+196 -39
View File
@@ -1,6 +1,8 @@
//! Clean up a straightened camera photo (Enhance ▸ Camera image ▸ enhancement level): even out
//! the lighting so the paper turns white (room-light colour casts, a hand's shadow, darker
//! edges go with it), then raise the contrast and drop isolated specks. Colour ink stays colour.
//! Clean up a straightened camera photo (Enhance ▸ Camera image ▸ enhancement level) or a
//! scanned page's picture (Enhance ▸ Scanned document, [`crate::scan_clean`]): even out the
//! lighting so the paper turns white (room-light colour casts, a hand's shadow, darker edges and
//! book gutters go with it), then raise the contrast and drop isolated specks. Colour ink stays
//! colour.
//!
//! The paper's illumination is estimated per channel on a coarse grid of blocks, from each
//! block's bright end (the paper, not the ink), smoothed into a continuous map and divided out.
@@ -8,6 +10,8 @@
/// The strongest a level can be.
pub const MAX_LEVEL: u8 = 100;
/// The level from which [`clean`] also drops isolated specks.
pub const DESPECKLE_FROM: u8 = 60;
/// How an enhanced camera page looks.
#[derive(Clone, Copy, Debug, Default, PartialEq, Eq)]
@@ -66,34 +70,158 @@ pub fn apply(rgb: &[u8], width: u32, height: u32, look: Look, level: u8) -> Opti
let even = clean(rgb, width, height, 50)?;
let t = 0.45 + 0.4 * level.min(MAX_LEVEL) as f32 / MAX_LEVEL as f32;
let mut grey: Vec<u8> = even.as_chunks::<3>().0.iter().map(|p| if (lum(p) as f32 / 255.0) < t { 0 } else { 255 }).collect();
despeckle_grey(&mut grey, w, h);
despeckle(&mut grey, 1, w, h);
Pixels::Grey(grey)
}
})
}
/// Clean `rgb` (RGB8, `width` × `height`) at `level` (0 = unchanged … 100 = strongest).
/// Returns `None` when the buffer doesn't match the size.
/// Clean `rgb` (RGB8, `width` × `height`) at `level` (0 = unchanged … 100 = strongest), from
/// [`DESPECKLE_FROM`] without isolated specks. Returns `None` when the buffer doesn't match the
/// size.
pub fn clean(rgb: &[u8], width: u32, height: u32, level: u8) -> Option<Vec<u8>> {
clean_samples(rgb, 3, width, height, CleanUp { level, edge_shadows: false, despeckle: level >= DESPECKLE_FROM })
}
/// What [`clean_samples`] does.
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub struct CleanUp {
/// 0 (unchanged) … 100: even out the paper's illumination and raise the contrast, as
/// [`clean`]'s level.
pub level: u8,
/// Even out shadows along the picture's edges first (a book's gutter, a copier's dark edge):
/// lightened line by line to the paper beside them, within the outer fifth of each side.
pub edge_shadows: bool,
/// Drop isolated dark specks.
pub despeckle: bool,
}
/// [`clean`] for samples of `n` components (1: grey, 3: RGB), with each step chosen apart (each
/// works at level 0 too). Returns `None` when `n` is neither or the buffer doesn't match the size.
pub fn clean_samples(px: &[u8], n: usize, width: u32, height: u32, how: CleanUp) -> Option<Vec<u8>> {
let (w, h) = (width as usize, height as usize);
if w == 0 || h == 0 || w.checked_mul(h)?.checked_mul(3)? != rgb.len() {
if !(n == 1 || n == 3) || w == 0 || h == 0 || w.checked_mul(h)?.checked_mul(n)? != px.len() {
return None;
}
let level = level.min(MAX_LEVEL);
if level == 0 {
return Some(rgb.to_vec());
let level = how.level.min(MAX_LEVEL);
let mut out = px.to_vec();
if how.edge_shadows {
even_edges(&mut out, n, w, h);
}
if level > 0 {
out = even_out(&out, n, w, h, level);
}
if how.despeckle {
despeckle(&mut out, n, w, h);
}
Some(out)
}
/// The share of the width (and of the height) at each side where [`even_edges`] works.
const EDGE_BAND: f32 = 0.2;
/// The most [`even_edges`] lightens a line by; a darker line (a scanner's black border, a
/// picture) is not taken for shadowed paper and stays as it is.
const MAX_EDGE_GAIN: f32 = 3.0;
/// Even out shadows that run along the picture's edges, in place: a book's gutter, a copier's
/// dark border, a page that curved away from the glass. Within the outer [`EDGE_BAND`] of each
/// side, each column (then each row) gets the paper's level along it (per channel, its 90th
/// percentile, so the text across it doesn't count), and is lightened to the brightest paper of
/// that band. The block grid of [`background`] is too coarse for such a narrow shadow: its
/// blocks borrow the brighter paper next to them.
fn even_edges(px: &mut [u8], n: usize, w: usize, h: usize) {
for vertical in [true, false] {
// Columns: lines across the width, each running down the height; rows the other way.
let (lines, along) = if vertical { (w, h) } else { (h, w) };
let band = ((lines as f32 * EDGE_BAND) as usize).min(lines / 2);
if band == 0 {
continue;
}
let at = |line: usize, i: usize| if vertical { i * w + line } else { line * w + i };
// The paper's level of every line in the bands and a few beyond them.
let reach = band.saturating_add(3).min(lines);
let measured = |line: usize| line < reach || line >= lines.saturating_sub(reach);
let mut level: Vec<Option<[f32; 3]>> = vec![None; lines];
let mut hist = [[0u32; 256]; 3];
for (line, slot) in level.iter_mut().enumerate().filter(|(line, _)| measured(*line)) {
hist.iter_mut().for_each(|hh| hh.fill(0));
for i in 0..along {
let p = at(line, i) * n;
for (c, hh) in hist.iter_mut().enumerate().take(n) {
if let Some(v) = px.get(p + c)
&& let Some(count) = hh.get_mut(*v as usize)
{
*count += 1;
}
}
}
let want = (along as f32 * 0.9) as u32;
*slot = Some(std::array::from_fn(|c| {
let mut seen = 0;
let v = hist.get(c).and_then(|hh| {
hh.iter().position(|&count| {
seen += count;
seen > want
})
});
v.unwrap_or(255) as f32 / 255.0
}));
}
// Smoothed: the median of five neighbouring lines (fewer at the very edge, centred all
// the same, so a gradient isn't shifted), so one odd line (a rule, a fold) doesn't stand
// out.
let smooth: Vec<Option<[f32; 3]>> = (0..lines)
.map(|line| {
level.get(line).copied().flatten()?;
let r = line.min(lines - 1 - line).min(2);
Some(std::array::from_fn(|c| {
let mut near: Vec<f32> = (line - r..=line + r).filter_map(|l| level.get(l).copied().flatten()).map(|v| v[c]).collect();
near.sort_by(f32::total_cmp);
near.get(near.len() / 2).copied().unwrap_or(1.0)
}))
})
.collect();
// Each side's band, and the lines its paper is measured on (the band and a few beyond).
for (shadowed, measured) in [(0..band, 0..reach), (lines - band..lines, lines.saturating_sub(reach)..lines)] {
// The brightest paper there: what the shadow is lightened to.
let paper: [f32; 3] =
std::array::from_fn(|c| measured.clone().filter_map(|l| smooth.get(l).copied().flatten()).map(|v| v[c]).fold(0.0, f32::max));
for line in shadowed {
let Some(own) = smooth.get(line).copied().flatten() else { continue };
let gain: [f32; 3] = std::array::from_fn(|c| {
let g = paper[c] / own[c].max(1e-3);
if (1.0..=MAX_EDGE_GAIN).contains(&g) { g } else { 1.0 }
});
if gain.iter().take(n).all(|&g| g == 1.0) {
continue;
}
for i in 0..along {
let p = at(line, i) * n;
for (c, g) in gain.iter().enumerate().take(n) {
if let Some(v) = px.get_mut(p + c) {
*v = (*v as f32 * g).round().clamp(0.0, 255.0) as u8;
}
}
}
}
}
}
}
/// The lighting evened out and the contrast raised at `level` (1–100); `px` holds `w` × `h`
/// pixels of `n` components.
fn even_out(px: &[u8], n: usize, w: usize, h: usize, level: u8) -> Vec<u8> {
let k = level as f32 / MAX_LEVEL as f32;
let background = background(rgb, w, h);
let background = background(px, n, w, h);
// Lighting is evened out fully from half strength on; contrast keeps growing to the top.
let flatten = (2.0 * k).min(1.0);
let white = 1.0 - 0.12 * k; // what reaches white
let black = 0.30 * k; // what reaches black
let mut out = Vec::with_capacity(rgb.len());
for (i, px) in rgb.as_chunks::<3>().0.iter().enumerate() {
let mut out = Vec::with_capacity(px.len());
for (i, p) in px.chunks_exact(n).enumerate() {
let (x, y) = (i % w, i / w);
let bg = background.at(x, y);
for (c, &v) in px.iter().enumerate() {
for (c, &v) in p.iter().enumerate() {
let v = v as f32 / 255.0;
let b = bg.get(c).copied().unwrap_or(1.0).max(0.05);
let even = (v / b).min(1.0);
@@ -102,10 +230,7 @@ pub fn clean(rgb: &[u8], width: u32, height: u32, level: u8) -> Option<Vec<u8>>
out.push((stretched * 255.0).round() as u8);
}
}
if level >= 60 {
despeckle(&mut out, w, h);
}
Some(out)
out
}
/// The paper's illumination: a grid of per-channel values, read back with bilinear
@@ -140,9 +265,10 @@ impl Background {
}
}
/// Estimate the paper's illumination: per block, the 90th percentile of each channel (bright
/// paper, not dark ink); blocks that are mostly ink borrow from their neighbours; then smoothed.
fn background(rgb: &[u8], w: usize, h: usize) -> Background {
/// Estimate the paper's illumination: per block, the 90th percentile of each of the `n` channels
/// (bright paper, not dark ink); blocks that are mostly ink borrow from their neighbours; then
/// smoothed. Channels past `n` stay white.
fn background(px: &[u8], n: usize, w: usize, h: usize) -> Background {
let block = (w.min(h) / 16).clamp(8, 128);
let (cols, rows) = (w.div_ceil(block), h.div_ceil(block));
let mut cells = Vec::with_capacity(cols * rows);
@@ -150,20 +276,20 @@ fn background(rgb: &[u8], w: usize, h: usize) -> Background {
for by in 0..rows {
for bx in 0..cols {
hist.iter_mut().for_each(|hh| hh.fill(0));
let mut n = 0u32;
let mut count = 0u32;
for y in by * block..((by + 1) * block).min(h) {
for x in bx * block..((bx + 1) * block).min(w) {
if let Some(p) = rgb.get((y * w + x) * 3..(y * w + x) * 3 + 3) {
if let Some(p) = px.get((y * w + x) * n..(y * w + x) * n + n) {
for (c, &v) in p.iter().enumerate() {
if let Some(slot) = hist.get_mut(c).and_then(|hh| hh.get_mut(v as usize)) {
*slot += 1;
}
}
n += 1;
count += 1;
}
}
}
let want = (n as f32 * 0.9) as u32;
let want = (count as f32 * 0.9) as u32;
cells.push(std::array::from_fn(|c| {
let mut seen = 0;
let v = hist.get(c).and_then(|hh| {
@@ -219,19 +345,15 @@ fn background(rgb: &[u8], w: usize, h: usize) -> Background {
Background { cols, rows, block: block as f32, cells }
}
/// [`despeckle`] for one grey byte per pixel.
fn despeckle_grey(grey: &mut [u8], w: usize, h: usize) {
let mut rgb: Vec<u8> = grey.iter().flat_map(|&v| [v, v, v]).collect();
despeckle(&mut rgb, w, h);
for (g, p) in grey.iter_mut().zip(rgb.as_chunks::<3>().0.iter()) {
*g = p[0];
}
}
/// Lighten dark pixels with at most one dark neighbour (specks), in place.
fn despeckle(rgb: &mut [u8], w: usize, h: usize) {
let lum = |p: &[u8; 3]| (p[0] as u32 * 299 + p[1] as u32 * 587 + p[2] as u32 * 114) / 1000;
let dark: Vec<bool> = rgb.as_chunks::<3>().0.iter().map(|p| lum(p) < 128).collect();
/// Lighten dark pixels with at most one dark neighbour (specks), in place; `px` holds `w` × `h`
/// pixels of `n` components (1 or 3).
fn despeckle(px: &mut [u8], n: usize, w: usize, h: usize) {
let lum = |p: &[u8]| match p {
[r, g, b] => (*r as u32 * 299 + *g as u32 * 587 + *b as u32 * 114) / 1000,
[v, ..] => *v as u32,
[] => 255,
};
let dark: Vec<bool> = px.chunks_exact(n.max(1)).map(|p| lum(p) < 128).collect();
let is_dark = |x: isize, y: isize| {
x >= 0 && y >= 0 && (x as usize) < w && (y as usize) < h && dark.get(y as usize * w + x as usize).copied().unwrap_or(false)
};
@@ -243,7 +365,7 @@ fn despeckle(rgb: &mut [u8], w: usize, h: usize) {
let neighbours =
(-1..=1).flat_map(|dy| (-1..=1).map(move |dx| (dx, dy))).filter(|&(dx, dy)| (dx, dy) != (0, 0) && is_dark(x + dx, y + dy)).count();
if neighbours <= 1
&& let Some(p) = rgb.get_mut((y as usize * w + x as usize) * 3..(y as usize * w + x as usize) * 3 + 3)
&& let Some(p) = px.get_mut((y as usize * w + x as usize) * n..(y as usize * w + x as usize) * n + n)
{
p.fill(255);
}
@@ -388,5 +510,40 @@ mod tests {
assert_eq!(clean(&[], 0, 0, 50), None);
assert_eq!(clean(&[0; 12], 2, 3, 50), None);
assert_eq!(clean(&[9; 3], 1, 1, 255).map(|v| v.len()), Some(3));
let every = CleanUp { level: 100, edge_shadows: true, despeckle: true };
assert_eq!(clean_samples(&[0; 12], 2, 2, 3, every), None, "two components");
assert_eq!(clean_samples(&[0; 5], 1, 2, 3, every), None);
for (w, h) in [(1, 1), (1, 9), (9, 1), (2, 2), (5, 3)] {
assert_eq!(clean_samples(&vec![100; w * h], 1, w as u32, h as u32, every).map(|v| v.len()), Some(w * h));
}
}
#[test]
fn edge_shadows_lighten_shadowed_paper_but_not_a_black_border_or_the_middle() {
// Grey paper (200) with a shadow darkening the left 20 columns to 120, a black scanner
// border in the right 8 columns, and a dark block in the middle.
let (w, h) = (100usize, 60usize);
let mut grey = vec![200u8; w * h];
for y in 0..h {
for x in 0..w {
let v = if x >= w - 8 {
5
} else if (40..60).contains(&x) && (20..40).contains(&y) {
30
} else if x < 20 {
120 + (80 * x / 20) as u8
} else {
200
};
grey[y * w + x] = v;
}
}
let out = clean_samples(&grey, 1, w as u32, h as u32, CleanUp { level: 0, edge_shadows: true, despeckle: false }).unwrap();
for x in [0, 5, 19] {
assert!(out[30 * w + x].abs_diff(200) <= 3, "column {x}: {} → {}", grey[30 * w + x], out[30 * w + x]);
}
assert_eq!(out[30 * w + w - 3], 5, "a black border is no shadow");
assert_eq!(out[30 * w + 50], 30, "the middle is untouched");
assert_eq!(out[30 * w + 30], 200);
}
}
crates/engine/src/catalog.rs
+1 -1
View File
@@ -265,7 +265,7 @@ pub static TOOL_GROUPS: &[ToolGroup] = &[
items: &[
item("In this file", "scan-text", "ocr.recognize", Ready),
item("In multiple files", "files", "ocr.recognize_batch", Ready),
item("Enhance scanned file", "sparkles", "ocr.enhance", Planned("M10")),
item("Enhance scanned file", "sparkles", "ocr.enhance", Ready),
item("Enhance camera image", "camera", "ocr.enhance_camera", Ready),
item("Correct recognized text", "text-select", "ocr.correct", Planned("M10")),
],
crates/engine/src/commands.rs
+5
View File
@@ -220,6 +220,10 @@ pub const COMMANDS: &[CommandSpec] = &[
// Alt-drag for moving windows (several Linux window managers).
// Copy events stay with the frontend so focused text inputs and page clipboard routing keep ownership.
c("edit.copy", "Copy", EDIT, None, Copying, "copy"),
// Comments travel between documents as XFDF (M14): ⌘C with comments selected copies them,
// ⌘V pastes the copies onto the current page; the strip's Copy/Paste in the menus too.
c("comment.copy", "Copy comments", EDIT, None, Document, "copy"),
c("comment.paste", "Paste comments", EDIT, None, Annotate, "clipboard-paste"),
c("edit.column_select", "Column select", EDIT, None, Document, "text-select"),
c("edit.snapshot", "Take a snapshot", EDIT, None, Copying, "camera"),
c("view.full_screen", "Full screen mode", VIEW, Some(Shortcut::cmd("L")), Document, "maximize"),
@@ -355,6 +359,7 @@ pub const COMMANDS: &[CommandSpec] = &[
c("tools.document_js", "Document JavaScripts…", None, None, Modification, "file-code"),
c("ocr.recognize", "Recognize text…", None, None, Modification, "scan-text"),
c("ocr.enhance_camera", "Enhance camera image", None, None, Modification, "camera"),
c("ocr.enhance", "Enhance scanned file…", None, None, Modification, "sparkles"),
c("ocr.recognize_batch", "Recognize text in multiple files…", None, None, Nothing, "files"),
c("a11y.check", "Check for accessibility…", None, None, Document, "accessibility"),
c("a11y.find_untagged", "Find untagged content", None, None, Document, "search"),
crates/engine/src/copy_image.rs
+1 -1
View File
@@ -112,7 +112,7 @@ pub(crate) fn added_picture(cos: &Cos, file: &[u8]) -> Option<ObjRef> {
}
/// The largest raster image Form `form` draws through its resources (nested Forms included).
fn picture_in_form(cos: &Cos, form: ObjRef) -> Option<ObjRef> {
pub(crate) fn picture_in_form(cos: &Cos, form: ObjRef) -> Option<ObjRef> {
let mut seen = HashSet::from([form]);
let mut best: Option<(u64, ObjRef)> = None;
let res = resources(cos, form)?;
crates/engine/src/dates.rs
+12
View File
@@ -582,6 +582,18 @@ pub fn date_language(code: &str) -> Option<&'static DateLanguage> {
DATE_LANGUAGES.iter().find(|l| l.code.eq_ignore_ascii_case(code.trim()))
}
/// Form date fields (`AFDate_FormatEx`, `util.printd`, `util.scand`) write month and weekday
/// names in date language `code` from now on, as a localized Acrobat writes its own language's:
/// `dd-mmm-yy` is `18-ott-26` in Italian (#5303). An unknown code is English.
pub fn use_form_date_language(code: &str) {
let l = date_language(code).unwrap_or(&DATE_LANGUAGES[0]);
let names = pdfcraft_forms::af::DateNames { months: l.months, months_short: l.months_short, days: l.days, days_short: l.days_short };
if pdfcraft_forms::af::date_names() != names {
pdfcraft_forms::af::set_date_names(names);
pdfcraft_js::set_date_names(l.months, l.months_short, l.days, l.days_short);
}
}
/// The characters of `text` that Fill & Sign can't write into a PDF yet, each once. Its text is
/// drawn in Helvetica with WinAnsi encoding (Western European letters), so anything else (`ř`,
/// Japanese, Chinese) would become `?` in the file.
crates/engine/src/edit.rs
+64 -4
View File
@@ -3,10 +3,11 @@
use std::sync::Arc;
use crate::{
AddedContent, AddedText, Background, BoxSpec, EditError, FieldProps, FieldValue, FillMark, HeaderFooter, Hidden, ImageEdit, InitialView, Label,
LineEnding, LinkAction, LinkStyle, MarkFile, MarkKind, Markup, NewAnnotation, NewField, NoteIcon, ObjectTarget, PageBox, Protection, ReviewState,
Rgb, Shape, Standard, Style, TabOrder, TransitionSpec, Watermark, a11y, copy_image, decode_data_file, embed_image, field_table, fit_picture,
image_content_type, is_xfa_data_file, js, mark_source, measure, open_source, strip_xfa, unrestricted, xfa, xfa_after_data_change,
AddedContent, AddedText, Background, BorderShape, BoxSpec, EditError, FieldProps, FieldValue, FillMark, HeaderFooter, Hidden, ImageEdit,
InitialView, Label, LineEnding, LinkAction, LinkStyle, MarkFile, MarkKind, Markup, NewAnnotation, NewField, NoteIcon, ObjectTarget, PageBox,
Protection, ReviewState, Rgb, Shape, Standard, Style, TabOrder, TransitionSpec, Watermark, a11y, copy_image, decode_data_file, embed_image,
field_table, fit_picture, image_content_type, is_xfa_data_file, js, mark_source, measure, model_poster, open_source, strip_xfa, unrestricted,
xfa, xfa_after_data_change,
};
mod labels;
@@ -119,6 +120,15 @@ pub enum Edit {
bold: Option<bool>,
italic: Option<bool>,
},
/// Write `bookmarks` as children of `parent` from child `index` on, or with `replace` in
/// place of the whole outline (`parent` must then be `[]`). A bookmark whose page this
/// document lacks is kept without a destination (see [`pdfcraft_organize::import_outline`]).
ImportBookmarks {
parent: Vec<usize>,
index: usize,
bookmarks: Vec<pdfcraft_organize::OutlineNode>,
replace: bool,
},
/// Bookmarks for the tagged headings (H, H1–H6), nested by level under a new first
/// "Untitled" bookmark (New Bookmarks from Structure).
BookmarksFromStructure,
@@ -157,6 +167,17 @@ pub enum Edit {
file: MarkFile,
author: String,
},
/// Rich Media ▸ Add 3D: a U3D or PRC `model` as a 3D annotation in `rect` on `page`, named
/// `name` (its alternate description). Its poster is the picture in `poster` (a PDF page or an
/// image) fitted inside `rect`, keeping its shape and centred, or a labelled frame without one.
AddModel3d {
page: usize,
rect: [f64; 4],
model: Arc<Vec<u8>>,
poster: Option<MarkFile>,
activation: pdfcraft_annot::Activation3d,
name: String,
},
/// Delete the comment at `index` in the page's `/Annots`, with its pop-up and replies.
DeleteAnnotation {
page: usize,
@@ -243,12 +264,35 @@ pub enum Edit {
index: usize,
fill: Option<Rgb>,
},
/// Draw a rectangle's, oval's, polygon's or text box's fill, or a text box's text, at an
/// opacity of its own (0–1), the line or border keeping the comment's; `None` leaves that
/// part as it is. `StyleAnnotation`'s opacity sets the whole comment's.
PartOpacity {
page: usize,
index: usize,
fill: Option<f64>,
text: Option<f64>,
},
/// Change the font size (pt) of a text box, typewriter text or callout.
FontSizeAnnotation {
page: usize,
index: usize,
size: f64,
},
/// Give a text box, typewriter text or callout a border colour of its own (`None`: the
/// text's again).
BorderColorAnnotation {
page: usize,
index: usize,
color: Option<Rgb>,
},
/// Give a text box or typewriter text an oval, a circle or a rectangle for its border: the
/// text box keeps its place, the oval or circle goes around it.
BorderShapeAnnotation {
page: usize,
index: usize,
shape: BorderShape,
},
/// Comment properties ▸ General / note icon.
SetAnnotationInfo {
page: usize,
@@ -386,6 +430,13 @@ pub enum Edit {
page: usize,
words: Vec<pdfcraft_ocr::PlacedWord>,
},
/// Scan & OCR ▸ Enhance ▸ Scanned document: put the scan images
/// [`ScanCleanJob::run`](crate::scan_clean::ScanCleanJob::run) cleaned in place of the old
/// ones (each object must still be an image). Built by
/// [`Session::apply_scan_clean`](crate::Session::apply_scan_clean).
EnhanceScan {
images: Vec<(pdfcraft_cos::ObjRef, pdfcraft_cos::Stream)>,
},
/// Edit a PDF ▸ Edit text: replace the text of line `line` (from `Document::text_lines`) on
/// `page`, in its own font when it can show it, else in Helvetica.
EditTextLine {
@@ -534,6 +585,15 @@ pub enum Edit {
name: String,
bytes: Arc<Vec<u8>>,
},
/// Paste copied comments (XFDF from [`Session::export_comments`]) onto `page` as new
/// comments: copies under names of their own, so nothing is replaced and pasting twice
/// gives two sets; replies keep their copied parent.
PasteComments {
xfdf: String,
page: usize,
/// Where the copies go: moved by (dx, dy) in user space, or at their own positions.
offset: Option<[f64; 2]>,
},
/// Remove Hidden Information: the chosen categories.
RemoveHidden {
which: Vec<Hidden>,
crates/engine/src/edit/labels.rs
+15 -3
View File
@@ -38,6 +38,7 @@ pub const EDIT_LABELS: &[&str] = &[
"Set bookmark destination",
"Colour",
"Style",
"Import bookmarks",
"New bookmarks from structure",
"New bookmarks from text styles",
"Number pages",
@@ -68,7 +69,9 @@ pub const EDIT_LABELS: &[&str] = &[
"Add inserted text",
"Add file attachment",
"Add stamp",
"Add 3D model",
"Delete comment",
"Paste comments",
"Edit comment",
"Reply",
"Set status None",
@@ -99,6 +102,7 @@ pub const EDIT_LABELS: &[&str] = &[
"Set alternate text",
"Mark figure as decorative",
"Recognize text",
"Enhance scanned file",
"Run JavaScript",
"Run form script",
"Save as {level}",
@@ -193,6 +197,7 @@ impl Edit {
// The bookmark Style and Colour menus' own names, already said in every language.
Edit::SetBookmarkColor { .. } => Label::of("Colour"),
Edit::SetBookmarkStyle { .. } => Label::of("Style"),
Edit::ImportBookmarks { .. } => Label::of("Import bookmarks"),
Edit::BookmarksFromStructure => Label::of("New bookmarks from structure"),
Edit::BookmarksFromTextStyles { .. } => Label::of("New bookmarks from text styles"),
Edit::NumberPages { .. } => Label::of("Number pages"),
@@ -205,7 +210,9 @@ impl Edit {
Edit::SetMeasurementScale { .. } => Label::of("Set measurement scale"),
Edit::AddAnnotation(a) => Label::of(add_annotation(&a.shape)),
Edit::AddCustomStamp { .. } => Label::of("Add stamp"),
Edit::AddModel3d { .. } => Label::of("Add 3D model"),
Edit::DeleteAnnotation { .. } => Label::of("Delete comment"),
Edit::PasteComments { .. } => Label::of("Paste comments"),
Edit::SetAnnotationContents { .. } => Label::of("Edit comment"),
Edit::ReplyToAnnotation { .. } => Label::of("Reply"),
// The state's name is part of the wording: a language may need to agree it with the rest.
@@ -225,9 +232,13 @@ impl Edit {
Edit::MoveAnnotation { .. } => Label::of("Move comment"),
Edit::ReshapeAnnotation { .. } => Label::of("Reshape comment"),
Edit::ResizeAnnotation { .. } => Label::of("Resize comment"),
Edit::StyleAnnotation { .. } | Edit::FillAnnotation { .. } | Edit::FontSizeAnnotation { .. } | Edit::SetAnnotationInfo { .. } => {
Label::of("Change comment properties")
}
Edit::StyleAnnotation { .. }
| Edit::FillAnnotation { .. }
| Edit::PartOpacity { .. }
| Edit::FontSizeAnnotation { .. }
| Edit::BorderColorAnnotation { .. }
| Edit::BorderShapeAnnotation { .. }
| Edit::SetAnnotationInfo { .. } => Label::of("Change comment properties"),
Edit::SetFieldValue { name, .. } => Label::with("Fill in {name}", name.as_str()),
Edit::SetFieldImage { name, .. } => Label::with("Set the image of {name}", name.as_str()),
Edit::ResetForm { .. } => Label::of("Clear form"),
@@ -248,6 +259,7 @@ impl Edit {
Edit::MarkDecorative { .. } => Label::of("Mark figure as decorative"),
Edit::MarkTagArtifact { .. } => Label::of("Mark content as artifact"),
Edit::AddOcrText { .. } => Label::of("Recognize text"),
Edit::EnhanceScan { .. } => Label::of("Enhance scanned file"),
Edit::ApplyScriptChanges { .. } => Label::of("Run JavaScript"),
Edit::XfaEvent { .. } => Label::of("Run form script"),
// The level is the standard's own name ("PDF/A-2b"), so it goes in as it stands.
crates/engine/src/edit/run.rs
+33 -1
View File
@@ -1,6 +1,6 @@
use super::{
AddedContent, Arc, Edit, EditError, FieldValue, ImageEdit, NewAnnotation, Shape, a11y, copy_image, decode_data_file, embed_image, fit_picture,
image_content_type, is_xfa_data_file, js, mark_source, measure, open_source, strip_xfa, unrestricted, xfa, xfa_after_data_change,
image_content_type, is_xfa_data_file, js, mark_source, measure, model_poster, open_source, strip_xfa, unrestricted, xfa, xfa_after_data_change,
};
use crate::SourceProblem;
use crate::sources::{comment_source, import_pdf_comments};
@@ -27,6 +27,7 @@ pub(crate) fn check_permission(edit: &Edit, p: &pdfcraft_cos::Permissions) -> Re
| Edit::SetBookmarkPage { .. }
| Edit::SetBookmarkColor { .. }
| Edit::SetBookmarkStyle { .. }
| Edit::ImportBookmarks { .. }
| Edit::BookmarksFromStructure
| Edit::NumberPages { .. } => {
if p.assemble() {
@@ -48,7 +49,9 @@ pub(crate) fn check_permission(edit: &Edit, p: &pdfcraft_cos::Permissions) -> Re
}
Edit::AddMeasurement(_)
| Edit::AddAnnotation(_)
| Edit::PasteComments { .. }
| Edit::AddCustomStamp { .. }
| Edit::AddModel3d { .. }
| Edit::DeleteAnnotation { .. }
| Edit::SetAnnotationContents { .. }
| Edit::ReplyToAnnotation { .. }
@@ -62,7 +65,10 @@ pub(crate) fn check_permission(edit: &Edit, p: &pdfcraft_cos::Permissions) -> Re
| Edit::ResizeAnnotation { .. }
| Edit::StyleAnnotation { .. }
| Edit::FillAnnotation { .. }
| Edit::PartOpacity { .. }
| Edit::FontSizeAnnotation { .. }
| Edit::BorderColorAnnotation { .. }
| Edit::BorderShapeAnnotation { .. }
| Edit::SetAnnotationInfo { .. }
| Edit::SetMeasurementScale { .. } => {
if p.annotate() {
@@ -136,6 +142,7 @@ pub(crate) fn check_permission(edit: &Edit, p: &pdfcraft_cos::Permissions) -> Re
| Edit::MarkDecorative { .. }
| Edit::MarkTagArtifact { .. }
| Edit::AddOcrText { .. }
| Edit::EnhanceScan { .. }
| Edit::SetDocumentScript { .. }
| Edit::SetFieldScript { .. }
| Edit::RemoveStandardId { .. }
@@ -454,6 +461,9 @@ pub(crate) fn run_edit(doc: &mut pdfcraft_cos::Document, edit: &Edit, cx: &mut E
Edit::SetBookmarkPage { path, page } => pdfcraft_organize::set_bookmark_page(doc, path, *page)?,
Edit::SetBookmarkColor { path, color } => pdfcraft_organize::set_bookmark_color(doc, path, *color)?,
Edit::SetBookmarkStyle { path, bold, italic } => pdfcraft_organize::set_bookmark_style(doc, path, *bold, *italic)?,
Edit::ImportBookmarks { parent, index, bookmarks, replace } => {
pdfcraft_organize::import_outline(doc, parent, *index, bookmarks, *replace)?;
}
Edit::BookmarksFromStructure => {
let entries: Vec<_> = pdfcraft_a11y::headings(doc)
.into_iter()
@@ -489,6 +499,14 @@ pub(crate) fn run_edit(doc: &mut pdfcraft_cos::Document, edit: &Edit, cx: &mut E
Edit::AddAnnotation(a) => {
pdfcraft_annot::add_annotation(doc, a, &cx.meta())?;
}
Edit::PasteComments { xfdf, page, offset } => {
// Names from the edit's own id generator: unique to this paste within the document.
let prefix = cx.meta().id;
let report = pdfcraft_xfdf::import_comments(doc, xfdf, Some(*page), Some(&prefix), *offset)?;
if report.comments == 0 {
return Err(EditError::Invalid("the clipboard holds no comments".into()));
}
}
Edit::AddCustomStamp { page, rect, name, file, author } => {
let src = mark_source(doc, file)?;
let (sw, sh) = (src.size.0.max(1.0), src.size.1.max(1.0));
@@ -509,6 +527,16 @@ pub(crate) fn run_edit(doc: &mut pdfcraft_cos::Document, edit: &Edit, cx: &mut E
let index = pdfcraft_annot::add_annotation(doc, &new, &cx.meta())?;
pdfcraft_annot::orient_image_stamp(doc, *page, index, rotation)?;
}
Edit::AddModel3d { page, rect, model, poster, activation, name } => {
// A file that is not a model is refused before its poster is brought in.
pdfcraft_annot::check_model(model)?;
let poster = match poster {
Some(file) => Some(model_poster(doc, *rect, file)?),
None => None,
};
let new = pdfcraft_annot::NewModel3d { page: *page, rect: *rect, model, activation: *activation, poster, contents: name };
pdfcraft_annot::add_3d_model(doc, &new, &cx.meta())?;
}
Edit::DeleteAnnotation { page, index } => pdfcraft_annot::delete_annotation(doc, *page, *index)?,
Edit::SetAnnotationContents { page, index, text } => pdfcraft_annot::set_contents(doc, *page, *index, text, &cx.meta())?,
Edit::ReplyToAnnotation { page, index, text, author } => {
@@ -559,7 +587,10 @@ pub(crate) fn run_edit(doc: &mut pdfcraft_cos::Document, edit: &Edit, cx: &mut E
)?;
}
Edit::FillAnnotation { page, index, fill } => pdfcraft_annot::set_fill(doc, *page, *index, *fill, &cx.meta())?,
Edit::PartOpacity { page, index, fill, text } => pdfcraft_annot::set_part_opacity(doc, *page, *index, *fill, *text, &cx.meta())?,
Edit::FontSizeAnnotation { page, index, size } => pdfcraft_annot::set_font_size(doc, *page, *index, *size, &cx.meta())?,
Edit::BorderColorAnnotation { page, index, color } => pdfcraft_annot::set_border_color(doc, *page, *index, *color, &cx.meta())?,
Edit::BorderShapeAnnotation { page, index, shape } => pdfcraft_annot::set_border_shape(doc, *page, *index, *shape, &cx.meta())?,
Edit::SetAnnotationInfo { page, index, author, subject, icon } => {
pdfcraft_annot::set_info(doc, *page, *index, author.as_deref(), subject.as_deref(), *icon, &cx.meta())?;
}
@@ -689,6 +720,7 @@ pub(crate) fn run_edit(doc: &mut pdfcraft_cos::Document, edit: &Edit, cx: &mut E
Edit::AddOcrText { page, words } => {
pdfcraft_edit::stamp(doc, *page, "OCR", pdfcraft_ocr::text_layer(words))?;
}
Edit::EnhanceScan { images } => crate::scan_clean::write_back(doc, images)?,
Edit::EditPageImage { page, index, change } => run_edit_page_image(doc, *page, *index, change)?,
Edit::MoveObjects { page, objects, offset } => pdfcraft_edit::move_objects(doc, *page, objects, *offset)?,
Edit::EditTextBlock { page, block, text, style } => {
crates/engine/src/history.rs
+6
View File
@@ -53,7 +53,9 @@ pub(crate) fn scope_of(edit: &Edit) -> Scope {
Edit::AddAnnotation(a) if matches!(a.shape, Shape::Attachment { .. }) => Scope::Full,
Edit::AddMeasurement(_)
| Edit::AddAnnotation(_)
| Edit::PasteComments { .. }
| Edit::AddCustomStamp { .. }
| Edit::AddModel3d { .. }
| Edit::DeleteAnnotation { .. }
| Edit::SetAnnotationContents { .. }
| Edit::ReplyToAnnotation { .. }
@@ -67,7 +69,10 @@ pub(crate) fn scope_of(edit: &Edit) -> Scope {
| Edit::ResizeAnnotation { .. }
| Edit::StyleAnnotation { .. }
| Edit::FillAnnotation { .. }
| Edit::PartOpacity { .. }
| Edit::FontSizeAnnotation { .. }
| Edit::BorderColorAnnotation { .. }
| Edit::BorderShapeAnnotation { .. }
| Edit::SetAnnotationInfo { .. } => Scope::Comments,
Edit::AddLink { .. } | Edit::SetLink { .. } | Edit::DeleteLink { .. } | Edit::RemoveLinks { .. } | Edit::AddLinks { .. } => Scope::Links,
Edit::SetFieldValue { .. } | Edit::ResetForm { .. } | Edit::SetFieldImage { .. } | Edit::ApplyScriptChanges { .. } => Scope::Form,
@@ -77,6 +82,7 @@ pub(crate) fn scope_of(edit: &Edit) -> Scope {
| Edit::EditTextLine { .. }
| Edit::EditTextBlock { .. }
| Edit::EditPageImage { .. }
| Edit::EnhanceScan { .. }
| Edit::AddText { .. }
| Edit::AddImage { .. }
| Edit::UpdateContent { .. }
crates/engine/src/js.rs
+25 -6
View File
@@ -199,6 +199,24 @@ impl Scripts for JsRunner {
/// Run field `target`'s script for `event` over the form's fields, keeping what it showed and
/// asked for; a refusal carries its last message.
fn run(&mut self, event: FieldEvent, script: &str, target: &Field, value: &str, fields: &[Field]) -> ScriptResult {
// As in Acrobat, a failing script leaves the value alone (the error goes to the console).
self.run_logged(event, script, target, value, fields, true).unwrap_or_else(|| ScriptResult {
rc: true,
value: value.to_string(),
changes: Vec::new(),
message: None,
})
}
fn rerun_with_english_dates(&mut self, event: FieldEvent, script: &str, target: &Field, value: &str, fields: &[Field]) -> Option<ScriptResult> {
pdfcraft_js::with_english_printd(true, || self.run_logged(event, script, target, value, fields, false))
}
}
impl JsRunner {
/// [`Scripts::run`], keeping what the script showed and asked for only when `log`; `None`
/// when the script failed.
fn run_logged(&mut self, event: FieldEvent, script: &str, target: &Field, value: &str, fields: &[Field], log: bool) -> Option<ScriptResult> {
let states: Vec<_> = fields
.iter()
.map(|f| {
@@ -212,17 +230,18 @@ impl Scripts for JsRunner {
let states = std::sync::Arc::new(states);
let ev = pdfcraft_js::Event::field(event.name(), &target.name, value);
let o = pdfcraft_js::run_shared(script, &ev, &self.doc, std::sync::Arc::clone(&states), &self.doc_scripts, Limits::default());
self.output.absorb(&o);
if o.error.is_some() {
// As in Acrobat, a failing script leaves the value alone (the error goes to the console).
return ScriptResult { rc: true, value: value.to_string(), changes: Vec::new(), message: None };
if log {
self.output.absorb(&o);
}
ScriptResult {
if o.error.is_some() {
return None;
}
Some(ScriptResult {
rc: o.rc,
value: o.value.clone(),
changes: changes(&states, &o.changed),
message: (!o.rc).then(|| o.alerts.last().map(|m| m.text.clone())).flatten(),
}
})
}
}
crates/engine/src/lib.rs
+11 -4
View File
@@ -35,6 +35,7 @@ mod history;
pub mod js;
mod layers;
pub mod links;
pub mod model3d;
pub mod ocr;
mod open;
pub mod optimizer;
@@ -43,6 +44,7 @@ mod pictures;
pub mod print_raster;
pub mod printcheck;
mod protection;
pub mod scan_clean;
mod scan_colours;
mod scan_fields;
mod session;
@@ -71,7 +73,10 @@ pub use signature_scan::{PhotoRect, ScanError, SignaturePhoto, Turn};
pub use stamp_preview::{CommentLayer, StampPreview};
pub use text_preview::{TextEditPreview, UnmarkedPages};
pub use pdfcraft_organize::{BoxSpec, FieldConflicts, PageBox, SplitBy, Transition, TransitionSpec, mime_for_name, split_ranges};
pub use pdfcraft_organize::{
BoxSpec, FieldConflicts, MAX_IMPORT_BOOKMARKS, OutlineNode, OutlineTree, PageBox, SplitBy, Transition, TransitionSpec, mime_for_name,
split_ranges, unmapped as outline_unmapped, view_arity as outline_view_arity,
};
pub use pdfcraft_organize::LabelStyle;
pub use pdfcraft_organize::Standard;
@@ -108,9 +113,10 @@ pub use pdfcraft_fonts::{MAX_SIGNATURE_CHARS, ScriptOutline, script_missing, scr
pub use pdfcraft_annot::appearance as annot_text;
pub use pdfcraft_annot::links::{Highlight as LinkHighlight, LinkAction, LinkItem, LinkStyle};
pub use pdfcraft_annot::{
AppearanceCache, AttachIcon, FillMark, LineEnding, Markup, NOTE_SIZE, NewAnnotation, NoteIcon, OverlayFont, OverlayLook, Props as CommentProps,
ReviewState, Rgb, Shape, StampGroup, StampKind, Style, rect_quad,
Activation3d, AppearanceCache, AttachIcon, BorderShape, FillMark, LineEnding, Markup, NOTE_SIZE, NewAnnotation, NoteIcon, OverlayFont,
OverlayLook, Props as CommentProps, ReviewState, Rgb, Shape, StampGroup, StampKind, Style, rect_quad,
};
pub use pdfcraft_annot::{MAX_MODEL_BYTES as MAX_3D_MODEL_BYTES, ModelFormat as Model3dFormat};
pub use pdfcraft_forms::detect;
pub use pdfcraft_optimize as optimize;
pub use pdfcraft_preflight as pdfa;
@@ -123,6 +129,7 @@ pub use pdfcraft_redact::sanitize::{HIDDEN, Hidden};
pub use pdfcraft_sign as sign;
pub use pdfcraft_sign::{SignOptions, SignatureInfo, Status as SignatureStatus, TrustStore};
pub use pdfcraft_xfdf::Format as DataFormat;
pub use pdfcraft_xfdf::{Outline as CommentOutline, comment_count, comment_outlines, comments_bounds, is_xfdf_comments};
use bookmarks::outline_rows;
pub use bookmarks::{BookmarkPage, BookmarkRow};
@@ -140,7 +147,7 @@ use layers::{apply_layer_state, set_default_layer_state, use_layer_choices};
pub use open::{OpenContext, PreparedDocument};
pub use page_filter::{PageOrientation, PageParity, filter_pages};
pub use pictures::{MarkFile, image_dimensions};
use pictures::{embed_image, fit_picture, mark_source};
use pictures::{embed_image, fit_picture, mark_source, model_poster};
use protection::unrestricted;
pub use protection::{Changes, Printing, Protection, SecuritySummary};
pub use session::Session;
crates/engine/src/model3d.rs
+19
View File
@@ -0,0 +1,19 @@
//! The 3D models on a page (ISO 32000-2 §13.6), read so a renderer can draw them.
pub use pdfcraft_annot::{MAX_MODEL_BYTES, Model3d, Model3dKind};
use crate::Document;
impl Document {
/// The `/Annots` positions of the 3D annotations on 0-based `page`, in order.
pub fn model_3d_annotations(&self, page: usize) -> Result<Vec<usize>, String> {
let editor = self.editor.as_ref().ok_or("the document can't be read")?;
pdfcraft_annot::model_3d_annotations(&editor.cos, page).map_err(|e| e.to_string())
}
/// The decoded model of the 3D annotation at `/Annots` position `index` on 0-based `page`.
pub fn model_3d(&self, page: usize, index: usize) -> Result<Model3d, String> {
let editor = self.editor.as_ref().ok_or("the document can't be read")?;
pdfcraft_annot::model_3d(&editor.cos, page, index).map_err(|e| e.to_string())
}
}
crates/engine/src/pictures.rs
+12
View File
@@ -60,6 +60,18 @@ pub(crate) fn mark_source(doc: &mut pdfcraft_cos::Document, f: &MarkFile) -> Res
}
}
/// A 3D model's poster: the picture in `file` brought into `doc` and fitted inside `rect`
/// (user space), keeping its shape, and centred. It is drawn in user space and not counter-rotated
/// (unlike a stamp), so on a page with `/Rotate` it turns with the page, as the model does.
pub(crate) fn model_poster(doc: &mut pdfcraft_cos::Document, rect: [f64; 4], file: &MarkFile) -> Result<pdfcraft_annot::PosterPicture, EditError> {
let src = mark_source(doc, file)?;
let (sw, sh) = src.size;
if !(sw.is_finite() && sh.is_finite() && sw > 0.0 && sh > 0.0) {
return Err(EditError::Source(format!("{}: the poster has no area", file.name)));
}
Ok(pdfcraft_annot::PosterPicture { xobject: src.xobject, image: src.image, size: (sw, sh), at: fit_picture(rect, sw, sh) })
}
/// The dimensions of an image from its bytes (JPEG, PNG, TIFF), without decoding the full raster.
pub fn image_dimensions(bytes: &[u8]) -> Option<(u32, u32)> {
let format = image::guess_format(bytes).ok()?;
crates/engine/src/scan_clean.rs
+308
View File
@@ -0,0 +1,308 @@
//! Scan & OCR ▸ Enhance ▸ Scanned document: clean up the pictures scanned pages are made of.
//!
//! Unlike Enhance ▸ Camera image, the page is not rendered and replaced: each scan image is
//! decoded, cleaned ([`crate::camera_clean::clean_samples`]) and written back **into the same
//! image object**, at the same size and in the same colour space. So everything else on the
//! page stays as it was: an invisible OCR text layer, comments, links, form fields, other
//! content, and the image's own soft mask, `/Interpolate` and other entries. A JPEG stays a JPEG
//! (at quality 90), any other grey or RGB image is stored losslessly (Flate), and a black and
//! white (1-bit) image stays 1-bit. An image several pages draw is cleaned once.
//!
//! A page is a scan when its pictures cover at least half of it (as for Recognize text); the
//! pictures each covering at least a tenth of the page are its scan. What is not cleaned is
//! reported, never skipped silently: pages that are no scan, and scan images in a form PdfCraft
//! can't rewrite yet (CMYK, indexed, JPEG 2000, JBIG2, CCITT fax, a stencil mask, a remapping
//! decode array, a scan drawn inside grouped Form artwork) or whose data is broken.
//!
//! Like Optimize, the work is a job: [`Session::scan_clean_job`] captures the document,
//! [`ScanCleanJob::run`] cleans anywhere (a worker thread), and [`Session::apply_scan_clean`]
//! makes the result one undoable step, "Enhance scanned file" (the name the Scan & OCR tool has).
use pdfcraft_cos::{ObjRef, Object, Stream};
use crate::camera_clean::{CleanUp, MAX_LEVEL, clean_samples};
use crate::edit::check_permission;
use crate::{DocId, Edit, EditError, Session, guard, optimize};
pub use crate::camera_clean::DESPECKLE_FROM;
/// The JPEG quality a cleaned JPEG scan is written back at (as Enhance camera image).
const QUALITY: u8 = 90;
/// On a scanned page, an image covering at least this share of the page is part of the scan;
/// smaller ones (a logo or a signature pasted on) are left alone.
const PART_OF_SCAN: f64 = 0.1;
/// Why a scan drawn through grouped Form artwork stays as it is.
const IN_FORM: &str = "it is drawn inside grouped artwork (a form XObject), which can't be cleaned up yet";
/// How scanned pages are cleaned up.
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub struct ScanCleanSettings {
/// Background removal, 0 (off) … 100 (strongest), as Enhance camera image's clean-up level:
/// the paper's illumination is evened out (fully from 50), so a grey, yellowed or unevenly lit
/// background turns white; then the contrast is raised so faint strokes darken. Colour ink
/// stays colour.
pub level: u8,
/// Edge shadow removal: shadows along the scan's edges (a book's gutter, a copier's dark
/// border) are lightened to the paper next to them, line by line, before the level applies.
pub edge_shadows: bool,
/// Despeckle: isolated dark specks (dust, scanner noise), dark pixels with at most one dark
/// neighbour, turn white. The only change a black and white (1-bit) scan takes.
pub despeckle: bool,
}
impl ScanCleanSettings {
/// The level `level` (capped at 100) with edge shadow removal (unless the level is 0), and
/// despeckling from [`DESPECKLE_FROM`] as the camera tool does.
pub fn at(level: u8) -> Self {
let level = level.min(MAX_LEVEL);
Self { level, edge_shadows: level > 0, despeckle: level >= DESPECKLE_FROM }
}
/// Whether these settings change anything at all.
pub fn changes_something(self) -> bool {
self.level > 0 || self.edge_shadows || self.despeckle
}
}
impl Default for ScanCleanSettings {
fn default() -> Self {
Self::at(50)
}
}
/// What a clean-up did.
#[derive(Clone, Debug, Default, PartialEq)]
pub struct ScanCleanReport {
/// Pages (0-based) whose scan was cleaned, in order.
pub pages: Vec<usize>,
/// Images rewritten. One drawn on several pages counts once.
pub images: usize,
/// Pages left alone because no picture covers most of them.
pub not_scanned: Vec<usize>,
/// Scan images left as they were, and why.
pub skipped: Vec<SkippedImage>,
}
/// A scan image that was not cleaned.
#[derive(Clone, Debug, PartialEq)]
pub struct SkippedImage {
/// The pages (0-based) that draw it, among those asked for.
pub pages: Vec<usize>,
/// Its resource name on the first of them.
pub name: String,
/// Why, in a sentence that follows "skipped because".
pub reason: String,
}
/// One image to clean: its object, its name and the pages that draw it.
struct Target {
object: ObjRef,
name: String,
pages: Vec<usize>,
}
/// What cleaning scanned pages needs, independent of the session.
pub struct ScanCleanJob {
cos: pdfcraft_cos::Document,
targets: Vec<Target>,
settings: ScanCleanSettings,
/// The document's edit generation it was taken at: a result is for that state only.
generation: u64,
report: ScanCleanReport,
}
/// The cleaned images, ready for [`Session::apply_scan_clean`], and the report.
pub struct ScanCleaned {
images: Vec<(ObjRef, Stream)>,
generation: u64,
pub report: ScanCleanReport,
}
impl Session {
/// Capture what cleaning up the scans on `pages` (0-based; empty = all) of document `id`
/// needs. Errors when the document can't be changed (read-only, or its permissions forbid
/// changes) or a page doesn't exist.
pub fn scan_clean_job(&self, id: DocId, pages: &[usize], settings: ScanCleanSettings) -> Result<ScanCleanJob, EditError> {
let doc = self.get(id).ok_or(EditError::NoDocument)?;
if let Some(why) = &doc.read_only_reason {
return Err(EditError::ReadOnly(why.clone()));
}
let editor = doc.editor.as_ref().ok_or_else(|| EditError::ReadOnly(String::new()))?;
if let Some(p) = editor.cos.permissions() {
check_permission(&Edit::EnhanceScan { images: Vec::new() }, &p)?;
}
let all = doc.info.pages.len();
let mut wanted: Vec<usize> = if pages.is_empty() { (0..all).collect() } else { pages.to_vec() };
if let Some(p) = wanted.iter().find(|p| **p >= all) {
return Err(EditError::Invalid(format!("there is no page {}: the document has {all}", p.saturating_add(1))));
}
wanted.sort_unstable();
wanted.dedup();
let mut report = ScanCleanReport::default();
let mut targets: Vec<Target> = Vec::new();
// Form artwork already reported, and where in `report.skipped`.
let mut forms_reported: Vec<(ObjRef, usize)> = Vec::new();
for page in wanted {
let Some(info) = doc.info.pages.get(page) else { continue };
// Ordinary images and scan-holding Forms are classified together, so a page that
// mixes them gets its images cleaned and its Forms reported.
let Some((images, forms, own_scan)) = crate::scan_fields::scan_parts(doc, page, info) else {
report.not_scanned.push(page);
continue;
};
let forms: Vec<_> = forms.into_iter().filter(|(_, share)| *share >= PART_OF_SCAN).collect();
if !own_scan && forms.is_empty() {
report.not_scanned.push(page);
continue;
}
// A scan drawn through a Form XObject is a scan all the same: say why it stays.
for (form, _) in forms {
let known = form.object.and_then(|o| forms_reported.iter().find(|(r, _)| *r == o)).and_then(|(_, k)| report.skipped.get_mut(*k));
match known {
Some(s) if !s.pages.contains(&page) => s.pages.push(page),
Some(_) => {}
None => {
if let Some(o) = form.object {
forms_reported.push((o, report.skipped.len()));
}
report.skipped.push(SkippedImage { pages: vec![page], name: form.name, reason: IN_FORM.into() });
}
}
}
for (image, _) in images.into_iter().filter(|(_, share)| *share >= PART_OF_SCAN) {
let Some(object) = image.object else {
report.skipped.push(SkippedImage {
pages: vec![page],
name: image.name,
reason: "it is stored inside the page's resources rather than as an object of its own".into(),
});
continue;
};
match targets.iter_mut().find(|t| t.object == object) {
Some(t) if !t.pages.contains(&page) => t.pages.push(page),
Some(_) => {}
None => targets.push(Target { object, name: image.name, pages: vec![page] }),
}
}
}
Ok(ScanCleanJob { cos: editor.cos.clone(), targets, settings, generation: doc.edit_generation(), report })
}
/// Write the images [`ScanCleanJob::run`] cleaned into document `id` as one undoable step,
/// "Enhance scanned file" (none when nothing was cleaned), and return the report. Refused
/// when the document changed since the job was taken.
pub fn apply_scan_clean(&mut self, id: DocId, cleaned: ScanCleaned) -> Result<ScanCleanReport, EditError> {
let doc = self.get(id).ok_or(EditError::NoDocument)?;
if doc.edit_generation() != cleaned.generation {
return Err(EditError::Invalid("the document changed while its scans were being cleaned up; nothing was changed, try again".into()));
}
if !cleaned.images.is_empty() {
self.apply(id, Edit::EnhanceScan { images: cleaned.images })?;
}
Ok(cleaned.report)
}
/// Enhance ▸ Scanned document in one call: [`Self::scan_clean_job`], run, then
/// [`Self::apply_scan_clean`] (the CLI and agents use this).
pub fn enhance_scanned_pages(&mut self, id: DocId, pages: &[usize], settings: ScanCleanSettings) -> Result<ScanCleanReport, EditError> {
let job = self.scan_clean_job(id, pages, settings)?;
let cleaned = job.run(|_, _| true)?;
self.apply_scan_clean(id, cleaned)
}
}
impl ScanCleanJob {
/// The images to clean (one drawn on several pages counts once).
pub fn total(&self) -> usize {
self.targets.len()
}
/// Clean every scan image. `progress(done, total)` is called before each image; returning
/// `false` stops with [`EditError::Cancelled`]. A panic is returned as an error, so a worker
/// thread always reports back.
pub fn run(self, mut progress: impl FnMut(usize, usize) -> bool) -> Result<ScanCleaned, EditError> {
guard(move || self.work(&mut progress)).map_err(|m| EditError::Invalid(format!("cleaning up the scans failed unexpectedly ({m})")))?
}
fn work(self, progress: &mut dyn FnMut(usize, usize) -> bool) -> Result<ScanCleaned, EditError> {
let ScanCleanJob { cos, targets, settings, generation, mut report } = self;
let total = targets.len();
let mut images = Vec::new();
let mut changed_pages: Vec<usize> = Vec::new();
// One decoding allowance for the whole run, sized by the file.
let mut budget = pdfcraft_cos::ImageBudget::new(&cos);
for (done, target) in targets.into_iter().enumerate() {
if !progress(done, total) {
return Err(EditError::Cancelled);
}
let mut skip = |reason: String| report.skipped.push(SkippedImage { pages: target.pages.clone(), name: target.name.clone(), reason });
match clean_image(&cos, target.object, settings, &mut budget) {
Ok(Some(stream)) => {
images.push((target.object, stream));
changed_pages.extend(target.pages.iter().copied());
}
Ok(None) => skip("it is already clean: nothing changes at these settings".into()),
Err(reason) => skip(reason),
}
}
changed_pages.sort_unstable();
changed_pages.dedup();
report.images = images.len();
report.pages = changed_pages;
report.skipped.sort_by_key(|s| s.pages.first().copied());
Ok(ScanCleaned { images, generation, report })
}
}
/// Image `object` cleaned with `settings`: the new stream, `None` when nothing changes, or why it
/// can't be cleaned.
fn clean_image(
cos: &pdfcraft_cos::Document,
object: ObjRef,
settings: ScanCleanSettings,
budget: &mut pdfcraft_cos::ImageBudget,
) -> Result<Option<Stream>, String> {
let obj = cos.get(object);
let Object::Stream(s) = &*obj else { return Err("it is not an image".into()) };
if s.dict.name(b"Subtype") != Some(b"Image") {
return Err("it is not an image".into());
}
// A colour-key mask names sample values to leave unpainted: cleaned samples would match others.
if matches!(s.dict.get(b"Mask").map(|m| cos.resolve(m)).as_deref(), Some(Object::Array(_))) {
return Err("it has a colour-key mask, which cleaning would change".into());
}
let samples = optimize::decode_samples(cos, s, budget).map_err(|e| e.to_string())?;
let cleaned = if samples.bilevel {
if !settings.despeckle {
return Err("it is black and white, which only Despeckle changes, and Despeckle is off".into());
}
clean_samples(&samples.data, 1, samples.width, samples.height, CleanUp { level: 0, edge_shadows: false, despeckle: true })
} else {
if !settings.changes_something() {
return Ok(None);
}
let how = CleanUp { level: settings.level, edge_shadows: settings.edge_shadows, despeckle: settings.despeckle };
clean_samples(&samples.data, samples.components, samples.width, samples.height, how)
}
.ok_or("its samples don't match its size")?;
if cleaned == samples.data {
return Ok(None);
}
let cleaned = optimize::Samples { data: cleaned, ..samples };
optimize::encode_samples(s, &cleaned, QUALITY).map(Some).ok_or_else(|| "it could not be written back".into())
}
/// [`Edit::EnhanceScan`]: put the cleaned images in place. Each object must still be an image.
pub(crate) fn write_back(doc: &mut pdfcraft_cos::Document, images: &[(ObjRef, Stream)]) -> Result<(), EditError> {
for (object, stream) in images {
let is_image = matches!(&*doc.get(*object), Object::Stream(s) if s.dict.name(b"Subtype") == Some(b"Image"));
if !is_image {
return Err(EditError::Invalid(format!("object {} is no longer an image; the scan was not changed", object.num)));
}
doc.set(*object, Object::Stream(stream.clone()));
}
Ok(())
}
#[cfg(test)]
mod tests;
crates/engine/src/scan_clean/tests.rs
+463
View File
@@ -0,0 +1,463 @@
use std::sync::Arc;
use pdfcraft_testkit::Pdf;
use super::*;
const W: usize = 300;
const H: usize = 400;
/// Isolated specks on the paper, away from the text rows.
const SPECKS: [(usize, usize); 3] = [(150, 25), (200, 85), (120, 125)];
/// The light falling on the paper at `x`: full in the middle, darker towards a shadowed left
/// edge (as a book's gutter) and a little at the right edge.
fn light(x: usize) -> f32 {
let left = (x as f32 / 60.0).min(1.0);
let right = ((W - 1 - x) as f32 / 30.0).min(1.0);
(0.55 + 0.45 * left) * (0.8 + 0.2 * right)
}
fn is_stroke(x: usize, y: usize) -> bool {
(60..360).contains(&y) && (y - 60) % 40 < 6 && (60..240).contains(&x)
}
/// A synthetic scan, `n` components (1 grey, 3 RGB): greyish yellowed paper, a shadowed gutter,
/// dark text strokes and a few specks.
fn scan_pixels(n: usize) -> Vec<u8> {
let mut out = Vec::with_capacity(W * H * n);
for y in 0..H {
for x in 0..W {
let ink = if is_stroke(x, y) || SPECKS.contains(&(x, y)) { 0.12 } else { 0.8 };
let v = ink * light(x);
let px = [v, v, v * 0.85];
for c in px.iter().take(n) {
out.push((c * 255.0).round().clamp(0.0, 255.0) as u8);
}
}
}
out
}
fn jpeg(px: &[u8], n: usize) -> Vec<u8> {
let mut out = Vec::new();
let ty = if n == 1 { image::ExtendedColorType::L8 } else { image::ExtendedColorType::Rgb8 };
image::ImageEncoder::write_image(image::codecs::jpeg::JpegEncoder::new_with_quality(&mut out, 95), px, W as u32, H as u32, ty).unwrap();
out
}
/// Three pages: 1 a scan (the image `image_dict`/`data` as object 4) with an invisible text
/// layer and a comment, 2 a page of real text, 3 the same scan image drawn again.
fn scan_pdf(image_dict: &str, data: &[u8]) -> Vec<u8> {
let resources = "/Resources << /XObject << /Im0 4 0 R >> /Font << /F1 5 0 R >> >>";
Pdf::new()
.obj("<< /Type /Catalog /Pages 2 0 R >>")
.obj("<< /Type /Pages /Kids [3 0 R 8 0 R 9 0 R] /Count 3 >>")
.obj(format!("<< /Type /Page /Parent 2 0 R /MediaBox [0 0 {W} {H}] {resources} /Contents 6 0 R /Annots [7 0 R] >>"))
.stream(&format!("/Type /XObject /Subtype /Image {image_dict}"), data)
.obj("<< /Type /Font /Subtype /Type1 /BaseFont /Helvetica >>")
.stream("", format!("q {W} 0 0 {H} 0 0 cm /Im0 Do Q BT 3 Tr /F1 12 Tf 20 380 Td (Invisible OCR words) Tj ET"))
.obj("<< /Type /Annot /Subtype /Square /Rect [10 10 50 50] /C [1 0 0] /Contents (A comment) >>")
.obj(format!("<< /Type /Page /Parent 2 0 R /MediaBox [0 0 {W} {H}] {resources} /Contents 10 0 R >>"))
.obj(format!("<< /Type /Page /Parent 2 0 R /MediaBox [0 0 {W} {H}] {resources} /Contents 11 0 R >>"))
.stream("", "BT /F1 12 Tf 20 380 Td (Real text) Tj ET")
.stream("", format!("q {W} 0 0 {H} 0 0 cm /Im0 Do Q"))
.build()
}
fn grey_scan() -> Vec<u8> {
scan_pdf(&format!("/Width {W} /Height {H} /ColorSpace /DeviceGray /BitsPerComponent 8 /Interpolate true"), &scan_pixels(1))
}
fn open(bytes: Vec<u8>) -> (Session, DocId) {
let mut s = Session::new().with_clock(|| 1_700_000_000);
let id = s.open("scan.pdf", None, Arc::new(bytes), None).unwrap();
(s, id)
}
fn object(s: &Session, id: DocId, num: u32) -> Arc<Object> {
s.get(id).unwrap().editor.as_ref().unwrap().cos.get(ObjRef::new(num, 0))
}
/// The samples of the image page 1 draws, now.
fn samples(s: &Session, id: DocId) -> optimize::Samples {
let doc = s.get(id).unwrap();
let object = doc.page_images(0)[0].object.unwrap();
let cos = &doc.editor.as_ref().unwrap().cos;
let Object::Stream(img) = &*cos.get(object) else { panic!("the image") };
optimize::decode_samples(cos, img, &mut pdfcraft_cos::ImageBudget::with_work(u64::MAX)).unwrap()
}
fn grey_at(px: &optimize::Samples, x: usize, y: usize) -> u8 {
px.data[(y * W + x) * px.components]
}
#[test]
fn a_scan_gets_white_paper_even_edges_dark_strokes_and_no_specks_in_place() {
let (mut s, id) = open(grey_scan());
let before = samples(&s, id);
let kept: Vec<Arc<Object>> = [3, 5, 6, 7, 8, 10, 11].iter().map(|&n| object(&s, id, n)).collect();
let text_before = s.get(id).unwrap().text_lines(0);
let original = s.get(id).unwrap().bytes.clone();
let report = s.enhance_scanned_pages(id, &[], ScanCleanSettings::at(100)).unwrap();
assert_eq!(report.pages, vec![0, 2], "the shared image counts for both its pages");
assert_eq!(report.images, 1);
assert_eq!(report.not_scanned, vec![1], "the text page is no scan");
assert!(report.skipped.is_empty(), "{:?}", report.skipped);
assert_eq!(s.get(id).unwrap().can_undo(), Some("Enhance scanned file"));
let after = samples(&s, id);
assert_eq!((after.width, after.height, after.components, after.jpeg), (W as u32, H as u32, 1, false));
// Paper, also in the shadowed gutter and at the right edge, turns (nearly) white.
for (x, y) in [(150, 40), (150, 380), (3, 40), (3, 380), (30, 200), (W - 2, 40)] {
assert!(grey_at(&before, x, y) < 210, "the fixture's paper is grey at ({x}, {y})");
assert!(grey_at(&after, x, y) >= 235, "paper at ({x}, {y}): {} → {}", grey_at(&before, x, y), grey_at(&after, x, y));
}
// The edge is evened out: the gutter is as white as the middle.
assert!(grey_at(&after, 150, 40).abs_diff(grey_at(&after, 3, 40)) <= 12);
// Strokes stay dark (darker, with the contrast), also in the gutter's light.
for (x, y) in [(150, 62), (61, 102), (238, 342)] {
assert!(grey_at(&after, x, y) <= 40, "stroke at ({x}, {y}): {}", grey_at(&after, x, y));
}
for (x, y) in SPECKS {
assert!(grey_at(&before, x, y) < 60);
assert!(grey_at(&after, x, y) >= 200, "speck at ({x}, {y}): {}", grey_at(&after, x, y));
}
// Everything but the image is as it was: the page, its invisible text layer, the comment, the
// font, the other pages. The image keeps its entries.
let now: Vec<Arc<Object>> = [3, 5, 6, 7, 8, 10, 11].iter().map(|&n| object(&s, id, n)).collect();
assert_eq!(kept, now);
assert_eq!(s.get(id).unwrap().info.pages.len(), 3);
let text_after = s.get(id).unwrap().text_lines(0);
assert!(text_after.iter().any(|l| l.text.contains("Invisible OCR words")), "{text_after:?}");
assert_eq!(text_before.len(), text_after.len());
let Object::Stream(img) = &*object(&s, id, 4) else { panic!("the image") };
assert_eq!(img.dict.get(b"Interpolate"), Some(&Object::Bool(true)));
assert_eq!(img.dict.name(b"ColorSpace"), Some(&b"DeviceGray"[..]));
// Undo brings the original image, and the original file, back.
s.undo(id).unwrap();
assert_eq!(samples(&s, id), before);
assert_eq!(s.get(id).unwrap().bytes, original);
assert_eq!(s.get(id).unwrap().can_undo(), None);
}
#[test]
fn a_jpeg_scan_stays_a_jpeg_in_colour() {
let px = scan_pixels(3);
let (mut s, id) = open(scan_pdf(&format!("/Width {W} /Height {H} /ColorSpace /DeviceRGB /BitsPerComponent 8 /Filter /DCTDecode"), &jpeg(&px, 3)));
let report = s.enhance_scanned_pages(id, &[0], ScanCleanSettings::at(80)).unwrap();
// Only page 1 was asked for; the image it shares with page 3 is changed for both.
assert_eq!((report.pages.as_slice(), report.images), (&[0][..], 1));
let after = samples(&s, id);
assert!(after.jpeg && after.components == 3);
let Object::Stream(img) = &*object(&s, id, 4) else { panic!("the image") };
assert_eq!(img.dict.name(b"Filter"), Some(&b"DCTDecode"[..]));
// Yellowed paper becomes white; the strokes stay dark.
let at = |x: usize, y: usize| after.data[(y * W + x) * 3..][..3].to_vec();
for (x, y) in [(150, 40), (5, 300)] {
assert!(at(x, y).iter().all(|&v| v >= 230), "paper at ({x}, {y}): {:?}", at(x, y));
}
assert!(at(150, 63).iter().all(|&v| v <= 60), "stroke: {:?}", at(150, 63));
// The file still renders.
let doc = s.get(id).unwrap();
let shot = pdfcraft_render::PageRenderer::new(doc.bytes.clone(), doc.config.clone()).render(pdfcraft_render::RenderRequest {
page: 0,
scale: 1.0,
..Default::default()
});
assert!(shot.error.is_none(), "{:?}", shot.error);
}
#[test]
fn despeckle_is_its_own_switch() {
let speck = |settings: ScanCleanSettings| {
let (mut s, id) = open(grey_scan());
s.enhance_scanned_pages(id, &[0], settings).unwrap();
let after = samples(&s, id);
SPECKS.iter().map(|&(x, y)| grey_at(&after, x, y)).max().unwrap_or(0)
};
assert!(speck(ScanCleanSettings { despeckle: false, ..ScanCleanSettings::at(80) }) < 100, "kept without despeckle");
assert!(speck(ScanCleanSettings::at(80)) >= 200);
// Despeckle alone, at level 0, changes nothing but the specks.
let (mut s, id) = open(grey_scan());
let before = samples(&s, id);
s.enhance_scanned_pages(id, &[], ScanCleanSettings { level: 0, edge_shadows: false, despeckle: true }).unwrap();
let after = samples(&s, id);
let changed: Vec<usize> = before.data.iter().zip(&after.data).enumerate().filter(|(_, (a, b))| a != b).map(|(i, _)| i).collect();
let mut specks: Vec<usize> = SPECKS.iter().map(|&(x, y)| y * W + x).collect();
specks.sort_unstable();
assert_eq!(changed, specks);
// Nothing to do: no undo step, and the report says why.
let (mut s, id) = open(grey_scan());
let report = s.enhance_scanned_pages(id, &[], ScanCleanSettings { level: 0, edge_shadows: false, despeckle: false }).unwrap();
assert_eq!(report.images, 0);
assert!(report.pages.is_empty());
assert_eq!(report.skipped.len(), 1);
assert!(report.skipped[0].reason.contains("already clean"), "{:?}", report.skipped);
assert_eq!(s.get(id).unwrap().can_undo(), None);
assert!(!s.get(id).unwrap().dirty);
}
#[test]
fn a_black_and_white_scan_stays_one_bit_and_loses_its_specks() {
// 64 × 64, one bit per pixel, 1 = white: a black bar and two specks.
let (w, h) = (64usize, 64usize);
let mut bits = vec![0xFFu8; w / 8 * h];
let mut black = |x: usize, y: usize| bits[y * w / 8 + x / 8] &= !(0x80 >> (x % 8));
for y in 20..30 {
for x in 10..50 {
black(x, y);
}
}
black(5, 5);
black(58, 58);
let bytes = Pdf::new()
.obj("<< /Type /Catalog /Pages 2 0 R >>")
.obj("<< /Type /Pages /Kids [3 0 R] /Count 1 >>")
.obj("<< /Type /Page /Parent 2 0 R /MediaBox [0 0 64 64] /Resources << /XObject << /Im0 4 0 R >> >> /Contents 5 0 R >>")
.stream("/Type /XObject /Subtype /Image /Width 64 /Height 64 /ColorSpace /DeviceGray /BitsPerComponent 1", &bits)
.stream("", "q 64 0 0 64 0 0 cm /Im0 Do Q")
.build();
let (mut s, id) = open(bytes);
// Only Despeckle changes a black and white scan: without it the image is reported.
let report = s.enhance_scanned_pages(id, &[], ScanCleanSettings { despeckle: false, ..ScanCleanSettings::at(100) }).unwrap();
assert_eq!(report.images, 0);
assert!(report.skipped.first().is_some_and(|k| k.reason.contains("black and white")), "{:?}", report.skipped);
let report = s.enhance_scanned_pages(id, &[], ScanCleanSettings::at(100)).unwrap();
assert_eq!(report.images, 1);
let Object::Stream(img) = &*object(&s, id, 4) else { panic!("the image") };
assert_eq!(img.dict.int(b"BitsPerComponent"), Some(1));
let px = samples(&s, id);
assert!(px.bilevel);
let at = |x: usize, y: usize| px.data[y * w + x];
assert_eq!((at(5, 5), at(58, 58)), (255, 255), "specks gone");
assert_eq!((at(10, 20), at(30, 25), at(49, 29)), (0, 0, 0), "the bar stays");
assert_eq!(px.data.iter().filter(|&&v| v == 0).count(), 40 * 10);
}
#[test]
fn hostile_images_are_reported_not_a_panic() {
// Each page draws one image over the whole page.
let images: [(&str, Vec<u8>); 6] = [
("/Width 100000 /Height 100000 /ColorSpace /DeviceGray /BitsPerComponent 8", vec![0; 64]),
("/Width 300 /Height 400 /ColorSpace /DeviceGray /BitsPerComponent 8 /Filter /DCTDecode", vec![0xFF, 0xD8, 0xFF, 0xE0, 0, 3, 9]),
("/Width 300 /Height 400 /ColorSpace /DeviceGray /BitsPerComponent 8", vec![200; 1000]),
("/Width 1 /Height 1 /ColorSpace /DeviceGray /BitsPerComponent 8", vec![128]),
("/Width 8 /Height 8 /ColorSpace /DeviceCMYK /BitsPerComponent 8", vec![0; 256]),
("/Width 300 /Height 400 /ColorSpace /DeviceGray /BitsPerComponent 8 /Mask [250 255]", vec![200; 300 * 400]),
];
let n = images.len();
let kids: String = (0..n).map(|i| format!("{} 0 R ", 3 + 3 * i)).collect();
let mut pdf = Pdf::new().obj("<< /Type /Catalog /Pages 2 0 R >>").obj(format!("<< /Type /Pages /Kids [{kids}] /Count {n} >>"));
for (i, (dict, data)) in images.iter().enumerate() {
let (img, contents) = (4 + 3 * i, 5 + 3 * i);
pdf = pdf
.obj(format!(
"<< /Type /Page /Parent 2 0 R /MediaBox [0 0 300 400] /Resources << /XObject << /Im0 {img} 0 R >> >> /Contents {contents} 0 R >>"
))
.stream(&format!("/Type /XObject /Subtype /Image {dict}"), data)
.stream("", "q 300 0 0 400 0 0 cm /Im0 Do Q");
}
let (mut s, id) = open(pdf.build());
let report = s.enhance_scanned_pages(id, &[], ScanCleanSettings::at(100)).unwrap();
assert_eq!(report.images, 0);
assert_eq!(report.not_scanned, vec![3], "a 1 × 1 image is no scan");
let why: Vec<(usize, &str)> = report.skipped.iter().map(|k| (k.pages[0], k.reason.as_str())).collect();
assert_eq!(why.len(), 5, "{why:?}");
assert!(why[0].0 == 0 && why[0].1.contains("80 million"), "{why:?}");
assert!(why[1].0 == 1 && why[1].1.contains("could not be decoded"), "{why:?}");
assert!(why[2].0 == 2 && why[2].1.contains("shorter"), "{why:?}");
assert!(why[3].0 == 4 && why[3].1.contains("colour space"), "{why:?}");
assert!(why[4].0 == 5 && why[4].1.contains("colour-key mask"), "{why:?}");
assert_eq!(s.get(id).unwrap().can_undo(), None);
// Pages that don't exist are an error, not a panic.
assert!(matches!(s.enhance_scanned_pages(id, &[99], ScanCleanSettings::default()), Err(EditError::Invalid(_))));
}
#[test]
fn a_scan_drawn_inside_a_form_xobject_is_reported_not_called_no_scan() {
// Pages 1 and 3 draw the scan through Form Fm0 (as some scanners write it), page 2 is text.
let page = |contents: usize| {
format!(
"<< /Type /Page /Parent 2 0 R /MediaBox [0 0 {W} {H}] /Resources << /XObject << /Fm0 4 0 R >> /Font << /F1 9 0 R >> >> /Contents {contents} 0 R >>"
)
};
let bytes = Pdf::new()
.obj("<< /Type /Catalog /Pages 2 0 R >>")
.obj("<< /Type /Pages /Kids [3 0 R 7 0 R 10 0 R] /Count 3 >>")
.obj(page(6))
.stream(
&format!("/Type /XObject /Subtype /Form /BBox [0 0 {W} {H}] /Resources << /XObject << /Im0 5 0 R >> >>"),
format!("q {W} 0 0 {H} 0 0 cm /Im0 Do Q"),
)
.stream(&format!("/Type /XObject /Subtype /Image /Width {W} /Height {H} /ColorSpace /DeviceGray /BitsPerComponent 8"), scan_pixels(1))
.stream("", "/Fm0 Do")
.obj(page(8))
.stream("", "BT /F1 12 Tf 20 380 Td (Real text) Tj ET")
.obj("<< /Type /Font /Subtype /Type1 /BaseFont /Helvetica >>")
.obj(page(11))
.stream("", "q /Fm0 Do Q")
.build();
let (mut s, id) = open(bytes);
let before = object(&s, id, 5);
let report = s.enhance_scanned_pages(id, &[], ScanCleanSettings::at(100)).unwrap();
assert!(report.pages.is_empty() && report.images == 0, "{report:?}");
assert_eq!(report.not_scanned, vec![1], "only the text page is no scan");
assert_eq!(report.skipped.len(), 1, "the form is reported once: {report:?}");
let k = &report.skipped[0];
assert_eq!((k.pages.as_slice(), k.name.as_str()), (&[0, 2][..], "Fm0"));
assert!(k.reason.contains("form XObject"), "{k:?}");
assert_eq!(object(&s, id, 5), before, "the picture is left as it was");
assert_eq!(s.get(id).unwrap().can_undo(), None);
// A page of vector artwork in a Form, holding only a small logo, is no scan.
let logo = Pdf::new()
.obj("<< /Type /Catalog /Pages 2 0 R >>")
.obj("<< /Type /Pages /Kids [3 0 R] /Count 1 >>")
.obj(format!("<< /Type /Page /Parent 2 0 R /MediaBox [0 0 {W} {H}] /Resources << /XObject << /Fm0 4 0 R >> >> /Contents 6 0 R >>"))
.stream(
&format!("/Type /XObject /Subtype /Form /BBox [0 0 {W} {H}] /Resources << /XObject << /Im0 5 0 R >> >>"),
format!("0 0 {W} {H} re f q 40 0 0 20 10 370 cm /Im0 Do Q"),
)
.stream("/Type /XObject /Subtype /Image /Width 40 /Height 20 /ColorSpace /DeviceGray /BitsPerComponent 8", [128; 800])
.stream("", "/Fm0 Do")
.build();
let (mut s, id) = open(logo);
let report = s.enhance_scanned_pages(id, &[], ScanCleanSettings::at(100)).unwrap();
assert_eq!((report.not_scanned.as_slice(), report.skipped.len()), (&[0][..], 0), "{report:?}");
}
#[test]
fn a_page_mixing_its_own_image_and_a_scan_in_a_form_cleans_the_one_and_reports_the_other() {
// Page 1: its own image covers 40% and a Form holding a scan 20%: a scan only together.
// Page 2: its own image covers the page and the Form 20% more. Both pages clean Im1 and
// report Fm0.
let page = |contents: usize| {
format!(
"<< /Type /Page /Parent 2 0 R /MediaBox [0 0 {W} {H}] /Resources << /XObject << /Fm0 4 0 R /Im1 6 0 R >> >> /Contents {contents} 0 R >>"
)
};
let gray = format!("/Type /XObject /Subtype /Image /Width {W} /Height {H} /ColorSpace /DeviceGray /BitsPerComponent 8");
let bytes = Pdf::new()
.obj("<< /Type /Catalog /Pages 2 0 R >>")
.obj("<< /Type /Pages /Kids [3 0 R 8 0 R] /Count 2 >>")
.obj(page(7))
.stream(
&format!("/Type /XObject /Subtype /Form /BBox [0 0 {W} {H}] /Resources << /XObject << /Im0 5 0 R >> >>"),
format!("q {W} 0 0 {H} 0 0 cm /Im0 Do Q"),
)
.stream(&gray, scan_pixels(1))
.stream(&gray, scan_pixels(1))
.stream("", format!("q {W} 0 0 160 0 240 cm /Im1 Do Q q 1 0 0 0.2 0 0 cm /Fm0 Do Q"))
.obj(page(9))
.stream("", format!("q {W} 0 0 {H} 0 0 cm /Im1 Do Q q 1 0 0 0.2 0 0 cm /Fm0 Do Q"))
.build();
let (mut s, id) = open(bytes);
let (form_picture, own) = (object(&s, id, 5), object(&s, id, 6));
let report = s.enhance_scanned_pages(id, &[], ScanCleanSettings::at(100)).unwrap();
assert_eq!((report.pages.as_slice(), report.images), (&[0, 1][..], 1), "Im1 is cleaned on both pages: {report:?}");
assert!(report.not_scanned.is_empty(), "{report:?}");
assert_eq!(report.skipped.len(), 1, "the form is reported once: {report:?}");
let k = &report.skipped[0];
assert_eq!((k.pages.as_slice(), k.name.as_str()), (&[0, 1][..], "Fm0"));
assert!(k.reason.contains("form XObject"), "{k:?}");
assert_ne!(object(&s, id, 6), own, "the page's own image is cleaned");
assert_eq!(object(&s, id, 5), form_picture, "the picture in the form is left as it was");
}
#[test]
fn protected_documents_keep_their_encryption_and_restrictions() {
let protect = |user: &str, permissions: i32| {
let mut doc = pdfcraft_cos::Document::open(Arc::new(grey_scan())).unwrap();
doc.set_encryption(&pdfcraft_cos::NewEncryption {
algorithm: pdfcraft_cos::Algorithm::Aes256,
user_password: user,
owner_password: "owner",
permissions,
encrypt_metadata: true,
seed: [7; 32],
})
.unwrap();
Arc::new(pdfcraft_cos::write_full(&doc, &Default::default()).unwrap())
};
// Opened with the password: cleaned, saved still encrypted, and the clean scan reads back.
let mut s = Session::new();
let id = s.open("x.pdf", None, protect("pw", -1), Some("pw")).unwrap();
assert_eq!(s.enhance_scanned_pages(id, &[], ScanCleanSettings::at(100)).unwrap().images, 1);
let saved = s.save_bytes(id).unwrap();
assert!(pdfcraft_cos::Document::open(saved.clone()).is_err(), "still protected");
let id2 = s.open("y.pdf", None, saved, Some("pw")).unwrap();
assert!(grey_at(&samples(&s, id2), 3, 40) >= 235);
// Changes withheld: refused before any work.
let id = s.open("r.pdf", None, protect("", 0b0100), None).unwrap();
assert_eq!(s.enhance_scanned_pages(id, &[], ScanCleanSettings::default()).unwrap_err(), EditError::NotPermitted("changes to the document"));
assert!(!s.get(id).unwrap().dirty);
}
#[test]
fn a_result_for_an_older_state_is_refused() {
let (mut s, id) = open(grey_scan());
let job = s.scan_clean_job(id, &[], ScanCleanSettings::at(70)).unwrap();
assert_eq!(job.total(), 1);
let cleaned = job.run(|_, _| true).unwrap();
s.apply(id, Edit::RotatePages { pages: vec![1], degrees: 90 }).unwrap();
assert!(matches!(s.apply_scan_clean(id, cleaned), Err(EditError::Invalid(_))));
assert_eq!(s.get(id).unwrap().can_undo(), Some("Rotate page"));
// Cancelled from the progress callback.
let job = s.scan_clean_job(id, &[], ScanCleanSettings::at(70)).unwrap();
assert!(matches!(job.run(|_, _| false), Err(EditError::Cancelled)));
}
#[test]
fn edge_shadow_removal_whitens_the_gutter_background_removal_alone_leaves() {
let gutter = |edge_shadows: bool| {
let (mut s, id) = open(grey_scan());
s.enhance_scanned_pages(id, &[0], ScanCleanSettings { edge_shadows, ..ScanCleanSettings::at(100) }).unwrap();
let after = samples(&s, id);
[(3, 40), (3, 380), (10, 200)].map(|(x, y)| grey_at(&after, x, y))
};
assert!(gutter(false).iter().any(|&v| v < 200), "{:?}", gutter(false));
assert!(gutter(true).iter().all(|&v| v >= 235), "{:?}", gutter(true));
// On its own, at level 0, it lightens the gutter's paper to the paper beside it.
let (mut s, id) = open(grey_scan());
let before = samples(&s, id);
s.enhance_scanned_pages(id, &[0], ScanCleanSettings { level: 0, edge_shadows: true, despeckle: false }).unwrap();
let after = samples(&s, id);
assert!(grey_at(&after, 3, 40).abs_diff(grey_at(&before, 150, 40)) <= 4, "{} vs {}", grey_at(&after, 3, 40), grey_at(&before, 150, 40));
assert_eq!(grey_at(&after, 150, 40), grey_at(&before, 150, 40), "the middle is left alone");
}
/// Two pages, each with its own scan image (objects 4 and 7), so cleaning one page can be told
/// apart from cleaning the other.
fn two_scans_pdf() -> Vec<u8> {
let dict = format!("/Type /XObject /Subtype /Image /Width {W} /Height {H} /ColorSpace /DeviceGray /BitsPerComponent 8");
let draw = format!("q {W} 0 0 {H} 0 0 cm /Im0 Do Q");
Pdf::new()
.obj("<< /Type /Catalog /Pages 2 0 R >>")
.obj("<< /Type /Pages /Kids [3 0 R 6 0 R] /Count 2 >>")
.obj(format!("<< /Type /Page /Parent 2 0 R /MediaBox [0 0 {W} {H}] /Resources << /XObject << /Im0 4 0 R >> >> /Contents 5 0 R >>"))
.stream(&dict, scan_pixels(1))
.stream("", &draw)
.obj(format!("<< /Type /Page /Parent 2 0 R /MediaBox [0 0 {W} {H}] /Resources << /XObject << /Im0 7 0 R >> >> /Contents 8 0 R >>"))
.stream(&dict, scan_pixels(1))
.stream("", &draw)
.build()
}
#[test]
fn only_the_chosen_pages_scan_is_cleaned() {
// Clean page 1, then page 2 of a fresh copy: the other page's image stays byte-identical.
for (chosen, image, other) in [(0usize, 4u32, 7u32), (1, 7, 4)] {
let (mut s, id) = open(two_scans_pdf());
let (image_before, other_before) = (object(&s, id, image), object(&s, id, other));
assert_eq!(s.get(id).unwrap().page_images(chosen)[0].object, Some(ObjRef::new(image, 0)), "the fixture");
let report = s.enhance_scanned_pages(id, &[chosen], ScanCleanSettings::at(100)).unwrap();
assert_eq!((report.pages.as_slice(), report.images), (&[chosen][..], 1), "page {}", chosen + 1);
assert_ne!(object(&s, id, image), image_before, "page {}'s image was not cleaned", chosen + 1);
assert_eq!(object(&s, id, other), other_before, "cleaning page {} changed the other page's image", chosen + 1);
let (Object::Stream(was), Object::Stream(now)) = (&*other_before, &*object(&s, id, other)) else { panic!("the images") };
assert_eq!(was.raw, now.raw, "the other image's stream bytes");
assert_eq!(s.get(id).unwrap().can_undo(), Some("Enhance scanned file"));
}
}
crates/engine/src/scan_fields.rs
+80 -16
View File
@@ -19,30 +19,94 @@ const SCALE: f32 = 2.0;
const MAX_SIDE: f32 = 3000.0;
/// A page is a scan when its images cover at least this share of it.
const COVERED: f64 = 0.5;
/// A picture inside Form artwork is taken for a scan when it has at least this resolution over
/// the area the artwork covers (a quarter of the pixels a 72 dpi scan would have).
const MIN_FORM_DPI: f64 = 36.0;
/// Recognised "words" less tall than this (points) are specks read as letters, and would make
/// the box they sit on look filled in.
const MIN_WORD_HEIGHT: f64 = 4.0;
/// Whether `page` is a picture of a page: images cover most of it.
pub(crate) fn is_scan(doc: &Document, page: usize, info: &PageInfo) -> bool {
let [cx0, cy0, cx1, cy1] = info.crop.map(f64::from);
let page_area = (cx1 - cx0) * (cy1 - cy0);
if !page_area.is_finite() || page_area <= 0.0 {
return false;
}
let covered: f64 = doc
scan_images(doc, page, info).is_some()
}
/// The raster images (larger than one pixel) `page` draws, each with the share of the page (its
/// crop box) it covers, when together they cover most of it ([`is_scan`]); `None` otherwise.
/// Enhance ▸ Scanned document cleans these up ([`crate::scan_clean`]).
pub(crate) fn scan_images(doc: &Document, page: usize, info: &PageInfo) -> Option<Vec<(pdfcraft_edit::PageImage, f64)>> {
let area = page_area(info)?;
let images: Vec<(pdfcraft_edit::PageImage, f64)> = doc
.page_images(page)
.iter()
.into_iter()
.filter(|i| !i.is_form && i.width > 1 && i.height > 1)
.map(|i| {
let r = i.visible_rect.unwrap_or(i.rect);
let w = (r[2].min(cx1) - r[0].max(cx0)).max(0.0);
let h = (r[3].min(cy1) - r[1].max(cy0)).max(0.0);
w * h
})
.filter(|a| a.is_finite())
.sum();
covered >= COVERED * page_area
.filter_map(|i| Some((share(info, area, &i)?, i)))
.map(|(share, i)| (i, share))
.collect();
let covered: f64 = images.iter().map(|(_, share)| share).sum();
(covered >= COVERED).then_some(images)
}
/// A page that is a scan, classified in one pass: its raster images (larger than one pixel) and
/// the grouped Form artwork that holds a raster picture, each with the share of the page it
/// covers, when together they cover most of the page; `None` otherwise. The flag says whether the
/// raster images alone do ([`scan_images`]). Enhance ▸ Scanned document cleans the images and
/// reports the Forms, which it can't clean yet, so a page mixing both gets both.
#[allow(clippy::type_complexity)]
pub(crate) fn scan_parts(
doc: &Document,
page: usize,
info: &PageInfo,
) -> Option<(Vec<(pdfcraft_edit::PageImage, f64)>, Vec<(pdfcraft_edit::PageImage, f64)>, bool)> {
let area = page_area(info)?;
let editor = doc.editor.as_ref();
let (mut images, mut forms) = (Vec::new(), Vec::new());
let (mut own, mut covered) = (0.0, 0.0);
for i in doc.page_images(page) {
let Some(share) = share(info, area, &i) else { continue };
if !i.is_form {
if i.width > 1 && i.height > 1 {
own += share;
covered += share;
images.push((i, share));
}
} else if let (Some(form), Some(editor)) = (i.object, editor)
&& holds_scan(&editor.cos, form, share * area)
{
covered += share;
forms.push((i, share));
}
}
(covered >= COVERED).then_some((images, forms, own >= COVERED))
}
/// Whether Form `form`, covering `points` square points of the page, draws a picture that could
/// be a scan of it: one with at least [`MIN_FORM_DPI`] over that area, not a logo inside a page
/// of vector artwork.
fn holds_scan(cos: &pdfcraft_cos::Document, form: pdfcraft_cos::ObjRef, points: f64) -> bool {
let Some(picture) = crate::copy_image::picture_in_form(cos, form) else { return false };
let obj = cos.get(picture);
let pdfcraft_cos::Object::Stream(s) = &*obj else { return false };
let px = |k: &[u8]| s.dict.int(k).unwrap_or(0).max(0) as f64;
let per_point = MIN_FORM_DPI / 72.0;
px(b"Width") * px(b"Height") >= points * per_point * per_point
}
/// The area of `info`'s crop box, when it is a usable one.
fn page_area(info: &PageInfo) -> Option<f64> {
let [cx0, cy0, cx1, cy1] = info.crop.map(f64::from);
let area = (cx1 - cx0) * (cy1 - cy0);
(area.is_finite() && area > 0.0).then_some(area)
}
/// The share of the crop box (of area `area`) that `image` covers, when it is a number.
fn share(info: &PageInfo, area: f64, image: &pdfcraft_edit::PageImage) -> Option<f64> {
let [cx0, cy0, cx1, cy1] = info.crop.map(f64::from);
let r = image.visible_rect.unwrap_or(image.rect);
let w = (r[2].min(cx1) - r[0].max(cx0)).max(0.0);
let h = (r[3].min(cy1) - r[1].max(cy0)).max(0.0);
let share = w * h / area;
share.is_finite().then_some(share)
}
/// The page as displayed, in points with y up: where detection on a scan works.
crates/engine/src/session/files.rs
+13
View File
@@ -114,6 +114,19 @@ impl Session {
.map_err(EditError::Write)
}
/// XFDF with these comments (0-based page, position in the page's `/Annots`) and their
/// replies: what Copy puts on the clipboard; [`Edit::PasteComments`] takes it back in.
pub fn export_comments(&self, id: DocId, which: &[(usize, usize)]) -> Result<String, EditError> {
let doc = self.get(id).ok_or(EditError::NoDocument)?;
let cos = match doc.editor.as_ref() {
Some(e) => e.cos.clone(),
None => {
pdfcraft_cos::Document::open_with_password(doc.bytes.clone(), doc.password.as_deref()).map_err(|e| EditError::Write(e.to_string()))?
}
};
guard(|| pdfcraft_xfdf::export_xfdf_comments(&cos, which, &doc.name)).map_err(EditError::Write)
}
/// The print-ready PDF for `settings` (sheets laid out for the paper; see `pdfcraft-print`).
pub fn print_pdf(&self, id: DocId, settings: &print::Settings) -> Result<Vec<u8>, EditError> {
let doc = self.get(id).ok_or(EditError::NoDocument)?;
crates/engine/src/session/view.rs
+8
View File
@@ -168,6 +168,14 @@ impl Session {
Some(BookmarkPage { rows, next: (end < all.len()).then_some(end), truncated: doc.info.outline_more })
}
/// The bookmark at `path` with everything under it, or the whole outline for `[]`, as a tree
/// another document can take ([`crate::Edit::ImportBookmarks`]). `None` when no document has
/// `id`, or its structure could not be read for editing.
pub fn outline_tree(&self, id: DocId, path: &[usize]) -> Option<Result<pdfcraft_organize::OutlineTree, pdfcraft_organize::OutlineError>> {
let editor = self.get(id)?.editor.as_ref()?;
Some(pdfcraft_organize::outline_tree(&editor.cos, path))
}
/// Top-level bookmarks as split points: (first page of each part, its bookmark's title).
pub fn bookmark_splits(&self, id: DocId) -> Vec<(usize, String)> {
let Some(doc) = self.get(id) else { return Vec::new() };
crates/engine/src/signature_image.rs
+7 -1
View File
@@ -55,7 +55,10 @@ impl Document {
if w <= 0 || h <= 0 || w > i64::from(MAX_SIDE) || h > i64::from(MAX_SIDE) || w.saturating_mul(h) > MAX_PIXELS as i64 {
return Err(SignatureImageError::PlacedSize.to_string());
}
let (_, bytes) = pdfcraft_create::image_file(&editor.cos, picture)?;
// Decoded within an allowance sized for the small picture a signature is, not the
// document's: a stream that inflates past it is refused before it is held whole.
let (_, bytes) =
pdfcraft_create::image_file_within(&editor.cos, picture, &mut pdfcraft_cos::ImageBudget::with_work(SIGNATURE_IMAGE_WORK))?;
let image = SignatureImage::from_bytes(&bytes).map_err(|e| e.to_string())?;
let opacity =
pdfcraft_annot::props(&editor.cos, page, index).map(|p| p.opacity).filter(|v| v.is_finite()).unwrap_or(1.0).clamp(0.0, 1.0) as f32;
@@ -90,6 +93,9 @@ pub fn upright_box(rotation: i64, at: [f64; 2], width: f64, height: f64) -> Opti
/// Saved signatures live in settings, so both source and normalized PNG are bounded.
pub const MAX_SIGNATURE_IMAGE_BYTES: usize = 4 << 20;
const MAX_PIXELS: u64 = 4 << 20;
/// Steps of decoding the signature's picture may take: every stage of its filters, for a
/// picture of at most [`MAX_PIXELS`] samples, fits in it many times over.
const SIGNATURE_IMAGE_WORK: u64 = 1 << 28;
const MAX_SIDE: u32 = 4096;
/// Larger sources (scans, phone photos) are decoded within these bounds, then scaled down.
const MAX_SOURCE_SIDE: u32 = 8192;
crates/engine/src/tests.rs
+249 -1
View File
@@ -549,6 +549,85 @@ fn pdf_page_stamps_fit_inside_an_explicit_box() {
assert!(content.contains("/Pic Do") && !content.contains("BI"), "{content}");
}
/// #1297: a 3D model on a page keeps the whole box as its rectangle and draws a PDF poster as a
/// form XObject fitted inside it, keeping its shape and centred.
#[test]
fn a_3d_model_gets_a_fitted_vector_poster_and_a_file_that_is_not_one_is_refused() {
let mut session = Session::new().with_clock(|| 1_700_000_000);
let blank = session.create_blank(595.0, 842.0, 1).unwrap();
let id = session.open_new("model.pdf", blank).unwrap();
let mut prc = b"PRC".to_vec();
prc.extend_from_slice(&[1; 64]);
let add = |model: Vec<u8>| Edit::AddModel3d {
page: 0,
rect: [56.0, 400.0, 539.0, 700.0],
model: Arc::new(model),
poster: Some(MarkFile { name: "poster.pdf".into(), bytes: Arc::new(half_red_page()), page: 0 }),
activation: Activation3d::default(),
name: "Bracket".into(),
};
assert!(session.apply(id, add(b"%PDF-1.7 not a model".to_vec())).is_err());
session.apply(id, add(prc)).unwrap();
let doc = pdfcraft_cos::Document::open(session.save_bytes(id).unwrap()).unwrap();
let page = &pdfcraft_model::pages(&doc)[0];
let annots = doc.resolve(page.dict.get(b"Annots").unwrap());
let [only] = annots.as_array().unwrap().as_slice() else { panic!("the refused model added nothing: {annots:?}") };
let annot = doc.get(only.as_ref().unwrap());
let annot = annot.as_dict().unwrap();
assert_eq!(annot.name(b"Subtype"), Some(&b"3D"[..]));
let rect: Vec<f64> = annot.get(b"Rect").unwrap().as_array().unwrap().iter().filter_map(pdfcraft_cos::Object::as_f64).collect();
assert_eq!(rect, [56.0, 400.0, 539.0, 700.0], "the whole box, not the fitted picture");
assert_eq!(doc.resolve(annot.get(b"3DA").unwrap()).as_dict().unwrap().name(b"A"), Some(&b"XA"[..]));
let ap = doc.resolve(annot.get(b"AP").unwrap());
let pdfcraft_cos::Object::Stream(normal) = doc.get(ap.as_dict().unwrap().reference(b"N").unwrap()).as_ref().clone() else {
panic!("an appearance stream")
};
let pic = doc.resolve(normal.dict.get(b"Resources").unwrap()).as_dict().unwrap().clone();
let pic = doc.resolve(pic.get(b"XObject").unwrap()).as_dict().unwrap().reference(b"Pic").unwrap();
let pdfcraft_cos::Object::Stream(picture) = doc.get(pic).as_ref().clone() else { panic!("the picture is a stream") };
assert_eq!(picture.dict.name(b"Subtype"), Some(&b"Form"[..]), "the page is kept as a form XObject");
// The 80 × 40 page in the 483 × 300 box: 483 × 241.5 from y 429.25.
let content = String::from_utf8(normal.decoded().unwrap().to_vec()).unwrap();
assert!(content.contains(" 56 429.25 cm /Pic Do"), "{content}");
}
/// A PDF poster whose page has no area is refused rather than drawn as nothing.
#[test]
fn a_3d_poster_page_without_area_is_refused() {
let mut session = Session::new().with_clock(|| 1_700_000_000);
let blank = session.create_blank(595.0, 842.0, 1).unwrap();
let id = session.open_new("model.pdf", blank).unwrap();
let flat = String::from_utf8(half_red_page()).unwrap().replace("[0 0 80 40]", "[0 0 80 0 ]");
let add = Edit::AddModel3d {
page: 0,
rect: [56.0, 400.0, 539.0, 700.0],
model: Arc::new(b"PRC rest".to_vec()),
poster: Some(MarkFile { name: "flat.pdf".into(), bytes: Arc::new(flat.into_bytes()), page: 0 }),
activation: Activation3d::default(),
name: "Bracket".into(),
};
assert!(matches!(session.apply(id, add), Err(EditError::Source(_))));
assert!(session.get(id).unwrap().info.annotations.is_empty(), "nothing was added");
}
/// Adding a 3D model is a comment change: refused without the comments permission, and only the
/// comments are rebuilt after it.
#[test]
fn a_3d_model_needs_the_comments_permission_and_changes_the_comments() {
let edit = Edit::AddModel3d {
page: 0,
rect: [0.0, 0.0, 100.0, 100.0],
model: Arc::new(b"PRC rest".to_vec()),
poster: None,
activation: Activation3d::default(),
name: String::new(),
};
let p = pdfcraft_cos::Permissions { bits: 0, owner: false };
assert_eq!(check_permission(&edit, &p), Err(EditError::NotPermitted("comments")));
assert_eq!(crate::history::scope_of(&edit), crate::history::Scope::Comments);
}
fn page_texts(s: &Session, id: DocId) -> Vec<String> {
let doc = s.get(id).unwrap();
let config = pdfcraft_render::RenderConfig { password: doc.password.as_deref().map(Arc::from), ..Default::default() };
@@ -2132,7 +2211,7 @@ fn rect_comment(page: usize, rect: [f64; 4]) -> Edit {
Edit::AddAnnotation(NewAnnotation {
page,
shape: Shape::Rectangle { rect },
style: Style { color: [1.0, 0.0, 0.0], opacity: 1.0, width: 2.0, fill: Some([1.0, 0.0, 0.0]) },
style: Style { color: [1.0, 0.0, 0.0], opacity: 1.0, width: 2.0, fill: Some([1.0, 0.0, 0.0]), ..Style::default() },
contents: "Look here".into(),
author: "Reviewer".into(),
})
@@ -10106,6 +10185,42 @@ fn need_appearances_buttons_are_displayed_from_their_captions() {
}
}
/// Copy and paste: the copies of a comment and its reply land on the asked-for page of another
/// document under names of their own (nothing replaced; a second paste gives a second set),
/// undo takes a paste away as one step, and a clipboard without comments is refused.
#[test]
fn pasting_copied_comments_adds_copies_that_undo_as_one_step() {
let (mut s, id) = session_with(2);
s.apply(id, rect_comment(1, [80.0, 80.0, 120.0, 120.0])).unwrap();
s.apply(id, Edit::ReplyToAnnotation { page: 1, index: 0, text: "Done".into(), author: "Ada".into() }).unwrap();
let xfdf = s.export_comments(id, &[(1, 0)]).unwrap();
assert!(xfdf.contains("<square") && xfdf.contains("inreplyto="), "{xfdf}");
let other = s.open("other.pdf", None, Arc::new(fixture(2)), None).expect("opens");
s.apply(other, Edit::PasteComments { xfdf: xfdf.clone(), page: 0, offset: None }).unwrap();
let a = s.get(other).unwrap().info.annotations.clone();
assert_eq!(a.len(), 2, "{a:?}");
assert!(a.iter().all(|x| x.page == 0), "on the asked-for page: {a:?}");
let parent = a.iter().find(|x| x.in_reply_to.is_none()).unwrap();
let reply = a.iter().find(|x| x.in_reply_to.is_some()).unwrap();
assert_eq!(reply.in_reply_to, parent.name, "the reply follows its copied parent");
assert_ne!(parent.name, s.get(id).unwrap().info.annotations[0].name, "a name of its own");
assert_eq!(parent.contents.as_deref(), Some("Look here"));
assert_eq!(s.get(other).unwrap().can_undo(), Some("Paste comments"));
s.apply(other, Edit::PasteComments { xfdf: xfdf.clone(), page: 1, offset: Some([30.0, -10.0]) }).unwrap();
let a = s.get(other).unwrap().info.annotations.clone();
assert_eq!(a.len(), 4, "a second set");
let placed = a.iter().find(|x| x.page == 1 && x.in_reply_to.is_none()).unwrap();
assert!((placed.rect[0] - 110.0).abs() < 0.5 && (placed.rect[1] - 70.0).abs() < 0.5, "moved by the offset: {:?}", placed.rect);
assert_eq!(comment_count(&xfdf), 1);
assert_eq!(comments_bounds(&xfdf).map(|b| b.map(|v| v.round())), Some([80.0, 80.0, 120.0, 120.0]));
s.undo(other).unwrap();
assert_eq!(s.get(other).unwrap().info.annotations.len(), 2);
assert!(s.apply(other, Edit::PasteComments { xfdf: "hello".into(), page: 0, offset: None }).is_err(), "no comments in it");
assert!(s.apply(other, Edit::PasteComments { xfdf, page: 9, offset: None }).is_err(), "no such page");
assert_eq!(s.get(other).unwrap().info.annotations.len(), 2, "a refused paste changes nothing");
assert_eq!(s.get(id).unwrap().info.annotations.len(), 2, "the source is untouched");
}
#[test]
fn repeated_backdrop_artwork_survives_background_undo_redo_and_both_save_modes() {
let white = "1 g";
@@ -10252,6 +10367,107 @@ fn image_transform_engine_refuses_singular_matrices_without_mutation() {
}
}
/// Optimize has JBIG2 images share a symbol dictionary only where redaction can store each again
/// without it. Of four pages of the same scan, the two in DeviceGray and the image mask share one;
/// the one in a DeviceN of one colourant, which redaction doesn't store again, is coded alone.
/// Redacting the first page then stores the second and the mask again without the dictionary,
/// which leaves the file.
#[test]
fn redacting_an_optimized_jbig2_page_stores_the_rest_of_its_group_again() {
use pdfcraft_cos::{Dict, Document, ObjRef, Object, SaveOptions, Stream, write_full};
let (w, h) = (600usize, 480usize);
// Lines of 26 letters, each 32 rows of a run from a stem down its left side, of 2 to 24 pixels,
// the same to the pixel wherever it is, picked and spaced at random.
let mut seed = 0x2545_F491_4F6C_DD1Du64;
let mut next = move || {
seed ^= seed << 13;
seed ^= seed >> 7;
seed ^= seed << 17;
seed as usize
};
let letters: Vec<Vec<usize>> = (0..26).map(|_| (0..16).map(|_| 1 + next() % 12).collect()).collect();
let stride = w.div_ceil(8);
let mut bits = vec![0xFFu8; stride * h];
for top in (8..h - 40).step_by(44) {
let mut x = 8 + next() % 8;
while x + 32 < w {
let letter = &letters[next() % 26];
for (row, run) in letter.iter().enumerate() {
for (px, py) in (x..x + 2 * run).flat_map(|px| [(px, top + 2 * row), (px, top + 2 * row + 1)]) {
bits[py * stride + px / 8] &= !(0x80 >> (px % 8));
}
}
x += 2 * letter.iter().max().unwrap_or(&1) + 4 + next() % 5;
}
}
let mut doc = Document::new_empty();
let pages = doc.root().and_then(|r| doc.get(r).as_dict().and_then(|d| d.reference(b"Pages"))).unwrap();
let device_n = Object::Array(vec![Object::name("DeviceN"), Object::Array(vec![Object::name("Black")]), Object::name("DeviceGray"), {
let mut tint = Dict::new();
tint.set(b"FunctionType".to_vec(), Object::Int(2));
tint.set(b"Domain".to_vec(), Object::Array(vec![Object::Int(0), Object::Int(1)]));
tint.set(b"N".to_vec(), Object::Int(1));
Object::Dict(tint)
}]);
let mut images: Vec<ObjRef> = Vec::new();
let grey = || ("ColorSpace", Object::name("DeviceGray"));
for (k, (key, value)) in [grey(), grey(), ("ColorSpace", device_n), ("ImageMask", Object::Bool(true))].into_iter().enumerate() {
let mut d = Dict::new();
d.set(b"Type".to_vec(), Object::name("XObject"));
d.set(b"Subtype".to_vec(), Object::name("Image"));
d.set(b"Width".to_vec(), Object::Int(w as i64));
d.set(b"Height".to_vec(), Object::Int(h as i64));
d.set(b"BitsPerComponent".to_vec(), Object::Int(1));
d.set(key.as_bytes().to_vec(), value);
let image = doc.add(Object::Stream(Stream::flate(d, &bits)));
images.push(image);
let content = doc.add(Object::Stream(Stream::from_raw(Dict::new(), b"q 144 0 0 115.2 36 600 cm /Im Do Q".to_vec())));
let mut xobjects = Dict::new();
xobjects.set(b"Im".to_vec(), Object::Ref(image));
let mut resources = Dict::new();
resources.set(b"XObject".to_vec(), Object::Dict(xobjects));
let mut page = Dict::new();
page.set(b"Type".to_vec(), Object::name("Page"));
page.set(b"Parent".to_vec(), Object::Ref(pages));
page.set(b"MediaBox".to_vec(), Object::Array(vec![0.into(), 0.into(), 612.into(), 792.into()]));
page.set(b"Resources".to_vec(), Object::Dict(resources));
page.set(b"Contents".to_vec(), Object::Ref(content));
if k == 0 {
let mut mark = Dict::new();
mark.set(b"Type".to_vec(), Object::name("Annot"));
mark.set(b"Subtype".to_vec(), Object::name("Redact"));
mark.set(b"Rect".to_vec(), Object::Array(vec![40.into(), 610.into(), 100.into(), 650.into()]));
let mark = doc.add(Object::Dict(mark));
page.set(b"Annots".to_vec(), Object::Array(vec![Object::Ref(mark)]));
}
let page = doc.add(Object::Dict(page));
doc.update_dict(pages, |d| {
let mut kids = d.get(b"Kids").and_then(|k| k.as_array().cloned()).unwrap_or_default();
kids.push(Object::Ref(page));
d.set(b"Count".to_vec(), Object::Int(kids.len() as i64));
d.set(b"Kids".to_vec(), Object::Array(kids));
})
.unwrap();
}
let mono = pdfcraft_optimize::MonoSettings {
compression: pdfcraft_optimize::MonoCompression::Jbig2,
downsample: false,
..pdfcraft_optimize::MonoSettings::REDUCE
};
pdfcraft_optimize::optimize(&mut doc, &pdfcraft_optimize::Settings { mono, ..Default::default() }).unwrap();
let globals = |doc: &Document, image: ObjRef| match &*doc.get(image) {
Object::Stream(s) => s.dict.get(b"DecodeParms").and_then(Object::as_dict).and_then(|d| d.reference(b"JBIG2Globals")),
_ => None,
};
let shared = globals(&doc, images[0]);
assert!(shared.is_some() && globals(&doc, images[1]) == shared && globals(&doc, images[3]) == shared, "DeviceGray and the mask share");
assert_eq!(globals(&doc, images[2]), None, "the DeviceN page is coded alone");
pdfcraft_redact::apply(&mut doc, None).unwrap();
let saved = Document::open(Arc::new(write_full(&doc, &SaveOptions::default()).unwrap())).unwrap();
let named = saved.scan_objects().filter(|(_, o)| matches!(&**o, Object::Stream(s) if saved.jbig2_globals_ref(s).is_some())).count();
assert_eq!(named, 0, "no image names the dictionary, which left with them");
}
#[test]
fn fill_sign_multiline_text_survives_engine_save_reopen_and_render() {
let (mut session, id) = rotated_blank(0);
@@ -10268,3 +10484,35 @@ fn fill_sign_multiline_text_survives_engine_save_reopen_and_render() {
assert_eq!(reopened.get(again).unwrap().info.annotations[0].contents.as_deref(), Some(text), "reopening preserves all blank lines");
assert_eq!(before.rgba, render_page(&reopened, again).rgba, "the saved appearance is identical");
}
/// A whole tree goes in as one undoable step, and a refused tree changes nothing.
#[test]
fn imported_bookmarks_undo_in_one_step() {
let (mut s, id) = session_with(2);
let node = |title: &str, page, children| pdfcraft_organize::OutlineNode {
title: title.into(),
page,
view: None,
color: None,
italic: false,
bold: false,
open: true,
children,
};
let bookmarks = vec![node("Intro", Some(0), vec![node("Detail", Some(5), vec![])]), node("End", Some(1), vec![])];
s.apply(id, Edit::ImportBookmarks { parent: vec![], index: 0, bookmarks: bookmarks.clone(), replace: false }).unwrap();
assert_eq!(outline_titles(&s.get(id).unwrap().info.outline), ["Intro→1[Detail→0]", "End→2"]);
assert_eq!(s.get(id).unwrap().can_undo(), Some("Import bookmarks"));
assert_eq!(s.outline_tree(id, &[]).unwrap().unwrap().nodes[0].title, "Intro");
s.undo(id).unwrap();
assert!(s.get(id).unwrap().info.outline.is_empty());
s.apply(id, Edit::ImportBookmarks { parent: vec![], index: 0, bookmarks: bookmarks.clone(), replace: false }).unwrap();
s.apply(id, Edit::ImportBookmarks { parent: vec![], index: 0, bookmarks: vec![node("New", Some(1), vec![])], replace: true }).unwrap();
assert_eq!(outline_titles(&s.get(id).unwrap().info.outline), ["New→2"]);
s.undo(id).unwrap();
assert_eq!(outline_titles(&s.get(id).unwrap().info.outline), ["Intro→1[Detail→0]", "End→2"]);
s.undo(id).unwrap();
let err = s.apply(id, Edit::ImportBookmarks { parent: vec![4], index: 0, bookmarks, replace: false }).unwrap_err();
assert!(err.to_string().contains("no bookmark"), "{err}");
assert!(s.get(id).unwrap().info.outline.is_empty());
}
crates/engine/tests/form_date_language.rs
+453
View File
@@ -0,0 +1,453 @@
//! Form date fields name months in the app's date language (#5303). Its own test binary: the
//! names are process-wide, so other tests must not see them change, and its tests take turns.
use std::sync::{Arc, Mutex, MutexGuard, PoisonError};
use pdfcraft_cos::{Document, Object};
use pdfcraft_engine::dates::use_form_date_language;
use pdfcraft_engine::js::JsRunner;
use pdfcraft_forms::FieldValue;
use pdfcraft_forms::af::{DateTime, Format, format_date, format_value, keystroke, parse_date};
use pdfcraft_js::{DocInfo, Event, Limits, run};
/// Held by each test while it changes the process-wide date names, which start as English (a
/// failed test may have left others).
fn names_lock() -> MutexGuard<'static, ()> {
static LOCK: Mutex<()> = Mutex::new(());
let lock = LOCK.lock().unwrap_or_else(PoisonError::into_inner);
use_form_date_language("en");
lock
}
/// `util.printd(fmt)` of 18 October 2026, then of what `util.scand(fmt, typed)` reads.
fn printd(fmt: &str, typed: &str) -> String {
let doc = DocInfo { file_name: "form.pdf".into(), num_pages: 1, page: 0, info: Vec::new(), xfa: false };
let script = format!("event.value = util.printd('{fmt}', new Date(2026, 9, 18)) + '|' + util.printd('{fmt}', util.scand('{fmt}', '{typed}'));");
let o = run(&script, &Event::field("Format", "date", ""), &doc, &[], &[], Limits::default());
assert_eq!(o.error, None);
o.value
}
#[test]
fn form_dates_follow_the_date_language() {
let _names = names_lock();
let oct18 = DateTime { y: 2026, m: 10, d: 18, hh: 0, mm: 0, ss: 0 };
let dd_mmm_yy = Format::Date("dd-mmm-yy".into());
// English until the app says otherwise.
assert_eq!(format_date(oct18, "dd-mmm-yy dddd ddd mmmm"), "18-Oct-26 Sunday Sun October");
use_form_date_language("it");
// AFDate_FormatEx("dd-mmm-yy"), what the issue's form used.
assert_eq!(format_value(&dd_mmm_yy, "18-10-26"), "18-ott-26");
assert_eq!(format_date(oct18, "dddd d mmmm yyyy, ddd"), "domenica 18 ottobre 2026, dom");
// What it writes reads back, and so does English typing.
assert_eq!(parse_date("18-ott-26", "dd-mmm-yy"), Some(oct18));
assert_eq!(parse_date("18 ottobre 2026", "dd mmmm yyyy"), Some(oct18));
assert_eq!(parse_date("18-Oct-26", "dd-mmm-yy"), Some(oct18));
assert!(keystroke(&dd_mmm_yy, "date", "18-ott-26").is_ok());
// util.printd writes Italian names; util.scand reads them.
assert_eq!(printd("dd-mmm-yy", "18-ott-26"), "18-ott-26|18-ott-26");
assert_eq!(printd("dd-mmm-yy", "18-Oct-26"), "18-ott-26|18-ott-26", "English typing reads too");
assert_eq!(printd("dddd d mmmm yyyy", "domenica 18 ottobre 2026"), "domenica 18 ottobre 2026|domenica 18 ottobre 2026");
// German short names end in a dot; they are written and read whole.
use_form_date_language("de");
assert_eq!(format_value(&dd_mmm_yy, "18-10-26"), "18-Okt.-26");
assert_eq!(parse_date("18-Okt.-26", "dd-mmm-yy"), Some(oct18));
// An unknown language is English again.
use_form_date_language("xx");
assert_eq!(format_value(&dd_mmm_yy, "18-10-26"), "18-Oct-26");
assert_eq!(printd("dd-mmm-yy", "18-Oct-26"), "18-Oct-26|18-Oct-26");
every_date_language_reads_back_what_it_writes();
// Hostile typing never panics and stays quick: long, odd and mixed-script text.
use_form_date_language("th");
let junk = ["", " ", "ม.", "ม.ค.ม.ค.ม.ค.", "\u{0}", "1月1月1月", "tháng tháng", "Ottobre", "🙂ott🙂", "İstanbul ott 26"];
for t in junk.iter().map(|s| s.to_string()).chain(["ม.ค.1".repeat(5000), "9".repeat(100_000)]) {
let started = std::time::Instant::now();
let _ = parse_date(&t, "dd-mmm-yy");
let _ = keystroke(&dd_mmm_yy, "date", &t);
assert!(started.elapsed() < std::time::Duration::from_secs(2), "{} chars took {:?}", t.len(), started.elapsed());
}
use_form_date_language("en");
}
/// An English weekday typed before the month fills the format's weekday, so a month name that
/// the date language shares with a weekday (Italian, Spanish, Romanian and Albanian "mar") is
/// still read as the month: "Tue 17 Mar 2026" read as "Tue 17 Tuesday 2026" (#5303).
#[test]
fn an_english_weekday_before_a_shared_month_name_reads_as_english() {
let _names = names_lock();
let doc = DocInfo { file_name: "form.pdf".into(), num_pages: 1, page: 0, info: Vec::new(), xfa: false };
let mar17 = DateTime { y: 2026, m: 3, d: 17, hh: 0, mm: 0, ss: 0 };
let mut bad = Vec::new();
for language in ["it", "es", "ro", "sq"] {
use_form_date_language(language);
let ddd_d_mmm = "ddd d mmm yyyy";
let typed = [
// What the language writes, then English, then English and localized mixed.
(ddd_d_mmm, format_date(mar17, ddd_d_mmm)),
(ddd_d_mmm, "Tue 17 Mar 2026".into()),
(ddd_d_mmm, "Tuesday 17 Mar 2026".into()),
(ddd_d_mmm, "tue 17 mar 2026".into()),
(ddd_d_mmm, format!("Tue 17 {}", format_date(mar17, "mmm yyyy"))),
(ddd_d_mmm, format!("{} 17 Mar 2026", format_date(mar17, "ddd"))),
("mmm d yyyy ddd", "Mar 17 2026 Tue".into()),
("dddd d mmmm yyyy", "Tuesday 17 March 2026".into()),
];
for (fmt, text) in typed {
if parse_date(&text, fmt) != Some(mar17) {
bad.push(format!("{language} forms {fmt}: {text:?} -> {:?}", parse_date(&text, fmt)));
}
let script = format!("var d = util.scand('{fmt}', {text:?}); event.value = d ? util.printd('yyyy-mm-dd', d) : 'null';");
let o = run(&script, &Event::field("Format", "date", ""), &doc, &[], &[], Limits::default());
if o.value != "2026-03-17" {
bad.push(format!("{language} scand {fmt}: {text:?} -> {:?} {:?}", o.value, o.error));
}
}
}
use_form_date_language("en");
assert!(bad.is_empty(), "{}", bad.join("; "));
}
/// Every date language: what a date field writes reads back as the same date, for each date
/// letter that writes a name, through the form path and through `util.scand`. Called from the
/// one test, which owns the process-wide names.
fn every_date_language_reads_back_what_it_writes() {
let doc = DocInfo { file_name: "form.pdf".into(), num_pages: 1, page: 0, info: Vec::new(), xfa: false };
let mut bad = Vec::new();
for l in pdfcraft_engine::dates::DATE_LANGUAGES.iter() {
use_form_date_language(l.code);
for m in 1..=12u32 {
let dt = DateTime { y: 2026, m, d: 18, hh: 0, mm: 0, ss: 0 };
for fmt in ["dd-mmm-yy", "d mmmm yyyy", "dddd, mmmm d, yyyy", "ddd d mmm yyyy", "mmm d, yyyy ddd", "yyyymmmdd"] {
let text = format_date(dt, fmt);
// A short month ending in a number ("Thg 1") glued to the day is ambiguous
// ("Thg 118": January 18 or November 8): nothing can read it back.
if fmt == "yyyymmmdd" && l.months_short.iter().any(|n| n.ends_with(|c: char| c.is_ascii_digit())) {
continue;
}
if parse_date(&text, fmt).map(|p| (p.y, p.m, p.d)) != Some((2026, m, 18)) {
bad.push(format!("{} forms {fmt}: {text:?}", l.code));
}
let script = format!("var d = util.scand('{fmt}', {text:?}); event.value = d ? util.printd('yyyy-mm-dd', d) : 'null';");
let o = run(&script, &Event::field("Format", "date", ""), &doc, &[], &[], Limits::default());
if o.value != format!("2026-{m:02}-18") {
bad.push(format!("{} scand {fmt}: {text:?} -> {:?} {:?}", l.code, o.value, o.error));
}
}
}
}
use_form_date_language("en");
assert!(bad.is_empty(), "{} failures:\n{}", bad.len(), bad.iter().take(40).cloned().collect::<Vec<_>>().join("\n"));
}
/// A form whose text fields are each `(name, font, format script)`: `/Helv` is Helvetica with
/// WinAnsi, `/Song` a Chinese font with a Unicode CMap, which also holds Cyrillic.
fn form(fields: &[(&str, &str, &str)]) -> Document {
let n = fields.len();
let first = 6;
let refs = |from: usize| (from..from + n).map(|i| format!("{i} 0 R")).collect::<Vec<_>>().join(" ");
let mut objs: Vec<String> = vec![
"<< /Type /Catalog /Pages 2 0 R /AcroForm 4 0 R >>".into(),
"<< /Type /Pages /Kids [3 0 R] /Count 1 /MediaBox [0 0 600 800] >>".into(),
format!("<< /Type /Page /Parent 2 0 R /Annots [{}] >>", refs(first)),
format!("<< /Fields [{}] /DA (/Helv 0 Tf 0 g) /DR << /Font << /Helv 5 0 R /Song {} 0 R >> >> >>", refs(first), first + n),
"<< /Type /Font /Subtype /Type1 /BaseFont /Helvetica /Encoding /WinAnsiEncoding >>".into(),
];
for (i, (name, font, js)) in fields.iter().enumerate() {
let js = js.replace('\\', "\\\\").replace('(', "\\(").replace(')', "\\)");
let y = 700 - 30 * i;
objs.push(format!(
"<< /Type /Annot /Subtype /Widget /FT /Tx /T ({name}) /DA (/{font} 10 Tf 0 g) /Rect [50 {y} 300 {}] /P 3 0 R /AA << /F << /S /JavaScript /JS ({js}) >> >> >>",
y + 20
));
}
objs.push("<< /Type /Font /Subtype /Type0 /BaseFont /STSong-Light /Encoding /UniGB-UCS2-H /DescendantFonts [<< /Type /Font /Subtype /CIDFontType0 /BaseFont /STSong-Light /CIDSystemInfo << /Registry (Adobe) /Ordering (GB1) /Supplement 4 >> >>] >>".into());
let mut out = b"%PDF-1.7
"
.to_vec();
let mut offsets = Vec::new();
for (i, o) in objs.iter().enumerate() {
offsets.push(out.len());
out.extend_from_slice(
format!(
"{} 0 obj
{o}
endobj
",
i + 1
)
.as_bytes(),
);
}
let xref = out.len();
out.extend_from_slice(
format!(
"xref
0 {}
0000000000 65535 f
",
objs.len() + 1
)
.as_bytes(),
);
for o in offsets {
out.extend_from_slice(
format!(
"{o:010} 00000 n
"
)
.as_bytes(),
);
}
out.extend_from_slice(
format!(
"trailer
<< /Size {} /Root 1 0 R >>
startxref
{xref}
%%EOF
",
objs.len() + 1
)
.as_bytes(),
);
Document::open(Arc::new(out)).expect("opens")
}
/// The content of field `name`'s normal appearance.
fn appearance(doc: &Document, name: &str) -> Vec<u8> {
let all = pdfcraft_forms::fields(doc);
let f = all.iter().find(|f| f.name == name).expect("field");
let widget = doc.get(f.widgets[0].obj);
let normal = widget.as_dict().and_then(|d| d.get(b"AP")).and_then(|a| doc.resolve(a).as_dict().and_then(|a| a.reference(b"N"))).expect("/AP /N");
match &*doc.get(normal) {
Object::Stream(s) => s.decoded().expect("decodes"),
other => panic!("not a stream: {other:?}"),
}
}
/// What a Unicode-CMap (UCS-2) font is given for `text`.
fn ucs2(text: &str) -> Vec<u8> {
text.encode_utf16().flat_map(u16::to_be_bytes).collect()
}
fn contains(hay: &[u8], needle: &[u8]) -> bool {
hay.windows(needle.len()).any(|w| w == needle)
}
/// Fills `form`'s field `name` with 18 October 2026 through the JavaScript runner.
fn fill(doc: &mut Document, name: &str) {
fill_with(doc, name, "18-10-26");
}
/// Fills `form`'s field `name` with `value` through the JavaScript runner.
fn fill_with(doc: &mut Document, name: &str, value: &str) {
let mut js = JsRunner::new(doc, "form.pdf");
pdfcraft_forms::set_value_with(doc, name, &FieldValue::Text(value.into()), &mut js).expect("fills");
}
/// Month names in the date language that no font of the field can draw are drawn in English,
/// never as `?` ("18-????.-26" in Russian); a font that holds them keeps them (#5303).
#[test]
fn date_names_no_field_font_can_draw_are_drawn_in_english() {
let _names = names_lock();
let date = "AFDate_FormatEx(\"dd-mmm-yy\");";
let fields = [("helv", "Helv", date), ("song", "Song", date)];
for (language, local) in [("ru", "окт."), ("cs", "říj"), ("ja", "10月"), ("el", "Οκτ")] {
use_form_date_language(language);
let mut doc = form(&fields);
fill(&mut doc, "helv");
fill(&mut doc, "song");
let helv = appearance(&doc, "helv");
assert!(contains(&helv, b"(18-Oct-26) Tj"), "{language}: {}", String::from_utf8_lossy(&helv));
assert!(!helv.contains(&b'?'), "{language}: {}", String::from_utf8_lossy(&helv));
assert!(pdfcraft_forms::undrawn_fields(&doc).is_empty(), "{language}");
let song = appearance(&doc, "song");
assert!(contains(&song, &ucs2(&format!("18-{local}-26"))), "{language} keeps its name: {song:?}");
}
// Names Helvetica holds are kept: Italian "ott".
use_form_date_language("it");
let mut doc = form(&fields);
fill(&mut doc, "helv");
assert!(contains(&appearance(&doc, "helv"), b"(18-ott-26) Tj"));
use_form_date_language("en");
}
/// A Format script's `util.printd` is run again with English names for a field whose fonts
/// can't draw the date language's, and kept where they can (#5303).
#[test]
fn format_scripts_print_dates_in_english_where_the_field_cannot_draw_them() {
let _names = names_lock();
let script = "event.value = event.value ? util.printd(\"dd-mmm-yy\", util.scand(\"dd-mm-yy\", event.value)) + \" !\" : \"\";";
let fields = [("helv", "Helv", script), ("song", "Song", script)];
use_form_date_language("ru");
let mut doc = form(&fields);
fill(&mut doc, "helv");
fill(&mut doc, "song");
let helv = appearance(&doc, "helv");
assert!(contains(&helv, b"(18-Oct-26 !) Tj"), "{}", String::from_utf8_lossy(&helv));
assert!(!helv.contains(&b'?'), "{}", String::from_utf8_lossy(&helv));
assert!(pdfcraft_forms::undrawn_fields(&doc).is_empty());
assert!(contains(&appearance(&doc, "song"), &ucs2("18-окт.-26 !")));
// A script whose text has no English either keeps it, reported as undrawn.
let greeting = [("hi", "Helv", "event.value = \"Привет \" + util.printd(\"mmm\", new Date(2026, 9, 18));")];
let mut doc = form(&greeting);
fill(&mut doc, "hi");
assert_eq!(pdfcraft_forms::undrawn_fields(&doc), [("hi".to_string(), 'П')]);
// The English run is the appearance's alone: scripts go on printing the date language.
assert_eq!(printd("dd-mmm-yy", "18-окт.-26"), "18-окт.-26|18-окт.-26");
use_form_date_language("en");
}
/// A Format script's English run is kept when it ran, even where it prints the stored value
/// again: a Helvetica field storing "18-Oct-26", printed again as `dd-mmm-yy`, showed
/// "18-????.-26" in Russian. A script that fails keeps what it showed, reported undrawn (#5303).
#[test]
fn an_english_run_that_prints_the_stored_value_is_drawn() {
let _names = names_lock();
let script = "event.value = util.printd(\"dd-mmm-yy\", util.scand(\"dd-mmm-yy\", event.value));";
use_form_date_language("ru");
let mut doc = form(&[("helv", "Helv", script), ("fails", "Helv", "throw new Error(\"no\");")]);
fill_with(&mut doc, "helv", "18-Oct-26");
let helv = appearance(&doc, "helv");
assert!(contains(&helv, b"(18-Oct-26) Tj"), "{}", String::from_utf8_lossy(&helv));
assert!(!helv.contains(&b'?'), "{}", String::from_utf8_lossy(&helv));
assert!(pdfcraft_forms::undrawn_fields(&doc).is_empty());
fill_with(&mut doc, "fails", "18-окт.-26");
assert_eq!(pdfcraft_forms::undrawn_fields(&doc), [("fails".to_string(), 'о')]);
use_form_date_language("en");
}
/// What `util.scand(fmt, text)` reads, as `yyyy-mm-dd`, or `null`.
fn scanned(fmt: &str, text: &str) -> String {
let doc = DocInfo { file_name: "form.pdf".into(), num_pages: 1, page: 0, info: Vec::new(), xfa: false };
let script = format!("var d = util.scand('{fmt}', {text:?}); event.value = d ? util.printd('yyyy-mm-dd', d) : 'null';");
let o = run(&script, &Event::field("Format", "date", ""), &doc, &[], &[], Limits::default());
assert_eq!(o.error, None);
o.value
}
/// A short weekday typed without its dot ("mar" for French "mar.", Tuesday) is the weekday where
/// the format has one, not the English "Mar" (March) that is the only other name of its length:
/// "mar 6 janv 2026" read as no date in French and Romanian (#5303).
#[test]
fn a_short_weekday_typed_without_its_dot_is_the_weekday() {
let _names = names_lock();
let mut bad = Vec::new();
let cases = [
("fr", "ddd d mmm yyyy", "Mar 6 janv 2026", "2026-01-06"),
("fr", "ddd d mmm yyyy", "mar 17 mars 2026", "2026-03-17"),
("fr", "mmm d yyyy ddd", "janv 6 2026 mar", "2026-01-06"),
("fr", "ddd dd/mm/yyyy", "mar 06/01/2026", "2026-01-06"),
("ro", "ddd d mmm yyyy", "mar 6 ian 2026", "2026-01-06"),
("ro", "ddd dd/mm/yyyy", "mar 06/01/2026", "2026-01-06"),
];
for (language, fmt, text, want) in cases {
use_form_date_language(language);
let got = scanned(fmt, text);
if got != want {
bad.push(format!("{language} scand {fmt}: {text:?} -> {got}"));
}
let read = parse_date(text, fmt).map(|d| format!("{:04}-{:02}-{:02}", d.y, d.m, d.d));
if read.as_deref() != Some(want) {
bad.push(format!("{language} forms {fmt}: {text:?} -> {read:?}"));
}
}
use_form_date_language("en");
assert!(bad.is_empty(), "{}", bad.join("; "));
}
/// A weekday left out of a format with one ("17 mar 2026" for "ddd d mmm yyyy") lets a name the
/// date language shares between a weekday and a month be the month: Italian, Spanish and
/// Albanian "mar" (Tuesday, March) and Albanian "sht" (Saturday, September) were read as the
/// weekday, leaving no month (#5303). Typed weekdays still go first.
#[test]
fn a_shared_name_is_the_month_when_the_weekday_is_left_out() {
let _names = names_lock();
let mut bad = Vec::new();
let cases = [
("it", "ddd d mmm yyyy", "17 Mar 2026", (2026, 3, 17)),
("es", "ddd d mmm yyyy", "17 Mar 2026", (2026, 3, 17)),
("sq", "ddd d mmm yyyy", "17 Mar 2026", (2026, 3, 17)),
("it", "ddd d mmm yyyy", "17 mar 2026", (2026, 3, 17)),
("it", "dddd, mmmm d, yyyy", "mar 17, 2026", (2026, 3, 17)),
("it", "dddd, mmm d, yyyy", "mar 17 2026", (2026, 3, 17)),
("sq", "ddd d mmm yyyy", "5 sht 2026", (2026, 9, 5)),
("it", "ddd d mmm yyyy", "mar 17 Mar 2026", (2026, 3, 17)),
("it", "ddd d mmm yyyy", "mar 17 mar 2026", (2026, 3, 17)),
];
for (language, fmt, text, want) in cases {
use_form_date_language(language);
let read = parse_date(text, fmt).map(|d| (d.y, d.m, d.d));
if read != Some(want) {
bad.push(format!("{language} forms {fmt}: {text:?} -> {read:?}"));
}
if keystroke(&Format::Date(fmt.into()), "date", text).is_err() {
bad.push(format!("{language} keystroke {fmt}: {text:?} refused"));
}
}
use_form_date_language("en");
assert!(bad.is_empty(), "{}", bad.join("; "));
}
/// Turkish names typed in capitals are read: lower-casing letter by letter made "MAYIS" "mayis"
/// (not "mayıs"), "SALI" "sali" and "EKİM" "eki̇m" (an i with a dot above), so no name matched.
/// Dotted, dotless and capital i are one letter when names are read (#5303).
#[test]
fn turkish_names_typed_in_capitals_are_read() {
let _names = names_lock();
use_form_date_language("tr");
let mut bad = Vec::new();
for (fmt, text, want) in [
("d mmmm yyyy", "1 MAYIS 2026", (2026, 5, 1)),
("d mmmm yyyy", "1 EKİM 2026", (2026, 10, 1)),
("d mmmm yyyy", "1 Ekim 2026", (2026, 10, 1)),
("d mmm yyyy", "1 EKİ 2026", (2026, 10, 1)),
("dddd d mmmm yyyy", "SALI 6 OCAK 2026", (2026, 1, 6)),
("dddd d mmmm yyyy", "Salı 6 Ocak 2026", (2026, 1, 6)),
] {
let read = parse_date(text, fmt).map(|d| (d.y, d.m, d.d));
if read != Some(want) {
bad.push(format!("forms {fmt}: {text:?} -> {read:?}"));
}
let (y, m, d) = want;
let got = scanned(fmt, text);
if got != format!("{y:04}-{m:02}-{d:02}") {
bad.push(format!("scand {fmt}: {text:?} -> {got}"));
}
}
// English is read as before, capitals and all.
use_form_date_language("en");
for (fmt, text) in [("d mmmm yyyy", "1 APRIL 2026"), ("dddd d mmmm yyyy", "FRIDAY 1 MAY 2026"), ("d mmm yyyy", "1 MAY 2026")] {
if parse_date(text, fmt).is_none() || scanned(fmt, text) == "null" {
bad.push(format!("en {fmt}: {text:?}"));
}
}
assert!(bad.is_empty(), "{}", bad.join("; "));
}
/// #5464 review: Greek names typed in capitals end in a capital Σ, which is the final ς of the
/// name ("ΙΑΝΟΥΆΡΙΟΣ" for "Ιανουάριος"); they read as they did before names were folded.
#[test]
fn greek_names_typed_in_capitals_are_read() {
let _names = names_lock();
use_form_date_language("el");
let mut bad = Vec::new();
for (fmt, text, want) in [
("d mmmm yyyy", "1 ΙΑΝΟΥΆΡΙΟΣ 2026", (2026, 1, 1)),
("d mmmm yyyy", "1 Ιανουάριος 2026", (2026, 1, 1)),
("d mmmm yyyy", "1 ιανουάριοσ 2026", (2026, 1, 1)),
("mmmm d, yyyy", "ΜΆΙΟΣ 5, 2026", (2026, 5, 5)),
("dddd, mmmm d, yyyy", "FRIDAY, ΙΑΝΟΥΆΡΙΟΣ 2, 2026", (2026, 1, 2)),
] {
let read = parse_date(text, fmt).map(|d| (d.y, d.m, d.d));
if read != Some(want) {
bad.push(format!("forms {fmt}: {text:?} -> {read:?}"));
}
let (y, m, d) = want;
let got = scanned(fmt, text);
if got != format!("{y:04}-{m:02}-{d:02}") {
bad.push(format!("scand {fmt}: {text:?} -> {got}"));
}
}
assert!(bad.is_empty(), "{}", bad.join("; "));
}
crates/engine/tests/it/ocr_edit.rs
+189
View File
@@ -207,3 +207,192 @@ fn visible_stretched_text_rewrites_unchanged() {
let i = ((y * out.width + x) * 4) as usize;
assert_eq!(&out.rgba[i..i + 3], &[255, 255, 255]);
}
/// A scan whose OCR layer is drawn first and the opaque page image over it, as some OCR tools
/// write it (#3973): invisible Helvetica `text` on the scanned word, then the image.
fn ocr_under_scan(text: &str) -> std::sync::Arc<Vec<u8>> {
under_scan(&format!("BT /F1 20 Tf 3 Tr 42 44 Td ({text}) Tj ET"))
}
/// A page whose `text` content is followed by the opaque 200 × 100 scan image.
fn under_scan(text: &str) -> std::sync::Arc<Vec<u8>> {
under_scan_with(text, |_| pdfcraft_cos::Dict::new())
}
/// As [`under_scan`], with the page's ExtGState resources made by `ext_gstates`.
fn under_scan_with(text: &str, ext_gstates: impl FnOnce(&mut pdfcraft_cos::Document) -> pdfcraft_cos::Dict) -> std::sync::Arc<Vec<u8>> {
use pdfcraft_cos::{Dict, Document, Object, SaveOptions, Stream};
let (w, h) = (200u32, 100u32);
let mut rgb = Vec::new();
for y in 0..h {
for x in 0..w {
let ink = (WORD[0]..WORD[2]).contains(&x) && (WORD[1]..WORD[3]).contains(&y);
rgb.extend(if ink { INK } else { PAPER });
}
}
let mut d = Document::new_empty();
let pages = d.add(Object::Dict(Dict::new()));
let content = format!("{text} q 200 0 0 100 0 0 cm /Im Do Q");
let contents = d.add(Object::Stream(Stream::flate(Dict::new(), content.as_bytes())));
let mut image = Dict::new();
image.set(b"Type".to_vec(), Object::name("XObject"));
image.set(b"Subtype".to_vec(), Object::name("Image"));
image.set(b"Width".to_vec(), Object::Int(i64::from(w)));
image.set(b"Height".to_vec(), Object::Int(i64::from(h)));
image.set(b"ColorSpace".to_vec(), Object::name("DeviceRGB"));
image.set(b"BitsPerComponent".to_vec(), Object::Int(8));
let image = d.add(Object::Stream(Stream::flate(image, &rgb)));
let mut font = Dict::new();
font.set(b"Type".to_vec(), Object::name("Font"));
font.set(b"Subtype".to_vec(), Object::name("Type1"));
font.set(b"BaseFont".to_vec(), Object::name("Helvetica"));
let mut fonts = Dict::new();
fonts.set(b"F1".to_vec(), Object::Dict(font));
let mut xobjects = Dict::new();
xobjects.set(b"Im".to_vec(), Object::Ref(image));
let mut res = Dict::new();
res.set(b"Font".to_vec(), Object::Dict(fonts));
res.set(b"XObject".to_vec(), Object::Dict(xobjects));
res.set(b"ExtGState".to_vec(), Object::Dict(ext_gstates(&mut d)));
let mut p = Dict::new();
p.set(b"Type".to_vec(), Object::name("Page"));
p.set(b"Parent".to_vec(), Object::Ref(pages));
p.set(b"Contents".to_vec(), Object::Ref(contents));
p.set(b"Resources".to_vec(), Object::Dict(res));
p.set(b"MediaBox".to_vec(), Object::Array([0, 0, 200, 100].map(Object::Int).to_vec()));
let page = d.add(Object::Dict(p));
d.update_dict(pages, |p| {
p.set(b"Type".to_vec(), Object::name("Pages"));
p.set(b"Kids".to_vec(), Object::Array(vec![Object::Ref(page)]));
p.set(b"Count".to_vec(), Object::Int(1));
})
.unwrap();
let mut root = Dict::new();
root.set(b"Type".to_vec(), Object::name("Catalog"));
root.set(b"Pages".to_vec(), Object::Ref(pages));
let root = d.add(Object::Dict(root));
d.trailer_mut().set(b"Root".to_vec(), Object::Ref(root));
std::sync::Arc::new(pdfcraft_cos::write_full(&d, &SaveOptions::default()).unwrap())
}
/// The scanned word's box shows the cover and new ink, not the scan.
fn assert_edit_shows(s: &Session, id: pdfcraft_engine::DocId, label: &str) {
let px = word_pixels(s, id);
let paper = px.iter().filter(|p| near(**p, PAPER)).count();
let ink = px.iter().filter(|p| near(**p, INK)).count();
assert!(paper * 10 > px.len() * 4, "{label}: the cover is drawn over the scan: {paper} of {} paper", px.len());
assert!(ink * 100 > px.len() * 3, "{label}: the new text shows: {ink} of {} ink", px.len());
}
fn texts(s: &Session, id: pdfcraft_engine::DocId) -> Vec<String> {
s.get(id).unwrap().text_blocks(0).iter().map(|b| b.text.replace('\n', " ")).collect()
}
fn edit_block(s: &mut Session, id: pdfcraft_engine::DocId, block: usize, text: &str) {
s.apply(id, Edit::EditTextBlock { page: 0, block, text: text.into(), style: Default::default() }).unwrap();
}
/// The OCR layer is under the scan image: the rewritten paragraph is drawn on top of the image,
/// not in the OCR text's place where the image would hide it (#3973). One line takes the
/// single-style rewrite, a typed line break the styled one. Undo restores the hidden layer, and
/// the edit survives redo, a save and reopen, and being edited again.
#[test]
fn an_ocr_layer_under_the_scan_image_shows_its_edit() {
for edit in ["Jane", "Jane\nDoe"] {
let mut s = Session::new();
let id = s.open("ocr-under-scan.pdf", None, ocr_under_scan("John"), None).unwrap();
let blocks = s.get(id).unwrap().text_blocks(0);
assert_eq!(blocks.len(), 1);
assert!(blocks[0].invisible, "the OCR layer is invisible text");
assert!(word_pixels(&s, id).iter().all(|p| near(*p, INK)), "the scanned word fills its box");
edit_block(&mut s, id, 0, edit);
assert_eq!(texts(&s, id), [edit.replace('\n', " ")], "{edit:?}");
assert_edit_shows(&s, id, edit);
s.undo(id).unwrap();
assert_eq!(texts(&s, id), ["John"], "{edit:?}: undo");
assert!(s.get(id).unwrap().text_blocks(0)[0].invisible, "{edit:?}: undo hides the layer again");
assert!(word_pixels(&s, id).iter().all(|p| near(*p, INK)), "{edit:?}: undo shows the scan again");
s.redo(id).unwrap();
assert_edit_shows(&s, id, &format!("{edit:?} redo"));
let saved = s.save_bytes(id).unwrap();
let mut reopened = Session::new();
let id = reopened.open("saved.pdf", None, saved, None).unwrap();
assert_eq!(texts(&reopened, id), [edit.replace('\n', " ")], "{edit:?}: reopened");
assert_edit_shows(&reopened, id, &format!("{edit:?} reopened"));
edit_block(&mut reopened, id, 0, "Ann");
assert_eq!(texts(&reopened, id), ["Ann"], "{edit:?}: edited again");
assert_edit_shows(&reopened, id, &format!("{edit:?} edited again"));
}
}
/// Visible text earlier in the same text object makes the rewrite split it; the OCR phrase still
/// goes on top of the scan, and the visible text stays.
#[test]
fn an_ocr_phrase_split_from_visible_text_under_the_scan_shows_its_edit() {
for edit in ["Jane", "Jane\nDoe"] {
let mut s = Session::new();
let page = under_scan("BT /F1 20 Tf 10 80 Td (Head) Tj 3 Tr 32 -36 Td (John) Tj ET");
let id = s.open("split.pdf", None, page, None).unwrap();
let blocks = s.get(id).unwrap().text_blocks(0);
let john = blocks.iter().position(|b| b.text == "John").unwrap();
assert!(blocks[john].invisible && blocks.iter().any(|b| b.text == "Head" && !b.invisible), "{blocks:?}");
edit_block(&mut s, id, john, edit);
let mut after = texts(&s, id);
after.sort();
let mut want = vec!["Head".to_string(), edit.replace('\n', " ")];
want.sort();
assert_eq!(after, want, "{edit:?}");
assert_edit_shows(&s, id, &format!("{edit:?} split"));
}
}
/// `/SM`: a luminosity soft mask that is white over 30..110 × 30..70 in the space it is selected
/// in, black elsewhere.
fn soft_mask_state(d: &mut pdfcraft_cos::Document) -> pdfcraft_cos::Dict {
use pdfcraft_cos::{Dict, Object, Stream};
let mut group = Dict::new();
group.set(b"S".to_vec(), Object::name("Transparency"));
group.set(b"CS".to_vec(), Object::name("DeviceRGB"));
let mut form = Dict::new();
form.set(b"Type".to_vec(), Object::name("XObject"));
form.set(b"Subtype".to_vec(), Object::name("Form"));
form.set(b"BBox".to_vec(), Object::Array([0, 0, 200, 100].map(Object::Int).to_vec()));
form.set(b"Group".to_vec(), Object::Dict(group));
let mut mask = Dict::new();
mask.set(b"Type".to_vec(), Object::name("Mask"));
mask.set(b"S".to_vec(), Object::name("Luminosity"));
mask.set(b"G".to_vec(), Object::Ref(d.add(Object::Stream(Stream::flate(form, b"1 g 30 30 80 40 re f")))));
let mut state = Dict::new();
state.set(b"Type".to_vec(), Object::name("ExtGState"));
state.set(b"SMask".to_vec(), Object::Dict(mask));
let mut states = Dict::new();
states.set(b"SM".to_vec(), Object::Dict(state));
states
}
/// The soft mask is selected before a translation, so it is placed by the CTM before it. The
/// paragraph lifted over the scan selects it again after its own `cm`: there it must be placed
/// as it was, not 100 pt to the right where it would hide the new text, so the edit renders
/// exactly as the same edit on the page without the mask. Both rewrite paths.
#[test]
fn a_soft_mask_selected_before_a_cm_stays_put_when_the_edit_is_lifted() {
for edit in ["Jane", "Jane\nDoe"] {
let mut plain = Session::new();
let page = under_scan("q 1 0 0 1 100 0 cm BT /F1 20 Tf 3 Tr -58 44 Td (John) Tj ET Q");
let plain_id = plain.open("plain.pdf", None, page, None).unwrap();
edit_block(&mut plain, plain_id, 0, edit);
let mut s = Session::new();
let page = under_scan_with("q /SM gs 1 0 0 1 100 0 cm BT /F1 20 Tf 3 Tr -58 44 Td (John) Tj ET Q", soft_mask_state);
let id = s.open("masked.pdf", None, page, None).unwrap();
assert_eq!(texts(&s, id), ["John"]);
edit_block(&mut s, id, 0, edit);
assert_eq!(texts(&s, id), [edit.replace('\n', " ")], "{edit:?}");
assert_edit_shows(&s, id, &format!("{edit:?} masked"));
let ink = |s: &Session, id| word_pixels(s, id).iter().filter(|p| near(**p, INK)).count();
assert_eq!(ink(&s, id), ink(&plain, plain_id), "{edit:?}: the mask does not hide the new text");
}
}
crates/filters/README.md
+9 -8
View File
@@ -27,10 +27,11 @@ encode(filter, params, data) -> Result<Vec<u8>, FilterError>
encode_flate(data) -> Vec<u8> // zlib, level 6
encode_g4(data, columns, rows) -> Result<Vec<u8>, FilterError> // 1-bit rows as CCITT Group 4; columns <= MAX_G4_COLUMNS
pub struct CcittParams { k, columns, rows, end_of_line, encoded_byte_align, end_of_block, black_is_1 } // Table 11; Default = its defaults
decode_ccitt(data, &params, max_rows) -> Result<Vec<u8>, FilterError> // CCITT fax as CCITTFaxDecode gives it; columns <= MAX_CCITT_COLUMNS
decode_jbig2(data, globals, columns, rows, max) -> Result<Vec<u8>, FilterError> // JBIG2 as JBIG2Decode gives it; a page of columns × rows, within max and MAX_JBIG2_PIXELS
encode_jbig2_generic(data, columns, rows) -> Result<Vec<u8>, FilterError> // 1-bit rows as lossless JBIG2 for JBIG2Decode; columns <= MAX_JBIG2_COLUMNS
encode_jbig2(data, columns, rows) -> Result<Vec<u8>, FilterError> // 1-bit rows as lossless JBIG2 for JBIG2Decode, symbol coded when that is smaller; columns <= MAX_JBIG2_COLUMNS
decode_ccitt_within(data, &params, max_rows, work: &mut u64) -> Result<Vec<u8>, FilterError> // CCITT fax as CCITTFaxDecode gives it; columns <= MAX_CCITT_COLUMNS; its rows paid for from work before they are decoded
decode_jbig2_within(data, globals, columns, rows, max, work: &mut u64) -> Result<Vec<u8>, FilterError> // JBIG2 as JBIG2Decode gives it; a page of columns × rows, within max and MAX_JBIG2_PIXELS; reading and decoding it spend from work, in the decoder's steps
encode_jbig2_generic(data, columns, rows) -> Result<Vec<u8>, FilterError> // 1-bit rows as lossless JBIG2 for JBIG2Decode; columns and rows <= MAX_JBIG2_COLUMNS (65,535) and MAX_JBIG2_PIXELS (2^28) samples, what the decoder takes
encode_jbig2(data, columns, rows) -> Result<Vec<u8>, FilterError> // 1-bit rows as lossless JBIG2 for JBIG2Decode, symbol coded when that is smaller; the same limits
encode_jbig2_shared(&[(data, columns, rows)]) -> Result<Option<SharedJbig2>, FilterError> // images symbol coded sharing one dictionary: the /JBIG2Globals data, and each image's stream or None; the same limits
```
- `chain` is applied first to last, as in a `/Filter` array.
@@ -50,8 +51,8 @@ encode_jbig2(data, columns, rows) -> Result<Vec<u8>, FilterError> // 1-bit row
| ASCIIHexDecode | white space ignored, odd digit padded with 0, `>` EOD (optional), other characters → Corrupt | upper case, newline every 64 digits, `>` |
| ASCII85Decode | `<~` prefix, `z`, white space, `~>` (or bare `~`, or none) EOD, final partial group; `z` inside a group, a 1-char final group, overflow or bad characters → Corrupt | `z` for zero groups, newline every 75 characters, `~>` |
| RunLengthDecode | 0–127 literal, 129–255 repeat, 128 EOD (optional); truncated run → Corrupt | runs of ≥2 repeat, literals up to 128, 128 |
| CCITTFaxDecode | not in a chain (`Unsupported`); `decode_ccitt` decodes Group 3 and Group 4 on its own, as rows of 1-bit samples each on a new byte, black 0 unless `black_is_1`, at most `max_rows` rows (fewer with `/Rows`), the rows before any damage, up to 1,048,576 samples a row; Group 3 rows on bytes are read by their end-of-line codes when the data starts with one and on bytes otherwise, whatever `end_of_line` says, so data with codes but none before the first row (not as T.4 sends them) is read on bytes, as the renderer draws it, and at `k` 0 may come out as other rows (`pdfcraft_cos::Document::decoded_image` calls it for a fax image, after the filters before it) | `encode_g4`: Group 4 (`/K -1`) with an end-of-block mark, 0 bits coded black so `/BlackIs1 false` decodes the same bits; rows packed MSB first, each on a new byte, padding bits and trailing bytes ignored; up to 32,767 samples a row (the encoder counts positions in 16 bits); no rows or columns, or too wide → Unsupported, too few bytes → Corrupt |
| JBIG2Decode | not in a chain (`Unsupported`); `decode_jbig2` decodes it on its own with `hayro-jbig2`, given what its `/JBIG2Globals` decode to, as rows of 1-bit samples each on a new byte, black 0; the page has to be `columns` × `rows` (PdfCraft's renderer draws the whole page into the image), within `max` bytes as such rows and 2^28 pixels (`MAX_JBIG2_PIXELS`, about the most the decoder takes), checked before it is decoded; regions inside it are bounded by the decoder alone (a region may claim more pixels than its page); damaged data, or globals that don't decode, are `Corrupt` (`pdfcraft_cos::Document::decoded_image` calls it for a JBIG2 image, after the filters before it) | `encode_jbig2_generic`: the embedded organisation (D.3) with a page information segment and an immediate lossless generic region (7.4.6), no `/JBIG2Globals`; template 0 with its nominal AT pixels and typical prediction (6.2.5), coded with the MQ arithmetic coder (Annex E, Table E.1) and ended by its 0xFF 0xAC marker. A 0 sample is coded as a JBIG2 1 (black), since `JBIG2Decode` gives black as 0, so the stream decodes to the same bits; rows packed MSB first, each on a new byte, padding bits and trailing bytes ignored; up to 1,048,576 samples a row; no rows or columns, or too wide → Unsupported, too few bytes → Corrupt. `encode_jbig2`: the smaller of that and symbol coding (`jbig2/symbols.rs`): each distinct mark (8-connected black pixels, at most 256 a side) stored once in a symbol dictionary (7.4.2; arithmetic coded, each bitmap a template 0 generic region without typical prediction, the bitmaps sharing their contexts) and placed by immediate lossless text regions referring to it (7.4.3; bottom-left reference corner, strips of 1, 2, 4 or 8 rows, whichever is smallest), each of at most 10,000 instances over a band of rows, as the decoder PdfCraft renders with (`hayro-jbig2`) takes no more in one region, with larger marks, and in a second plan marks that occur once, in a generic region over their box. Only exact copies share a symbol (no refinement), so the stream decodes to the same bits; fewer than two symbols, more than 2^20 black runs, 65,536 distinct marks or 16 MiB of their pixels, or more than 2^24 rows, and the generic coding is used; a plan of more than 32,767 symbols (that decoder refuses more) isn't made |
| CCITTFaxDecode | not in a chain (`Unsupported`); `decode_ccitt_within` decodes Group 3 and Group 4 on its own, as rows of 1-bit samples each on a new byte, black 0 unless `black_is_1`, at most `max_rows` rows (fewer with `/Rows`), the rows before any damage, up to 1,048,576 samples a row; Group 3 rows on bytes are read by their end-of-line codes when the data starts with one and on bytes otherwise, whatever `end_of_line` says, so data with codes but none before the first row (not as T.4 sends them) is read on bytes, as the renderer draws it, and at `k` 0 may come out as other rows (`pdfcraft_cos::ImageBudget::decode` calls it for a fax image, after the filters before it) | `encode_g4`: Group 4 (`/K -1`) with an end-of-block mark, 0 bits coded black so `/BlackIs1 false` decodes the same bits; rows packed MSB first, each on a new byte, padding bits and trailing bytes ignored; up to 32,767 samples a row (the encoder counts positions in 16 bits); no rows or columns, or too wide → Unsupported, too few bytes → Corrupt |
| JBIG2Decode | not in a chain (`Unsupported`); `decode_jbig2_within` decodes it on its own with `hayro-jbig2`, given what its `/JBIG2Globals` decode to, as rows of 1-bit samples each on a new byte, black 0; the page has to be `columns` × `rows` (PdfCraft's renderer draws the whole page into the image), within `max` bytes as such rows and 2^28 pixels (`MAX_JBIG2_PIXELS`, the most the decoder takes), checked before it is decoded; regions inside it are bounded by the decoder alone (a region may claim more pixels than its page); damaged data, or globals that don't decode, are `Corrupt` (`pdfcraft_cos::ImageBudget::decode` calls it for a JBIG2 image, after the filters before it) | `encode_jbig2_generic`: the embedded organisation (D.3) with a page information segment and an immediate lossless generic region (7.4.6), no `/JBIG2Globals`; template 0 with its nominal AT pixels and typical prediction (6.2.5), coded with the MQ arithmetic coder (Annex E, Table E.1) and ended by its 0xFF 0xAC marker. A 0 sample is coded as a JBIG2 1 (black), since `JBIG2Decode` gives black as 0, so the stream decodes to the same bits; rows packed MSB first, each on a new byte, padding bits and trailing bytes ignored; up to 65,535 samples on a side and 2^28 in all, where `hayro-jbig2` stops; no rows or columns, or too large → Unsupported, too few bytes → Corrupt. `encode_jbig2`: the smaller of that and symbol coding (`jbig2/symbols.rs`): each distinct mark (8-connected black pixels, at most 256 a side) stored once in a symbol dictionary (7.4.2; arithmetic coded, each bitmap a template 0 generic region without typical prediction, the bitmaps sharing their contexts) and placed by immediate lossless text regions referring to it (7.4.3; bottom-left reference corner, strips of 1, 2, 4 or 8 rows, whichever is smallest), each of at most 10,000 instances over a band of rows, as the decoder PdfCraft renders with (`hayro-jbig2`) takes no more in one region, with larger marks, and in a second plan marks that occur once, in a generic region over their box. Only exact copies share a symbol (no refinement), so the stream decodes to the same bits; fewer than two symbols, more than 2^20 black runs, 65,536 distinct marks or 16 MiB of their pixels, and the generic coding is used; a plan of more than 32,767 symbols (that decoder refuses more) isn't made. `encode_jbig2_shared`: the marks found in more than one image of a group (at most 16,383, those in the most images first, within the same limits on distinct marks and their pixels as one image's) go once in a symbol dictionary segment associated with no page, number 0, which is the `/JBIG2Globals` data (7.2.6, D.3); each image's stream numbers its segments from 1 (a segment refers only to lower numbers, 7.2.5) and its text regions refer to that dictionary first, then to one of its own for the marks only it repeats, so the shared symbols' IDs come first (Table 31's SBSYMS); the rest is coded as `encode_jbig2`'s symbol coding codes it, the smaller of its two plans. An image that places none of the shared marks gets no stream, to be stored alone; whether the rest save more than the globals take is the caller's to weigh |
## Tests
@@ -61,8 +62,8 @@ encode_jbig2(data, columns, rows) -> Result<Vec<u8>, FilterError> // 1-bit row
- `ccitt.rs`: `decode_ccitt` gives back the rows `encode_g4` was given (with `/BlackIs1`, `/Rows` and a row limit), refuses bad widths, and bounds its output for any input; byte-aligned Group 3 in each layout, over 32 seeded pictures, decodes to its picture, and a byte short to fewer rows. Group 4 output decoded by `hayro-ccitt` (an independent decoder) matches the input bits: a text-like page, padding bits, the widest row, and property tests over bitmaps of noise and of runs; any input returns rather than panics.
- `mq.rs`: the arithmetic coder reproduces Annex H.2's 30 bytes from its 256 decisions.
- `jbig2.rs`: output decoded by `hayro-jbig2` (dev-dependency, an independent decoder) matches the input bits: a text-like page (smaller than Group 4), edges, padding, all-white and all-black regions (typical rows), and property tests over bitmaps of noise and of runs; any input returns rather than panics.
- `jbig2/decode.rs`: `decode_jbig2` gives back the rows of both codings, padding bits 0; refuses a page of another width or height, past `max` or past `MAX_JBIG2_PIXELS` before decoding it (2^28 pixels take about 34 MB, a bit each), globals that aren't JBIG2 and damaged data, and returns rather than panics on every prefix of a stream.
- `jbig2/symbols.rs`: the integer coders against Annex A's example and Table A.1; symbol-coded output decoded by `hayro-jbig2` matches the input bits for every plan and strip size: text whose letters repeat (under half the generic coding's size, the dictionary retained and referred to), large and lone marks left to a generic region, marks on the edges, touching at corners and inside each other's boxes, and property tests over glyphs scattered over each other and the edges with stray pixels (never larger than the generic coding); one shape, a blank page and noise are coded generically; marks joined at a corner, either way, are one mark, and 256 pixels a side is a symbol while 257 wide or tall is left out; black padding bits are ignored; coordinates past 4,435 use Annex A's 32-bit range and PREV stays below 512; past 2^20 runs, 2^24 rows, 65,536 distinct marks or 2^24 symbol pixels the image is coded generically, and at each cap it is still looked at; a plan codes at most 32,767 symbols, which decode, and past that the second plan still codes the repeated ones; more than 10,000 marks are placed by several text regions, numbered in turn, the dictionary retained for each but the last; any input returns rather than panics.
- `jbig2/decode.rs`: `decode_jbig2_within` gives back the rows of both codings, padding bits 0; refuses a page of another width or height, past `max` or past `MAX_JBIG2_PIXELS` before decoding it (2^28 pixels take about 34 MB, a bit each), globals that aren't JBIG2 and damaged data, and returns rather than panics on every prefix of a stream.
- `jbig2/symbols.rs`: the integer coders against Annex A's example and Table A.1; symbol-coded output decoded by `hayro-jbig2` matches the input bits for every plan and strip size: text whose letters repeat (under half the generic coding's size, the dictionary retained and referred to), large and lone marks left to a generic region, marks on the edges, touching at corners and inside each other's boxes, and property tests over glyphs scattered over each other and the edges with stray pixels (never larger than the generic coding); one shape, a blank page and noise are coded generically; marks joined at a corner, either way, are one mark, and 256 pixels a side is a symbol while 257 wide or tall is left out; black padding bits are ignored; coordinates past 4,435 use Annex A's 32-bit range and PREV stays below 512; past 2^20 runs, 65,536 distinct marks or 2^24 symbol pixels the image is coded generically, and at each cap it is still looked at; a plan codes at most 32,767 symbols, which decode, and past that the second plan still codes the repeated ones; more than 10,000 marks are placed by several text regions, numbered in turn, the dictionary retained for each but the last; pages in one font share one dictionary, segment 0 of no page, retained, which each page's text regions refer to first, taking less together than alone, a page's own repeated marks in a dictionary referred to after it, a page sharing nothing left alone, images with fewer than two marks in common sharing nothing, and past 16,383 shared marks those in the most images kept, with a property test over groups of scattered pages; any input returns rather than panics.
- `fuzz/`: a detached cargo-fuzz crate (`cargo +nightly fuzz run decode`, run from `crates/filters/`).
## Status
crates/filters/src/jbig2.rs
+121 -25
View File
@@ -1,8 +1,9 @@
//! JBIG2 encoding (ITU-T T.88 | ISO/IEC 14492), for `JBIG2Decode` (ISO 32000-2 §7.4.7), lossless:
//! a generic region, template 0 with its nominal adaptive pixels, typical prediction on, MQ
//! arithmetic coding ([`crate::mq`]); or the image's marks as symbols ([`symbols`]). Written from
//! the standard's 6.2.5 (coding a generic region), 7.2 (segment headers), 7.4.1, 7.4.6 and 7.4.8
//! (the segments) and D.3 (the embedded organisation PDF uses). Decoding is in [`decode`].
//! arithmetic coding ([`crate::mq`]); or the image's marks as symbols ([`symbols`]), which a group
//! of images may share through `/JBIG2Globals`. Written from the standard's 6.2.5 (coding a generic
//! region), 7.2 (segment headers), 7.4.1, 7.4.6 and 7.4.8 (the segments) and D.3 (the embedded
//! organisation PDF uses). Decoding is in [`decode`].
mod decode;
mod symbols;
@@ -12,9 +13,11 @@ pub use decode::{MAX_JBIG2_PIXELS, decode_jbig2_within};
use crate::FilterError;
use crate::mq::{Context, MqEncoder};
/// The widest row [`encode_jbig2_generic`] encodes. JBIG2 allows 32-bit widths; this keeps a
/// row's buffers small and its positions in range.
pub const MAX_JBIG2_COLUMNS: u32 = 1 << 20;
/// The widest row, and the most rows, [`encode_jbig2_generic`] encodes: what `hayro-jbig2`, the
/// decoder behind `JBIG2Decode` here, takes on a side. JBIG2 allows 32-bit sizes, but past this
/// nothing here could read the stream, and the optimizer, which doesn't read back what it stores,
/// replaced such an image with one that draws nothing.
pub const MAX_JBIG2_COLUMNS: u32 = 65_535;
/// Template 0's pixels (Figure 3), as (x, y) offsets from the pixel being coded, each a bit of the
/// context in this order: the four on the current row, the five fixed ones on the row above, the
@@ -54,9 +57,7 @@ const PAD: usize = 4;
/// Errors: as [`encode_jbig2_generic`].
pub fn encode_jbig2(data: &[u8], columns: u32, rows: u32) -> Result<Vec<u8>, FilterError> {
let generic = encode_jbig2_generic(data, columns, rows)?;
// `encode_jbig2_generic` checked that the rows are there.
let stride = (columns as usize).div_ceil(8);
let samples = usize::try_from(rows).ok().and_then(|r| r.checked_mul(stride)).and_then(|n| data.get(..n));
let samples = rows_to_encode(data, columns, rows).ok();
Ok(match samples.and_then(|s| symbols::encode(s, columns, rows)) {
Some(coded) if coded.len() < generic.len() => coded,
_ => generic,
@@ -72,18 +73,13 @@ pub fn encode_jbig2(data: &[u8], columns: u32, rows: u32) -> Result<Vec<u8>, Fil
/// samples have it. Bytes after the last row are ignored, as are the bits that pad a row to a whole
/// byte.
///
/// Errors: no rows or columns, or rows wider than [`MAX_JBIG2_COLUMNS`]
/// ([`FilterError::Unsupported`]); fewer bytes than the rows need ([`FilterError::Corrupt`]).
/// Errors: no rows or columns, more of either than [`MAX_JBIG2_COLUMNS`], or more samples than
/// [`MAX_JBIG2_PIXELS`], what the decoder takes ([`FilterError::Unsupported`]); fewer bytes than the
/// rows need ([`FilterError::Corrupt`]).
pub fn encode_jbig2_generic(data: &[u8], columns: u32, rows: u32) -> Result<Vec<u8>, FilterError> {
if columns == 0 || rows == 0 || columns > MAX_JBIG2_COLUMNS {
return Err(FilterError::Unsupported(format!("JBIG2 encoding of {columns} × {rows} samples")));
}
let samples = rows_to_encode(data, columns, rows)?;
let width = columns as usize;
let stride = width.div_ceil(8);
let samples = usize::try_from(rows).ok().and_then(|r| r.checked_mul(stride)).and_then(|n| data.get(..n)).ok_or_else(|| FilterError::Corrupt {
filter: "JBIG2Decode",
detail: format!("{} bytes for {rows} rows of {columns} samples", data.len()),
})?;
let mut contexts: Vec<Context> = vec![0; 1 << 16];
let mut mq = MqEncoder::new();
let height = samples.len() / stride;
@@ -101,6 +97,48 @@ pub fn encode_jbig2_generic(data: &[u8], columns: u32, rows: u32) -> Result<Vec<
Ok(out)
}
/// The samples of `rows` rows of `columns` 1-bit samples in `data`, when the encoders here take
/// that size; the errors are [`encode_jbig2_generic`]'s.
fn rows_to_encode(data: &[u8], columns: u32, rows: u32) -> Result<&[u8], FilterError> {
let too_large = columns > MAX_JBIG2_COLUMNS || rows > MAX_JBIG2_COLUMNS || u64::from(columns) * u64::from(rows) > MAX_JBIG2_PIXELS;
if columns == 0 || rows == 0 || too_large {
return Err(FilterError::Unsupported(format!("JBIG2 encoding of {columns} × {rows} samples")));
}
let stride = (columns as usize).div_ceil(8);
usize::try_from(rows)
.ok()
.and_then(|r| r.checked_mul(stride))
.and_then(|n| data.get(..n))
.ok_or_else(|| FilterError::Corrupt { filter: "JBIG2Decode", detail: format!("{} bytes for {rows} rows of {columns} samples", data.len()) })
}
/// A group of images coded by [`encode_jbig2_shared`]: the data of the `/JBIG2Globals` stream
/// they share, and each image's own stream, in the order given, or `None` for one that shares
/// nothing with the others.
#[derive(Debug)]
pub struct SharedJbig2 {
pub globals: Vec<u8>,
pub images: Vec<Option<Vec<u8>>>,
}
/// Images, each `rows` rows of `columns` 1-bit PDF samples, as lossless JBIG2 sharing one symbol
/// dictionary: the marks found in more than one of them are stored once, in a symbol dictionary
/// segment not associated with a page, which is the `/JBIG2Globals` stream's data, and each image's
/// stream, in the embedded organisation, places them from there, with its other marks coded as
/// [`encode_jbig2`]'s symbol coding codes them (see [`symbols`]). With the globals, each decodes to
/// its own bits. An image that holds none of those marks, or that symbol coding doesn't take, has
/// no stream here; it is stored alone. `None` when the images share fewer than two marks.
///
/// What each image's stream saves is left for the caller to weigh against what the globals take,
/// which the images that use them share.
///
/// Errors: as [`encode_jbig2_generic`], for the first image it would refuse.
pub fn encode_jbig2_shared(images: &[(&[u8], u32, u32)]) -> Result<Option<SharedJbig2>, FilterError> {
let images =
images.iter().map(|(data, columns, rows)| Ok((rows_to_encode(data, *columns, *rows)?, *columns, *rows))).collect::<Result<Vec<_>, _>>()?;
Ok(symbols::encode_shared(&images).map(|(globals, images)| SharedJbig2 { globals, images }))
}
/// A page information segment's data (7.4.8): the page's size, resolution unknown, eventually
/// lossless, default pixel 0 (white), OR combination, not striped.
fn page_information(columns: u32, rows: u32) -> Vec<u8> {
@@ -141,12 +179,18 @@ fn at_pixels(out: &mut Vec<u8>) {
}
}
/// A segment (7.2): its header (number, type, the segments it refers to, page 1, data length),
/// then its data. `retained`: a later segment refers to this one. `refers_to` holds each
/// A segment (7.2) of page 1: its header (number, type, the segments it refers to, the page, data
/// length), then its data. `retained`: a later segment refers to this one. `refers_to` holds each
/// referred-to segment's number and whether a segment after this one refers to it too (its
/// retain bit, 7.2.4). Segment numbers here stay below 256, so each referred-to one takes a byte
/// (7.2.5); a segment refers to at most four, so their count and retain bits take one byte too.
fn segment(out: &mut Vec<u8>, number: u32, kind: u8, refers_to: &[(u8, bool)], retained: bool, data: &[u8]) {
segment_of(out, 1, number, kind, refers_to, retained, data);
}
/// [`segment`], of page `page`: 0 for a segment associated with no page (7.2.6), such as the
/// global ones `/JBIG2Globals` holds, which a page's segments may refer to.
fn segment_of(out: &mut Vec<u8>, page: u8, number: u32, kind: u8, refers_to: &[(u8, bool)], retained: bool, data: &[u8]) {
out.extend_from_slice(&number.to_be_bytes());
out.push(kind);
// At most four, so the count fits its three bits.
@@ -154,8 +198,8 @@ fn segment(out: &mut Vec<u8>, number: u32, kind: u8, refers_to: &[(u8, bool)], r
let kept = refers_to.iter().enumerate().fold(0u8, |bits, (k, (_, kept))| bits | u8::from(*kept) << (k + 1));
out.push((refers_to.len() as u8) << 5 | kept | u8::from(retained));
out.extend(refers_to.iter().map(|(n, _)| *n));
out.push(1);
// A segment's data never comes near 4 GiB here: a row is at most 1 MiB wide.
out.push(page);
// A segment's data never comes near 4 GiB here: an image is at most 2^28 samples.
out.extend_from_slice(&u32::try_from(data.len()).unwrap_or(u32::MAX).to_be_bytes());
out.extend_from_slice(data);
}
@@ -212,6 +256,11 @@ mod tests {
/// The rows an independent decoder (`hayro-jbig2`) reads from `data`, as PDF samples: true
/// for white (a JBIG2 0).
pub(super) fn decode(data: &[u8]) -> Vec<Vec<bool>> {
decode_with(data, None)
}
/// [`decode`], with `/JBIG2Globals`.
pub(super) fn decode_with(data: &[u8], globals: Option<&[u8]>) -> Vec<Vec<bool>> {
struct Rows(Vec<Vec<bool>>, Vec<bool>, usize);
impl hayro_jbig2::Decoder for Rows {
fn push_pixel(&mut self, black: bool) {
@@ -226,7 +275,7 @@ mod tests {
self.0.push(row);
}
}
let image = hayro_jbig2::Image::new_embedded(data, None).unwrap();
let image = hayro_jbig2::Image::new_embedded(data, globals).unwrap();
let mut rows = Rows(Vec::new(), Vec::new(), image.width() as usize);
image.decode(&mut rows).unwrap();
rows.0
@@ -343,9 +392,56 @@ mod tests {
fn what_it_cannot_encode_is_an_error() {
assert!(matches!(encode_jbig2_generic(&[0; 8], 0, 1), Err(FilterError::Unsupported(_))));
assert!(matches!(encode_jbig2_generic(&[0; 8], 8, 0), Err(FilterError::Unsupported(_))));
assert!(matches!(encode_jbig2_generic(&[0; 8], MAX_JBIG2_COLUMNS + 1, 1), Err(FilterError::Unsupported(_))));
assert!(matches!(encode_jbig2_generic(&[0; 3], 16, 2), Err(FilterError::Corrupt { .. })), "two rows of 16 need 4 bytes");
assert!(matches!(encode_jbig2_generic(&[], 1, u32::MAX), Err(FilterError::Corrupt { .. })));
assert!(matches!(encode_jbig2_generic(&[], 1, 2), Err(FilterError::Corrupt { .. })));
}
/// The encoder stops where the decoder does, on each side and in total: past that it would
/// write a stream nothing here can read. 16,384 squared is exactly the bound on samples.
/// `encode_jbig2_shared` refuses what `encode_jbig2_generic` refuses: an image taller than
/// the decoder takes is `Unsupported`, one at the bound isn't refused. (Its one tall mark is
/// too large to be a symbol, so nothing is shared: `Ok(None)`.)
#[test]
fn shared_coding_takes_only_what_the_decoder_takes() {
let past = vec![0x7F; MAX_JBIG2_COLUMNS as usize + 1];
let tall = (&past[..], 1, MAX_JBIG2_COLUMNS + 1);
assert!(matches!(encode_jbig2_shared(&[tall, tall]), Err(FilterError::Unsupported(_))), "too tall");
let at = (&past[..MAX_JBIG2_COLUMNS as usize], 1, MAX_JBIG2_COLUMNS);
assert!(matches!(encode_jbig2_shared(&[at, at]), Ok(None)), "at the bound");
}
#[test]
fn it_encodes_only_what_the_decoder_takes() {
let unsupported = |columns, rows| matches!(encode_jbig2_generic(&[], columns, rows), Err(FilterError::Unsupported(_)));
assert!(unsupported(MAX_JBIG2_COLUMNS + 1, 1), "too wide");
assert!(unsupported(1, MAX_JBIG2_COLUMNS + 1), "too tall");
assert!(unsupported(16_385, 16_384), "too many samples, each side within the bound");
assert!(unsupported(u32::MAX, u32::MAX), "the product doesn't overflow");
// At the bounds what it writes reads back, through the decoder that has to read it.
let wide = vec![0x0F; (MAX_JBIG2_COLUMNS as usize).div_ceil(8) * 4];
assert_eq!(decode(&encode_jbig2_generic(&wide, MAX_JBIG2_COLUMNS, 4).unwrap()), unpack(&wide, MAX_JBIG2_COLUMNS, 4));
let tall = vec![0x7F; MAX_JBIG2_COLUMNS as usize];
assert_eq!(decode(&encode_jbig2_generic(&tall, 1, MAX_JBIG2_COLUMNS).unwrap()), unpack(&tall, 1, MAX_JBIG2_COLUMNS));
struct Rows(u32, u64);
impl hayro_jbig2::Decoder for Rows {
fn push_pixel(&mut self, black: bool) {
self.1 += u64::from(black);
}
fn push_pixel_chunk(&mut self, black: bool, chunks: u32) {
self.1 += u64::from(black) * u64::from(chunks) * 8;
}
fn next_line(&mut self) {
self.0 += 1;
}
}
// One black pixel, so a page that came back blank wouldn't pass.
let mut largest = vec![0xFF; 2_048 * 16_384];
largest[0] = 0x7F;
let largest = encode_jbig2_generic(&largest, 16_384, 16_384).unwrap();
let mut rows = Rows(0, 0);
hayro_jbig2::Image::new_embedded(&largest, None).unwrap().decode(&mut rows).unwrap();
assert_eq!(rows.0, 16_384);
assert_eq!(rows.1, 1);
}
proptest::proptest! {
crates/filters/src/jbig2/decode.rs
+273 -1
View File
@@ -4,7 +4,8 @@ use crate::FilterError;
use crate::ccitt::Rows;
/// The largest JBIG2 page [`decode_jbig2_within`] decodes, in pixels. The decoder decodes a page whole,
/// and takes pages of up to about this many (20,000 × 10,000 decodes).
/// and takes pages of up to this many (20,000 × 10,000 decodes). It is also the most
/// [`encode_jbig2_generic`](crate::encode_jbig2_generic) encodes, so what it stores reads back.
pub const MAX_JBIG2_PIXELS: u64 = 1 << 28;
/// A `JBIG2Decode` stream's `data`, with what its `/JBIG2Globals` stream decodes to, decoded as the
@@ -143,6 +144,143 @@ mod tests {
assert_eq!(decode_jbig2(&symbols, None, w, h, usize::MAX).unwrap(), marks);
}
/// A region's location is unsigned (ISO/IEC 14492 5.4.4), so one past `i32::MAX` is off the
/// page. The decoder took it for a signed one: a region at 0xFFFFFFFC came back onto the page
/// four pixels left of it, and a refinement anywhere past `i32::MAX` was refused, and the page
/// with it. Each stream is a page of 16 × 12 covered by a generic region, then a region of 8 × 6
/// of one kind past `i32::MAX`, coded by the standard's procedures, and decodes to that page
/// alone; moved onto the page, the region draws. Vendored hayro-jbig2 patch 14.
#[test]
fn a_region_located_past_i32_max_is_off_the_page() {
const PAGE: &str = "0000000030000100000013000000100000000c00000000000000004200000000000126000100000035000000100000000c0000000000000000040003fffdff02fefefed906aa2cb6ee4ccb4bc975d02669ca7a9c14c20c638f5a5e1bffac";
let cases = [
// At 0xFFFFFFFC, 1.
("generic", "00000002260001000000220000000800000006fffffffc00000001000003fffdff02fefefe17c876ae6f77ffac"),
// At 1, 0xFFFFFFFE, with its symbol dictionary.
(
"text",
"000000020000010000001e000003fffdff02fefefe00000001000000014fdda14149cec477ed5fffac00000003062002010000001b000000080000000600000001fffffffe00001000000001a943ffac",
),
// At 0xFFFFFFFC, 1, with its pattern dictionary.
(
"halftone",
"000000021000010000000a000101000000011fffac00000003162002010000002e0000000800000006fffffffc0000000100000000000800000006000000000000000001000000d09301774485ffac",
),
// At 0xFFFFFFFC, 1, reading the page there (7.4.7.4): 0s.
("refinement of the page", "000000022a00010000001e0000000800000006fffffffc000000010400ffffffff0bd98f188428ffac"),
// Of an intermediate region at the same place (7.4.7.5), 1, 0xFFFFFFFE.
(
"refinement of a region",
"0000000224000100000023000000080000000600000001fffffffe000003fffdff02fefefe3d81322da16b7fffac000000032a2002010000001e000000080000000600000001fffffffe0000ffffffffe29946b82229ffac",
),
// Of an intermediate refinement of an intermediate region, all at 0xFFFFFFFC, 1.
(
"refinement of a refinement",
"00000002240001000000230000000800000006fffffffc00000001000003fffdff02fefefecdfd921ec8c47fffac00000003282002010000001e0000000800000006fffffffc000000010000ffffffff903cdb447a47ffac000000042a2003010000001e0000000800000006fffffffc000000010000ffffffffdcd6078cab86ffac",
),
];
let hex = |text: &str| (0..text.len() / 2).map(|i| u8::from_str_radix(&text[i * 2..i * 2 + 2], 16).unwrap()).collect::<Vec<u8>>();
let decode = |segments: &str| decode_jbig2(&[hex(PAGE), hex(segments)].concat(), None, 16, 12, usize::MAX).unwrap();
let page = decode("");
assert!(page.iter().any(|&byte| byte != 0xFF), "the page has ink");
for (kind, segments) in cases {
assert_eq!(decode(segments), page, "{kind}");
// Moved onto the page, to 4, 1 or 1, 2, the same regions draw: they decode, and are
// left out for where they are.
let (off, on) = [("fffffffc00000001", "0000000400000001"), ("00000001fffffffe", "0000000100000002")]
.into_iter()
.find(|(off, _)| segments.contains(off))
.unwrap();
assert_ne!(decode(&segments.replace(off, on)), page, "{kind} on the page");
}
// A refinement at 2, 0 of an intermediate region at 0x80000000, 0. A refinement of a
// region refines it where it is, "GRREFERENCEDX 0, GRREFERENCEDY 0" (Table 35), and the two
// should be at one place (7.4.7.5). Where they aren't, it decodes as it does where they are.
const ELSEWHERE: &str = "000000022400010000002300000008000000068000000000000000000003fffdff02fefefe1bca3149568cdfffac000000032a2002010000001e000000080000000600000002000000000000ffffffff784a909cc761ffac";
let refined = decode(ELSEWHERE);
assert_eq!(refined, decode(&ELSEWHERE.replace("8000000000000000", "0000000200000000")), "the region at 2, 0 too");
assert_ne!(refined, page, "drawn");
}
/// A refinement of the page refines "the current contents of the page buffer, restricted to the
/// area of the page buffer specified by this segment's region segment information field"
/// (ISO/IEC 14492 7.4.7.4), at an offset of 0 (Table 35), and reads 0 outside that area
/// (6.3.5.2). The decoder read the page around the area too, and refused a refinement wider or
/// taller than the page. A refinement of another region refines that region, not the page. And
/// a reference 32 or more left of a refinement (GRREFERENCEDX below −31, as a symbol refined
/// with an RDX below −31 has, Table 18) is read from where each row starts in it: each row's
/// first context read 0 for the reference column it starts at, and from 33 left for the one
/// before it too. Each stream is coded by the standard's procedures and decodes to its
/// picture, `#` for black. Vendored hayro-jbig2 patch 16.
#[test]
fn a_refinement_refines_the_page_in_its_area_and_any_reference_where_it_is() {
let cases: [(&str, &str, &[&str]); 5] = [
(
"a 48 × 6 page a generic region has inked, refined at 33: ink on both sides of the area, template 0",
"000000003000010000001300000030000000060000000000000000420000000000012600010000004100000030000000060000000000000000040003fffdff02fefefe331433558a5c5056cf412003ec9a38164df39bd4463b9331502c4fb1c267ab77b665aa5962ffac000000022a0001000000210000000c0000000600000021000000000400ffffffff9a6c35bbeee221f687ffac",
&[
".#..##.######..#.#..#####.##..#........##.##..##",
".#..####.####.#...###.##.##.#.#.....#.#.####.###",
"##.#......#..#....#..#.###.#..#.#####.###.######",
"......#....###.#.##..#.##........#.#.#...#####..",
"#.##..###.#.#.##.....#...#.#..#.##.#..#...#...#.",
".###...#..#..#..##..#.#..#####.#.##.......###.##",
],
),
(
"a 24 × 8 inked page refined at 5, 1, template 1",
"000000003000010000001300000018000000080000000000000000420000000000012600010000003500000018000000080000000000000000040003fffdff02fefefeb1e133cb8bcbb8510e81d219bb880dfb40a311dc5c298819efffac000000022a00010000001b0000000a0000000500000005000000010401e55c0078dcd30fffac",
&[
"#.##.##...##....##.##..#",
"#.##.##...#.####..##..#.",
"##..##.##...#...###..#.#",
".....#.#.##.#...###....#",
"..#.#...#.#...#.#...##.#",
"....######...#.####....#",
"..######..##.#.#..##.#..",
"...#..##.####.#..##...#.",
],
),
(
"a 16 × 5 inked page refined by a region 4 wider than it, template 0",
"000000003000010000001300000010000000050000000000000000420000000000012600010000002600000010000000050000000000000000040003fffdff02fefefe22e5e253a429cbf2c701ffac000000022a000100000023000000140000000500000000000000000400ffffffff3afa516532bbe3d481926fffac",
&[".#.#.#.#..#.#.#.", "##.####..###...#", "..####.#..####.#", ".###.#.....#..#.", "#.###..#########"],
),
(
"an intermediate region at 4, 1 on an inked 20 × 8 page, refined there (OR): the region, not the page, template 0",
"000000003000010000001300000014000000080000000000000000420000000000012600010000003000000014000000080000000000000000040003fffdff02fefefeb91e434e1726ca89ed06c2bffb0fe5fa7fccc80fffac00000002240001000000230000000a000000050000000400000001000003fffdff02fefefece7a9826ba81bfffac000000032a2002010000001e0000000a0000000500000004000000010000ffffffffd046ee6f6190ffac",
&[
"#.#.#....#.#...#....",
"##..#.#.#.##.....##.",
"##.#.#.#.#.###.##...",
"###.##.####..###.#.#",
"...#.###..###.#.#..#",
"#...#######.#...####",
"..##.##...#..##..#.#",
"#...##...#.##....#..",
],
),
(
"a symbol refined in a symbol dictionary from a 40 × 6 one, RDX −33: 33 left, template 1",
"0000000030000100000013000000070000000600000000000000004200000000000100000100000036000003fffdff02fefefe00000001000000014f913296209e2187f37bddb8589c8beb7d70109de7124ebe47d04914ef7e655c3c4cffac00000002002001010000001f100203fffdff02fefefe00000001000000014fe451a27bddf49f7ef7e7ffac00000003062002010000001b0000000700000006000000000000000004001000000001a943ffac",
&[".####.#", ".##..##", "......#", "...#..#", "#.#.###", "....##."],
),
];
let hex = |text: &str| (0..text.len() / 2).map(|i| u8::from_str_radix(&text[i * 2..i * 2 + 2], 16).unwrap()).collect::<Vec<u8>>();
for (form, stream, picture) in cases {
let (w, h) = (picture[0].len(), picture.len());
// As the filter gives it: black as 0, each row padded to a byte with 0s.
let mut rows = vec![0u8; w.div_ceil(8) * h];
for (y, row) in picture.iter().enumerate() {
for (x, _) in row.bytes().enumerate().filter(|(_, c)| *c == b'.') {
rows[y * w.div_ceil(8) + x / 8] |= 0x80 >> (x % 8);
}
}
assert_eq!(decode_jbig2(&hex(stream), None, w as u32, h as u32, usize::MAX).unwrap(), rows, "{form}");
}
}
/// What a page costs is the same however much a run has left. Given exactly that it decodes,
/// and less stops it: spending all that's left, and what comes after isn't read. A page that
/// fails still spends what it took.
@@ -182,6 +320,57 @@ mod tests {
assert_eq!(u64::MAX - work, 4 * 14);
}
/// A refinement whose reference is offset right of it (GRREFERENCEDX above 0, as a refined text
/// instance has from a wider bitmap, ISO/IEC 14492 Table 12, and a refined symbol from an RDX
/// above 0, Table 18) reads the reference from its first column. Left of the reference the
/// decoder emptied what it had read of it, and at its first column didn't read it again, so its
/// first 32 columns read as 0: such a refinement decoded to other pixels than were coded, or
/// failed. Each stream is a refinement coded by the standard's procedures on a page of its own
/// size, and decodes to its picture, `#` for black. Vendored hayro-jbig2 patch 15.
#[test]
fn a_refinement_reads_a_reference_right_of_it_from_its_first_column() {
let cases: [(&str, &str, &[&str]); 4] = [
(
"a text instance of an 8 × 6 symbol refined 6 wider: 3 right, template 0",
"00000000300001000000130000000e000000060000000000000000420000000000010000010000001d000003fffdff02fefefe00000001000000014fdd3dba48bdddde35ffac00000002062001010000002c0000000e000000060000000000000000040012ffffffff00000001a96b4a6b78b7fc094fadde1b3e881fffac",
&[".#.#.#....#...", "##.###....##.#", ".###...###..#.", "#.......#...#.", "#.####.#...#..", "#.##...###.#.."],
),
(
"refined 2 wider and 2 taller, RDX and RDY 1: 2 right, 2 down, template 1",
"00000000300001000000130000000a000000080000000000000000420000000000010000010000001d000003fffdff02fefefe00000001000000014fdd9c52d3ea827a43ffac0000000206200101000000280000000a00000008000000000000000004801200000001a96c00cdac6dd5f6f848ce872411cfffac",
&["####.##...", "####..####", "####..##..", "...##...#.", "#......#..", "..#.....#.", "##....####", ".#........"],
),
(
"a symbol refined in a symbol dictionary, RDX 5, RDY −1, template 0 with its AT pixels moved",
"00000000300001000000130000000c000000060000000000000000420000000000010000010000001e000003fffdff02fefefe00000001000000014fdda13efc4029efc95fffac000000020020010100000027000203fffdff02fefefefeff010100000001000000014fc98c352dd84a906f120ffb49793fffac00000003062002010000001b0000000c00000006000000000000000004001000000001a943ffac",
&["#.#.##.#.###", "..#.#.....##", "##..#...#...", "..##....##..", "##.#.##.####", ".#..#####..#"],
),
(
"a text instance of a 40 × 5 symbol refined 2 wider, RDX 1: 2 right, across a word, template 0",
"00000000300001000000130000002a0000000500000000000000004200000000000100000100000031000003fffdff02fefefe00000001000000015240f590d953ae3f4069dede672090d4d61b377906d3f41aafa4f77107ffac00000002062001010000003c0000002a000000050000000000000000040012ffffffff00000001a96c0b9386d4b140486bdc3673dbd031490e468c7f7813d19068ab04a79cbfffac",
&[
"##.#.#..##.##.#####.#....#.#.#..#..######.",
".#.#..###....#..##.#..#.####.#..#.###...##",
".#.#......#....#####....#...###...#####...",
"#...##.##.......#...###.###...##..##.###.#",
"..#..###..###..#...#.......##.#........##.",
],
),
];
let hex = |text: &str| (0..text.len() / 2).map(|i| u8::from_str_radix(&text[i * 2..i * 2 + 2], 16).unwrap()).collect::<Vec<u8>>();
for (form, stream, picture) in cases {
let (w, h) = (picture[0].len(), picture.len());
// As the filter gives it: black as 0, each row padded to a byte with 0s.
let mut rows = vec![0u8; w.div_ceil(8) * h];
for (y, row) in picture.iter().enumerate() {
for (x, _) in row.bytes().enumerate().filter(|(_, c)| *c == b'.') {
rows[y * w.div_ceil(8) + x / 8] |= 0x80 >> (x % 8);
}
}
assert_eq!(decode_jbig2(&hex(stream), None, w as u32, h as u32, usize::MAX).unwrap(), rows, "{form}");
}
}
/// The page has to be the image's size, and is bounded before it is decoded, being decoded
/// whole: within `max` as rows of one bit, and within [`MAX_JBIG2_PIXELS`]. Globals are given
/// to the decoder: ones it skips decode, ones that aren't JBIG2 are an error, as is damaged
@@ -213,4 +402,87 @@ mod tests {
let _ = decode_jbig2(&data[..end], None, w, h, usize::MAX);
}
}
/// Pixels past a bitmap's edge are 0 (ISO/IEC 14492 6.3.5.2). A page whose default pixel is 1,
/// and an intermediate text or halftone region whose default pixel is 1, kept 1s past their
/// width in their last word of each row, and a refinement, which reads its reference a word at
/// a time, read them as pixels: of rows not a multiple of 32 wide, it decoded to other pixels
/// than were coded. Each stream is a refinement coded by the standard's procedures from a
/// reference read as 0 past its edge, and decodes to the picture given, `#` for black.
/// Vendored hayro-jbig2 patch 13.
#[test]
fn a_refinement_reads_nothing_past_its_references_width() {
// A page of 7 × 6, then of 33 × 6, whose default pixel is 1, refined whole.
const PAGE_7: &str = "000000003000010000001300000007000000060000000000000000460000000000012a00010000001e000000070000000600000000000000000400ffffffffe42d7e889ebcffac";
const PAGE_33: &str = "000000003000010000001300000021000000060000000000000000460000000000012a000100000034000000210000000600000000000000000400ffffffff314a8d598037eb3f915ab3d526ccae915ad6a6a7b814111b124ae60fffac";
// An intermediate text region of 31 × 6 whose default pixel is 1, then one of 7 × 6 on a
// page of 40, each refined into the page.
const TEXT_31: &str = "00000000300001000000130000001f0000000600000000000000004200000000000100000100000030000003fffdff02fefefe00000001000000014f96dceb76c7f1ff585971969cd9f88e538261a6f713be68d77e8e1fffac00000002042001010000001b0000001f00000006000000000000000000029000000001a943ffac000000032a2002010000002f0000001f0000000600000000000000000400ffffffff9ac311688adbc8f570a2f96aef93d332e3b1f4b27b704bffac";
const TEXT_7: &str = "000000003000010000001300000028000000060000000000000000420000000000010000010000001d000003fffdff02fefefe00000001000000014fe1f79f4a95c5b77fffac00000002042001010000001b0000000700000006000000000000000000029000000001a943ffac000000032a2002010000001e000000070000000600000000000000000400ffffffffa71dae12bd7fffac";
// An intermediate halftone region of 33 × 6 whose default pixel is 1, refined into the page.
const HALFTONE_33: &str = "000000003000010000001300000021000000060000000000000000420000000000011000010000000a000101000000011fffac00000002142001010000004100000021000000060000000000000000009000000021000000060000000000000000010000004ea6618f551d48ff24c2c5c98124d449f2adea6fca76e4297fffac000000032a20020100000032000000210000000600000000000000000400ffffffffe87db809393a3251e565e7a749e6470ab37f7aef186ceaf5167bffac";
let cases: [(&str, &str, [&str; 6]); 5] = [
("page 7", PAGE_7, ["###.#.#", ".......", "#.####.", "######.", "#.#...#", ".#...##"]),
(
"page 33",
PAGE_33,
[
".#..##........###..####....#.##.#",
".#.###.###.##.##..#..#...#.....##",
".....##..#....#.##.######.###..#.",
"##...#..##.###..###.##..#.#.#.#.#",
".##.#........##.#.#.....###......",
"...#.#####.#..#.###......###..###",
],
),
(
"text 31",
TEXT_31,
[
"....##....###......#.##.#.#....",
".#.....##.......##..###.#...#..",
".#.#.#.####...#####.####...#.#.",
"......#...#...#.##...##...#..#.",
".#..##.###.#.###..####...#...#.",
"#...#...#..#.#....#.##.#.....#.",
],
),
(
"text 7 on a page of 40",
TEXT_7,
[
"......#.................................",
".....##.................................",
"##...##.................................",
"..#.....................................",
"##......................................",
"..#.....................................",
],
),
(
"halftone 33",
HALFTONE_33,
[
"##...#.###.#.###.....#.####..####",
"##.##.#...#.##.##.##.#...#.#.###.",
"###..#.#....##.###...#..##.##..##",
"##..#...###...##..#.####.###.#...",
"####.###.###.#..#..#.#.####...###",
"#.#..#........#....#.#.##.##..#.#",
],
),
];
let hex = |text: &str| (0..text.len() / 2).map(|i| u8::from_str_radix(&text[i * 2..i * 2 + 2], 16).unwrap()).collect::<Vec<u8>>();
for (form, stream, picture) in cases {
let w = picture[0].len();
// As the filter gives it: black as 0, each row padded to a byte with 0s.
let mut rows = vec![0u8; w.div_ceil(8) * picture.len()];
for (y, row) in picture.iter().enumerate() {
for (x, _) in row.bytes().enumerate().filter(|(_, c)| *c == b'.') {
rows[y * w.div_ceil(8) + x / 8] |= 0x80 >> (x % 8);
}
}
assert_eq!(decode_jbig2(&hex(stream), None, w as u32, 6, usize::MAX).unwrap(), rows, "{form}");
}
}
}
crates/filters/src/jbig2/symbols.rs
+367 -67
View File
@@ -11,7 +11,7 @@ use std::collections::HashMap;
use std::hash::{DefaultHasher, Hash, Hasher};
use std::ops::Range;
use super::{at_pixels, code_generic, generic_region, page_information, region_information, segment};
use super::{at_pixels, code_generic, generic_region, page_information, region_information, segment, segment_of};
use crate::mq::{Context, MqEncoder};
/// The longest side of a mark stored as a symbol. A larger one is left to the generic region.
@@ -27,7 +27,9 @@ const MAX_SYMBOLS: usize = 1 << 16;
const MAX_SYMBOL_PIXELS: usize = 1 << 24;
/// The most symbols one plan codes. The decoder PdfCraft renders JBIG2 with (`hayro-jbig2`)
/// refuses 32,768 symbols or more, and others may too, so a plan with more isn't made.
/// refuses a dictionary whose new and exported symbols pass 65,535 together, and a plan exports
/// every symbol it codes, so it refuses 32,768 or more; others may too, so a plan with more isn't
/// made.
const MAX_CODED_SYMBOLS: usize = (1 << 15) - 1;
/// The most symbol instances one text region places. The decoder PdfCraft renders JBIG2 with
@@ -35,14 +37,14 @@ const MAX_CODED_SYMBOLS: usize = (1 << 15) - 1;
/// placed by several regions, each over a band of rows, all referring to the one dictionary.
const MAX_REGION_INSTANCES: usize = 10_000;
// Segment numbers stay at most 256, so a referred-to one takes a byte (7.2.5): the page, the
// dictionary, a text region for each MAX_REGION_INSTANCES of at most MAX_RUNS marks, and the
// generic region.
const _: () = assert!(MAX_RUNS.div_ceil(MAX_REGION_INSTANCES) + 3 <= 256);
/// The most symbols a group's shared dictionary holds ([`encode_shared`]): half of what a page may
/// place, so each page keeps room for its own.
const MAX_SHARED_SYMBOLS: usize = MAX_CODED_SYMBOLS / 2;
/// The most rows symbol coded. With rows at most 2^20 samples wide, every coordinate and
/// difference coded then fits the 32-bit integers Annex A codes.
const MAX_ROWS: u32 = 1 << 24;
// Segment numbers stay at most 256, so a referred-to one takes a byte (7.2.5): the shared
// dictionary, the page, its own dictionary, a text region for each MAX_REGION_INSTANCES of at most
// MAX_RUNS marks, and the generic region.
const _: () = assert!(MAX_RUNS.div_ceil(MAX_REGION_INSTANCES) + 4 <= 256);
/// The integer coders' ranges (Table A.1): how many 1 bits pick the range, how many bits give the
/// offset into it, and its first value. The last range has no 0 after its prefix.
@@ -99,20 +101,111 @@ struct Instance {
height: u32,
}
/// A group of images coded by [`encode_shared`]: the globals' data, and each image's stream.
type SharedCoding = (Vec<u8>, Vec<Option<Vec<u8>>>);
/// The symbols of a group's shared dictionary, as one image's [`plan`] places them: the ID each of
/// the image's symbols has there, if it is there, and how many symbols that dictionary exports.
struct Shared {
ids: Vec<Option<u32>>,
count: u32,
}
/// `rows` rows of `columns` 1-bit PDF samples (black 0), symbol coded: the smaller of two plans,
/// every mark of at most [`MAX_SYMBOL_SIDE`] a side as a symbol, or only those with a copy.
/// `None` when neither plan has from two to [`MAX_CODED_SYMBOLS`] symbols, or the image is too
/// large or too busy to look at.
pub(super) fn encode(samples: &[u8], columns: u32, rows: u32) -> Option<Vec<u8>> {
let page = page(samples, columns, rows)?;
[1, 2].into_iter().filter_map(|copies| plan(&page, columns, rows, copies, None)).min_by_key(Vec::len)
let page = page(samples, columns)?;
[1, 2].into_iter().filter_map(|copies| plan(&page, columns, rows, copies, None, None)).min_by_key(Vec::len)
}
/// Images, each as [`encode`] takes one, symbol coded sharing one dictionary: the globals' data, a
/// symbol dictionary segment associated with no page (number 0) holding the distinct marks found
/// in more than one image, and each image's stream placing them from there, the smaller of
/// [`encode`]'s two plans for its other marks. The group's marks are held within the limits one
/// image's are ([`MAX_SYMBOLS`], [`MAX_SYMBOL_PIXELS`]): past them, later marks aren't shared. Past
/// [`MAX_SHARED_SYMBOLS`] shared marks, those in the most images are kept.
///
/// An image too large or too busy to look at, or that places none of the shared marks, gets `None`;
/// `None` overall when fewer than two marks are shared.
pub(super) fn encode_shared(images: &[(&[u8], u32, u32)]) -> Option<SharedCoding> {
// The group's distinct marks, each with how many images hold it (as its uses), looked at an
// image at a time.
let mut table: Vec<Symbol> = Vec::new();
let mut by_hash: HashMap<u64, Vec<u32>> = HashMap::new();
let mut held = 0usize;
for (samples, columns, _) in images {
let Some(page) = page(samples, *columns) else { continue };
// An image's symbols are distinct, so each counts an image once.
for symbol in page.symbols {
match find(&table, &by_hash, &symbol) {
Some(t) => {
if let Some(s) = table.get_mut(t as usize) {
s.uses = s.uses.saturating_add(1);
}
}
None => {
let Some(more) = held.checked_add(symbol.pixels.len()).filter(|h| *h <= MAX_SYMBOL_PIXELS && table.len() < MAX_SYMBOLS) else {
continue;
};
held = more;
let t = u32::try_from(table.len()).ok()?;
by_hash.entry(fingerprint(&symbol)).or_default().push(t);
table.push(Symbol { uses: 1, ..symbol });
}
}
}
}
let uses = |t: &u32| table.get(*t as usize).map_or(0, |s| s.uses);
let mut shared: Vec<u32> = (0..u32::try_from(table.len()).ok()?).filter(|t| uses(t) >= 2).collect();
if shared.len() < 2 {
return None;
}
shared.sort_by_key(|t| std::cmp::Reverse(uses(t)));
shared.truncate(MAX_SHARED_SYMBOLS);
// The dictionary's order, as `plan` keeps it.
shared.sort_by_key(|t| table.get(*t as usize).map(|s| (s.height, s.width)));
let mut global = vec![None; table.len()];
for (k, t) in shared.iter().enumerate() {
if let Some(slot) = global.get_mut(*t as usize) {
*slot = u32::try_from(k).ok();
}
}
let symbols: Vec<&Symbol> = shared.iter().filter_map(|t| table.get(*t as usize)).collect();
let mut globals = Vec::new();
segment_of(&mut globals, 0, 0, 0, &[], true, &symbol_dictionary(&symbols)?);
let count = u32::try_from(symbols.len()).ok()?;
let streams = images
.iter()
.map(|(samples, columns, rows)| {
let page = page(samples, *columns)?;
let ids = page.symbols.iter().map(|s| find(&table, &by_hash, s).and_then(|t| global.get(t as usize).copied().flatten())).collect();
let shared = Shared { ids, count };
[1, 2].into_iter().filter_map(|copies| plan(&page, *columns, *rows, copies, None, Some(&shared))).min_by_key(Vec::len)
})
.collect();
Some((globals, streams))
}
/// What sorts a symbol's bitmap ([`symbols`] and [`encode_shared`] look exact copies up by it).
fn fingerprint(s: &Symbol) -> u64 {
let mut hasher = DefaultHasher::new();
(s.width, s.height, &s.pixels).hash(&mut hasher);
hasher.finish()
}
/// The index in `table` of the symbol with `s`'s bitmap, looked up by its [`fingerprint`].
fn find(table: &[Symbol], by_hash: &HashMap<u64, Vec<u32>>, s: &Symbol) -> Option<u32> {
by_hash
.get(&fingerprint(s))?
.iter()
.copied()
.find(|t| table.get(*t as usize).is_some_and(|t| (t.width, t.height) == (s.width, s.height) && t.pixels == s.pixels))
}
/// The image's marks and symbols; `None` past the limits above.
fn page(samples: &[u8], columns: u32, rows: u32) -> Option<Page> {
if rows > MAX_ROWS {
return None;
}
fn page(samples: &[u8], columns: u32) -> Option<Page> {
let width = usize::try_from(columns).ok()?;
let stride = width.div_ceil(8);
if stride == 0 {
@@ -272,23 +365,18 @@ fn symbols(page: &mut Page) -> Option<()> {
page.large.push(m);
continue;
}
let pixels = pixels(mark, &page.runs, &page.order);
let mut hasher = DefaultHasher::new();
(width, height, &pixels).hash(&mut hasher);
let same = by_hash.entry(hasher.finish()).or_default();
let symbol = Symbol { width, height, pixels: pixels(mark, &page.runs, &page.order), uses: 0 };
let symbols = &mut page.symbols;
let known =
same.iter().copied().find(|s| symbols.get(*s as usize).is_some_and(|s| (s.width, s.height) == (width, height) && s.pixels == pixels));
let id = match known {
let id = match find(symbols, &by_hash, &symbol) {
Some(id) => id,
None => {
kept = kept.checked_add(pixels.len())?;
kept = kept.checked_add(symbol.pixels.len())?;
if kept > MAX_SYMBOL_PIXELS || symbols.len() == MAX_SYMBOLS {
return None;
}
let id = u32::try_from(symbols.len()).ok()?;
symbols.push(Symbol { width, height, pixels, uses: 0 });
same.push(id);
by_hash.entry(fingerprint(&symbol)).or_default().push(id);
symbols.push(symbol);
id
}
};
@@ -305,18 +393,28 @@ fn symbols(page: &mut Page) -> Option<()> {
/// the other marks in a generic region. `None` with fewer than two such symbols (their IDs would
/// take no bits at all, SBSYMCODELEN 0, a corner of the standard left alone here) or more than
/// [`MAX_CODED_SYMBOLS`].
fn plan(page: &Page, columns: u32, rows: u32, copies: u32, log_strips: Option<u32>) -> Option<Vec<u8>> {
///
/// With `shared`, the symbols a group's dictionary holds are placed from there, whatever their uses
/// here, and the page's own dictionary holds only the others (none, when it has none), as
/// [`encode_shared`] writes them; `None` too for a page that places none of the shared ones.
fn plan(page: &Page, columns: u32, rows: u32, copies: u32, log_strips: Option<u32>, shared: Option<&Shared>) -> Option<Vec<u8>> {
let symbol = |s: u32| page.symbols.get(s as usize);
let global = |s: u32| shared.and_then(|g| g.ids.get(s as usize).copied().flatten());
let globals = shared.map_or(0, |g| g.count);
// The dictionary's order (6.5.5): height classes from the shortest, each narrowest first.
let mut kept: Vec<u32> = (0..page.symbols.len() as u32).filter(|s| symbol(*s).is_some_and(|s| s.uses >= copies)).collect();
if kept.len() < 2 || kept.len() > MAX_CODED_SYMBOLS {
let mut kept: Vec<u32> =
(0..page.symbols.len() as u32).filter(|s| global(*s).is_none() && symbol(*s).is_some_and(|s| s.uses >= copies)).collect();
let count = usize::try_from(globals).ok()?.checked_add(kept.len())?;
if !(2..=MAX_CODED_SYMBOLS).contains(&count) {
return None;
}
kept.sort_by_key(|s| symbol(*s).map(|s| (s.height, s.width)));
let mut id = vec![None; page.symbols.len()];
// The shared symbols' IDs come first, as their dictionary is referred to first (Table 31's
// SBSYMS).
let mut id: Vec<Option<u32>> = (0..page.symbols.len() as u32).map(global).collect();
for (k, s) in kept.iter().enumerate() {
if let Some(slot) = id.get_mut(*s as usize) {
*slot = u32::try_from(k).ok();
*slot = u32::try_from(k).ok().and_then(|k| k.checked_add(globals));
}
}
let mut instances = Vec::new();
@@ -329,9 +427,12 @@ fn plan(page: &Page, columns: u32, rows: u32, copies: u32, log_strips: Option<u3
_ => left.push(*m),
}
}
if shared.is_some() && !instances.iter().any(|i| i.id < globals) {
return None;
}
let kept: Vec<&Symbol> = kept.iter().filter_map(|s| symbol(*s)).collect();
let dictionary = symbol_dictionary(&kept)?;
let count = u32::try_from(kept.len()).ok()?;
let dictionary = if kept.is_empty() { None } else { Some(symbol_dictionary(&kept)?) };
let count = u32::try_from(count).ok()?;
// Bands of at most MAX_REGION_INSTANCES, from the top.
instances.sort_by_key(|i| (i.t, i.s));
let mut regions = Vec::new();
@@ -344,12 +445,21 @@ fn plan(page: &Page, columns: u32, rows: u32, copies: u32, log_strips: Option<u3
});
}
let mut out = Vec::new();
segment(&mut out, 0, 48, &[], false, &page_information(columns, rows));
segment(&mut out, 1, 0, &[], true, &dictionary);
let mut number = 2;
// The shared dictionary is segment 0, in the globals, and a segment refers only to lower
// numbers (7.2.5): the page's own follow it.
let mut number = u32::from(shared.is_some());
segment(&mut out, number, 48, &[], false, &page_information(columns, rows));
number += 1;
let mut dictionaries: Vec<u8> = shared.map(|_| 0).into_iter().collect();
if let Some(dictionary) = &dictionary {
segment(&mut out, number, 0, &[], true, dictionary);
dictionaries.push(u8::try_from(number).ok()?);
number += 1;
}
for (k, text) in regions.iter().enumerate() {
// The dictionary stays retained for the next region, until the last (7.2.4).
segment(&mut out, number, 7, &[(1, k + 1 < regions.len())], false, text);
// The dictionaries stay retained for the next region, until the last (7.2.4).
let refers_to: Vec<(u8, bool)> = dictionaries.iter().map(|n| (*n, k + 1 < regions.len())).collect();
segment(&mut out, number, 7, &refers_to, false, text);
number += 1;
}
if !left.is_empty() {
@@ -543,8 +653,9 @@ fn encode_id(mq: &mut MqEncoder, contexts: &mut [Context], length: u32, id: u32)
#[cfg(test)]
mod tests {
use super::*;
use crate::jbig2::tests::{decode, samples, unpack};
use crate::jbig2::{encode_jbig2, encode_jbig2_generic};
use crate::FilterError;
use crate::jbig2::tests::{decode, decode_with, samples, unpack};
use crate::jbig2::{encode_jbig2, encode_jbig2_generic, encode_jbig2_shared};
type Bitmap = Vec<Vec<u8>>;
@@ -665,11 +776,11 @@ mod tests {
fn every_plan_round_trips(picture: &Bitmap) -> usize {
let (columns, rows) = (picture[0].len() as u32, picture.len() as u32);
let data = samples(picture);
let page = page(&data, columns, rows).unwrap();
let page = page(&data, columns).unwrap();
let mut plans = 0;
for copies in [1, 2] {
for log_strips in 0..4 {
if let Some(coded) = plan(&page, columns, rows, copies, Some(log_strips)) {
if let Some(coded) = plan(&page, columns, rows, copies, Some(log_strips), None) {
assert_eq!(decode(&coded), unpack(&data, columns, rows), "{copies} copies, strips of {}", 1 << log_strips);
plans += 1;
}
@@ -735,10 +846,10 @@ mod tests {
row[stride - 1] &= 0xFF << (8 * stride - columns);
}
let (c, r) = (columns as u32, 120);
let page = page(&data, c, r).unwrap();
let page = page(&data, c).unwrap();
assert!(page.marks.iter().all(|m| m.x1 as usize <= columns), "{columns} columns");
for copies in [1, 2] {
if let Some(coded) = plan(&page, c, r, copies, None) {
if let Some(coded) = plan(&page, c, r, copies, None, None) {
assert_eq!(decode(&coded), unpack(&data, c, r), "{columns} columns");
}
}
@@ -771,7 +882,7 @@ mod tests {
row[290] = 1;
}
let data = samples(&picture);
let page = page(&data, 300, 300).unwrap();
let page = page(&data, 300).unwrap();
let boxes: Vec<(u32, u32, u32, u32)> = page.marks.iter().map(|m| (m.x0, m.y0, m.x1 - m.x0, m.y1 - m.y0)).collect();
assert_eq!(
boxes,
@@ -792,18 +903,14 @@ mod tests {
assert_eq!(every_plan_round_trips(&picture), 8);
}
/// Past 2^20 black runs or 2^24 rows the image is left to generic coding; at either cap it is
/// still looked at.
/// Past 2^20 black runs the image is left to generic coding; at the cap it is still looked at.
#[test]
fn busy_or_tall_images_are_left_to_generic_coding() {
fn busy_images_are_left_to_generic_coding() {
// 0x55: four black pixels a byte, each its own run, under the same ones on the next row.
let columns = 1 << 20;
let row = vec![0x55u8; columns / 8];
assert!(page(&row.repeat(2), columns as u32, 2).is_some(), "2^20 runs");
assert!(page(&row.repeat(3), columns as u32, 3).is_none(), "1.5 × 2^20 runs");
let tall = vec![0xFFu8; MAX_ROWS as usize + 1];
assert!(page(&tall, 1, MAX_ROWS + 1).is_none());
assert!(page(&tall, 1, MAX_ROWS).is_some());
assert!(page(&row.repeat(2), columns as u32).is_some(), "2^20 runs");
assert!(page(&row.repeat(3), columns as u32).is_none(), "1.5 × 2^20 runs");
}
/// A plan codes at most 32,767 symbols, which an independent decoder reads back; with more,
@@ -828,18 +935,18 @@ mod tests {
(samples(&picture), columns as u32, rows as u32)
};
let (data, columns, rows) = codes(MAX_CODED_SYMBOLS, 0);
let at_cap = page(&data, columns, rows).unwrap();
assert_eq!(decode(&plan(&at_cap, columns, rows, 1, None).unwrap()), unpack(&data, columns, rows));
let at_cap = page(&data, columns).unwrap();
assert_eq!(decode(&plan(&at_cap, columns, rows, 1, None, None).unwrap()), unpack(&data, columns, rows));
// One more distinct mark, each once: neither plan is made.
let (data, columns, rows) = codes(MAX_CODED_SYMBOLS + 1, 0);
let past = page(&data, columns, rows).unwrap();
assert!(plan(&past, columns, rows, 1, None).is_none() && plan(&past, columns, rows, 2, None).is_none());
let past = page(&data, columns).unwrap();
assert!(plan(&past, columns, rows, 1, None, None).is_none() && plan(&past, columns, rows, 2, None, None).is_none());
assert_eq!(encode_jbig2(&data, columns, rows).unwrap(), encode_jbig2_generic(&data, columns, rows).unwrap());
// With two of them repeated, the second plan codes those two.
let (data, columns, rows) = codes(MAX_CODED_SYMBOLS + 1, 200);
let past = page(&data, columns, rows).unwrap();
assert!(plan(&past, columns, rows, 1, None).is_none());
assert_eq!(decode(&plan(&past, columns, rows, 2, None).unwrap()), unpack(&data, columns, rows));
let past = page(&data, columns).unwrap();
assert!(plan(&past, columns, rows, 1, None, None).is_none());
assert_eq!(decode(&plan(&past, columns, rows, 2, None, None).unwrap()), unpack(&data, columns, rows));
// Above a page of text whose letters repeat, that second plan is the one `encode_jbig2`
// keeps: the first, of every mark, would code too many symbols.
let mut picture: Bitmap = vec![vec![0u8; 256 * 18]; 512];
@@ -857,10 +964,10 @@ mod tests {
let coded = encode_jbig2(&data, columns, rows).unwrap();
assert_eq!(segment_types(&coded)[1], 0, "symbol coded");
assert_eq!(decode(&coded), unpack(&data, columns, rows));
let (data, columns, rows) = codes(MAX_SYMBOLS, 0);
assert_eq!(page(&data, columns, rows).map(|p| p.symbols.len()), Some(MAX_SYMBOLS));
let (data, columns, rows) = codes(MAX_SYMBOLS + 1, 0);
assert!(page(&data, columns, rows).is_none());
let (data, columns, _) = codes(MAX_SYMBOLS, 0);
assert_eq!(page(&data, columns).map(|p| p.symbols.len()), Some(MAX_SYMBOLS));
let (data, columns, _) = codes(MAX_SYMBOLS + 1, 0);
assert!(page(&data, columns).is_none());
// `n` black squares of 256 × 256, each with a white pixel in a different place: 2^16
// pixels a symbol, so 256 of them are exactly the cap.
let squares = |n: usize| {
@@ -876,11 +983,11 @@ mod tests {
(samples(&picture), columns as u32, rows as u32)
};
let (data, columns, rows) = squares(256);
let at_cap = page(&data, columns, rows).unwrap();
let at_cap = page(&data, columns).unwrap();
assert_eq!(at_cap.symbols.len(), 256);
assert_eq!(decode(&plan(&at_cap, columns, rows, 1, None).unwrap()), unpack(&data, columns, rows));
let (data, columns, rows) = squares(257);
assert!(page(&data, columns, rows).is_none());
assert_eq!(decode(&plan(&at_cap, columns, rows, 1, None, None).unwrap()), unpack(&data, columns, rows));
let (data, columns, _) = squares(257);
assert!(page(&data, columns).is_none());
}
/// More than 10,000 marks are placed by several text regions, each of at most 10,000 (more
@@ -993,7 +1100,199 @@ mod tests {
}
}
/// Each segment's page association (7.2.6): the byte after the numbers it refers to.
fn associations(data: &[u8]) -> Vec<u8> {
let (mut pages, mut at) = (Vec::new(), 0);
while at < data.len() {
let refers = usize::from(data[at + 5] >> 5);
pages.push(data[at + 6 + refers]);
let length_at = at + 6 + refers + 1;
at = length_at + 4 + u32::from_be_bytes(data[length_at..length_at + 4].try_into().unwrap()) as usize;
}
pages
}
/// `pictures`, `columns` × `rows` each, coded sharing a dictionary; each image's stream, which
/// has to be there, decodes with the globals to its picture.
fn shared_round_trip(pictures: &[Bitmap]) -> (Vec<u8>, Vec<Vec<u8>>) {
let (columns, rows) = (pictures[0][0].len() as u32, pictures[0].len() as u32);
let data: Vec<Vec<u8>> = pictures.iter().map(|p| samples(p)).collect();
let images: Vec<(&[u8], u32, u32)> = data.iter().map(|d| (&d[..], columns, rows)).collect();
let shared = encode_jbig2_shared(&images).unwrap().unwrap();
let (globals, coded) = (shared.globals, shared.images);
let coded: Vec<Vec<u8>> = coded.into_iter().map(Option::unwrap).collect();
for (d, c) in data.iter().zip(&coded) {
assert_eq!(decode_with(c, Some(&globals)), unpack(d, columns, rows));
}
(globals, coded)
}
/// `font` with a stem 20 pixels tall down each glyph's left side: a font sharing no glyph with
/// any [`glyphs`] gives, whose glyphs are at most 18 pixels tall.
fn stemmed(font: &[Bitmap]) -> Vec<Bitmap> {
font.iter()
.map(|g| {
(0..g.len().max(20)).map(|y| std::iter::once(1).chain(g.get(y).cloned().unwrap_or_else(|| vec![0; g[0].len()])).collect()).collect()
})
.collect()
}
/// Mark `k` of a set of distinct ones: a stroke 33 pixels long with a pixel above every other
/// column for each 1 bit of `k`, below 2^17.
fn comb(k: usize) -> Bitmap {
let mut g = vec![vec![0u8; 33]; 2];
g[1].fill(1);
for bit in 0..17 {
g[0][2 * bit] = u8::from(k >> bit & 1 == 1);
}
g
}
/// The combs `ks`, 120 a row, the rows four pixels apart.
fn combs(ks: impl ExactSizeIterator<Item = usize>) -> Bitmap {
let mut page = vec![vec![0u8; 120 * 36]; ks.len().div_ceil(120) * 4];
for (i, k) in ks.enumerate() {
draw(&mut page, &comb(k), (i % 120 * 36) as i64, (i / 120 * 4) as i64);
}
page
}
/// The symbols a symbol dictionary segment exports (SDNUMEXSYMS, 7.4.2.1.4), in a stream
/// that is that segment alone: after its header, flags and adaptive pixels.
fn exported(globals: &[u8]) -> usize {
u32::from_be_bytes(globals[21..25].try_into().unwrap()) as usize
}
/// Pages in one font share its letters: one symbol dictionary, segment 0, associated with no
/// page and retained (7.2.4, 7.2.6), which each page's text region refers to first, its own
/// segments numbered after it (7.2.5). Together they take less than they take alone.
#[test]
fn pages_in_one_font_share_one_dictionary() {
let font = glyphs(40, 7);
let pictures: Vec<Bitmap> = (0..4).map(|seed| text(1200, 300, &font, 20 + seed)).collect();
let (globals, coded) = shared_round_trip(&pictures);
assert_eq!(segments(&globals), [(0, 0, true, vec![])]);
assert_eq!(associations(&globals), [0]);
let (mut shared, mut alone) = (globals.len(), 0);
for (picture, c) in pictures.iter().zip(&coded) {
let segments = segments(c);
assert_eq!(segments.first().map(|s| s.0), Some(1), "the page's own segments follow");
assert!(segments.iter().filter(|s| s.1 == 7).all(|s| s.3.first() == Some(&(0, false))), "{segments:?}");
assert!(associations(c).iter().all(|page| *page == 1));
shared += c.len();
alone += encode_jbig2(&samples(picture), 1200, 300).unwrap().len();
}
assert!(shared < alone, "shared {shared} bytes, alone {alone}");
}
/// Two pages alike place every mark from the shared dictionary and have none of their own; a
/// page with letters of its own as well keeps them in its own dictionary, referred to after the
/// shared one (its symbols' IDs follow), and the page with nothing shared has no stream.
#[test]
fn what_a_page_has_alone_stays_its_own() {
let (font, other) = (glyphs(30, 7), stemmed(&glyphs(20, 99)));
let same = text(900, 240, &font, 3);
let mut more = text(900, 240, &font, 4);
more.extend(text(900, 120, &other, 5));
let (globals, coded) = shared_round_trip(&[same.clone(), same.clone()]);
assert!(!globals.is_empty());
assert!(coded.iter().all(|c| segment_types(c) == [48, 7]), "{:?}", coded.iter().map(|c| segment_types(c)).collect::<Vec<_>>());
let with_more = |p: &Bitmap| p.iter().cloned().chain(std::iter::repeat_n(vec![0u8; 900], 120)).collect::<Bitmap>();
let (_, coded) = shared_round_trip(&[with_more(&same), more]);
assert_eq!(segment_types(&coded[1])[..3], [48, 0, 7]);
assert_eq!(segments(&coded[1])[2].3, [(0, false), (2, false)]);
let alone = text(900, 360, &other, 6);
let data: Vec<Vec<u8>> = [&same, &same, &alone].iter().map(|p| samples(p)).collect();
let images: Vec<(&[u8], u32, u32)> = vec![(&data[0], 900, 240), (&data[1], 900, 240), (&data[2], 900, 360)];
let (_, coded) = encode_shared(&images).unwrap();
assert!(coded[0].is_some() && coded[1].is_some() && coded[2].is_none());
}
/// Images that share fewer than two marks share nothing; two are enough.
#[test]
fn images_with_nothing_in_common_share_nothing() {
let font = glyphs(30, 7);
let (a, b) = (samples(&text(600, 200, &font, 1)), samples(&text(600, 200, &stemmed(&font), 1)));
assert!(encode_shared(&[(&a, 600, 200), (&b, 600, 200)]).is_none());
assert!(encode_shared(&[(&a, 600, 200)]).is_none(), "one image");
assert!(encode_shared(&[]).is_none());
// The two commonest letters of the first page, below the last line of the second.
let mut two = text(600, 200, &stemmed(&font), 1);
draw(&mut two, &font[0], 10, 180);
draw(&mut two, &font[1], 100, 180);
let two = samples(&two);
let (globals, coded) = encode_shared(&[(&a, 600, 200), (&two, 600, 200)]).unwrap();
assert_eq!(exported(&globals), 2);
assert_eq!(decode_with(coded[1].as_ref().unwrap(), Some(&globals)), unpack(&two, 600, 200));
}
/// The images of a group are checked as `encode_jbig2_generic` checks one, and an image it
/// would refuse refuses the group.
#[test]
fn a_group_with_an_image_the_encoder_refuses_is_refused() {
let page = samples(&text(600, 200, &glyphs(30, 7), 1));
assert!(matches!(encode_jbig2_shared(&[(&page, 600, 200), (&page[..10], 600, 200)]), Err(FilterError::Corrupt { .. })));
assert!(matches!(encode_jbig2_shared(&[(&page, 600, 200), (&page, 0, 200)]), Err(FilterError::Unsupported(_))));
assert!(encode_jbig2_shared(&[(&page, 600, 200), (&page, 600, 200)]).unwrap().is_some());
}
/// A group holds its distinct marks within what one image's are held to: past 65,536 a mark
/// first seen later isn't held, so isn't shared, while those held count every image they're
/// in. Ten marks of the first image are in two images after it, with 464 first seen there.
#[test]
fn past_the_marks_a_group_holds_later_ones_arent_shared() {
let (a, b) = (samples(&combs(0..40_000)), samples(&combs(40_000..65_536)));
let late = combs((0..10).chain(65_536..66_000).collect::<Vec<_>>().into_iter());
let (rows, late) = (late.len() as u32, samples(&late));
let (columns, a_rows, b_rows) = (120 * 36, 40_000u32.div_ceil(120) * 4, 25_536u32.div_ceil(120) * 4);
let images = [(&a[..], columns, a_rows), (&b[..], columns, b_rows), (&late[..], columns, rows), (&late[..], columns, rows)];
let (globals, coded) = encode_shared(&images).unwrap();
assert_eq!(exported(&globals), 10);
for c in &coded[2..] {
assert_eq!(decode_with(c.as_ref().unwrap(), Some(&globals)), unpack(&late, columns, rows));
}
}
/// Past what one dictionary shares, it keeps the marks the most images hold: 20,000 shared,
/// the first 10,000 in all three images and the rest in two. The third image then places all
/// its marks from it, and the other two code the rest themselves (in a dictionary of their own
/// or a generic region, whichever is smaller).
#[test]
fn past_the_shared_limit_the_most_widely_held_marks_are_shared() {
let (both, first) = (combs(0..20_000), combs(0..10_000));
let rows = both.len();
let first: Bitmap = first.iter().cloned().chain(std::iter::repeat_n(vec![0u8; 120 * 36], rows - first.len())).collect();
let (globals, coded) = shared_round_trip(&[both.clone(), both, first]);
assert_eq!(exported(&globals), MAX_SHARED_SYMBOLS);
assert_eq!(segment_types(&coded[2]), [48, 7]);
assert!(segment_types(&coded[0]).iter().any(|t| *t == 0 || *t == 39), "{:?}", segment_types(&coded[0]));
}
proptest::proptest! {
/// Glyphs scattered over pages of one font: whatever is shared, every page with a stream
/// decodes with the globals to its picture.
#[test]
fn shared_groups_round_trip(
columns in 1usize..200,
rows in 1usize..100,
shapes in 2usize..8,
n in 0usize..60,
pages in 2usize..5,
seed in proptest::num::u64::ANY,
) {
let font = glyphs(shapes, seed);
let data: Vec<Vec<u8>> = (0..pages).map(|p| samples(&scatter(columns, rows, &font, seed ^ p as u64, n))).collect();
let (c, r) = (columns as u32, rows as u32);
let images: Vec<(&[u8], u32, u32)> = data.iter().map(|d| (&d[..], c, r)).collect();
if let Some((globals, coded)) = encode_shared(&images) {
for (d, coded) in data.iter().zip(coded) {
if let Some(coded) = coded {
proptest::prop_assert_eq!(decode_with(&coded, Some(&globals)), unpack(d, c, r));
}
}
}
}
/// Glyphs anywhere, over the edges and each other, with stray pixels: every plan decodes to
/// the picture, and `encode_jbig2` is never larger than the generic coding.
#[test]
@@ -1024,6 +1323,7 @@ mod tests {
fn any_input_returns(data in proptest::collection::vec(0u8..=255, 0..600), columns in 0u32..5000, rows in 0u32..80) {
let _ = encode_jbig2(&data, columns, rows);
let _ = encode(&data, columns, rows);
let _ = encode_shared(&[(&data, columns, rows), (&data, columns, rows.saturating_sub(1))]);
}
}
}
crates/filters/src/lib.rs
+3 -2
View File
@@ -27,7 +27,8 @@
//! ```
//!
//! 1-bit samples can be encoded as CCITT Group 4 ([`encode_g4`]) or lossless JBIG2
//! ([`encode_jbig2`]).
//! ([`encode_jbig2`]), a group of images sharing their marks through `/JBIG2Globals`
//! ([`encode_jbig2_shared`]).
//!
//! Every decoder is bounded by a caller-supplied `max_output` (decompression-bomb
//! defence) and never panics on malformed input.
@@ -47,7 +48,7 @@ mod runlength;
use std::borrow::Cow;
pub use ccitt::{CcittParams, MAX_CCITT_COLUMNS, MAX_G4_COLUMNS, decode_ccitt_within, encode_g4};
pub use jbig2::{MAX_JBIG2_COLUMNS, MAX_JBIG2_PIXELS, decode_jbig2_within, encode_jbig2, encode_jbig2_generic};
pub use jbig2::{MAX_JBIG2_COLUMNS, MAX_JBIG2_PIXELS, SharedJbig2, decode_jbig2_within, encode_jbig2, encode_jbig2_generic, encode_jbig2_shared};
/// A PDF stream filter, identified by its `/Filter` name.
#[derive(Clone, Debug, PartialEq, Eq, Hash)]
crates/forms/README.md
+4 -1
View File
@@ -50,7 +50,10 @@ let cells: Vec<table::Cell> = grid.cells_for(names)?; // "Table1_1", "Table
keeps it: `/V` is stored as typed, the appearance draws `?` for those characters, the widget is
marked with the first one (`pdfcraft_fonts::UNDRAWN_MARK`, cleared once all of it can be drawn)
and the form sets `/NeedAppearances`, so other viewers draw the field from its value with fonts
that have them. Not in a document that declares PDF/A, which forbids that flag (`declares_pdfa`).
that have them. Not in a document that declares PDF/A, which forbids that flag (`declares_pdfa`:
the bytes `pdfaid:part` in XMP metadata each filter stage of which decodes within 8 MiB, the cap
preflight's Standards panel reads it with; preflight reads the identifier by its namespace, so a
packet that gives that namespace another prefix declares PDF/A there and not here).
`undrawn_fields` lists the marked fields; the engine reports each edit that draws one, and
flattening refuses to make that `?` permanent.
- Displaying a form that leaves its appearances to the viewer (`/NeedAppearances`, or a text or
crates/forms/src/af.rs
+4 -1
View File
@@ -20,7 +20,10 @@ mod script;
#[cfg(test)]
use calc::{Notation, eval_notation, notation_name, notation_tokens, unique};
pub use calc::{calculate, calculation_reads, notation_js, reads_built_names, rename_in_calculation};
pub use format::{DATE_PRESETS, DateTime, TIME_PRESETS, format_date, format_value, js_number, keystroke, make_number, parse_date, validate};
pub use format::{
DATE_PRESETS, DateNames, DateTime, ENGLISH_DATE_NAMES, TIME_PRESETS, date_names, format_date, format_date_in, format_value, format_value_in,
js_number, keystroke, make_number, parse_date, set_date_names, validate,
};
use script::{Arg, call};
pub use script::{NOTATION_NOT_STORABLE, calculate_js, format_js, notation_storable, parse_calculate, parse_format, parse_validate, validate_js};
crates/forms/src/af/format.rs
+191 -11
View File
@@ -89,6 +89,165 @@ const MONTHS: [&str; 12] = ["January", "February", "March", "April", "May", "Jun
const DAYS: [&str; 7] = ["Sunday", "Monday", "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday"];
/// The month and weekday names `mmmm`, `mmm`, `dddd` and `ddd` write (weekdays from Sunday), as
/// a localized Acrobat writes them in its own language (#5303).
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub struct DateNames {
pub months: [&'static str; 12],
pub months_short: [&'static str; 12],
pub days: [&'static str; 7],
pub days_short: [&'static str; 7],
}
/// English, the names until the app sets others.
pub const ENGLISH_DATE_NAMES: DateNames = DateNames {
months: MONTHS,
months_short: ["Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec"],
days: DAYS,
days_short: ["Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat"],
};
static DATE_NAMES: std::sync::RwLock<DateNames> = std::sync::RwLock::new(ENGLISH_DATE_NAMES);
/// Sets the names date formats write from now on (the app's date language); typed dates are read
/// in them and in English.
pub fn set_date_names(names: DateNames) {
*DATE_NAMES.write().unwrap_or_else(std::sync::PoisonError::into_inner) = names;
}
/// The names date formats write.
pub fn date_names() -> DateNames {
*DATE_NAMES.read().unwrap_or_else(std::sync::PoisonError::into_inner)
}
/// `text` with the month and weekday names of `lists` (in the order months, short months,
/// weekdays, short weekdays) written as their English names, so a parser that reads English
/// reads them. The longest name wins at each place, case ignored, and only a whole name: one
/// that doesn't continue a word, nor end in a digit that a number continues. A name may hold
/// digits ("10月"), spaces ("tháng 10"), dots ("ม.ค.") or combining vowel signs, which a
/// word-by-word reader would split (#5303). Where a weekday and a month share a name (Italian
/// "mar"), the order of the format's name fields (`mmm`, `ddd`) decides. English names are
/// found too, so that one typed ("Tue" before "Mar") fills its field of the format as a
/// localized one does.
fn english_names(text: &str, fmt: &str, lists: [&[&str]; 4]) -> String {
const EN: [&[&str]; 4] = [&MONTHS, &MONTHS, &DAYS, &DAYS];
// No date is this long: hostile text is left as it is rather than scanned name by name.
if text.len() > MAX_DATE_TEXT_BYTES {
return text.to_string();
}
let chars: Vec<char> = text.chars().collect();
let typed_english: [&[&str]; 4] = [&MONTHS, &ENGLISH_DATE_NAMES.months_short, &DAYS, &ENGLISH_DATE_NAMES.days_short];
// Each name as folded chars, whether it is a month, and its English name.
let mut names: Vec<(Vec<String>, bool, &str)> = Vec::new();
for (k, (list, en)) in lists.into_iter().zip(EN).chain(typed_english.into_iter().zip(EN)).enumerate() {
for (name, english) in list.iter().zip(en.iter()) {
// A short name ending in a dot ("mar.") is also read typed without it, as the same
// kind of name, so that where it equals another ("Mar", March) the format decides.
let bare = name.trim_end_matches('.');
for name in std::iter::once(*name).chain((bare.len() < name.len()).then_some(bare)) {
let n: Vec<String> = name.chars().map(fold).collect();
if !n.is_empty() && n.len() <= MAX_NAME_CHARS {
names.push((n, k % 4 < 2, english));
}
}
}
}
names.sort_by_key(|(n, ..)| std::cmp::Reverse(n.len()));
// The format's name fields in order: true for a month, false for a weekday.
let f: Vec<char> = fmt.chars().collect();
let mut expected = Vec::new();
let mut j = 0;
while let Some(&c) = f.get(j) {
let run = f.get(j..).map_or(1, |r| r.iter().take_while(|x| **x == c).count()).max(1);
if run >= 3 && (c == 'm' || c == 'd') {
expected.push(c == 'm');
}
j = j.saturating_add(run);
}
let mut slot = 0;
// A name's first letter may not continue a word; its last may not run on into a letter, nor
// (a digit) into a number.
let letter = |c: Option<&char>| c.is_some_and(|c| c.is_alphabetic());
let mut out = String::with_capacity(text.len());
let mut i = 0;
while let Some(&c) = chars.get(i) {
let fits = |n: &Vec<String>| {
let end = i.saturating_add(n.len());
let first = n.first().and_then(|e| e.chars().next());
let last = n.last().and_then(|e| e.chars().next());
chars.get(i..end).is_some_and(|t| t.iter().map(|c| fold(*c)).eq(n.iter().cloned()))
&& !(first.is_some_and(char::is_alphabetic) && letter(i.checked_sub(1).and_then(|p| chars.get(p))))
&& !match last {
Some(l) if l.is_ascii_digit() => chars.get(end).is_some_and(char::is_ascii_digit),
Some(l) if l.is_alphabetic() => letter(chars.get(end)),
_ => false,
}
};
let longest = names.iter().find(|(n, ..)| fits(n)).map(|(n, ..)| n.len());
let hit = longest.and_then(|len| {
let mut same = names.iter().filter(|(n, ..)| n.len() == len && fits(n));
let want = expected.get(slot).copied();
let all: Vec<_> = same.by_ref().collect();
all.iter().find(|(_, month, _)| Some(*month) == want).or_else(|| all.first()).copied()
});
match hit {
Some((n, month, english)) => {
if expected.get(slot) == Some(month) {
slot += 1;
}
// Apart from what touches it ("10月18"), but no space added at either end.
if out.ends_with(char::is_alphanumeric) {
out.push(' ');
}
out.push_str(english);
i = i.saturating_add(n.len());
if chars.get(i).is_some_and(|c| c.is_alphanumeric()) {
out.push(' ');
}
}
None => {
out.push(c);
i += 1;
}
}
}
out
}
/// `c` as names are compared: lower case, with the capital I and the Turkish dotted and dotless
/// i all "i", and a combining dot above (U+0307) dropped, so that "EKİM", "SALI" and "MAYIS" are
/// Turkish "Ekim", "Salı" and "Mayıs", which lower-casing alone makes "ekim" with a dotted i,
/// "sali" and "mayis" (#5303). The Greek final sigma is a sigma, since a capital Σ, which
/// lower-cases to σ on its own, ends "ΙΑΝΟΥΆΡΙΟΣ" for "Ιανουάριος". One piece for each char, so
/// positions in the text stay aligned.
fn fold(c: char) -> String {
match c {
'I' | 'ı' | 'İ' => "i".to_string(),
'ς' => "σ".to_string(),
_ => c.to_lowercase().filter(|&l| l != '\u{307}').collect(),
}
}
/// `s` [`fold`]ed char by char.
fn folded(s: &str) -> String {
s.chars().map(fold).collect()
}
/// Typed dates longer than this are not searched for names.
const MAX_DATE_TEXT_BYTES: usize = 1024;
/// Names longer than this are not looked for (settings are trusted, but a cap keeps the scan cheap).
const MAX_NAME_CHARS: usize = 32;
/// The month (1 to 12) a typed word names: a start of at least three letters of a month's name,
/// or its whole short name, in `names` or in English.
fn month_named(word: &str, names: &DateNames) -> Option<u32> {
let w = folded(word);
let starts = |list: &[&str; 12]| list.iter().position(|m| w.chars().count() >= 3 && folded(m).starts_with(&w));
let whole = |list: &[&str; 12]| list.iter().position(|m| folded(m).trim_end_matches('.') == w);
starts(&names.months).or_else(|| whole(&names.months_short)).or_else(|| starts(&MONTHS)).map(|i| i as u32 + 1)
}
/// A date and time read from text.
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub struct DateTime {
@@ -202,7 +361,20 @@ fn order(fmt: &str) -> Vec<Part> {
/// Parse typed text against a date/time format (AFDate_KeystrokeEx's leniency: any
/// separators, month names, two-digit years). `None` when it isn't a valid date.
pub fn parse_date(text: &str, fmt: &str) -> Option<DateTime> {
let t = text.trim();
let names = date_names();
let lists: [&[&str]; 4] = [&names.months, &names.months_short, &names.days, &names.days_short];
// Weekdays are shown, not typed (see `order`): where the format has one but none was typed, a
// name the language shares between a weekday and a month (Italian "mar") is the month, as
// the names are read again with no weekday expected (#5303).
read_date(&english_names(text, fmt, lists), fmt, &names).or_else(|| {
let no_weekday = fmt.replace("dddd", " ").replace("ddd", " ");
(no_weekday != fmt).then(|| read_date(&english_names(text, &no_weekday, lists), fmt, &names)).flatten()
})
}
/// [`parse_date`] of `english`, typed text whose month and weekday names are English.
fn read_date(english: &str, fmt: &str, names: &DateNames) -> Option<DateTime> {
let t = english.trim();
if t.is_empty() {
return None;
}
@@ -223,10 +395,8 @@ pub fn parse_date(text: &str, fmt: &str) -> Option<DateTime> {
}
if cur.starts_with(|c: char| c.is_ascii_digit()) {
nums.push_back(std::mem::take(cur));
} else if cur.len() >= 3
&& let Some(i) = MONTHS.iter().position(|m| m.to_lowercase().starts_with(&cur.to_lowercase()))
{
*month = Some(i as u32 + 1);
} else if let Some(m) = month_named(cur, names) {
*month = Some(m);
}
cur.clear();
};
@@ -337,6 +507,11 @@ pub fn parse_date(text: &str, fmt: &str) -> Option<DateTime> {
/// Format a date/time with Acrobat's date format letters.
pub fn format_date(dt: DateTime, fmt: &str) -> String {
format_date_in(dt, fmt, &date_names())
}
/// [`format_date`] with month and weekday names from `names`.
pub fn format_date_in(dt: DateTime, fmt: &str, names: &DateNames) -> String {
let chars: Vec<char> = fmt.chars().collect();
let mut out = String::new();
let mut i = 0;
@@ -354,9 +529,9 @@ pub fn format_date(dt: DateTime, fmt: &str) -> String {
('y', _) => {
let _ = write!(out, "{:02}", dt.y.rem_euclid(100));
}
('m', 3..) => match MONTHS.get((dt.m as usize).wrapping_sub(1)) {
Some(name) if n >= 4 => out.push_str(name),
Some(name) => out.push_str(name.get(..3).unwrap_or(name)),
('m', 3..) => match (dt.m as usize).checked_sub(1).and_then(|i| names.months.get(i).zip(names.months_short.get(i))) {
Some((name, _)) if n >= 4 => out.push_str(name),
Some((_, short)) => out.push_str(short),
None => {
let _ = write!(out, "{}", dt.m);
}
@@ -368,8 +543,8 @@ pub fn format_date(dt: DateTime, fmt: &str) -> String {
let _ = write!(out, "{}", dt.m);
}
('d', 3..) => {
let name = DAYS.get(weekday(dt.y, dt.m, dt.d)).copied().unwrap_or_default();
out.push_str(if n >= 4 { name } else { name.get(..3).unwrap_or(name) });
let list = if n >= 4 { &names.days } else { &names.days_short };
out.push_str(list.get(weekday(dt.y, dt.m, dt.d)).copied().unwrap_or_default());
}
('d', 2) => {
let _ = write!(out, "{:02}", dt.d);
@@ -451,6 +626,11 @@ fn apply_mask(mask: &str, text: &str) -> Option<String> {
/// The text shown for a value (the Format event).
pub fn format_value(f: &Format, value: &str) -> String {
format_value_in(f, value, &date_names())
}
/// [`format_value`] with month and weekday names from `names`.
pub fn format_value_in(f: &Format, value: &str, names: &DateNames) -> String {
if value.is_empty() {
return String::new();
}
@@ -479,7 +659,7 @@ pub fn format_value(f: &Format, value: &str) -> String {
),
None => value.to_string(),
},
Format::Date(fmt) | Format::Time(fmt) => parse_date(value, fmt).map_or_else(|| value.to_string(), |dt| format_date(dt, fmt)),
Format::Date(fmt) | Format::Time(fmt) => parse_date(value, fmt).map_or_else(|| value.to_string(), |dt| format_date_in(dt, fmt, names)),
Format::Special(n) => apply_mask(SPECIAL_MASKS[(*n).min(3) as usize], value).unwrap_or_else(|| value.to_string()),
Format::Mask(m) => apply_mask(m, value).unwrap_or_else(|| value.to_string()),
}
crates/forms/src/appearance.rs
+16 -3
View File
@@ -739,8 +739,18 @@ pub fn value_overflow(doc: &Document, f: &Field, w: &Widget) -> Option<Overflow>
field_appearance_fit(doc, f, w, &f.value, true).2
}
/// Whether `text` draws in widget `wd` of `f` with no `?`: in the field's `/DR` font when that
/// holds it, as [`field_appearance_fit`] picks, or else in the Helvetica (WinAnsi) stand-in.
fn draws(doc: &Document, f: &Field, wd: &Dict, text: &str) -> bool {
let da = parse_da(wd.get(b"DA").and_then(|o| doc.resolve(o).as_string().map(|s| s.to_text())).as_deref().unwrap_or(&f.da));
let enc = dr_font(doc, &da.font).map(|(_, enc)| enc).filter(|enc| enc.can_encode(text)).unwrap_or(FontText::WinAnsi);
enc.first_unencodable(text).is_none()
}
/// [`field_appearance_drawn`], and whether the text drawn is cut off by the box.
fn field_appearance_fit(doc: &Document, f: &Field, w: &Widget, values: &[String], format: bool) -> (Stream, Option<char>, Option<Overflow>) {
let wobj = doc.get(w.obj);
let wd = wobj.as_dict().cloned().unwrap_or_default();
// The Format event: what is shown, not what is stored.
let formatted: Vec<String>;
let values = if format
@@ -748,13 +758,16 @@ fn field_appearance_fit(doc: &Document, f: &Field, w: &Widget, values: &[String]
&& values.len() == 1
&& f.actions.format != crate::af::Format::None
{
formatted = vec![crate::af::format_value(&f.actions.format, &values[0])];
let shown = crate::af::format_value(&f.actions.format, &values[0]);
// Month and weekday names in the date language that no font here can draw ("окт." in
// Helvetica) are drawn in English rather than as `?` (#5303).
let english = || crate::af::format_value_in(&f.actions.format, &values[0], &crate::af::ENGLISH_DATE_NAMES);
let dated = matches!(f.actions.format, crate::af::Format::Date(_) | crate::af::Format::Time(_));
formatted = vec![if !dated || draws(doc, f, &wd, &shown) { shown } else { english() }];
&formatted[..]
} else {
values
};
let wobj = doc.get(w.obj);
let wd = wobj.as_dict().cloned().unwrap_or_default();
let (page_w, page_h) = ((w.rect[2] - w.rect[0]).max(1.0), (w.rect[3] - w.rect[1]).max(1.0));
let place = placement(mk_rotation(doc, &wd), page_w, page_h);
let (width, height) = (place.layout_w, place.layout_h);
crates/forms/src/appearances.rs
+13 -2
View File
@@ -212,13 +212,24 @@ impl FieldAppearanceCache {
}
}
/// Whether the document's XMP metadata declares PDF/A conformance (`pdfaid:part`).
/// The most each filter stage of the XMP metadata stream may decode for its PDF/A declaration to
/// be read, the cap preflight's Standards panel reads it with: past it, metadata declares nothing
/// there either.
pub(crate) const MAX_XMP_BYTES: usize = 8 << 20;
/// Whether the document's XMP metadata declares PDF/A conformance (the bytes `pdfaid:part`).
/// Metadata a filter stage of which decodes past [`MAX_XMP_BYTES`] doesn't.
pub fn declares_pdfa(doc: &Document) -> bool {
let Some(root) = doc.root() else { return false };
let Some(m) = doc.get(root).as_dict().and_then(|c| c.get(b"Metadata").cloned()) else { return false };
let m = doc.resolve(&m);
match &*m {
Object::Stream(s) => s.decoded().is_ok_and(|b| b.windows(11).any(|w| w == b"pdfaid:part")),
// Not a `DecodeBudget`, which counts every stage against one allowance: preflight bounds
// each stage, and the two must agree on metadata stored through a chain of filters.
// `/DecodeParms` is read only when direct, as preflight and organize read metadata too:
// PDF/A asks for it unfiltered, so only a file that breaks that reaches parameters here,
// and the readers go to `with_direct_parms` together or they disagree.
Object::Stream(s) => s.decoded_within(MAX_XMP_BYTES).is_ok_and(|b| b.windows(11).any(|w| w == b"pdfaid:part")),
_ => false,
}
}
Loaded 100 of 328 files, more files were not shown because too many files have changed in this diff. Show more