From 3275f853c3a941872d3d3738aa5dc15be4139654 Mon Sep 17 00:00:00 2001 From: mtlcl410 <2368091+mtlcl410@users.noreply.github.com> Date: Mon, 2 Nov 2026 02:02:45 +0800 Subject: [PATCH] M2: an Indexed image whose lookup table is short or whose hival is out of range is drawn as other readers draw it, with a warning (ISO 32000-2 8.6.6.3) Patch (60) skipped such an image. A reviewer's side-by-side tests on #4393 (seven cases against main and Chrome's viewer) showed the skip dropped colours other readers draw, and regressed a real hival that main draws. Indexed::new now recovers instead: - a lookup table shorter than m x (hival + 1) bytes, missing, or not a string or decodable stream is padded with zero bytes; - a hival past 255 is taken as 255; - a real hival is truncated toward zero (as upstream read it). The space records that it was recovered. ImageXObject::new raises MalformedIndexedImage from the space it built; a cs/CS colour in such a space reports ColorFallback, as the upstream DeviceGray fallback did. A hival below zero or not a number has no reading: the space falls back as upstream, and the image reports the warning. Index values past hival stay clamped (8.6.6.3). The separate is_malformed_indexed check is gone, so no lookup table is read twice: inline images naming an Indexed resource read its table once through the resource cache (CodeRabbit, interpret/mod.rs:945). The read count is exposed as InterpreterCache::indexed_lookup_reads. Tests: indexed_lookup.rs rewritten as draw tests (17, cases A-G and bounds); raster::tests::indexed_lookup_cache. On b6f93a3ae logic 14 of the 17 fail (the three that pass are guards: full table, case C, device name) and the cache test reads [21, 41] instead of [1, 1]. Co-Authored-By: Claude Opus 5.5 Claude-Session: https://claude.ai/code/session_01Pq7HyCfwzyRXj4VfQAVVdv --- apps/pdfcraft-cli/src/commands.rs | 5 +- crates/automation/src/pages.rs | 2 +- crates/automation/tests/it/render_warnings.rs | 6 +- crates/render/src/raster.rs | 8 +- crates/render/src/raster/tests.rs | 4 + .../src/raster/tests/indexed_lookup_cache.rs | 63 ++++ crates/render/tests/it/indexed_lookup.rs | 305 +++++++++++------- vendor/README.md | 3 +- vendor/hayro-interpret/src/cache.rs | 22 +- vendor/hayro-interpret/src/color.rs | 148 +++++---- vendor/hayro-interpret/src/context.rs | 7 + vendor/hayro-interpret/src/interpret/mod.rs | 33 +- vendor/hayro-interpret/src/x_object.rs | 28 +- 13 files changed, 424 insertions(+), 210 deletions(-) create mode 100644 crates/render/src/raster/tests/indexed_lookup_cache.rs diff --git a/apps/pdfcraft-cli/src/commands.rs b/apps/pdfcraft-cli/src/commands.rs index c8e449933c..896d41ca72 100644 --- a/apps/pdfcraft-cli/src/commands.rs +++ b/apps/pdfcraft-cli/src/commands.rs @@ -408,7 +408,10 @@ pub(crate) fn render(args: &[String]) -> Result<(), CliError> { writeln!(std::io::stderr().lock(), "warning: page {page}: a colour that could not be computed was painted in a fallback colour"); } pdfcraft_render::RenderWarning::MalformedIndexedImage => { - let _ = writeln!(std::io::stderr().lock(), "warning: page {page}: an image with a malformed Indexed colour table was not drawn"); + let _ = writeln!( + std::io::stderr().lock(), + "warning: page {page}: an image with a malformed Indexed colour table was drawn, and its colours may be wrong" + ); } } } diff --git a/crates/automation/src/pages.rs b/crates/automation/src/pages.rs index 6559935f83..17bb14cee7 100644 --- a/crates/automation/src/pages.rs +++ b/crates/automation/src/pages.rs @@ -345,7 +345,7 @@ fn render_warning_text(warning: pdfcraft_render::RenderWarning) -> &'static str match warning { RenderWarning::ContentTruncated => "content past the page's safety budget was skipped", RenderWarning::ColourFallback => "a colour that could not be computed was painted in a fallback colour", - RenderWarning::MalformedIndexedImage => "an image with a malformed Indexed colour table was not drawn", + RenderWarning::MalformedIndexedImage => "an image with a malformed Indexed colour table was drawn, and its colours may be wrong", } } diff --git a/crates/automation/tests/it/render_warnings.rs b/crates/automation/tests/it/render_warnings.rs index dd9c158d6b..a7f68e9314 100644 --- a/crates/automation/tests/it/render_warnings.rs +++ b/crates/automation/tests/it/render_warnings.rs @@ -96,10 +96,10 @@ fn a_clean_page_renders_to_the_png_alone() { let _ = std::fs::remove_dir_all(dir); } -/// An image whose Indexed lookup table is too short (ISO 32000-2 §8.6.6.3) is skipped, and +/// An image whose Indexed lookup table is too short (ISO 32000-2 §8.6.6.3) is drawn padded, and /// `page_render` says so in the words `pdfcraft-cli render` prints. #[test] -fn page_render_reports_a_skipped_malformed_indexed_image() { +fn page_render_reports_a_recovered_malformed_indexed_image() { let dir = std::env::temp_dir().join(format!("pdfcraft-render-warnings-indexed-{}", std::process::id())); let _ = std::fs::remove_dir_all(&dir); std::fs::create_dir_all(&dir).unwrap(); @@ -119,7 +119,7 @@ fn page_render_reports_a_skipped_malformed_indexed_image() { let doc = open(&mut a, "indexed.pdf"); let out = a.call("page_render", &json!({ "doc": doc, "page": 1, "dpi": 72 })).unwrap(); let Some(Content::Json(report)) = out.get(1) else { panic!("expected the warnings as JSON: {out:?}") }; - assert_eq!(report["warnings"], json!(["an image with a malformed Indexed colour table was not drawn"])); + assert_eq!(report["warnings"], json!(["an image with a malformed Indexed colour table was drawn, and its colours may be wrong"])); let _ = std::fs::remove_dir_all(dir); } diff --git a/crates/render/src/raster.rs b/crates/render/src/raster.rs index 413d6773a0..f9e0622133 100644 --- a/crates/render/src/raster.rs +++ b/crates/render/src/raster.rs @@ -260,9 +260,11 @@ pub enum RenderWarning { /// A colour could not be computed (a colour space that cannot be read, or a Separation or /// DeviceN tint transform that cannot be evaluated) and was painted in a fallback colour. ColourFallback, - /// An image was not drawn because its `/Indexed` colour space is malformed (ISO 32000-2 - /// §8.6.6.3): `hival` not an integer in 0..=255, or a lookup table shorter than - /// m × (hival + 1) bytes. Reported once per page however many images it affects. + /// An image's `/Indexed` colour space is malformed (ISO 32000-2 §8.6.6.3) and the image was + /// drawn recovered, so its colours may be wrong: a lookup table shorter than m × (hival + 1) + /// bytes padded with zero bytes, a `hival` past 255 taken as 255, a real one truncated, or + /// (a `hival` below zero or not a number) the image drawn in DeviceGray. Reported once per + /// page however many images it affects. MalformedIndexedImage, } diff --git a/crates/render/src/raster/tests.rs b/crates/render/src/raster/tests.rs index df6896b980..3e808ed388 100644 --- a/crates/render/src/raster/tests.rs +++ b/crates/render/src/raster/tests.rs @@ -6275,6 +6275,10 @@ fn decoded_images_serve_every_size_they_are_drawn_at() { // ICC profiles whose stream dictionaries are the same (vendored hayro-interpret patch 48). mod icc_profile_cache; +// An Indexed resource named by many inline images reads its lookup once (vendored +// hayro-interpret patch 60). +mod indexed_lookup_cache; + // A symbolic TrueType font draws through its (3, 0) cmap subtable first (#4156). mod symbolic_truetype_cmap; diff --git a/crates/render/src/raster/tests/indexed_lookup_cache.rs b/crates/render/src/raster/tests/indexed_lookup_cache.rs new file mode 100644 index 0000000000..af772b0126 --- /dev/null +++ b/crates/render/src/raster/tests/indexed_lookup_cache.rs @@ -0,0 +1,63 @@ +//! An `/Indexed` colour space named by a resource is built once, and its lookup table read +//! once, however many inline images name it (vendored hayro-interpret patch 60). An earlier +//! version of the patch checked the table again before every inline image, decoding a lookup +//! stream once per image outside the colour-space cache. + +use super::image_cache_pdf; + +/// A page that draws `count` 2 × 1 inline images in the resource colour space `/P`, an Indexed +/// space over DeviceRGB with `hival` whose lookup is stream object 5 holding `table`. +fn page(count: usize, hival: u8, table: &[u8]) -> Vec { + let image = "q 20 0 0 20 0 0 cm BI /W 2 /H 1 /BPC 8 /CS /P /F /AHx ID 0001> EI Q"; + let content = vec![image; count].join("\n"); + let resources = format!("<< /ColorSpace << /P [/Indexed /DeviceRGB {hival} 5 0 R] >> >>"); + let dict = format!("<< /Length {} >>", table.len()); + image_cache_pdf(&content, &resources, &[(&dict, table)]) +} + +fn lookup_reads(pdf: Vec) -> [usize; 2] { + use hayro::hayro_interpret::font::Glyph; + use hayro::hayro_interpret::hayro_syntax::Pdf; + use hayro::hayro_interpret::{ + BlendMode, ClipPath, Context, Device, GlyphDrawMode, Image, InterpreterCache, InterpreterSettings, Paint, PathDrawMode, SoftMask, + TransformExt, interpret_page, + }; + use kurbo::{Affine, BezPath, Rect}; + use std::sync::Arc; + + /// Draws nothing: only the colour spaces the page reads matter here. + struct Nothing; + impl<'a> Device<'a> for Nothing { + fn set_soft_mask(&mut self, _: Option>) {} + fn set_blend_mode(&mut self, _: BlendMode) {} + fn draw_path(&mut self, _: &BezPath, _: Affine, _: &Paint<'a>, _: &PathDrawMode) {} + fn push_clip_path(&mut self, _: &ClipPath) {} + fn push_transparency_group(&mut self, _: f32, _: Option>, _: BlendMode) {} + fn draw_glyph(&mut self, _: &Glyph<'a>, _: Affine, _: Affine, _: &Paint<'a>, _: &GlyphDrawMode) {} + fn draw_image(&mut self, _: Image<'a, '_>, _: Affine) {} + fn pop_clip_path(&mut self) {} + fn pop_transparency_group(&mut self) {} + } + + let pdf = Pdf::new(Arc::new(pdf)).expect("parses"); + let page = &pdf.pages()[0]; + let cache = InterpreterCache::new(); + [(); 2].map(|()| { + let initial = page.initial_transform(true).to_kurbo(); + let mut ctx = Context::new(initial, Rect::new(0.0, 0.0, 130.0, 70.0), &cache, page.xref(), InterpreterSettings::default()); + interpret_page(page, &mut ctx, &mut Nothing); + cache.indexed_lookup_reads() + }) +} + +/// Twenty inline images naming one resource read its lookup stream once, in the first render +/// and none in a second render through the same cache: for a full table, and for a short one +/// that is padded (whose images an earlier version of the patch checked, and skipped, one by +/// one). +#[test] +fn inline_images_naming_one_indexed_resource_read_its_lookup_once() { + let full = page(20, 1, b"\xFF\x00\x00\x00\xFF\x00"); + assert_eq!(lookup_reads(full), [1, 1], "a full table: read once, then cached"); + let short = page(20, 7, b"\xFF\x00\x00\x00\xFF\x00"); + assert_eq!(lookup_reads(short), [1, 1], "a short table: read once, then cached"); +} diff --git a/crates/render/tests/it/indexed_lookup.rs b/crates/render/tests/it/indexed_lookup.rs index 53f7551bca..6c89da629e 100644 --- a/crates/render/tests/it/indexed_lookup.rs +++ b/crates/render/tests/it/indexed_lookup.rs @@ -1,14 +1,36 @@ -//! An `/Indexed` image whose colour space ISO 32000-2 §8.6.6.3 rules out (a lookup table -//! shorter than m × (hival + 1) bytes, or a `hival` that is not an integer in 0..=255) is not -//! drawn, and the render reports `RenderWarning::MalformedIndexedImage`. It used to fall back to -//! DeviceGray and paint its indices as near-black, even indices whose entries were present. +//! An `/Indexed` image whose colour space breaks ISO 32000-2 §8.6.6.3 (a lookup table shorter +//! than m × (hival + 1) bytes, or a `hival` past 255 or with a fraction) is drawn as other +//! readers draw it, and the render reports `RenderWarning::MalformedIndexedImage`: +//! - a short table is padded with zero bytes, so the entries it has keep their colours; +//! - a `hival` past 255 is taken as 255; +//! - a real `hival` is truncated toward zero. +//! +//! A `hival` that is negative or not a number keeps the upstream handling (the image falls back +//! to DeviceGray), and is reported too. Upstream fell back to DeviceGray for every one of these, +//! painting the indices as near-black grey levels, even those whose entries were present. +//! +//! The images are 8 × 1, samples 0 to 7, each column 25 px wide on a 200 × 200 page (the layout +//! of the seven cases A to G a reviewer compared across readers on #4393). use std::sync::Arc; use pdfcraft_render::{PageRenderer, RenderConfig, RenderRequest, RenderWarning, RenderedPage, RequestKind}; -/// A 200 × 200 page whose `/Im` image is two samples wide (indices 0 and 1) and covers the page, -/// with `/P` naming `colour_space` as a resource. `extra` objects are numbered from 6. +/// The eight entries the tests' tables hold, in order. +const TABLE: &str = "FF0000 00FF00 0000FF FFFF00 FF00FF 00FFFF 808080 000000"; +const COLOURS: [[u8; 3]; 8] = [[255, 0, 0], [0, 255, 0], [0, 0, 255], [255, 255, 0], [255, 0, 255], [0, 255, 255], [128, 128, 128], [0, 0, 0]]; +const BLACK: [u8; 3] = [0, 0, 0]; +const WHITE: [u8; 3] = [255, 255, 255]; +const DRAW_IMAGE: &str = "q 200 0 0 200 0 0 cm /Im Do Q"; +const MALFORMED: [RenderWarning; 1] = [RenderWarning::MalformedIndexedImage]; + +/// The first `n` hex entries of `TABLE`. +fn entries(n: usize) -> String { + TABLE.split(' ').take(n).collect::>().join(" ") +} + +/// A 200 × 200 page whose `/Im` image is eight samples wide (indices 0 to 7) and covers the +/// page, with `/P` naming `colour_space` as a resource. `extra` objects are numbered from 6. fn pdf(colour_space: &str, content: &str, extra: &[&[u8]]) -> Vec { pdf_with_resource("P", colour_space, content, extra) } @@ -25,8 +47,8 @@ fn pdf_with_resource(resource: &str, colour_space: &str, content: &str, extra: & .into_bytes(), format!("<< /Length {} >>\nstream\n{content}\nendstream", content.len()).into_bytes(), format!( - "<< /Type /XObject /Subtype /Image /Width 2 /Height 1 /BitsPerComponent 8 \ - /ColorSpace {colour_space} /Filter /ASCIIHexDecode /Length 5 >>\nstream\n0001>\nendstream" + "<< /Type /XObject /Subtype /Image /Width 8 /Height 1 /BitsPerComponent 8 \ + /ColorSpace {colour_space} /Filter /ASCIIHexDecode /Length 17 >>\nstream\n0001020304050607>\nendstream" ) .into_bytes(), ]; @@ -63,169 +85,232 @@ fn render(bytes: Vec) -> RenderedPage { page } -/// The left (index 0) and right (index 1) halves' centre pixels, RGB. -fn halves(page: &RenderedPage) -> [[u8; 3]; 2] { - [50usize, 150].map(|x| { - let at = (100 * page.width as usize + x) * 4; +/// The centre pixel of each of the eight columns (indices 0 to 7), RGB. +fn columns(page: &RenderedPage) -> [[u8; 3]; 8] { + std::array::from_fn(|i| { + let at = (100 * page.width as usize + 12 + 25 * i) * 4; let px = page.rgba.get(at..at + 3).expect("pixel inside the page"); [px[0], px[1], px[2]] }) } -const WHITE: [[u8; 3]; 2] = [[255, 255, 255], [255, 255, 255]]; -const RED_GREEN: [[u8; 3]; 2] = [[255, 0, 0], [0, 255, 0]]; -const DRAW_IMAGE: &str = "q 200 0 0 200 0 0 cm /Im Do Q"; +/// The first `n` of `COLOURS`, then black: a table of `n` entries padded with zero bytes. +fn padded(n: usize) -> [[u8; 3]; 8] { + std::array::from_fn(|i| if i < n { COLOURS[i] } else { BLACK }) +} -fn assert_skipped(colour_space: &str, extra: &[&[u8]]) { +/// Index `i` drawn as the grey level `i` / 255: the DeviceGray fallback. +fn grey_fallback() -> [[u8; 3]; 8] { + std::array::from_fn(|i| [i as u8; 3]) +} + +#[track_caller] +fn assert_page(colour_space: &str, extra: &[&[u8]], want: [[u8; 3]; 8], warned: bool) { let page = render(pdf(colour_space, DRAW_IMAGE, extra)); - assert_eq!(halves(&page), WHITE, "{colour_space}: the image must not be drawn"); - assert_eq!(page.warnings, vec![RenderWarning::MalformedIndexedImage], "{colour_space}: the skip is reported"); -} - -fn assert_drawn(colour_space: &str, extra: &[&[u8]]) { - let page = render(pdf(colour_space, DRAW_IMAGE, extra)); - assert_eq!(halves(&page), RED_GREEN, "{colour_space}: palette entries 0 and 1"); - assert!(page.warnings.is_empty(), "{colour_space}: {:?}", page.warnings); + assert_eq!(columns(&page), want, "{colour_space}: the columns"); + let want_warnings: &[RenderWarning] = if warned { &MALFORMED } else { &[] }; + assert_eq!(page.warnings, want_warnings, "{colour_space}: the warnings"); } +/// Case A: a full table draws its eight colours, with no warning. #[test] -fn a_lookup_shorter_than_hival_needs_skips_the_image_with_a_warning() { - // hival 7 over DeviceRGB needs 24 bytes; 6 are given. Indices 0 and 1 used to draw black. - assert_skipped("[/Indexed /DeviceRGB 7 ]", &[]); - // One byte short of two entries. - assert_skipped("[/Indexed /DeviceRGB 1 ]", &[]); +fn a_full_table_draws_its_colours() { + assert_page(&format!("[/Indexed /DeviceRGB 7 <{TABLE}>]"), &[], COLOURS, false); + assert_page(&format!("[/Indexed /DeviceRGB 7.0 <{TABLE}>]"), &[], COLOURS, false); } +/// Cases B and F: a short table is padded with zero bytes. The entries it has keep their +/// colours, and the rest are black over DeviceRGB. #[test] -fn a_complete_lookup_draws_its_colours() { - assert_drawn("[/Indexed /DeviceRGB 1 ]", &[]); - assert_drawn("[/Indexed /DeviceRGB 7 ]", &[]); +fn a_short_table_is_padded_with_zero_bytes() { + // B: hival 7 needs 24 bytes; 6 are given. Indices 0 and 1 used to draw black. + assert_page(&format!("[/Indexed /DeviceRGB 7 <{}>]", entries(2)), &[], padded(2), true); + // F: one entry short. + assert_page(&format!("[/Indexed /DeviceRGB 7 <{}>]", entries(7)), &[], padded(7), true); + // One byte short: entry 1 is 00FF plus a zero byte, green; indices past hival take entry 1. + let mut want = [COLOURS[1]; 8]; + want[0] = COLOURS[0]; + assert_page("[/Indexed /DeviceRGB 1 ]", &[], want, true); } +/// Case C: an index past `hival` takes the nearest entry (ISO 32000-2 §8.6.6.3: "shall be +/// adjusted to the nearest value within that range"); a table longer than needed is fine. #[test] -fn a_lookup_longer_than_needed_uses_its_first_entries() { - assert_drawn("[/Indexed /DeviceRGB 1 ]", &[]); +fn an_index_past_hival_takes_the_last_entry() { + let mut want = COLOURS; + want[4..].fill(COLOURS[3]); + assert_page(&format!("[/Indexed /DeviceRGB 3 <{TABLE}>]"), &[], want, false); } +/// Case D: a `hival` past 255 is taken as 255, and the table padded to 256 entries. #[test] -fn a_hival_outside_0_to_255_or_not_an_integer_skips_the_image() { - let long = format!("", "00".repeat(3 * 300)); - assert_skipped(&format!("[/Indexed /DeviceRGB 300 {long}]"), &[]); - assert_skipped("[/Indexed /DeviceRGB -1 ]", &[]); - assert_skipped("[/Indexed /DeviceRGB 1.5 ]", &[]); - assert_skipped("[/Indexed /DeviceRGB /One ]", &[]); - // 255, the largest hival, with a full table still draws. - assert_drawn(&format!("[/Indexed /DeviceRGB 255 ]", "00".repeat(3 * 254)), &[]); +fn a_hival_past_255_is_clamped() { + assert_page(&format!("[/Indexed /DeviceRGB 300 <{TABLE}>]"), &[], COLOURS, true); + // With a table long enough for 301 entries, the clamp alone is reported. + let long = format!("<{TABLE}{}>", "00".repeat(3 * 293)); + assert_page(&format!("[/Indexed /DeviceRGB 300 {long}]"), &[], COLOURS, true); + // 256, one past. + assert_page(&format!("[/Indexed /DeviceRGB 256 <{TABLE}{}>]", "00".repeat(3 * 249)), &[], COLOURS, true); + // 255, the largest, with a full table draws without a warning. + assert_page(&format!("[/Indexed /DeviceRGB 255 <{TABLE}{}>]", "00".repeat(3 * 248)), &[], COLOURS, false); } +/// Case E: a real `hival` is truncated toward zero, as hayro read it before patch (60). #[test] -fn an_empty_or_missing_lookup_skips_the_image_without_panicking() { - assert_skipped("[/Indexed /DeviceRGB 1 <>]", &[]); - assert_skipped("[/Indexed /DeviceRGB 1 ()]", &[]); - assert_skipped("[/Indexed /DeviceRGB 1]", &[]); - assert_skipped("[/Indexed /DeviceRGB 0 null]", &[]); +fn a_real_hival_is_truncated() { + assert_page(&format!("[/Indexed /DeviceRGB 7.5 <{TABLE}>]"), &[], COLOURS, true); + // 6.9 is 6, not 7: index 7 takes entry 6. + let mut want = COLOURS; + want[7] = COLOURS[6]; + assert_page(&format!("[/Indexed /DeviceRGB 6.9 <{TABLE}>]"), &[], want, true); + // -0.5 is 0 (toward zero; a real above -1 truncates to 0): every index takes entry 0. + assert_page(&format!("[/Indexed /DeviceRGB -0.5 <{TABLE}>]"), &[], [COLOURS[0]; 8], true); + // 255.5 is 255; 255.0 written as a real is the integer 255. + let full = format!("<{TABLE}{}>", "00".repeat(3 * 248)); + assert_page(&format!("[/Indexed /DeviceRGB 255.5 {full}]"), &[], COLOURS, true); + assert_page(&format!("[/Indexed /DeviceRGB 255.0 {full}]"), &[], COLOURS, false); } +/// Case G: inline images, named through the resources or given in place, are padded too. #[test] -fn a_lookup_stream_is_measured_after_decoding() { - let short: &[u8] = b"<< /Length 14 /Filter /ASCIIHexDecode >>\nstream\nFF0000 00FF00>\nendstream"; - assert_skipped("[/Indexed /DeviceRGB 7 6 0 R]", &[short]); - assert_drawn("[/Indexed /DeviceRGB 1 6 0 R]", &[short]); - let empty: &[u8] = b"<< /Length 0 >>\nstream\n\nendstream"; - assert_skipped("[/Indexed /DeviceRGB 1 6 0 R]", &[empty]); -} - -#[test] -fn inline_images_with_a_short_lookup_are_skipped_too() { - let inline = |cs: &str| format!("q 200 0 0 200 0 0 cm BI /W 2 /H 1 /BPC 8 /CS {cs} /F /AHx ID 0001> EI Q"); +fn inline_images_with_a_short_table_are_padded_too() { + let inline = |cs: &str| format!("q 200 0 0 200 0 0 cm BI /W 8 /H 1 /BPC 8 /CS {cs} /F /AHx ID 0001020304050607> EI Q"); + let short = format!("[/Indexed /DeviceRGB 7 <{}>]", entries(2)); // Named through the page's resources. - let page = render(pdf("[/Indexed /DeviceRGB 7 ]", &inline("/P"), &[])); - assert_eq!(halves(&page), WHITE); - assert_eq!(page.warnings, vec![RenderWarning::MalformedIndexedImage]); + let page = render(pdf_with_resource("CS0", &short, &inline("/CS0"), &[])); + assert_eq!(columns(&page), padded(2)); + assert_eq!(page.warnings, MALFORMED); // Given in place, with the inline abbreviations. - let page = render(pdf("/DeviceRGB", &inline("[/I /RGB 7 ]"), &[])); - assert_eq!(halves(&page), WHITE); - assert_eq!(page.warnings, vec![RenderWarning::MalformedIndexedImage]); - // And a well-formed one still draws. - let page = render(pdf("[/Indexed /DeviceRGB 1 ]", &inline("/P"), &[])); - assert_eq!(halves(&page), RED_GREEN); + let page = render(pdf("/DeviceRGB", &inline(&format!("[/I /RGB 7 <{}>]", entries(2))), &[])); + assert_eq!(columns(&page), padded(2)); + assert_eq!(page.warnings, MALFORMED); + // A `hival` with no reading falls back and is reported, named or in place. + let page = render(pdf(&format!("[/Indexed /DeviceRGB -1 <{TABLE}>]"), &inline("/P"), &[])); + assert_eq!((columns(&page), page.warnings), (grey_fallback(), MALFORMED.to_vec())); + let page = render(pdf("/DeviceRGB", &inline(&format!("[/I /RGB -1 <{TABLE}>]")), &[])); + assert_eq!((columns(&page), page.warnings), (grey_fallback(), MALFORMED.to_vec())); + // A well-formed one draws without a warning. + let page = render(pdf(&format!("[/Indexed /DeviceRGB 7 <{TABLE}>]"), &inline("/P"), &[])); + assert_eq!(columns(&page), COLOURS); assert!(page.warnings.is_empty(), "{:?}", page.warnings); } #[test] fn an_inline_device_space_name_ignores_a_resource_of_the_same_name() { - // ISO 32000-2 §8.6.8 (the CS operator): the device colour space names never refer to resources in the - // ColorSpace subdictionary, so a malformed Indexed resource that shares the name (here the - // inline abbreviation `/RGB`, and `/DeviceRGB`) does not stop the image from drawing. + // ISO 32000-2 §8.6.8 (the CS operator): the device colour space names never refer to + // resources in the ColorSpace subdictionary, so a malformed Indexed resource that shares the + // name (here the inline abbreviation `/RGB`, and `/DeviceRGB`) is not used. + let data: String = TABLE.split(' ').collect(); for (resource, cs) in [("RGB", "/RGB"), ("DeviceRGB", "/DeviceRGB")] { - let content = format!("q 200 0 0 200 0 0 cm BI /W 2 /H 1 /BPC 8 /CS {cs} /F /AHx ID FF000000FF00> EI Q"); + let content = format!("q 200 0 0 200 0 0 cm BI /W 8 /H 1 /BPC 8 /CS {cs} /F /AHx ID {data}> EI Q"); let page = render(pdf_with_resource(resource, "[/Indexed /DeviceRGB 7 ]", &content, &[])); - assert_eq!(halves(&page), RED_GREEN, "{cs}: the device space draws"); + assert_eq!(columns(&page), COLOURS, "{cs}: the device space draws"); assert!(page.warnings.is_empty(), "{cs}: {:?}", page.warnings); } } #[test] -fn a_page_with_several_malformed_images_reports_the_warning_once_and_draws_the_rest() { +fn a_page_with_several_malformed_images_reports_the_warning_once() { let content = format!("{DRAW_IMAGE} {DRAW_IMAGE} 0 0 1 rg 0 0 10 10 re f"); - let page = render(pdf("[/Indexed /DeviceRGB 7 ]", &content, &[])); - assert_eq!(page.warnings, vec![RenderWarning::MalformedIndexedImage]); + let page = render(pdf(&format!("[/Indexed /DeviceRGB 7 <{}>]", entries(2)), &content, &[])); + assert_eq!(page.warnings, MALFORMED); + assert_eq!(columns(&page), padded(2)); let at = (195 * page.width as usize + 5) * 4; - assert_eq!(page.rgba.get(at..at + 3), Some(&[0u8, 0, 255][..]), "content after the skipped images still draws"); + assert_eq!(page.rgba.get(at..at + 3), Some(&[0u8, 0, 255][..]), "content after the images still draws"); } #[test] fn a_lookup_string_given_by_reference_is_read() { // A lookup is a byte string or a stream (ISO 32000-2 §8.6.6.3), and either may be indirect. // A string by reference used to be missed: trying it as a stream consumed the reference. - let full: &[u8] = b""; - assert_drawn("[/Indexed /DeviceRGB 1 6 0 R]", &[full]); - let short: &[u8] = b""; - assert_skipped("[/Indexed /DeviceRGB 1 6 0 R]", &[short]); - // Something that is neither a string nor a stream is no table. - let number: &[u8] = b"42"; - assert_skipped("[/Indexed /DeviceRGB 1 6 0 R]", &[number]); + let full = format!("<{TABLE}>"); + assert_page("[/Indexed /DeviceRGB 7 6 0 R]", &[full.as_bytes()], COLOURS, false); + let short = format!("<{}>", entries(2)); + assert_page("[/Indexed /DeviceRGB 7 6 0 R]", &[short.as_bytes()], padded(2), true); + // Something that is neither a string nor a stream is no table: every entry is padding. + assert_page("[/Indexed /DeviceRGB 7 6 0 R]", &[b"42"], padded(0), true); } #[test] -fn the_size_check_holds_at_the_bounds_of_hival_and_the_base() { - // Four components (DeviceCMYK): hival 1 needs exactly 8 bytes. - let cmyk = |table: &str| format!("[/Indexed /DeviceCMYK 1 <{table}>]"); - let page = render(pdf(&cmyk("FF000000 00FF0000"), DRAW_IMAGE, &[])); - assert!(page.warnings.is_empty(), "exact length: {:?}", page.warnings); - assert_ne!(halves(&page), WHITE, "exact length draws"); - assert_skipped(&cmyk("FF000000 00FF00"), &[]); - // One component (DeviceGray): hival 0 needs one byte, and none is not enough. - let page = render(pdf("[/Indexed /DeviceGray 0 <00>]", DRAW_IMAGE, &[])); - assert!(page.warnings.is_empty(), "{:?}", page.warnings); - assert_skipped("[/Indexed /DeviceGray 0 <>]", &[]); +fn padding_holds_at_the_bounds_of_hival_and_the_base() { + // Four components (DeviceCMYK): hival 1 needs 8 bytes. With 7, the missing byte is entry 1's + // black, and zero there is what the full table holds, so both draw the same. + let full = render(pdf("[/Indexed /DeviceCMYK 1 ]", DRAW_IMAGE, &[])); + assert!(full.warnings.is_empty(), "{:?}", full.warnings); + let short = render(pdf("[/Indexed /DeviceCMYK 1 ]", DRAW_IMAGE, &[])); + assert_eq!(short.warnings, MALFORMED); + assert_eq!(columns(&short), columns(&full)); + assert_ne!(columns(&full)[0], WHITE, "the image is drawn"); + // One component (DeviceGray): hival 0 needs one byte, and an empty table pads it. + assert_page("[/Indexed /DeviceGray 0 <80>]", &[], [[128; 3]; 8], false); + assert_page("[/Indexed /DeviceGray 0 <>]", &[], [BLACK; 8], true); // hival 255 over DeviceRGB needs 768 bytes: 767 is one short. - assert_skipped(&format!("[/Indexed /DeviceRGB 255 ]", "00".repeat(3 * 254 - 1)), &[]); + assert_page(&format!("[/Indexed /DeviceRGB 255 <{TABLE}{}>]", "00".repeat(3 * 248 - 1)), &[], COLOURS, true); } #[test] -fn a_hival_that_would_overflow_a_size_is_ruled_out_without_panicking() { +fn a_hival_that_would_overflow_a_size_is_clamped_without_panicking() { // Values whose (hival + 1) × components would overflow a 32-bit or 64-bit size, or that do - // not fit an integer at all, never reach the multiplication: they fail the 0..=255 test. - for hival in ["4294967295", "4294967296", "18446744073709551615", "99999999999999999999999", "-2147483649", "256", "255.5"] { - assert_skipped(&format!("[/Indexed /DeviceRGB {hival} ]"), &[]); + // not fit an integer at all, are taken as 255. + for hival in ["4294967295", "4294967296", "18446744073709551615", "99999999999999999999999"] { + assert_page(&format!("[/Indexed /DeviceRGB {hival} <{}>]", entries(2)), &[], padded(2), true); + } +} + +/// A `hival` that is negative or not a number has no reading to recover: the image keeps the +/// upstream handling (it falls back to DeviceGray), and the warning is still reported. +#[test] +fn a_negative_or_non_numeric_hival_falls_back_with_a_warning() { + for hival in ["-1", "-1.5", "-2147483649", "/One", "null"] { + assert_page(&format!("[/Indexed /DeviceRGB {hival} <{TABLE}>]"), &[], grey_fallback(), true); } - // 255.0 written as a real is the integer 255. - assert_drawn(&format!("[/Indexed /DeviceRGB 255.0 ]", "00".repeat(3 * 254)), &[]); } #[test] -fn an_unusable_hival_is_judged_even_when_the_base_cannot_be_read() { - // The base is unknown, so the table size cannot be judged, but hival 300 is out of range - // whatever the base. - assert_skipped("[/Indexed /NoSuchSpace 300 <00>]", &[]); - // With a usable hival, an unreadable base keeps its existing handling: no Indexed warning. - let page = render(pdf("[/Indexed /NoSuchSpace 1 <00>]", DRAW_IMAGE, &[])); - assert!(!page.warnings.contains(&RenderWarning::MalformedIndexedImage), "{:?}", page.warnings); +fn an_empty_or_missing_lookup_is_all_padding() { + for lookup in ["<>", "()", "", "null"] { + assert_page(&format!("[/Indexed /DeviceRGB 1 {lookup}]"), &[], padded(0), true); + } } #[test] -fn a_lookup_stream_that_cannot_be_decoded_is_no_table() { +fn a_lookup_stream_is_measured_after_decoding() { + let short: &[u8] = b"<< /Length 14 /Filter /ASCIIHexDecode >>\nstream\nFF0000 00FF00>\nendstream"; + assert_page("[/Indexed /DeviceRGB 7 6 0 R]", &[short], padded(2), true); + let mut want = [COLOURS[1]; 8]; + want[0] = COLOURS[0]; + assert_page("[/Indexed /DeviceRGB 1 6 0 R]", &[short], want, false); + let empty: &[u8] = b"<< /Length 0 >>\nstream\n\nendstream"; + assert_page("[/Indexed /DeviceRGB 1 6 0 R]", &[empty], padded(0), true); +} + +#[test] +fn a_lookup_stream_that_cannot_be_decoded_is_all_padding() { let broken: &[u8] = b"<< /Length 6 /Filter /FlateDecode >>\nstream\nnot-zl\nendstream"; - assert_skipped("[/Indexed /DeviceRGB 1 6 0 R]", &[broken]); + assert_page("[/Indexed /DeviceRGB 1 6 0 R]", &[broken], padded(0), true); +} + +/// An Indexed space whose base cannot be read keeps the upstream handling (DeviceGray), whatever +/// its `hival`; only a `hival` with no reading is reported. +#[test] +fn an_unreadable_base_keeps_its_fallback() { + for (hival, warned) in [("1", false), ("300", false), ("-1", true)] { + assert_page(&format!("[/Indexed /NoSuchSpace {hival} <{TABLE}>]"), &[], grey_fallback(), warned); + } +} + +/// A fill colour set in such a space (`cs`, `sc`) draws through the same recovered table, and +/// the recovery is reported as `ColourFallback`, as the upstream DeviceGray fallback was. +#[test] +fn a_fill_colour_in_a_short_table_is_padded_and_reported() { + let fill = |index: u8| format!("/P cs {index} sc 0 0 200 200 re f"); + let short = format!("[/Indexed /DeviceRGB 7 <{}>]", entries(2)); + let page = render(pdf(&short, &fill(1), &[])); + assert_eq!(columns(&page)[0], COLOURS[1], "entry 1 is green"); + assert_eq!(page.warnings, vec![RenderWarning::ColourFallback]); + let page = render(pdf(&short, &fill(5), &[])); + assert_eq!(columns(&page)[0], BLACK, "entry 5 is padding"); + let page = render(pdf(&format!("[/Indexed /DeviceRGB 7 <{TABLE}>]"), &fill(1), &[])); + assert_eq!(columns(&page)[0], COLOURS[1]); + assert!(page.warnings.is_empty(), "{:?}", page.warnings); } diff --git a/vendor/README.md b/vendor/README.md index 066d8e67d6..f0101a08b1 100644 --- a/vendor/README.md +++ b/vendor/README.md @@ -39,7 +39,8 @@ Other resource categories retain their existing parent-chain lookup. | Crate | Version | Licence | Patches | Test | |---|---|---|---|---| -| hayro-interpret | 0.7.0 | Apache-2.0 OR MIT | (1) select `/AP /N` appearance states via `/AS` (checkboxes, radios, stateful icons) and honour the NoView flag; (2) Type3 Unicode fallback to `/Encoding` glyph names, then printable ASCII; (3) `Type3Glyph::advance_width`; (4) `InterpreterSettings::ocg_overrides` for viewer layer toggles; (5) soft masks whose group lacks `/CS` were dropped (Chrome gradient text); (6) paint nesting cap `MAX_PAINT_NESTING` for tiling patterns and Type 3 glyphs (self-referencing content overflowed the stack); (7) CID width ranges in `/W` and `/W2` clamped to `MAX_CID` (a `0 4294967295 w` range hung); (8) `InterpreterSettings::hide_comments` skips markup annotations but keeps widgets and links (View ▸ Hide all comments); (9) `image_size_ok`: images and masks over 2^28 pixels are not decoded (a fuzzed `/W 4294967295` stencil mask allocated 4 GiB); (10) `MAX_NESTED_PAINTS`: nested paints (forms, Type 3 glyphs and tiling patterns inside one another) share a per-page budget of 50,000 (a glyph showing eight glyphs of its own font, sixteen deep, hung); (11) depends on skrifa 0.47 instead of 0.42 (upstream read-fonts fixed a Type 1 tokenizer that recursed through runs of numbers: a fuzzed font hung or overflowed the stack), with the one API change (`Charset::glyph_id` returns `Option`); (12) annotation appearances whose `/BBox` or `/Matrix` collapses to a line or a point are skipped (mapping them onto `/Rect` divided by zero and handed the device infinite or NaN matrices); (13) a Highlight annotation without a usable normal appearance is drawn from its `/QuadPoints`, `/C` and `/CA` with Multiply blending, as pdfcraft-annot draws highlights (it was not drawn at all); (14) simple-font `/Widths` (`read_widths`): entries only for codes 0–255 (one entry per code below `/FirstChar` was allocated: `4294967295` asked for 34 GB), an entry that isn't a number leaves its code missing instead of ending the array, and a `/LastChar` below `/FirstChar` gives no widths instead of discarding the font for a standard one; (15) the appearance box is mapped onto `/Rect` corner to corner: the translation now accounts for the scale (an appearance whose box didn't start at the origin and differed in size from `/Rect` was drawn shifted); (16) a clip set by `W`/`W*` takes effect after whichever path-painting operator ends the path (`S`, `f`, `B`, … as well as `n`; ISO 32000-2 §8.5.4): it was applied only on `n`, so `re W* S` clipped nothing and the clip stayed pending for a later, unrelated `n`; (17) `InterpreterSettings::cancelled`: once set, interpretation stops before the next content operator (a render pool sets it for every worker when dropped, after an edit or on closing, and for one worker when the watchdog gives up on its render, so renders whose answers nobody can receive stop instead of running to the end); (18) `MAX_PAGE_CONTENT`: the content a page holds at once (its contents, and the forms, tiling patterns and Type 3 glyph procedures being painted) stays within 512 MiB (`decode_content`; once it runs out, `InterpreterWarning::ContentTruncated` is reported (once per page by pdfcraft-render) so a truncated page is not a clean render; each level gives its bytes back when it finishes): forms nest fifty deep and each level kept its content while the next ran, so a form painting itself around a stream inflating to 256 MiB would have held over 12 GB; and Type 0 sampled functions take at most `MAX_SAMPLE_ENTRIES` (2^21) table points and `MAX_SAMPLE_VALUES` (2^23) samples, read no further than `/Size` asks (256 MiB of 1-bit samples asked for over 8 GB); (19) mesh shadings (types 4–7) are sampled only inside the area being drawn (`ShadingPattern::encode_within`, which the vendored hayro renderer calls with that area plus a pixel), within 8 visits per pixel of it plus 4 per triangle (`TriangleSampler`), and patches are triangulated one at a time: points far off the page were sampled over their whole bounding box, up to 65536 × 65536 entries (a fuzzed 147-byte patch stream; a synthetic page passed 6 GB in six seconds), ten thousand overlapping triangles ran for minutes, and every patch became 722 triangles before any limit; meshes hold at most `MAX_MESH_PATCHES` (2^16) patches and `MAX_MESH_TRIANGLES` (2^20) triangles, a lattice needs two to 2^19 vertices per row (`/VerticesPerRow 0` pushed empty rows for ever; one huge row read the whole stream), and a page samples at most `MAX_PAGE_MESH_TRIANGLES` (2^25) mesh triangles in all (a mesh painted a thousand times was triangulated a thousand times); each cut logs a warning; (20) PDF function construction is limited to 64 levels and 10,000 total nodes with cycle detection; calculator parsing and evaluation share procedure depth/work budgets, integer division/bitshift/rotation/index are panic-free, and fixed-size stack errors propagate; (21) a Type 0 font drawn with a substitute that isn't CID-keyed (it isn't embedded, or its program doesn't load) draws `.notdef` for a CID of an Adobe character collection (Japan1, GB1, CNS1, Korea1) that it can't find by Unicode: hayro took the CID for a glyph index of the substitute, which numbers its glyphs its own way, so the character was drawn as an unrelated one (𠮷 as "Ｑ" in BIZ UDPGothic); (22) a Type 0 font drawn with a substitute finds a character given with a variation selector (`/ToUnicode`, or Adobe-Japan1-UCS2, which maps the JIS X 0208 forms of the kanji JIS X 0213:2004 redrew that way: CID 1247 is 噂 U+E0100) through the substitute's variation sequences, else as the character alone: such a CID found no glyph and was drawn as an unrelated one; (23) backport of hayro 0.8's `gradient` module: `EncodedShadingPattern::as_svg_gradient` turns an axial or radial shading into stops of an sRGB gradient (within 0.01 per channel, a stop at every stitching-function bound via `Function::stitching_bounds`) when a native gradient draws it exactly (extended past the box drawn in; a radial one with no background colour); (24) `DecodedImageCache`: decoded image pixels shared across renders through the interpreter cache within a 64 MiB budget (an image-heavy page decoded its thousands of images again per tile and per scale), keyed by every decode input (`DecodedImageKey`: the stream's place and dictionary bytes, its object, mask/stencil/raster output, interpolation, decode target size, preview or final); images under a transfer function and inline images are not cached, an image over the budget is never kept, and a poisoned lock decodes uncached; (25) `Device::actual_text` hands the device a marked content sequence's `/ActualText` (inline or named in `/Properties`), so text extraction keeps the Thai, Indic and Arabic clusters Chrome marks this way (their shaped tone marks and vowels have no Unicode of their own); (26) non-embedded simple fonts drawn with a standard-font stand-in build the Latin Extended-A letters the stand-in lacks (`/Gbreve`, `/Idotaccent`, `/Scedilla`, `/aogonek`, `/ccaron`…) from the base letter and accent it has (`StandardKind::composite_glyph`, `ACCENTED_LETTERS`): PDFium's Foxit stand-ins hold only the Standard/WinAnsi/Mac Roman glyphs, so iText letters with Times New Roman unembedded showed blanks for Ğ, İ, Ş, ğ and ş; the comma-below letters (`/scommaaccent`, `/Tcommaaccent`, `/kcommaaccent`…, found by character, so `/uni0136` and the glyph list's `/Kcedilla` too) are the letter and the font's comma, smaller and under the baseline, since the stand-ins have no `commaaccent` (Romanian ș drew nothing and ț a cedilla in forms laid out from XFA; ģ, whose comma is turned and above, is not built); (27) an embedded TrueType font without a base encoding reads its codes as WinAnsi for text extraction (`EmbeddedKind::code_to_unicode_name`), matching pdftotext's output, instead of Standard: OCR text layers of scans write WinAnsi codes with no `/Encoding`, and Standard has no ü, ö or ß and reads é as Ø (glyph selection unchanged); (28) an image whose pixels hold at most 16 bits of samples (an 8-bit Separation or DeviceN image, a 1-, 2- or 4-bit gray or indexed one) is converted once per possible pixel value through a lookup table instead of once per pixel (`rgb_via_lookup`): a spot-colour tint transform is often a PostScript calculator program, run for every pixel, so a fillable character sheet's 3304 × 4105 DeviceN background took four seconds and 430 MB to draw a page; each entry is converted on its own (`convert_each`), so a tint transform giving the wrong number of components for a value the image never uses cannot shift the colours of the values it does; (29) `Device::begin_annotations`/`end_annotations` bracket the page's annotations, so a device can tell comment paints from document paints (reading colours recolour the document, never a comment, #1548); (30) the Annex D.2 alias codes read as the ink they draw: WinAnsiEncoding 0xA0 and 0xAD, and MacRomanEncoding 0xCA, carry the glyph names `space` and `hyphen` -- the names those tables also give 0x20 and 0x2D -- not `nbspace` and `sfthyphen`, whose glyph list values U+00A0 and U+00AD are invisible formatting characters. Extraction hid drawn ink behind them, so search and redaction matched neither (`XiaoBiaoSong.pdf` in the pdf.js corpus drew 104 soft hyphens; across the 989 files the soft hyphens we extract fall from 117 to 13 and the non-breaking spaces from 30 to 26, what remains coming from `/ToUnicode` maps, which say so explicitly). The base table only, in the text direction only: a `/Differences` name keeps its glyph list value and glyph lookup is untouched, both as pdftotext does; (31) a Type 0 sampled function keeps its samples in one flat array indexed by strides (`Type0::entry`) instead of a hash map keyed by each grid point: the map was built again, one hashed insert per sample, every time a shading using the function was painted (a 2,097,152-sample ramp painted 30 times took 38 s; `bug1721218` in the pdf.js corpus spent most of a 1.5 s page on it); the interpolation is unchanged, so results are bit for bit the same, a sample missing from a short stream still gives no value, and the maximum sample value is computed without overflowing at 32 bits per sample; (32) the decoded-image cache (24) keys a decode by the size the image is drawn at only when a JPX stream decodes it or its soft mask or mask (`decode_reads_target`): JPX is the only filter that decodes a reduced resolution for a size, so a JPEG or Flate scan was decoded again at every zoom step and kept once per zoom level; one decode now serves every size; `InterpreterCache::decoded_image_decodes` counts the cache's decodes; (33) Coons and tensor-product patches copy their vertices' colours less: each point of a patch's 20 × 20 triangle grid is interpolated once (`generate_patch_triangles`; each cell interpolated its four corners itself, 1,444 interpolations a patch for 400 points), a cell's second triangle takes the vertices its first copied, and the mesh sampler moves a patch's triangles into device space instead of copying each vertex (`TriangleSampler::sample_owned`); the same values, 10–16% faster on mesh-gradient pages; (34) `color::tint_transform_runs`: how many colours the calling thread has run through a Separation or DeviceN tint transform (counted once per batch in `convert_f32`), so a test counts patch (28)'s once-per-sample-value runs instead of timing them (its 10 s deadline failed on a busy FreeBSD VM in CI); (35) `EncodedShadingPattern::extent`: where a mesh shading (types 4–7) can paint at all, the pixels its triangles touched when it was encoded (`None` for the types that paint everywhere, and when the shading has a `/Background`), so the renderer samples a texture there alone and skips a mesh with nothing on the canvas; (36) a mesh shading's pixel whose corner no triangle holds strictly (on an edge two triangles share, or on the mesh's side) takes the triangle that covers its centre, edges included (`Triangle::covers`): testing only the corner left a white seam along every shared edge, and the colour is still taken at the corner; (38) the `/DefaultGray`, `/DefaultRGB` and `/DefaultCMYK` colour spaces of a resource dictionary (ISO 32000-2 §8.6.5.6) replace the device space content selects: with `g`/`G`, `rg`/`RG`, `k`/`K` and `cs`/`CS` (by name, or through a `/ColorSpace` resource that is a device space), and as the `/ColorSpace` of an image XObject, an inline image or a shading painted by `sh` or as a shading pattern (`DefaultColorSpaces`, looked up once per content window in the resources and those they inherit, so a form XObject without a default of its own draws in its page's, as Poppler and MuPDF draw it); a default that is a Lab, Indexed, Pattern or device space, has another number of components or is no colour space is ignored and device colour is drawn as before; device spaces inside other spaces (an Indexed base, a Separation or DeviceN alternate, a Pattern's underlying space), shading dictionaries, masks and transparency group spaces are unchanged, and the decoded-image cache (24) keys a decode by the default it was drawn in. hayro drew device colour as it was, so pages that calibrate it this way (common in PDF/A) came out in the wrong colours; (39) knockout groups (#2971): `Device::set_knockout` right after a `/K true` group is pushed; `Device::knockout_image` for an unmasked image before its group is pushed, when `Device::in_knockout_group`: the device clears its placement rectangle only if it will draw the image (on the canvas, not a preview's tiny image, and it decodes), decoded at its draw size so the draw reuses that decode; and, when `Device::knockout_erase_wanted`, a nested group's content drawn once between `begin_knockout_erase` and `end_knockout_erase` before the group itself, which then always gets a layer of its own; (40) an optional content membership dictionary's visibility expression (`/VE`: `/And`, `/Or`, `/Not`, nested up to 32 deep) decides its visibility ahead of `/OCGs` and `/P` (ISO 32000-2 §8.11.2.2); it was ignored, so content under it stayed visible with its layers hidden; a malformed expression, including one naming an object the document doesn't list in `/OCProperties /OCGs`, falls back to `/OCGs` and `/P` (#2986); (41) a transparency group that is neither isolated nor knockout, entered at full opacity with no soft mask and Normal blending, is drawn straight onto the page instead of into a layer of its own (§11.4.7: its elements composite with the backdrop; the layer made a `/BM /Multiply` highlight appearance an opaque block, #1968); (42) text in a Type 3 font is drawn filled whatever the text rendering mode, except mode 3 (invisible): only that mode has any effect on Type 3 fonts (ISO 32000-1 §9.3.6). Stroking ran the glyph procedure a second time, so `2 Tr` made bitmap fonts look bold, `1 Tr` painted an uncoloured glyph in the stroking colour, and `7 Tr` hid it (`7 Tr` since follows 64); (43) Lab colour spaces honour their own `/WhitePoint`: L\*a\*b\* goes to XYZ relative to that white, is Bradford-adapted to D65 and converted to sRGB (ISO 32000-2 §8.6.5.4) instead of through `LAB.icc` as if every white were D50 (a D65 blue came out 45 levels too red; the profile is removed); components normalised to [0, 1] (`manual_scale`, an `/Indexed` palette over Lab) map onto L\* 0..100 and the a\*/b\* `/Range`; (44) a Type 0 font whose `/ToUnicode` is the name `/Identity-H` or `/Identity-V` (not a CMap stream; WSCAD writes it) reads each code as its Unicode value, as Acrobat shows it (`Type0Font::identity_to_unicode`); the identity CMap the name resolved to maps codes to CIDs only, so such pages had no text to select or find; (46) an image XObject's group applies the fill opacity, so the fill alpha is reset to 1 before a stencil mask takes the fill colour as its paint, and a pattern paint drops the opacity it captured at `scn` (`Pattern::without_opacity`) (it was applied twice: a mask at `ca 0.5` was drawn at a quarter, #2975); (47) `path_as_rect`: a five-element path is a rectangle only when its four points are distinct, each step from one to the next runs along a single axis, and the last returns to the first. The length test admits one MoveTo and four LineTo as well as one MoveTo, three LineTo and a ClosePath, and asking only that every corner of the bounding box was visited says nothing about the order or the steps between, so a five-point polyline whose ends reached the corners was painted as that box -- a fourth side the path does not have -- and a clip of that shape was discarded; the coordinates are compared exactly, since the matrix is applied after this check and a `cm` of 10,000 turns `is_nearly_equal`'s 256th of a unit into tens of device units (#3327); (48) an `/ICCBased` colour space is cached by what its profile is read from (`IccProfileKey`): the stream object it names, looked up without reading the stream, and then the stream's dictionary and its own bytes (decrypted, not decoded), so a byte-for-byte copy of a profile in another object is read once, and a profile stream written directly in its array, which reports the object holding the array as every other stream written there does, is told apart by its bytes. hayro keyed it by the stream dictionary's bytes alone, so two different profiles with the same `/N` and `/Length` shared whichever was read first, and colours in the second were drawn through the first. `InterpreterCache::icc_profile_reads` counts the profiles read; (49) `OpenTypeFontBlob::new` also accepts a `post` table version 1.0, 2.0 or 3.0 written with its 16-bit halves swapped (`00 00 00 03`), as ScanSoft PDF Create! embeds every font; such fonts were refused and their text drawn with a substitute (Hebrew came out garbled or missing, #2946); `glyph_names` reads such a table from a copy with the version written correctly, so a 2.0 table's names serve `/Differences`. Any other unknown version is still refused (pdf.js 9462); (50) `CffFontBlob::glyph_index_by_name` looks a glyph name up in a name to glyph map built on first use (shared between clones) instead of searching the whole charset, with a string lookup per glyph, for every character code of a font; the first glyph with a name still wins; (51) at most 4096 graphics states saved at once, holding at most 64 MiB together (each `q` copies the whole state, dash array and clip list included, so short operators multiplied one large state); a `q` past either limit is counted but not copied, and its `Q` still removes the clips added since (from storytold/craft-libs#239); (52) a `/W2` run whose first CID is past `MAX_CID` stops there, as in `/W` (7), checked before the run's metrics are read: a first CID of 4294967295 inserted one entry and then overflowed, discarding the whole `/W2` table (from storytold/craft-libs#254); (55) a Separation or DeviceN tint transform that cannot be read keeps its colour space, and one that cannot be read or evaluated paints the alternate space's initial colour, as evaluation failures already did; that fallback, and a `cs`/`CS` colour space that cannot be read (drawn in DeviceGray as before), are now reported once per content stream as `InterpreterWarning::ColorFallback` (`RenderWarning::ColourFallback` in `pdfcraft-render`) instead of silently (#4260); (56) an annotation whose Invisible flag (bit 1) is set and whose `/Subtype` is none of the standard types of ISO 32000-2 Table 171 (`is_standard_annotation_type`; no `/Subtype` counts as unknown) is not drawn (Table 167), and a `/Popup` annotation's appearance is never painted on the page (§12.5.6.14: a popup has no appearance stream of its own; the viewer shows it as a window for its parent annotation). Both were drawn from their `/AP`; (57) a simple font drawn with a standard-font stand-in because the document names a font it doesn't embed (`StandardKind` with `fallback`) narrows a glyph to a narrower `/Widths` advance but no longer widens it to a wider one: the advance already places the next glyph (ISO 32000-2 §9.2.4; /Widths in Table 109), and widening drew a non-embedded `Arial,Bold` with 600-unit widths as a heavy fixed-pitch face whose `i` and `l` were solid blocks twice their width; an exact standard-14 font still matches its standard metrics both ways; (58) the fallback request for a Type 0 font that is not embedded (`FallbackFontQuery::with_descriptor`) takes its serif, italic and small-cap flags, `/FontWeight`, `/FontFamily` and `/FontStretch` from the descendant CIDFont's font descriptor (ISO 32000-2 §9.7.4.1, §9.7.6.1; §9.8): it read them from the Type 0 dictionary, which has none, so every such request asked for a regular sans-serif face and no resolver could tell a Song or Mincho font from a Hei or Gothic one, or see `/FontWeight 700`; (59) the NoRotate flag (`/F` bit 5) is honoured: the appearance, fitted to `/Rect`, is turned against the page's `/Rotate` about the rectangle's upper-left corner, which stays where it is, so it reads upright as the page is displayed (ISO 32000-2 §12.5.3). Sticky notes, which are saved NoRotate, were drawn turned with the page (#2583). Form fields and links stay in `/Rect`, where PdfCraft takes their click targets from it; (64) text in a Type 3 font follows ISO 32000-2 §9.3.6 in the clipping modes: under `7 Tr` it is not rendered (reported to the device as clipping text, `GlyphDrawMode::Clip`, so extraction keeps it), and in modes 4 to 7 a Type 3 glyph adds nothing to the clipping path (`clip_glyph`), so text shown only in a Type 3 font leaves no clip at `ET`. Since (42) `7 Tr` painted the glyphs in the fill colour (black by default), so a fill meant to follow showed them wherever it didn't cover them; 4 to 6 still render filled; (66) a Separation colour space whose colourant is the special name `All` paints every colourant at once (ISO 32000-2 §8.6.6.4): on screen and in an RGB print raster its tint t is the gray 1 − t, complemented because the device is additive, whatever its alternate space and tint transform say (hayro drew it through the tint transform, so a `/DeviceCMYK` transform giving cyan drew registration marks light cyan); `/None` still paints nothing, any other name still goes through its alternate, and a DeviceN space naming `All` keeps its alternate, since the clause gives the name that meaning in a Separation space; (63) a luminosity soft mask whose group colour space is `/DeviceCMYK` (by name, not an ICC-based or calibrated space) reads a `/DeviceCMYK` fill or stroke colour drawn in it, and its `/BC`, as their gray `1 - min(1, 0.3 c + 0.59 m + 0.11 y + k)` (ISO 32000-2 §10.4.2.3, §11.5.3: device colour converts to gray with no colour calibration; `Context::luminosity_in_device_cmyk`, `color::device_cmyk_to_gray`): it went through the CMYK display profile, where 100% K is a dark grey, so a 100% K mask let about an eighth of the masked paint through. CMYK images, shadings and tiling patterns inside such a mask, and CMYK on the page, still use the profile; (60) an image whose `/Indexed` colour space ISO 32000-2 §8.6.6.3 rules out (`hival` not an integer in 0..=255, or a lookup table missing or shorter than m × (hival + 1) bytes; a longer one is fine) is not drawn and reports `InterpreterWarning::MalformedIndexedImage`, surfaced as `RenderWarning::MalformedIndexedImage`; upstream fell back to DeviceGray and painted the indices near-black, even ones whose entries were present. Inline images naming such a space as a resource are skipped the same way; fill and stroke colours in such a space are unchanged. A lookup string given by an indirect reference is read (upstream consumed the reference trying it as a stream and found no table) | (1) `raster::tests::widget_appearance_states`; (2, 3) `xtask text-oracle` on corpus/pdfjs issue918; (4) `raster::tests::layer_override_hides_content`; (5) `raster::tests::soft_mask_without_group_colour_space_is_applied`; (6) `raster::tests::self_referencing_tiling_pattern_terminates`, `raster::tests::self_referencing_type3_glyph_terminates`; (7) `raster::tests::huge_cid_width_ranges_terminate`; (8) `raster::tests::hiding_comments_keeps_fields`; (9) `raster::tests::absurd_mask_and_fax_sizes_are_refused_before_decoding`; (10) `raster::tests::type3_glyphs_that_fan_out_into_themselves_finish`; (11) `raster::tests::type1_font_programs_with_long_runs_of_numbers_load_quickly`; (12) `raster::tests::degenerate_annotation_appearances_are_skipped`; (13) `raster::tests::highlights_without_an_appearance_are_drawn`; (14) `raster::tests::malformed_simple_font_widths_keep_later_widths_and_the_font`, `raster::tests::huge_simple_font_first_char_terminates`; (15) `raster::tests::annotation_appearance_box_maps_onto_rect`; (16) `raster::tests::clipping_path_applies_after_any_painting_operator`; (17) `raster::tests::dropping_the_pool_cancels_its_renders`, `raster::tests::the_watchdog_stops_the_render_it_gives_up_on`; (18) `raster::tests::page_contents_and_nested_forms_share_one_budget`, `raster::tests::content_painted_again_and_again_keeps_drawing`, `raster::tests::sampled_functions_with_billions_of_samples_are_refused`; (19) `raster::tests::mesh_shadings_far_off_the_page_render`, `raster::tests::many_overlapping_mesh_triangles_finish`, `raster::tests::mesh_shadings_with_many_patches_or_no_rows_finish`; (20) `raster::tests::function_limits_calculator_arithmetic_and_depth`, `raster::tests::function_limits_public_construction_work_and_depth`, `raster::tests::function_limits_stitching_cycles_are_refused_and_shared_children_work`, `raster::tests::function_limits_invalid_calculator_keeps_the_page_renderable`; (21) `raster::tests::cids_a_substitute_lacks_are_not_drawn_as_other_glyphs`; (22) `raster::tests::variation_sequences_draw_their_kanji_in_a_substitute`; (23) `raster::tests::axial_and_radial_shadings_draw_as_native_gradients`; (24) `raster::tests::tiles_match_full_render`, `raster::tests::skipping_paths_off_the_canvas_changes_no_pixel` (both decode through the cache), `raster::tests::decoded_images_with_identical_dictionaries_keep_their_own_pixels`, `raster::tests::decoded_images_under_a_transfer_function_are_not_shared`; (25) `text::tests::actual_text_replaces_the_glyphs_it_marks_in_every_writing_direction`, `text::tests::actual_text_follows_the_pdf_rules_for_marked_content`; (26) `raster::tests::accented_letters_the_standard_stand_ins_lack_are_built_from_their_parts`, `raster::tests::comma_below_letters_in_standard_fonts_draw_a_comma_under_the_letter`; (27) `text::tests::truetype_without_an_encoding_reads_its_codes_as_win_ansi`; (28) `raster::tests::spot_colour_images_run_their_tint_transform_once_per_sample_value`, `raster::tests::images_converted_through_a_table_keep_every_samples_colour`, `raster::tests::a_malformed_tint_for_an_unused_sample_keeps_the_used_colours`; (29) `raster::tests::reading_colors_keep_highlight_annotations_readable`; (30) `text::tests::the_annex_d2_alias_codes_read_as_an_ordinary_hyphen_and_space`; (31) `raster::tests::shadings_with_large_sampled_functions_paint_quickly`, `raster::tests::sampled_functions_interpolate_their_grid_as_before`; (32) `raster::tests::decoded_images_serve_every_size_they_are_drawn_at`; (33) `raster::tests::patch_meshes_keep_their_corner_colours`; (34) `raster::tests::spot_colour_images_run_their_tint_transform_once_per_sample_value`; (35) `mesh_extent_tests::{a_mesh_paints_its_triangles_alone, a_mesh_keeps_the_graphics_state_alpha, a_mesh_off_the_canvas_paints_nothing}` (`cargo test -p hayro`); (36) `raster::tests::mesh_triangles_sharing_an_edge_leave_no_seam`; (38) `raster::tests::default_colour_spaces::device_rgb_operators_draw_in_default_rgb`, `raster::tests::default_colour_spaces::device_cmyk_operators_draw_in_default_cmyk`, `raster::tests::default_colour_spaces::device_gray_operators_draw_in_default_gray`, `raster::tests::default_colour_spaces::device_images_draw_in_the_default_colour_spaces`, `raster::tests::default_colour_spaces::device_shadings_draw_in_the_default_colour_spaces`, `raster::tests::default_colour_spaces::an_image_drawn_with_and_without_a_default_keeps_both_colours`, `raster::tests::default_colour_spaces::forms_inherit_the_default_colour_spaces_they_do_not_override`, `raster::tests::default_colour_spaces::unusable_defaults_leave_device_colour_alone`; (39) `raster::tests::knockout_groups_replace_what_their_elements_cover`; (40) `raster::tests::visibility_expressions_decide_what_a_membership_dictionary_shows`; (41) `raster::tests::a_multiply_highlight_inside_a_transparency_group_keeps_the_text`; (42) `raster::tests::rendering_modes_but_3_and_7_draw_type3_glyphs_as_filled`; (43) `raster::tests::lab_honours_the_space_white_point`, `raster::tests::indexed_palettes_over_lab_keep_their_colours`; (44) `text::tests::identity_to_unicode_names_give_text`; (46) `raster::tests::stencil_masks_are_drawn_at_the_fill_opacity_once`; (47) `raster::tests::a_five_point_polyline_is_not_painted_as_its_bounding_box`, `raster::tests::a_path_that_visits_every_corner_out_of_order_is_not_a_rectangle`, `raster::tests::a_rectangle_is_still_painted_as_a_rectangle`; (48) `raster::tests::icc_profile_cache::icc_profiles_with_identical_dictionaries_draw_their_own_colours`, `raster::tests::icc_profile_cache::icc_profiles_used_again_are_read_once`; (49) `raster::tests::a_font_whose_post_version_has_swapped_halves_is_still_used`; (50) `font::blob::tests::glyph_names_resolve_to_the_first_glyph_with_that_name`; (51) `context::saved_state_tests::repeated_saves_stay_bounded_and_balanced`; (52) `font::cid::tests::w2_run_past_the_largest_cid_keeps_the_rest_of_the_table` (hayro-interpret); (55) `raster::tests::devicen_inks::a_tint_transform_that_cannot_be_evaluated_draws_the_rest_of_the_page_and_warns`, `raster::tests::devicen_inks::a_tint_transform_that_cannot_be_read_falls_back_and_warns`; (56) `raster::tests::invisible_unknown_annotations_and_popups_paint_nothing`; (57) `substitute_widths::a_substitute_keeps_its_glyph_shapes_at_wider_advances`, `substitute_widths::a_substitute_still_narrows_glyphs_to_narrower_advances` (pdfcraft-render `tests/it`); (58) `cid_fallback_query::a_cid_font_fallback_request_carries_its_descendants_style`, `cid_fallback_query::a_cid_font_without_a_descriptor_asks_for_the_default_style` (pdfcraft-render integration tests); (59) `raster::tests::no_rotate_notes_stay_upright_at_their_corner_on_rotated_pages` (hayro-interpret); (64) `type3_text_clip::{clip_only_type3_text_is_not_rendered_and_leaves_no_clip, invisible_type3_text_is_not_rendered, other_modes_render_type3_glyphs_filled_and_never_clip, only_outline_glyphs_shape_a_clip_shared_with_type3_text, self_nesting_type3_text_in_a_clip_mode_stays_bounded}` (`cargo test -p pdfcraft-render --test it`), `raster::tests::rendering_modes_but_3_and_7_draw_type3_glyphs_as_filled`, `text::tests::clipping_mode_type3_text_is_kept_by_the_ocr_filter`; (66) `separation_all::{separation_all_fills_paint_the_complemented_tint_as_gray, separation_all_strokes_paint_the_complemented_tint_as_gray, separation_all_images_paint_each_sample_as_the_complemented_gray, separation_all_shadings_paint_the_complemented_gray, separation_none_paints_nothing, a_named_separation_still_paints_through_its_alternate, a_device_n_space_naming_all_still_paints_through_its_alternate}` (`cargo test -p pdfcraft-render --test it`); (63) `color::device_cmyk_gray_tests::{each_ink_darkens_by_its_iso_weight, ink_past_full_coverage_is_black_and_bad_components_read_as_zero}` (hayro-interpret), `it::luminosity_group_space::{black_ink_in_a_device_cmyk_group_masks_fully_and_no_ink_shows_fully, each_process_ink_in_a_device_cmyk_group_masks_by_its_weight, a_device_cmyk_backdrop_colour_masks_by_its_gray, gray_and_rgb_groups_mask_as_before, device_cmyk_on_the_page_keeps_its_display_colour}` (pdfcraft-render); (60) `crates/render/tests/it/indexed_lookup.rs` (`cargo test -p pdfcraft-render --test it indexed_lookup`) | +| hayro-interpret | 0.7.0 | Apache-2.0 OR MIT | (1) select `/AP /N` appearance states via `/AS` (checkboxes, radios, stateful icons) and honour the NoView flag; (2) Type3 Unicode fallback to `/Encoding` glyph names, then printable ASCII; (3) `Type3Glyph::advance_width`; (4) `InterpreterSettings::ocg_overrides` for viewer layer toggles; (5) soft masks whose group lacks `/CS` were dropped (Chrome gradient text); (6) paint nesting cap `MAX_PAINT_NESTING` for tiling patterns and Type 3 glyphs (self-referencing content overflowed the stack); (7) CID width ranges in `/W` and `/W2` clamped to `MAX_CID` (a `0 4294967295 w` range hung); (8) `InterpreterSettings::hide_comments` skips markup annotations but keeps widgets and links (View ▸ Hide all comments); (9) `image_size_ok`: images and masks over 2^28 pixels are not decoded (a fuzzed `/W 4294967295` stencil mask allocated 4 GiB); (10) `MAX_NESTED_PAINTS`: nested paints (forms, Type 3 glyphs and tiling patterns inside one another) share a per-page budget of 50,000 (a glyph showing eight glyphs of its own font, sixteen deep, hung); (11) depends on skrifa 0.47 instead of 0.42 (upstream read-fonts fixed a Type 1 tokenizer that recursed through runs of numbers: a fuzzed font hung or overflowed the stack), with the one API change (`Charset::glyph_id` returns `Option`); (12) annotation appearances whose `/BBox` or `/Matrix` collapses to a line or a point are skipped (mapping them onto `/Rect` divided by zero and handed the device infinite or NaN matrices); (13) a Highlight annotation without a usable normal appearance is drawn from its `/QuadPoints`, `/C` and `/CA` with Multiply blending, as pdfcraft-annot draws highlights (it was not drawn at all); (14) simple-font `/Widths` (`read_widths`): entries only for codes 0–255 (one entry per code below `/FirstChar` was allocated: `4294967295` asked for 34 GB), an entry that isn't a number leaves its code missing instead of ending the array, and a `/LastChar` below `/FirstChar` gives no widths instead of discarding the font for a standard one; (15) the appearance box is mapped onto `/Rect` corner to corner: the translation now accounts for the scale (an appearance whose box didn't start at the origin and differed in size from `/Rect` was drawn shifted); (16) a clip set by `W`/`W*` takes effect after whichever path-painting operator ends the path (`S`, `f`, `B`, … as well as `n`; ISO 32000-2 §8.5.4): it was applied only on `n`, so `re W* S` clipped nothing and the clip stayed pending for a later, unrelated `n`; (17) `InterpreterSettings::cancelled`: once set, interpretation stops before the next content operator (a render pool sets it for every worker when dropped, after an edit or on closing, and for one worker when the watchdog gives up on its render, so renders whose answers nobody can receive stop instead of running to the end); (18) `MAX_PAGE_CONTENT`: the content a page holds at once (its contents, and the forms, tiling patterns and Type 3 glyph procedures being painted) stays within 512 MiB (`decode_content`; once it runs out, `InterpreterWarning::ContentTruncated` is reported (once per page by pdfcraft-render) so a truncated page is not a clean render; each level gives its bytes back when it finishes): forms nest fifty deep and each level kept its content while the next ran, so a form painting itself around a stream inflating to 256 MiB would have held over 12 GB; and Type 0 sampled functions take at most `MAX_SAMPLE_ENTRIES` (2^21) table points and `MAX_SAMPLE_VALUES` (2^23) samples, read no further than `/Size` asks (256 MiB of 1-bit samples asked for over 8 GB); (19) mesh shadings (types 4–7) are sampled only inside the area being drawn (`ShadingPattern::encode_within`, which the vendored hayro renderer calls with that area plus a pixel), within 8 visits per pixel of it plus 4 per triangle (`TriangleSampler`), and patches are triangulated one at a time: points far off the page were sampled over their whole bounding box, up to 65536 × 65536 entries (a fuzzed 147-byte patch stream; a synthetic page passed 6 GB in six seconds), ten thousand overlapping triangles ran for minutes, and every patch became 722 triangles before any limit; meshes hold at most `MAX_MESH_PATCHES` (2^16) patches and `MAX_MESH_TRIANGLES` (2^20) triangles, a lattice needs two to 2^19 vertices per row (`/VerticesPerRow 0` pushed empty rows for ever; one huge row read the whole stream), and a page samples at most `MAX_PAGE_MESH_TRIANGLES` (2^25) mesh triangles in all (a mesh painted a thousand times was triangulated a thousand times); each cut logs a warning; (20) PDF function construction is limited to 64 levels and 10,000 total nodes with cycle detection; calculator parsing and evaluation share procedure depth/work budgets, integer division/bitshift/rotation/index are panic-free, and fixed-size stack errors propagate; (21) a Type 0 font drawn with a substitute that isn't CID-keyed (it isn't embedded, or its program doesn't load) draws `.notdef` for a CID of an Adobe character collection (Japan1, GB1, CNS1, Korea1) that it can't find by Unicode: hayro took the CID for a glyph index of the substitute, which numbers its glyphs its own way, so the character was drawn as an unrelated one (𠮷 as "Ｑ" in BIZ UDPGothic); (22) a Type 0 font drawn with a substitute finds a character given with a variation selector (`/ToUnicode`, or Adobe-Japan1-UCS2, which maps the JIS X 0208 forms of the kanji JIS X 0213:2004 redrew that way: CID 1247 is 噂 U+E0100) through the substitute's variation sequences, else as the character alone: such a CID found no glyph and was drawn as an unrelated one; (23) backport of hayro 0.8's `gradient` module: `EncodedShadingPattern::as_svg_gradient` turns an axial or radial shading into stops of an sRGB gradient (within 0.01 per channel, a stop at every stitching-function bound via `Function::stitching_bounds`) when a native gradient draws it exactly (extended past the box drawn in; a radial one with no background colour); (24) `DecodedImageCache`: decoded image pixels shared across renders through the interpreter cache within a 64 MiB budget (an image-heavy page decoded its thousands of images again per tile and per scale), keyed by every decode input (`DecodedImageKey`: the stream's place and dictionary bytes, its object, mask/stencil/raster output, interpolation, decode target size, preview or final); images under a transfer function and inline images are not cached, an image over the budget is never kept, and a poisoned lock decodes uncached; (25) `Device::actual_text` hands the device a marked content sequence's `/ActualText` (inline or named in `/Properties`), so text extraction keeps the Thai, Indic and Arabic clusters Chrome marks this way (their shaped tone marks and vowels have no Unicode of their own); (26) non-embedded simple fonts drawn with a standard-font stand-in build the Latin Extended-A letters the stand-in lacks (`/Gbreve`, `/Idotaccent`, `/Scedilla`, `/aogonek`, `/ccaron`…) from the base letter and accent it has (`StandardKind::composite_glyph`, `ACCENTED_LETTERS`): PDFium's Foxit stand-ins hold only the Standard/WinAnsi/Mac Roman glyphs, so iText letters with Times New Roman unembedded showed blanks for Ğ, İ, Ş, ğ and ş; the comma-below letters (`/scommaaccent`, `/Tcommaaccent`, `/kcommaaccent`…, found by character, so `/uni0136` and the glyph list's `/Kcedilla` too) are the letter and the font's comma, smaller and under the baseline, since the stand-ins have no `commaaccent` (Romanian ș drew nothing and ț a cedilla in forms laid out from XFA; ģ, whose comma is turned and above, is not built); (27) an embedded TrueType font without a base encoding reads its codes as WinAnsi for text extraction (`EmbeddedKind::code_to_unicode_name`), matching pdftotext's output, instead of Standard: OCR text layers of scans write WinAnsi codes with no `/Encoding`, and Standard has no ü, ö or ß and reads é as Ø (glyph selection unchanged); (28) an image whose pixels hold at most 16 bits of samples (an 8-bit Separation or DeviceN image, a 1-, 2- or 4-bit gray or indexed one) is converted once per possible pixel value through a lookup table instead of once per pixel (`rgb_via_lookup`): a spot-colour tint transform is often a PostScript calculator program, run for every pixel, so a fillable character sheet's 3304 × 4105 DeviceN background took four seconds and 430 MB to draw a page; each entry is converted on its own (`convert_each`), so a tint transform giving the wrong number of components for a value the image never uses cannot shift the colours of the values it does; (29) `Device::begin_annotations`/`end_annotations` bracket the page's annotations, so a device can tell comment paints from document paints (reading colours recolour the document, never a comment, #1548); (30) the Annex D.2 alias codes read as the ink they draw: WinAnsiEncoding 0xA0 and 0xAD, and MacRomanEncoding 0xCA, carry the glyph names `space` and `hyphen` -- the names those tables also give 0x20 and 0x2D -- not `nbspace` and `sfthyphen`, whose glyph list values U+00A0 and U+00AD are invisible formatting characters. Extraction hid drawn ink behind them, so search and redaction matched neither (`XiaoBiaoSong.pdf` in the pdf.js corpus drew 104 soft hyphens; across the 989 files the soft hyphens we extract fall from 117 to 13 and the non-breaking spaces from 30 to 26, what remains coming from `/ToUnicode` maps, which say so explicitly). The base table only, in the text direction only: a `/Differences` name keeps its glyph list value and glyph lookup is untouched, both as pdftotext does; (31) a Type 0 sampled function keeps its samples in one flat array indexed by strides (`Type0::entry`) instead of a hash map keyed by each grid point: the map was built again, one hashed insert per sample, every time a shading using the function was painted (a 2,097,152-sample ramp painted 30 times took 38 s; `bug1721218` in the pdf.js corpus spent most of a 1.5 s page on it); the interpolation is unchanged, so results are bit for bit the same, a sample missing from a short stream still gives no value, and the maximum sample value is computed without overflowing at 32 bits per sample; (32) the decoded-image cache (24) keys a decode by the size the image is drawn at only when a JPX stream decodes it or its soft mask or mask (`decode_reads_target`): JPX is the only filter that decodes a reduced resolution for a size, so a JPEG or Flate scan was decoded again at every zoom step and kept once per zoom level; one decode now serves every size; `InterpreterCache::decoded_image_decodes` counts the cache's decodes; (33) Coons and tensor-product patches copy their vertices' colours less: each point of a patch's 20 × 20 triangle grid is interpolated once (`generate_patch_triangles`; each cell interpolated its four corners itself, 1,444 interpolations a patch for 400 points), a cell's second triangle takes the vertices its first copied, and the mesh sampler moves a patch's triangles into device space instead of copying each vertex (`TriangleSampler::sample_owned`); the same values, 10–16% faster on mesh-gradient pages; (34) `color::tint_transform_runs`: how many colours the calling thread has run through a Separation or DeviceN tint transform (counted once per batch in `convert_f32`), so a test counts patch (28)'s once-per-sample-value runs instead of timing them (its 10 s deadline failed on a busy FreeBSD VM in CI); (35) `EncodedShadingPattern::extent`: where a mesh shading (types 4–7) can paint at all, the pixels its triangles touched when it was encoded (`None` for the types that paint everywhere, and when the shading has a `/Background`), so the renderer samples a texture there alone and skips a mesh with nothing on the canvas; (36) a mesh shading's pixel whose corner no triangle holds strictly (on an edge two triangles share, or on the mesh's side) takes the triangle that covers its centre, edges included (`Triangle::covers`): testing only the corner left a white seam along every shared edge, and the colour is still taken at the corner; (38) the `/DefaultGray`, `/DefaultRGB` and `/DefaultCMYK` colour spaces of a resource dictionary (ISO 32000-2 §8.6.5.6) replace the device space content selects: with `g`/`G`, `rg`/`RG`, `k`/`K` and `cs`/`CS` (by name, or through a `/ColorSpace` resource that is a device space), and as the `/ColorSpace` of an image XObject, an inline image or a shading painted by `sh` or as a shading pattern (`DefaultColorSpaces`, looked up once per content window in the resources and those they inherit, so a form XObject without a default of its own draws in its page's, as Poppler and MuPDF draw it); a default that is a Lab, Indexed, Pattern or device space, has another number of components or is no colour space is ignored and device colour is drawn as before; device spaces inside other spaces (an Indexed base, a Separation or DeviceN alternate, a Pattern's underlying space), shading dictionaries, masks and transparency group spaces are unchanged, and the decoded-image cache (24) keys a decode by the default it was drawn in. hayro drew device colour as it was, so pages that calibrate it this way (common in PDF/A) came out in the wrong colours; (39) knockout groups (#2971): `Device::set_knockout` right after a `/K true` group is pushed; `Device::knockout_image` for an unmasked image before its group is pushed, when `Device::in_knockout_group`: the device clears its placement rectangle only if it will draw the image (on the canvas, not a preview's tiny image, and it decodes), decoded at its draw size so the draw reuses that decode; and, when `Device::knockout_erase_wanted`, a nested group's content drawn once between `begin_knockout_erase` and `end_knockout_erase` before the group itself, which then always gets a layer of its own; (40) an optional content membership dictionary's visibility expression (`/VE`: `/And`, `/Or`, `/Not`, nested up to 32 deep) decides its visibility ahead of `/OCGs` and `/P` (ISO 32000-2 §8.11.2.2); it was ignored, so content under it stayed visible with its layers hidden; a malformed expression, including one naming an object the document doesn't list in `/OCProperties /OCGs`, falls back to `/OCGs` and `/P` (#2986); (41) a transparency group that is neither isolated nor knockout, entered at full opacity with no soft mask and Normal blending, is drawn straight onto the page instead of into a layer of its own (§11.4.7: its elements composite with the backdrop; the layer made a `/BM /Multiply` highlight appearance an opaque block, #1968); (42) text in a Type 3 font is drawn filled whatever the text rendering mode, except mode 3 (invisible): only that mode has any effect on Type 3 fonts (ISO 32000-1 §9.3.6). Stroking ran the glyph procedure a second time, so `2 Tr` made bitmap fonts look bold, `1 Tr` painted an uncoloured glyph in the stroking colour, and `7 Tr` hid it (`7 Tr` since follows 64); (43) Lab colour spaces honour their own `/WhitePoint`: L\*a\*b\* goes to XYZ relative to that white, is Bradford-adapted to D65 and converted to sRGB (ISO 32000-2 §8.6.5.4) instead of through `LAB.icc` as if every white were D50 (a D65 blue came out 45 levels too red; the profile is removed); components normalised to [0, 1] (`manual_scale`, an `/Indexed` palette over Lab) map onto L\* 0..100 and the a\*/b\* `/Range`; (44) a Type 0 font whose `/ToUnicode` is the name `/Identity-H` or `/Identity-V` (not a CMap stream; WSCAD writes it) reads each code as its Unicode value, as Acrobat shows it (`Type0Font::identity_to_unicode`); the identity CMap the name resolved to maps codes to CIDs only, so such pages had no text to select or find; (46) an image XObject's group applies the fill opacity, so the fill alpha is reset to 1 before a stencil mask takes the fill colour as its paint, and a pattern paint drops the opacity it captured at `scn` (`Pattern::without_opacity`) (it was applied twice: a mask at `ca 0.5` was drawn at a quarter, #2975); (47) `path_as_rect`: a five-element path is a rectangle only when its four points are distinct, each step from one to the next runs along a single axis, and the last returns to the first. The length test admits one MoveTo and four LineTo as well as one MoveTo, three LineTo and a ClosePath, and asking only that every corner of the bounding box was visited says nothing about the order or the steps between, so a five-point polyline whose ends reached the corners was painted as that box -- a fourth side the path does not have -- and a clip of that shape was discarded; the coordinates are compared exactly, since the matrix is applied after this check and a `cm` of 10,000 turns `is_nearly_equal`'s 256th of a unit into tens of device units (#3327); (48) an `/ICCBased` colour space is cached by what its profile is read from (`IccProfileKey`): the stream object it names, looked up without reading the stream, and then the stream's dictionary and its own bytes (decrypted, not decoded), so a byte-for-byte copy of a profile in another object is read once, and a profile stream written directly in its array, which reports the object holding the array as every other stream written there does, is told apart by its bytes. hayro keyed it by the stream dictionary's bytes alone, so two different profiles with the same `/N` and `/Length` shared whichever was read first, and colours in the second were drawn through the first. `InterpreterCache::icc_profile_reads` counts the profiles read; (49) `OpenTypeFontBlob::new` also accepts a `post` table version 1.0, 2.0 or 3.0 written with its 16-bit halves swapped (`00 00 00 03`), as ScanSoft PDF Create! embeds every font; such fonts were refused and their text drawn with a substitute (Hebrew came out garbled or missing, #2946); `glyph_names` reads such a table from a copy with the version written correctly, so a 2.0 table's names serve `/Differences`. Any other unknown version is still refused (pdf.js 9462); (50) `CffFontBlob::glyph_index_by_name` looks a glyph name up in a name to glyph map built on first use (shared between clones) instead of searching the whole charset, with a string lookup per glyph, for every character code of a font; the first glyph with a name still wins; (51) at most 4096 graphics states saved at once, holding at most 64 MiB together (each `q` copies the whole state, dash array and clip list included, so short operators multiplied one large state); a `q` past either limit is counted but not copied, and its `Q` still removes the clips added since (from storytold/craft-libs#239); (52) a `/W2` run whose first CID is past `MAX_CID` stops there, as in `/W` (7), checked before the run's metrics are read: a first CID of 4294967295 inserted one entry and then overflowed, discarding the whole `/W2` table (from storytold/craft-libs#254); (55) a Separation or DeviceN tint transform that cannot be read keeps its colour space, and one that cannot be read or evaluated paints the alternate space's initial colour, as evaluation failures already did; that fallback, and a `cs`/`CS` colour space that cannot be read (drawn in DeviceGray as before), are now reported once per content stream as `InterpreterWarning::ColorFallback` (`RenderWarning::ColourFallback` in `pdfcraft-render`) instead of silently (#4260); (56) an annotation whose Invisible flag (bit 1) is set and whose `/Subtype` is none of the standard types of ISO 32000-2 Table 171 (`is_standard_annotation_type`; no `/Subtype` counts as unknown) is not drawn (Table 167), and a `/Popup` annotation's appearance is never painted on the page (§12.5.6.14: a popup has no appearance stream of its own; the viewer shows it as a window for its parent annotation). Both were drawn from their `/AP`; (57) a simple font drawn with a standard-font stand-in because the document names a font it doesn't embed (`StandardKind` with `fallback`) narrows a glyph to a narrower `/Widths` advance but no longer widens it to a wider one: the advance already places the next glyph (ISO 32000-2 §9.2.4; /Widths in Table 109), and widening drew a non-embedded `Arial,Bold` with 600-unit widths as a heavy fixed-pitch face whose `i` and `l` were solid blocks twice their width; an exact standard-14 font still matches its standard metrics both ways; (58) the fallback request for a Type 0 font that is not embedded (`FallbackFontQuery::with_descriptor`) takes its serif, italic and small-cap flags, `/FontWeight`, `/FontFamily` and `/FontStretch` from the descendant CIDFont's font descriptor (ISO 32000-2 §9.7.4.1, §9.7.6.1; §9.8): it read them from the Type 0 dictionary, which has none, so every such request asked for a regular sans-serif face and no resolver could tell a Song or Mincho font from a Hei or Gothic one, or see `/FontWeight 700`; (59) the NoRotate flag (`/F` bit 5) is honoured: the appearance, fitted to `/Rect`, is turned against the page's `/Rotate` about the rectangle's upper-left corner, which stays where it is, so it reads upright as the page is displayed (ISO 32000-2 §12.5.3). Sticky notes, which are saved NoRotate, were drawn turned with the page (#2583). Form fields and links stay in `/Rect`, where PdfCraft takes their click targets from it; (64) text in a Type 3 font follows ISO 32000-2 §9.3.6 in the clipping modes: under `7 Tr` it is not rendered (reported to the device as clipping text, `GlyphDrawMode::Clip`, so extraction keeps it), and in modes 4 to 7 a Type 3 glyph adds nothing to the clipping path (`clip_glyph`), so text shown only in a Type 3 font leaves no clip at `ET`. Since (42) `7 Tr` painted the glyphs in the fill colour (black by default), so a fill meant to follow showed them wherever it didn't cover them; 4 to 6 still render filled; (66) a Separation colour space whose colourant is the special name `All` paints every colourant at once (ISO 32000-2 §8.6.6.4): on screen and in an RGB print raster its tint t is the gray 1 − t, complemented because the device is additive, whatever its alternate space and tint transform say (hayro drew it through the tint transform, so a `/DeviceCMYK` transform giving cyan drew registration marks light cyan); `/None` still paints nothing, any other name still goes through its alternate, and a DeviceN space naming `All` keeps its alternate, since the clause gives the name that meaning in a Separation space; (63) a luminosity soft mask whose group colour space is `/DeviceCMYK` (by name, not an ICC-based or calibrated space) reads a `/DeviceCMYK` fill or stroke colour drawn in it, and its `/BC`, as their gray `1 - min(1, 0.3 c + 0.59 m + 0.11 y + k)` (ISO 32000-2 §10.4.2.3, §11.5.3: device colour converts to gray with no colour calibration; `Context::luminosity_in_device_cmyk`, `color::device_cmyk_to_gray`): it went through the CMYK display profile, where 100% K is a dark grey, so a 100% K mask let about an eighth of the masked paint through. CMYK images, shadings and tiling patterns inside such a mask, and CMYK on the page, still use the profile; (60) an `/Indexed` colour space that breaks ISO 32000-2 §8.6.6.3 is drawn as other readers draw it: a lookup table shorter than m × (hival + 1) bytes (or missing, or not a string or decodable stream) is padded with zero bytes, a `hival` past 255 is taken as 255, and a real `hival` is truncated toward zero; an image drawn so reports `InterpreterWarning::MalformedIndexedImage`, surfaced as `RenderWarning::MalformedIndexedImage`, and a fill or stroke colour in such a space reports `ColorFallback`. Upstream treated a short table as an unreadable space and fell back to DeviceGray, painting the indices near-black, even ones whose entries were present. A `hival` below zero or not a number has no reading: the space falls back as upstream, and an image in it reports the warning too. A lookup string given by an indirect reference is read (upstream consumed the reference trying it as a stream and found no table). The warning is taken from the space as built, so no table is read twice: inline images naming the space through a resource read its table once however many there are (`InterpreterCache::indexed_lookup_reads`) + | (1) `raster::tests::widget_appearance_states`; (2, 3) `xtask text-oracle` on corpus/pdfjs issue918; (4) `raster::tests::layer_override_hides_content`; (5) `raster::tests::soft_mask_without_group_colour_space_is_applied`; (6) `raster::tests::self_referencing_tiling_pattern_terminates`, `raster::tests::self_referencing_type3_glyph_terminates`; (7) `raster::tests::huge_cid_width_ranges_terminate`; (8) `raster::tests::hiding_comments_keeps_fields`; (9) `raster::tests::absurd_mask_and_fax_sizes_are_refused_before_decoding`; (10) `raster::tests::type3_glyphs_that_fan_out_into_themselves_finish`; (11) `raster::tests::type1_font_programs_with_long_runs_of_numbers_load_quickly`; (12) `raster::tests::degenerate_annotation_appearances_are_skipped`; (13) `raster::tests::highlights_without_an_appearance_are_drawn`; (14) `raster::tests::malformed_simple_font_widths_keep_later_widths_and_the_font`, `raster::tests::huge_simple_font_first_char_terminates`; (15) `raster::tests::annotation_appearance_box_maps_onto_rect`; (16) `raster::tests::clipping_path_applies_after_any_painting_operator`; (17) `raster::tests::dropping_the_pool_cancels_its_renders`, `raster::tests::the_watchdog_stops_the_render_it_gives_up_on`; (18) `raster::tests::page_contents_and_nested_forms_share_one_budget`, `raster::tests::content_painted_again_and_again_keeps_drawing`, `raster::tests::sampled_functions_with_billions_of_samples_are_refused`; (19) `raster::tests::mesh_shadings_far_off_the_page_render`, `raster::tests::many_overlapping_mesh_triangles_finish`, `raster::tests::mesh_shadings_with_many_patches_or_no_rows_finish`; (20) `raster::tests::function_limits_calculator_arithmetic_and_depth`, `raster::tests::function_limits_public_construction_work_and_depth`, `raster::tests::function_limits_stitching_cycles_are_refused_and_shared_children_work`, `raster::tests::function_limits_invalid_calculator_keeps_the_page_renderable`; (21) `raster::tests::cids_a_substitute_lacks_are_not_drawn_as_other_glyphs`; (22) `raster::tests::variation_sequences_draw_their_kanji_in_a_substitute`; (23) `raster::tests::axial_and_radial_shadings_draw_as_native_gradients`; (24) `raster::tests::tiles_match_full_render`, `raster::tests::skipping_paths_off_the_canvas_changes_no_pixel` (both decode through the cache), `raster::tests::decoded_images_with_identical_dictionaries_keep_their_own_pixels`, `raster::tests::decoded_images_under_a_transfer_function_are_not_shared`; (25) `text::tests::actual_text_replaces_the_glyphs_it_marks_in_every_writing_direction`, `text::tests::actual_text_follows_the_pdf_rules_for_marked_content`; (26) `raster::tests::accented_letters_the_standard_stand_ins_lack_are_built_from_their_parts`, `raster::tests::comma_below_letters_in_standard_fonts_draw_a_comma_under_the_letter`; (27) `text::tests::truetype_without_an_encoding_reads_its_codes_as_win_ansi`; (28) `raster::tests::spot_colour_images_run_their_tint_transform_once_per_sample_value`, `raster::tests::images_converted_through_a_table_keep_every_samples_colour`, `raster::tests::a_malformed_tint_for_an_unused_sample_keeps_the_used_colours`; (29) `raster::tests::reading_colors_keep_highlight_annotations_readable`; (30) `text::tests::the_annex_d2_alias_codes_read_as_an_ordinary_hyphen_and_space`; (31) `raster::tests::shadings_with_large_sampled_functions_paint_quickly`, `raster::tests::sampled_functions_interpolate_their_grid_as_before`; (32) `raster::tests::decoded_images_serve_every_size_they_are_drawn_at`; (33) `raster::tests::patch_meshes_keep_their_corner_colours`; (34) `raster::tests::spot_colour_images_run_their_tint_transform_once_per_sample_value`; (35) `mesh_extent_tests::{a_mesh_paints_its_triangles_alone, a_mesh_keeps_the_graphics_state_alpha, a_mesh_off_the_canvas_paints_nothing}` (`cargo test -p hayro`); (36) `raster::tests::mesh_triangles_sharing_an_edge_leave_no_seam`; (38) `raster::tests::default_colour_spaces::device_rgb_operators_draw_in_default_rgb`, `raster::tests::default_colour_spaces::device_cmyk_operators_draw_in_default_cmyk`, `raster::tests::default_colour_spaces::device_gray_operators_draw_in_default_gray`, `raster::tests::default_colour_spaces::device_images_draw_in_the_default_colour_spaces`, `raster::tests::default_colour_spaces::device_shadings_draw_in_the_default_colour_spaces`, `raster::tests::default_colour_spaces::an_image_drawn_with_and_without_a_default_keeps_both_colours`, `raster::tests::default_colour_spaces::forms_inherit_the_default_colour_spaces_they_do_not_override`, `raster::tests::default_colour_spaces::unusable_defaults_leave_device_colour_alone`; (39) `raster::tests::knockout_groups_replace_what_their_elements_cover`; (40) `raster::tests::visibility_expressions_decide_what_a_membership_dictionary_shows`; (41) `raster::tests::a_multiply_highlight_inside_a_transparency_group_keeps_the_text`; (42) `raster::tests::rendering_modes_but_3_and_7_draw_type3_glyphs_as_filled`; (43) `raster::tests::lab_honours_the_space_white_point`, `raster::tests::indexed_palettes_over_lab_keep_their_colours`; (44) `text::tests::identity_to_unicode_names_give_text`; (46) `raster::tests::stencil_masks_are_drawn_at_the_fill_opacity_once`; (47) `raster::tests::a_five_point_polyline_is_not_painted_as_its_bounding_box`, `raster::tests::a_path_that_visits_every_corner_out_of_order_is_not_a_rectangle`, `raster::tests::a_rectangle_is_still_painted_as_a_rectangle`; (48) `raster::tests::icc_profile_cache::icc_profiles_with_identical_dictionaries_draw_their_own_colours`, `raster::tests::icc_profile_cache::icc_profiles_used_again_are_read_once`; (49) `raster::tests::a_font_whose_post_version_has_swapped_halves_is_still_used`; (50) `font::blob::tests::glyph_names_resolve_to_the_first_glyph_with_that_name`; (51) `context::saved_state_tests::repeated_saves_stay_bounded_and_balanced`; (52) `font::cid::tests::w2_run_past_the_largest_cid_keeps_the_rest_of_the_table` (hayro-interpret); (55) `raster::tests::devicen_inks::a_tint_transform_that_cannot_be_evaluated_draws_the_rest_of_the_page_and_warns`, `raster::tests::devicen_inks::a_tint_transform_that_cannot_be_read_falls_back_and_warns`; (56) `raster::tests::invisible_unknown_annotations_and_popups_paint_nothing`; (57) `substitute_widths::a_substitute_keeps_its_glyph_shapes_at_wider_advances`, `substitute_widths::a_substitute_still_narrows_glyphs_to_narrower_advances` (pdfcraft-render `tests/it`); (58) `cid_fallback_query::a_cid_font_fallback_request_carries_its_descendants_style`, `cid_fallback_query::a_cid_font_without_a_descriptor_asks_for_the_default_style` (pdfcraft-render integration tests); (59) `raster::tests::no_rotate_notes_stay_upright_at_their_corner_on_rotated_pages` (hayro-interpret); (64) `type3_text_clip::{clip_only_type3_text_is_not_rendered_and_leaves_no_clip, invisible_type3_text_is_not_rendered, other_modes_render_type3_glyphs_filled_and_never_clip, only_outline_glyphs_shape_a_clip_shared_with_type3_text, self_nesting_type3_text_in_a_clip_mode_stays_bounded}` (`cargo test -p pdfcraft-render --test it`), `raster::tests::rendering_modes_but_3_and_7_draw_type3_glyphs_as_filled`, `text::tests::clipping_mode_type3_text_is_kept_by_the_ocr_filter`; (66) `separation_all::{separation_all_fills_paint_the_complemented_tint_as_gray, separation_all_strokes_paint_the_complemented_tint_as_gray, separation_all_images_paint_each_sample_as_the_complemented_gray, separation_all_shadings_paint_the_complemented_gray, separation_none_paints_nothing, a_named_separation_still_paints_through_its_alternate, a_device_n_space_naming_all_still_paints_through_its_alternate}` (`cargo test -p pdfcraft-render --test it`); (63) `color::device_cmyk_gray_tests::{each_ink_darkens_by_its_iso_weight, ink_past_full_coverage_is_black_and_bad_components_read_as_zero}` (hayro-interpret), `it::luminosity_group_space::{black_ink_in_a_device_cmyk_group_masks_fully_and_no_ink_shows_fully, each_process_ink_in_a_device_cmyk_group_masks_by_its_weight, a_device_cmyk_backdrop_colour_masks_by_its_gray, gray_and_rgb_groups_mask_as_before, device_cmyk_on_the_page_keeps_its_display_colour}` (pdfcraft-render); (60) `crates/render/tests/it/indexed_lookup.rs` (`cargo test -p pdfcraft-render --test it indexed_lookup`), `raster::tests::indexed_lookup_cache` | | hayro | 0.7.1 | Apache-2.0 OR MIT | (1) `RenderSettings::{x_offset, y_offset}` to render one tile of a large page; (2) images over `MAX_IMAGE_PIXELS` (2^28) are skipped (a fuzzed `/W 4294967295` hung in resampling); (3) `tiling_cell_scale` keeps a tiling pattern's cell pixmap within 3000 px a side whatever its `/XStep`/`/YStep` (a fuzzed step far beyond `/BBox` allocated 2 GB); (4) stroke widths clamped to 4 × (canvas width + height) device pixels (a fuzzed `/LW 9223372036854775807` allocated 10 GB in stroke expansion); (5) `usable_dash_pattern`: a stroke whose dash pattern could cut it into more than `MAX_DASHES_PER_STROKE` (250,000) pieces, counting kurbo's restart of the pattern at every subpath and `ClosePath`, is drawn solid (a fuzzed `/D [[11] 0]` on a line from x = 92234775807 allocated over 5 GB, `[0.000001] 0 d` on a 20 pt line over 10 GB); (6) a dash pattern with a negative or non-finite entry, or a period that isn't positive, is drawn solid (`[-1 -1] 0 d` looped for ever in kurbo); (7) `render_size` and `render_into`: render into a buffer the caller allocates (PdfCraft renders into pixel-aligned memory that the GUI takes over as texture data); `render` is `render_into` on a new pixmap; (8) checked image source/target/RGBA, alpha-premultiplication scratch and intermediate sizes before resampling, finite transforms and exact source/mask buffers; sources wider than u16 may shrink before backend limits, with a 2^20 source-axis cap and conservative 128 MiB filter-planning estimate; an over-limit or failed resize keeps the original image resolution and geometry when the source fits the backend, otherwise skips it, with a logged warning; staging, resampling output/scratch, RGBA conversion and padding buffers use fallible reserves; (9) `may_paint_canvas`: a fill or stroke whose control-point box, grown by how far its paint reaches past the path (half the stroke width times the larger of the miter limit and √2) plus 0.5 user units and 2 pixels, misses the canvas is not drawn, so a tile no longer strokes and fills every path of its page (item 5 of #623); `set_skip_offscreen_paths` turns this off for the regression test; (10) `/Interpolate` governs the pre-resize too (#624): a minified image that does not request interpolation is pre-resized with nearest-neighbour, planned without the Catmull-Rom filter limits (`image_resampling_size` takes the flag); a Type 3 glyph image is sampled bicubically only when interpolated; a minified image with a matching soft mask is nearest-sampled straight into premultiplied RGBA; (11) `may_paint_rect`: an image whose stream-size box, mapped to the device, misses the canvas is skipped before decoding (decoded data can only shrink against that box), so a tile of an image-heavy page no longer decodes, resamples and rasterizes every image on it; (12) preview renders (`InterpreterSettings::preview`: backdrops, thumbnails, grid cells, signature-drag backgrounds) skip images under `PREVIEW_MIN_IMAGE_AREA` device pixels, while final renders always draw them; (13) reading colours (#897): `RenderSettings::color_replacement` takes a `ColorReplacement` (text colour, background, `only_black`, `line_art`) that remaps solid text and line-art paint while rendering, glyph outlines counting as text and stencil-mask ink as line art; images, image alpha masks, soft masks and pattern cells keep their colours; `render_pdf` gains a `color_replacement: Option` argument (pass `None` for true colours); the upstream-only `[dev-dependencies.log]` entry is removed from `Cargo.toml` (log is a normal dependency since the logging patch), so `cargo test -p hayro` runs this crate's unit tests; (14) a shrunk, interpolated, axis-aligned image is snapped to whole device pixels: shrunk once to its rounded device rectangle and drawn 1:1 at the rounded corner, instead of a ceil-sized raster resampled again at a fractional scale and offset (`snap_minified_image`; scans looked soft next to Acrobat); (15) a soft mask of another size than its image is filtered when shrunk, whatever its `/Interpolate` (#1649): such a mask carries the shape on its own (Word saves text as a tiny black image with a glyph-bitmap mask), and nearest-neighbour dropped every stroke thinner than the shrink factor; (16) under a soft mask or a blend mode, axial and radial shadings are sampled instead of drawn as native gradients (#2128): vello_cpu 0.0.8 copies an opaque gradient over fully covered tiles without the mask or blend mode, so PowerPoint's faded gradient fills came out solid; (17) the renderer keeps annotation paints out of the reading-colours mapping between `begin_annotations` and `end_annotations` (#1548); (18) the page renders as one transparency group; under reading colours its background is painted inside it as the initial backdrop, so a Multiply highlight blends with the dark background instead of staying an opaque block, and otherwise the background stays outside, so a blend mode over an empty part of the page keeps its colour instead of blending with the white paper (#2970); (19) `set_paint` samples a mesh shading's texture over its triangles alone (`EncodedShadingPattern::extent`, padded rather than repeated) and skips a mesh with none on the canvas: a BricsCAD plan's 310 one-triangle `sh` each covered the whole tile, 2 s per 1024 px tile at any zoom (the watchdog gave the page up on a busy machine), now about 70 ms; (20) a tiling pattern whose step is too large for a cell pixmap, where what is painted (a fill, or a stroke with its width) can only show the first cell (PowerPoint's `/XStep 2147480100` for a picture fill drawn once), has its `/BBox` drawn at full scale with a transparent border and padded instead of tiled (`shows_one_cell`; a cell whose scale isn't finite and positive, from a `/BBox` too thin to divide by, keeps the old path: `usable_cell_scale`); shrinking the cell to fit the 3000 px limit of (3) left the picture under a pixel, so network diagrams lost their pictures (#3349); (22) knockout groups: while one is innermost, fills, strokes, glyph outlines and shadings are drawn with `Compose::Copy` (weighted by coverage), so each element replaces what earlier elements of the group drew instead of compositing over it; an image or nested group, drawn into a layer of its own, first has what is under its shape cleared (`knockout_image` for an unmasked image's rectangle, only when `draw_image` will draw it, decoded at the size it is drawn; a knockout erase pass that draws a nested group's content with every paint clearing instead of painting, a shading fill's clear clipped to the shading's `/BBox` as its paint is) and then composites over it (#2971); (23) backport of hayro 0.8: shadings that `as_svg_gradient` can represent are painted as a `vello_cpu` linear or two-point radial gradient instead of a texture sampled pixel by pixel through the shading function (InDesign fades are luminosity soft masks of one full-page axial `sh`: 40 on a report page took 15 s at 144 dpi, now under 1 s); without upstream's half-pixel shift, which suits vello_cpu 0.3 but moves 0.0.8's gradients off pixel centres; (25) a tiling cell is drawn at the scale that makes one `/XStep` (`/YStep`) exactly the cell pixmap's whole number of pixels (`tiling_cell_pixels`, at least 1 and at most the cap), so the repeated pixmap puts cells at whole multiples of the step at any device scale; drawing at the unrounded scale repeated the cell up to half a pixel off at every step, and the error grew with distance from the pattern's origin (ISO 32000-2 8.7.3.1) | (1) `raster::tests::tiles_match_full_render`; (2) `raster::tests::absurd_image_dimensions_are_skipped`; (3) `raster::tests::tiling_cells_stay_small_whatever_the_step`; (4) `raster::tests::absurd_predictor_columns_and_line_widths_render`; (5) `raster::tests::dash_patterns_that_cut_a_stroke_into_billions_of_pieces_render`; (6) `raster::tests::invalid_dash_patterns_terminate_and_are_drawn_solid`; (7) `raster::tests::pixels_are_a_plain_renders_bytes_in_a_buffer_a_gui_can_take_over`; (8) `raster::tests::image_resampling_rejects_anisotropic_target_growth`, `raster::tests::image_resampling_bounds_intermediate_before_planning`, `raster::tests::image_resampling_rejects_invalid_sources_and_scales`, `raster::tests::image_resampling_keeps_valid_sizes_and_exact_limits`, `raster::tests::resampling_fallback_preserves_original_geometry_and_pixels`, `raster::tests::image_resampling_wide_sources_shrink_before_backend_side_limits`, `raster::tests::image_resampling_wide_transparent_sources_still_shrink`, `raster::tests::image_resampling_padded_backend_dimensions_stay_checked`, `raster::tests::image_resampling_mismatched_alpha_mask_keeps_pixels_and_placement`; (9) `raster::tests::skipping_paths_off_the_canvas_changes_no_pixel`; (10) `raster::tests::non_interpolated_images_are_not_smoothed_when_minified`, `raster::tests::non_interpolated_gray_images_are_not_smoothed_when_minified`, `raster::tests::non_integer_minification_samples_the_same_columns`, `raster::tests::non_integer_minification_samples_colour_and_mask_from_the_same_columns`, `raster::tests::non_interpolated_alpha_images_keep_their_colours_and_mask_when_minified`, `raster::tests::non_interpolated_type3_glyph_images_are_not_smoothed_when_minified`, `raster::tests::image_resampling_nearest_accepts_wide_sources_the_filter_refuses`, `raster::tests::wide_non_interpolated_sources_draw_after_shrinking`; (11) `raster::tests::skipping_paths_off_the_canvas_changes_no_pixel`; (12) `raster::tests::preview_renders_skip_tiny_images_but_finals_draw_everything`; (13) `color_replacement_tests::{black_maps_to_text_colour, white_maps_to_background, midgrey_interpolates, alpha_copied_verbatim, only_black_threshold, line_art_false_skips_paths_but_not_text}`, `color_replacement_render_tests::{color_replacement_maps_text_and_line_art, color_replacement_maps_stencil_mask_ink}`; (14) `raster::tests::interpolated_minified_image_edges_snap_to_whole_pixels` (`cargo test -p hayro`); (15) `raster::tests::shrunk_soft_mask_larger_than_its_image_keeps_thin_strokes`; (16) `raster::tests::soft_mask_applies_to_a_shading_pattern_fill`; (17), (18) `raster::tests::reading_colors_keep_highlight_annotations_readable`; (19) `mesh_extent_tests::{a_mesh_paints_its_triangles_alone, a_mesh_keeps_the_graphics_state_alpha, a_mesh_off_the_canvas_paints_nothing}` (`cargo test -p hayro`); (20) `raster::tests::a_tiling_pattern_drawn_once_shows_its_cell`, `raster::tests::a_large_step_still_tiles_where_the_next_cell_shows`, `raster::tests::a_tiling_cell_too_thin_to_scale_keeps_the_old_path`; (22) `raster::tests::knockout_groups_replace_what_their_elements_cover`; (23) `raster::tests::axial_and_radial_shadings_draw_as_native_gradients`; (25) `tiling_cell_placement::{far_cells_of_a_fill_sit_at_whole_steps, far_cells_sit_at_whole_steps_under_a_turned_scaled_matrix, far_cells_of_a_stroke_sit_at_whole_steps, a_cell_pixmap_spans_exactly_one_step}` (`cargo test -p pdfcraft-render --test it`) | | hayro-jbig2 | 0.3.0 | Apache-2.0 OR MIT | (1) at most 2^28 pixels per bitmap, page bitmap, halftone grid and grayscale image (each side could be 65535: 4.3 billion pixels decoded one by one); (2) at most 2^26 symbol pixels per symbol dictionary (a fuzzed dictionary decoded for minutes); (3) a symbol dictionary's height classes are capped at one per new symbol and its export runs at `2 × symbols + 1`, each plus `EXTRA_EMPTY_STEPS` (1024): an empty height class or a zero-length export run doesn't advance, and past the end of its data the arithmetic decoder can decode them for ever (one changed byte hung the renderer); (4) a segment header's long-form referred-to segment count is capped at `u16::MAX` and a larger count is refused before its retention or referred-to bytes are read: the 29-bit count went straight to `Vec::with_capacity`, a reservation of about 2 GiB whose failure aborts the process (from storytold/craft-libs#176) | (1) `raster::tests::jbig2_regions_of_billions_of_pixels_are_refused`; (2) the nightly `cargo xtask fuzz` finding, which now finishes in about a second; (3) `raster::tests::jbig2_symbol_dictionaries_that_never_advance_terminate`; (4) `segment::tests::test_referred_to_count_is_capped` (`cargo test --manifest-path vendor/hayro-jbig2/Cargo.toml`) | | lopdf | 0.45.0 | MIT | (1) PNG predictor rows longer than the data are refused before their buffers are allocated (a fuzzed `/Columns 4294967295` allocated two 4 GiB rows), and empty frames return without row buffers; (2) predictor row widths use saturating arithmetic or reject overflow; (3) sub-byte TIFF Predictor 2 scratch is bounded by the samples present in the input instead of the declared `/Colors`, aligned with [lopdf#589](https://github.com/J-F-Liu/lopdf/pull/589); (4) the object lookup keeps its cycle history through compressed objects, requires an ordinary in-file container, and caps the chain of indirect lookups at 16 (a chain of new objects, each a `/Length` compressed in a container whose own `/Length` names the next, or a stream whose `/Length` names the next stream, overflowed the stack: a dev build's 2 MiB thread within 800 hops); (5) object-stream members are parsed within their own extent (up to the next distinct offset), an entry whose offset repeats an earlier entry's is skipped (so a repeated object number can resolve to its earlier entry), and offsets are processed in index order; (6) `find_latest_trailer`: a rebuild takes a `/Type /XRef` stream dictionary as the trailer when the file has no `trailer` keyword (the latest in the file wins, as a later `trailer` would), so a concatenated or truncated file keeps its /Info and /ID instead of failing to load; a candidate whose /Root is the catalog beats one whose /Root is another dictionary, and an in-file /Root must match its generation (#2365) | (1) `inspect::tests::png_predictor_rows_longer_than_the_data_are_refused_before_allocating`, `inspect::tests::empty_png_predictor_frames_do_not_allocate_rows`; (2) `inspect::tests::tiff_subbyte_predictor_row_width_overflow_is_refused`; (3) `inspect::tests::tiff_subbyte_predictor_scratch_is_bounded_by_available_samples`, `inspect::tests::tiff_subbyte_predictors_keep_components_rows_and_padding`; (4) `inspect::tests::compressed_object_self_container_does_not_recurse`, `inspect::tests::compressed_object_mutual_containers_do_not_recurse`, `inspect::tests::object_stream_length_cycle_does_not_recurse`, `inspect::tests::compressed_object_container_chain_does_not_recurse`, `inspect::tests::compressed_length_chain_does_not_recurse`, `inspect::tests::plain_stream_length_chain_does_not_recurse`, `inspect::tests::compressed_objects_and_indirect_stream_length_still_load`, `inspect::tests::compressed_length_in_a_container_with_an_indirect_length_still_loads`; (5) `inspect::tests::object_stream_member_extents_bound_owned_strings`, `inspect::tests::object_stream_member_extents_preserve_index_order_and_whitespace`; (6) `inspect::tests::rebuilt_file_keeps_the_info_of_its_cross_reference_stream`, `inspect::tests::rebuilt_file_prefers_a_trailer_whose_root_is_the_catalog` | diff --git a/vendor/hayro-interpret/src/cache.rs b/vendor/hayro-interpret/src/cache.rs index bbc26e163f..c727722c28 100644 --- a/vendor/hayro-interpret/src/cache.rs +++ b/vendor/hayro-interpret/src/cache.rs @@ -9,9 +9,10 @@ use std::sync::{Arc, Mutex}; type CacheMap = HashMap>>; /// The objects read from a document, by key. PdfCraft patch (48): the second field counts the -/// ICC profiles read through it ([`Cache::icc_profile_reads`]). +/// ICC profiles read through it ([`Cache::icc_profile_reads`]). PdfCraft patch (60): the third +/// counts the `/Indexed` lookup tables read ([`Cache::indexed_lookup_reads`]). #[derive(Clone)] -pub(crate) struct Cache(Arc>, Arc); +pub(crate) struct Cache(Arc>, Arc, Arc); impl Default for Cache { fn default() -> Self { Self::new() @@ -20,7 +21,11 @@ impl Default for Cache { impl Cache { pub(crate) fn new() -> Self { - Self(Arc::new(Mutex::new(HashMap::new())), Arc::default()) + Self( + Arc::new(Mutex::new(HashMap::new())), + Arc::default(), + Arc::default(), + ) } /// PdfCraft patch (48): count one ICC profile read, the colour space of an `/ICCBased` @@ -34,6 +39,17 @@ impl Cache { self.1.load(Ordering::Relaxed) } + /// PdfCraft patch (60): count one `/Indexed` lookup table read (a stream decoded or a + /// string copied). + pub(crate) fn count_indexed_lookup_read(&self) { + self.2.fetch_add(1, Ordering::Relaxed); + } + + /// PdfCraft patch (60): how many `/Indexed` lookup tables were read through this cache. + pub(crate) fn indexed_lookup_reads(&self) -> usize { + self.2.load(Ordering::Relaxed) + } + pub(crate) fn get_or_insert_with( &self, id: u128, diff --git a/vendor/hayro-interpret/src/color.rs b/vendor/hayro-interpret/src/color.rs index 638eae97bd..2255545eda 100644 --- a/vendor/hayro-interpret/src/color.rs +++ b/vendor/hayro-interpret/src/color.rs @@ -336,6 +336,12 @@ impl ColorSpace { matches!(self.0.as_ref(), ColorSpaceType::Indexed(_)) } + /// PdfCraft patch (60): whether this is an indexed colour space that broke ISO 32000-2 + /// §8.6.6.3 and was recovered (see `Indexed::new`). + pub(crate) fn is_recovered_indexed(&self) -> bool { + matches!(self.0.as_ref(), ColorSpaceType::Indexed(i) if i.recovered) + } + /// Get the default decode array for the color space. pub(crate) fn default_decode_arr(&self, n: f32) -> SmallVec<[(f32, f32); 4]> { match self.0.as_ref() { @@ -985,106 +991,112 @@ impl ToRgb for Lab { #[derive(Debug, Clone)] pub(crate) struct Indexed { + /// `hival + 1` entries of the base's component count each. values: Vec>, hival: u8, base: Box, + /// PdfCraft patch (60): the space broke ISO 32000-2 §8.6.6.3 and was recovered (see + /// `Indexed::new`). + recovered: bool, } impl Indexed { + /// PdfCraft patch (60): ISO 32000-2 §8.6.6.3 says `hival` "shall be an integer" no greater + /// than 255 and the lookup table "shall be m × (hival + 1) bytes long", but not how a reader + /// recovers when a file breaks this. Other readers draw such a space, and so does this: + /// - a real `hival` is truncated toward zero (upstream read it so too); + /// - a `hival` past 255 is taken as 255; + /// - a lookup table that is short, missing, or not a string or decodable stream is padded + /// with zero bytes to m × (hival + 1), so the entries it has keep their colours. + /// + /// Any of these sets `recovered`, which the image path reports. A `hival` below zero (once + /// truncated), or not a number, has no reading: the space cannot be read, as upstream, and + /// `has_unusable_hival` lets the image path report it. + /// + /// Upstream treated a short table or a `hival` past `u32` as unreadable, and an image that + /// named the space fell back to DeviceGray: its indices were drawn as near-black grey + /// levels, even indices whose entries were present. fn new(array: &Array<'_>, cache: &Cache) -> Option { let mut iter = array.flex_iter(); // Skip name let _ = iter.next::>()?; let base_color_space = ColorSpace::new(iter.next::>()?, cache)?; - let hival = iter.next::()?.min(u8::MAX as u32) as u8; + let (hival, hival_recovered) = indexed_hival(iter.next::()?)?; - let values = { - // PdfCraft patch (60): read once and then by type, so a string lookup given by an - // indirect reference is found (a failed stream read had already consumed the - // reference). - let data = iter.next::>().and_then(indexed_lookup)?; + let num_components = usize::from(base_color_space.num_components()); + // At most 256 entries of at most 255 components: 65 280 values whatever the input. + let entries = usize::from(hival) + 1; - let num_components = base_color_space.num_components(); - - let mut byte_iter = data.iter().copied(); - - let mut vals = vec![]; - for _ in 0..=hival { - let mut temp = vec![]; - - for _ in 0..num_components { - temp.push(byte_iter.next()? as f32 / 255.0); - } - - vals.push(temp); - } - - vals - }; + // PdfCraft patch (60): read once and then by type, so a string lookup given by an + // indirect reference is found (a failed stream read had already consumed the + // reference). + let data = iter + .next::>() + .and_then(|o| indexed_lookup(o, cache)); + let data: &[u8] = data.as_deref().unwrap_or_default(); + let mut bytes = data.iter().copied(); + let values: Vec> = (0..entries) + .map(|_| { + (0..num_components) + .map(|_| f32::from(bytes.next().unwrap_or(0)) / 255.0) + .collect() + }) + .collect(); + // `entries × num_components` cannot overflow (see above); a product that somehow did + // not fit counts as short, and the table is padded all the same. + let short = entries + .checked_mul(num_components) + .is_none_or(|needed| data.len() < needed); Some(Self { values, hival, base: Box::new(base_color_space), + recovered: hival_recovered || short, }) } } -/// PdfCraft patch (60): whether `object` is an `/Indexed` colour space that ISO 32000-2 -/// §8.6.6.3 rules out: `hival` is not an integer in 0..=255, or the lookup table is missing or -/// shorter than m × (hival + 1) bytes for a base space of m components. A longer table is fine -/// (its first bytes are used). The lookup may be a string or a stream, given in place or by -/// reference. An Indexed space whose base cannot be read is judged only by its `hival` (the -/// table's size depends on the base), so with a usable `hival` its existing handling is -/// unchanged. -/// -/// Upstream treated such a space as unreadable, and an image that named it then fell back to -/// DeviceGray: its indices were drawn as near-black grey levels, even indices whose entries -/// were present. The image path uses this to skip the image with a warning instead. -pub(crate) fn is_malformed_indexed(object: &Object<'_>, cache: &Cache) -> bool { +/// PdfCraft patch (60): an Indexed space's `hival` (ISO 32000-2 §8.6.6.3) as drawn, and whether +/// it had to be recovered: a real is truncated toward zero and a value past 255 is taken as 255. +/// `None` for a value below zero once truncated, or one that is not finite. +fn indexed_hival(number: object::Number) -> Option<(u8, bool)> { + let value = number.as_f64(); + let whole = value.trunc(); + // NaN fails this test too. + if !(whole >= 0.0 && whole.is_finite()) { + return None; + } + let clamped = whole.min(f64::from(u8::MAX)); + // `clamped` is a whole number in 0..=255, so the cast is exact. + Some((clamped as u8, clamped != value)) +} + +/// PdfCraft patch (60): whether `object` is an `/Indexed` colour space array whose `hival` has no +/// reading (below zero, or not a number), which `Indexed::new` cannot recover. Reads the array's +/// first and third entries only; the lookup table is not touched. +pub(crate) fn has_unusable_hival(object: &Object<'_>) -> bool { let Object::Array(array) = object else { return false; }; let mut iter = array.flex_iter(); - let Some(family) = iter.next::>() else { - return false; - }; - if !matches!(family.deref(), INDEXED | I) { + if !iter + .next::>() + .is_some_and(|family| matches!(family.deref(), INDEXED | I)) + { return false; } - let base = iter.next::>(); - let Some(hival) = iter.next::().and_then(|n| { - let value = n.as_f64(); - // A real with a fraction, a negative number or one past 255 is not a usable `hival` - // (NaN and the infinities fail the range test too). - (value.fract() == 0.0 && (0.0..=255.0).contains(&value)).then(|| n.as_i64()) - }) else { - // Judged before the base: an unusable `hival` rules the space out whatever the base. - return true; - }; - let Some(base) = base.and_then(|o| ColorSpace::new(o, cache)) else { - return false; - }; - // `hival` is at most 255 here, so this cannot overflow; it is checked all the same, and a - // size that did not fit would rule the space out rather than wrap. - let Some(needed) = usize::try_from(hival) - .ok() - .and_then(|h| h.checked_add(1)) - .and_then(|entries| entries.checked_mul(usize::from(base.num_components()))) - else { - return true; - }; - let lookup_len = iter - .next::>() - .and_then(indexed_lookup) - .map(|d| d.len()); - lookup_len.is_none_or(|len| len < needed) + let _base = iter.next::>(); + iter.next::() + .and_then(indexed_hival) + .is_none() } /// PdfCraft patch (60): an Indexed space's lookup table, a byte string or a stream (ISO 32000-2 /// §8.6.6.3), with a stream's filters applied. `None` for any other object or a stream that -/// cannot be decoded. -fn indexed_lookup(lookup: Object<'_>) -> Option> { +/// cannot be decoded. Each call is counted on `cache` (`Cache::indexed_lookup_reads`). +fn indexed_lookup<'a>(lookup: Object<'a>, cache: &Cache) -> Option> { + cache.count_indexed_lookup_read(); match lookup { Object::Stream(s) => s.decoded().ok(), Object::String(s) => Some(Cow::Owned(s.to_vec())), diff --git a/vendor/hayro-interpret/src/context.rs b/vendor/hayro-interpret/src/context.rs index d41f35ee83..1018fec395 100644 --- a/vendor/hayro-interpret/src/context.rs +++ b/vendor/hayro-interpret/src/context.rs @@ -153,6 +153,13 @@ impl<'a> InterpreterCache<'a> { pub fn icc_profile_reads(&self) -> usize { self.object_cache.icc_profile_reads() } + + /// PdfCraft patch (60): how many `/Indexed` lookup tables were read through this cache, for + /// diagnostics and tests. A colour space named by a resource is built, and its table read, + /// once however many images or colours use it. + pub fn indexed_lookup_reads(&self) -> usize { + self.object_cache.indexed_lookup_reads() + } } /// A per-page interpretation context that borrows shared data from an [`InterpreterCache`]. diff --git a/vendor/hayro-interpret/src/interpret/mod.rs b/vendor/hayro-interpret/src/interpret/mod.rs index 270aae9d58..b0e22346d0 100644 --- a/vendor/hayro-interpret/src/interpret/mod.rs +++ b/vendor/hayro-interpret/src/interpret/mod.rs @@ -161,9 +161,10 @@ pub enum InterpreterWarning { /// read, or a Separation or DeviceN tint transform that cannot be evaluated) and was painted /// in a fallback colour. Reported at most once per content stream. ColorFallback, - /// PdfCraft patch (60): an image was not drawn because its `/Indexed` colour space is - /// malformed (ISO 32000-2 §8.6.6.3): `hival` not an integer in 0..=255, or a lookup table - /// shorter than its entries need. + /// PdfCraft patch (60): an image's `/Indexed` colour space is malformed (ISO 32000-2 + /// §8.6.6.3) and the image was drawn recovered: a lookup table shorter than its entries need + /// padded with zero bytes, a `hival` past 255 taken as 255, or a real one truncated. A `hival` + /// below zero or not a number cannot be recovered, and the image falls back to DeviceGray. MalformedIndexedImage, } @@ -660,6 +661,11 @@ fn interpret_window<'a>( crate::color::note_color_fallback(); ColorSpace::device_gray() }); + // PdfCraft patch (60): so is an Indexed space drawn recovered (a short lookup + // table padded, a `hival` clamped or truncated; see `Indexed::new`). + if cs.is_recovered_indexed() { + crate::color::note_color_fallback(); + } if !cs.is_pattern() { context.get_mut().graphics_state.stroke_pattern = None; @@ -676,6 +682,11 @@ fn interpret_window<'a>( crate::color::note_color_fallback(); ColorSpace::device_gray() }); + // PdfCraft patch (60): so is an Indexed space drawn recovered (a short lookup + // table padded, a `hival` clamped or truncated; see `Indexed::new`). + if cs.is_recovered_indexed() { + crate::color::note_color_fallback(); + } if !cs.is_pattern() { context.get_mut().graphics_state.non_stroke_pattern = None; @@ -931,21 +942,23 @@ fn interpret_window<'a>( let warning_sink = context.settings.warning_sink.clone(); let transfer_function = context.get().graphics_state.transfer_function.clone(); let cache = context.interpreter_cache.object_cache.clone(); - // PdfCraft patch (60): an inline image naming a resource colour space that is a - // malformed Indexed space is skipped with a warning, as `ImageXObject::new` does - // for one given in place. A device space name (or its inline abbreviation) never - // refers to a resource (ISO 32000-2 §8.6.8), and `ImageXObject::new` takes it - // first, so a resource of that name is not consulted. + // PdfCraft patch (60): an inline image naming a resource colour space that is an + // Indexed space whose `hival` has no reading is reported, as `ImageXObject::new` + // reports one given in place; the image falls back as upstream. Only the array's + // name and `hival` are read here, never its lookup table: a recovered space is + // reported by `ImageXObject::new` from the space it builds through the resource + // cache. A device space name (or its inline abbreviation) never refers to a + // resource (ISO 32000-2 §8.6.8), and `ImageXObject::new` takes it first, so a + // resource of that name is not consulted. let inline_dict = i.0.dict(); if inline_dict .get::>(hayro_syntax::object::dict::keys::CS) .or_else(|| inline_dict.get::>(hayro_syntax::object::dict::keys::COLORSPACE)) .filter(|name| ColorSpace::new_from_name(name).is_none()) .and_then(|name| resources.get_color_space(&name)) - .is_some_and(|cs| crate::color::is_malformed_indexed(&cs, &cache)) + .is_some_and(|cs| crate::color::has_unusable_hival(&cs)) { warning_sink(InterpreterWarning::MalformedIndexedImage); - continue; } // PdfCraft patch: inline images are not cached (see `DecodedImageKey`). // PdfCraft patch (38): an inline image in a device space, named directly or by a diff --git a/vendor/hayro-interpret/src/x_object.rs b/vendor/hayro-interpret/src/x_object.rs index 9f6685b40e..e7350b882c 100644 --- a/vendor/hayro-interpret/src/x_object.rs +++ b/vendor/hayro-interpret/src/x_object.rs @@ -508,6 +508,7 @@ impl<'a> ImageXObject<'a> { is_mask |= is_stencil_mask; let mut default_space = None; + let mut malformed_indexed = false; let image_cs = if is_mask { // Masks are always single-channel. Some(ColorSpace::device_gray()) @@ -524,17 +525,12 @@ impl<'a> ImageXObject<'a> { if let Some((space, key)) = default { default_space = Some(key); Some(space) - } else if cs_obj - .as_ref() - .is_some_and(|c| crate::color::is_malformed_indexed(c, cache)) - { - // PdfCraft patch (60): an Indexed space that ISO 32000-2 §8.6.6.3 rules out - // (a short lookup table, an unusable `hival`) leaves no colour for some or all - // samples. The image is not drawn, and the caller is told, rather than falling - // back to DeviceGray and painting its indices as near-black. - warning_sink(InterpreterWarning::MalformedIndexedImage); - return None; } else { + // PdfCraft patch (60): an Indexed space given in place whose `hival` has no + // reading cannot be built and falls back below, as upstream; it is reported. + malformed_indexed = cs_obj + .as_ref() + .is_some_and(crate::color::has_unusable_hival); cs_obj .clone() .and_then(|c| ColorSpace::new(c, cache)) @@ -546,6 +542,14 @@ impl<'a> ImageXObject<'a> { }) } }; + // PdfCraft patch (60): an Indexed space that ISO 32000-2 §8.6.6.3 rules out (a short + // lookup table, a `hival` past 255 or with a fraction) is drawn recovered (see + // `Indexed::new`), and reported. The check asks the space that was built, so it reads + // no lookup table of its own (an inline image naming a resource builds it once, through + // the resource cache). + malformed_indexed |= image_cs + .as_ref() + .is_some_and(ColorSpace::is_recovered_indexed); let interpolate = dict .get::(I) @@ -559,6 +563,10 @@ impl<'a> ImageXObject<'a> { return None; } + if malformed_indexed { + warning_sink(InterpreterWarning::MalformedIndexedImage); + } + Some(Self { width, cache: cache.clone(),