M2: zero-area fills and clips: exact zero only, snap only within one row or column

`degenerate_fill_area` took a side under 1/256 unit as zero, so a thin
rectangle with area that the CTM scales up to whole pixels (100 x 1/1024
under a 4096x vertical scale covers four device rows) was clipped and
filled as a one-pixel band along its bounding-box diagonal. The fill
fallback had the same threshold and stroked it as a hairline. Both now
treat only an exactly zero side as degenerate; a box with area is
filled and clipped as the shape it is.

A slanted zero-height line less than a pixel high was snapped to its
whole bounding box, marking both rows along its full length when it
crossed a row boundary. Whole-pixel snapping is now used only when the
line stays within one device row or column (a line under 1/256 pixel
across is taken at its middle); otherwise it is the one-pixel band, so
each row is marked only where the line runs through it (ISO 32000-2
§10.7.4, half-open pixel squares the shape intersects).

Tests: crates/render/tests/it/zero_area_fill_clip.rs
a_thin_rectangle_scaled_up_to_whole_pixels_keeps_its_area and
a_slanted_zero_height_line_crossing_a_row_marks_each_row_only_where_it_runs
each fail with their part of the fix reverted.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0175FWj4xFhMhSH3UpMU8DSC

crates/render/tests/it/zero_area_fill_clip.rs

@@ -198,3 +198,37 @@ fn the_clip_is_restored_after_q() {
     let rgba = render(&format!("q 10 50 100 0 re W n Q {FLOOD}"));
     assert_eq!(red_pixels(&rgba).len(), (SIZE * SIZE) as usize);
 }
+
+#[test]
+fn a_thin_rectangle_scaled_up_to_whole_pixels_keeps_its_area() {
+    // 1/1024 pt high is under the rasteriser's "nearly zero" threshold, but the rectangle has
+    // area: scaled 4096 times vertically it covers device y 164..168, four whole rows, and is
+    // painted as the rectangle it is, not as a line across its bounding box.
+    let fill = render("q 1 0 0 4096 0 0 cm 1 0 0 rg 10 0.0078125 100 0.0009765625 re f Q");
+    assert_only_red_run(&fill, (10, 110), (164, 168), "scaled-up thin fill");
+    let clip = render(&format!("q 1 0 0 4096 0 0 cm 10 0.0078125 100 0.0009765625 re W n {FLOOD} Q"));
+    assert_only_red_run(&clip, (10, 110), (164, 168), "scaled-up thin clip");
+}
+
+#[test]
+fn a_slanted_zero_height_line_crossing_a_row_marks_each_row_only_where_it_runs() {
+    // A slight skew lifts the line from device y 149.52 at x = 10 to 148.74 at x = 110: under a
+    // pixel high in all, but crossing the boundary between rows 148 and 149 near x = 75. Each
+    // row is marked where the line runs through it, not along the whole length.
+    let skew = "1 0.008 0 1 0 0 cm";
+    let fill = render(&format!("q {skew} 1 0 0 rg 10 50.4 100 0 re f Q"));
+    let clip = render(&format!("q {skew} 10 50.4 100 0 re W n {FLOOD} Q"));
+    for (rgba, what) in [(&fill, "fill"), (&clip, "clip")] {
+        for x in [12, 15, 20] {
+            assert!(px(rgba, x, 149)[1] < 128, "{what}: ({x}, 149) on the line is {:?}", px(rgba, x, 149));
+            assert!(px(rgba, x, 148)[1] > 128, "{what}: ({x}, 148) off the line is {:?}", px(rgba, x, 148));
+        }
+        for x in [100, 105, 108] {
+            assert!(px(rgba, x, 148)[1] < 128, "{what}: ({x}, 148) on the line is {:?}", px(rgba, x, 148));
+            assert!(px(rgba, x, 149)[1] > 128, "{what}: ({x}, 149) off the line is {:?}", px(rgba, x, 149));
+        }
+        // One pixel wide: its ink adds up to its length, about 100 pixels, not two rows' worth.
+        let ink: f64 = rgba.as_chunks::<4>().0.iter().map(|p| 1.0 - f64::from(p[1]) / 255.0).sum();
+        assert!((90.0..=110.0).contains(&ink), "{what}: ink {ink}");
+    }
+}

vendor/README.md

@@ -39,7 +39,7 @@ 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 `.nLine truncated
+| 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 `.nLine truncated
 | 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<ColorReplacement>` 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 thLine truncated
 | 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` |

vendor/hayro-interpret/src/interpret/path.rs

@@ -58,22 +58,25 @@ pub(crate) fn apply_pending_clip<'a>(context: &mut Context<'a>, device: &mut imp
 }
 
 /// PdfCraft patch (65): the device-space area a fill marks for a path whose bounding box `bbox`
-/// (user space) has no width or no height, but not both, under `ctm`. ISO 32000-2 §10.7.4 paints
-/// "any pixel whose half-open square region intersects the shape, no matter how small the
-/// intersection is", and its example says "A zero-width or zero-height rectangle paints a line
-/// 1 pixel wide". Such a path encloses no area, so the rasteriser marks nothing: the fill used
-/// to fall back to a thinnest-line stroke along the box (half-covering two rows when the line
-/// fell on a pixel boundary), and a clip of it hid everything painted after it.
+/// (user space) has exactly no width or exactly no height, but not both, under `ctm`.
+/// ISO 32000-2 §10.7.4 paints "any pixel whose half-open square region intersects the shape, no
+/// matter how small the intersection is", and its example says "A zero-width or zero-height
+/// rectangle paints a line 1 pixel wide". Such a path encloses no area, so the rasteriser marks
+/// nothing: the fill used to fall back to a thinnest-line stroke along the box (half-covering two
+/// rows when the line fell on a pixel boundary), and a clip of it hid everything painted after it.
+/// A box with area, however thin, is not one of these: scaled up it can cover many pixels.
 ///
-/// When the line lies along a device axis (less than a pixel across the other way), the area is
-/// the run of whole pixels it touches, by the half-open rule: from the floor of its low ends to
-/// the ceiling of its high ends, at least one pixel either way. Otherwise (a rotated or skewed
-/// `ctm`) it is a band one device pixel wide centred on the line, like the thinnest line of
-/// §8.4.3.2. `None` for a box with area, a point (both sides zero), a non-finite box, or a `ctm`
-/// that collapses the line to a point: those paint as before.
+/// When the line stays within one device row or column (counting a line less than 1/256 of a
+/// pixel across as lying exactly along the axis, at its middle), the area is the run of whole
+/// pixels it touches, by the half-open rule: from the floor of its low end to the ceiling of its
+/// high end, at least one pixel. Otherwise (a rotated or skewed `ctm` that carries it across
+/// rows and columns) it is a band one device pixel wide centred on the line, like the thinnest
+/// line of §8.4.3.2, so each row or column is marked only where the line runs through it. `None`
+/// for a box with area, a point (both sides zero), a non-finite box, or a `ctm` that collapses
+/// the line to a point: those paint as before.
 pub(crate) fn degenerate_fill_area(bbox: Rect, ctm: Affine) -> Option<BezPath> {
-    let flat_x = (bbox.width() as f32).is_nearly_zero();
-    let flat_y = (bbox.height() as f32).is_nearly_zero();
+    let flat_x = bbox.width() == 0.0;
+    let flat_y = bbox.height() == 0.0;
     if flat_x == flat_y {
         return None;
     }
@@ -88,11 +91,9 @@ pub(crate) fn degenerate_fill_area(bbox: Rect, ctm: Affine) -> Option<BezPath> {
         return None;
     }
     let device = Rect::from_points(a, b);
-    if device.width() < 1.0 || device.height() < 1.0 {
-        let x0 = device.x0.floor();
-        let y0 = device.y0.floor();
-        let x1 = device.x1.ceil().max(x0 + 1.0);
-        let y1 = device.y1.ceil().max(y0 + 1.0);
+    let (x0, x1) = touched_pixels(device.x0, device.x1);
+    let (y0, y1) = touched_pixels(device.y0, device.y1);
+    if x1 - x0 <= 1.0 || y1 - y0 <= 1.0 {
         return Some(Rect::new(x0, y0, x1, y1).to_path(0.1));
     }
     let half = Vec2::new(-along.y, along.x) * (0.5 / length);
@@ -105,6 +106,21 @@ pub(crate) fn degenerate_fill_area(bbox: Rect, ctm: Affine) -> Option<BezPath> {
     Some(band)
 }
 
+/// PdfCraft patch (65): the device pixel boundaries `(first, end)` of the pixels a span `lo..=hi`
+/// (`lo <= hi`, finite) touches along one axis by the half-open rule, at least one pixel. A span
+/// under 1/256 of a pixel is taken at its middle, so a line along a device axis that a CTM's
+/// rounding tilts slightly still marks one row or column.
+fn touched_pixels(lo: f64, hi: f64) -> (f64, f64) {
+    let (lo, hi) = if ((hi - lo) as f32).is_nearly_zero() {
+        let mid = lo + (hi - lo) / 2.0;
+        (mid, mid)
+    } else {
+        (lo, hi)
+    };
+    let first = lo.floor();
+    (first, hi.ceil().max(first + 1.0))
+}
+
 pub(crate) fn close_path(context: &mut Context<'_>) {
     // This is necessary to prevent artifacts (see for example issue 157),
     // but it does cause some weird lines (maybe conflation artifacts) in
@@ -148,10 +164,10 @@ pub(crate) fn fill_path_impl<'a>(
             return;
         }
 
-        match (
-            (bbox.width() as f32).is_nearly_zero(),
-            (bbox.height() as f32).is_nearly_zero(),
-        ) {
+        // PdfCraft patch (65): only a box with exactly no width or height (by now a point) is
+        // stroked. A box with area, however thin, is filled as the shape it is: a "nearly zero"
+        // side scaled up by the CTM can cover many device pixels.
+        match (bbox.width() == 0.0, bbox.height() == 0.0) {
             (false, false) => {
                 let draw_mode = PathDrawMode::Fill(fill_rule);
                 if let Some(rect) = path_as_rect(path) {