Commit Graph
52 Commits
Author SHA1 Message Date
Behdad EsfahbodandClaude Opus 4.8 a3ad47c776 [ft] Report zero color stops when skipping to start fails
hb_color_line_get_color_stops() defines *count as in/out: capacity on
entry, number of stops written on exit.  The skip-to-start early return
added in #6121 (to avoid an infinite loop when FT_Get_Colorline_Stops()
refuses a stop) writes no stops but left *count untouched, so an embedder
paging a malformed ColorLine with start > 0 reads uninitialized
hb_color_stop_t entries and draws the gradient from its own buffer.

Set *count = 0 on that exit, matching the start >= num_color_stops exit
a few lines above and the hb-ot ColorLine::get_color_stops().  Well-formed
fonts are unaffected; no API/ABI change.

Adds test_color_stops_short_ft with a generated 392-byte font whose
ColorLine declares one more stop than it holds.

Reported by Anthropic's OSS Scanner as ordinary bug ANT-2026-O5RJ72EZ.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-10-31 22:45:12 -06:00
Behdad EsfahbodandClaude Opus 4.8 b732218126 [paint] Clip sweep-gradient tiles to 0..2π in hb_paint_sweep_gradient_tiles
hb_paint_sweep_gradient_tiles() documents that it iterates the 0..2π
sweep, and all three consumers (hb-cairo mesh patches, the SVG and PDF
writers) size their per-tile loops by tile width on that assumption.
Several emit sites (the start==end PAD case, the first PAD stop past 0,
and repeat/reflect segments straddling 0 or 2π) passed the angles
straight through, and COLRv1 variation deltas can place them hundreds of
thousands of radians outside the circle, so one tile could expand to
millions of patches: a 504-byte variable font drawn off its default
allocated ~1.6 GiB through hb-cairo.

#6157 bounded the number of sweep tiles; this bounds the width of each
one.  Route every emit through a small clipping lambda inside the tiler
that clamps the tile to [0, 2π], lerps the end colours at the clip
points, and drops tiles entirely outside, so the contract holds at the
single place it is established rather than in each backend.
HB_PAINT_MAX_SWEEP_TILES and the repeat-loop bounds are untouched;
external callers need no change.

Adds a regression test calling the public function with far-out angles
for PAD/REPEAT/REFLECT; it fails without the fix.

Follow-up to #6157.

Discovered by Anthropic's OSS Scanner, as vulnerability ANT-2026-GCNW3SND.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-10-30 16:15:07 -06:00
Behdad EsfahbodandClaude Opus 4.8 292942ceca [COLR] Charge gradient color stops to the paint session work budget
The COLRv1 paint walk bounds paint-table visits (HB_MAX_NESTING_LEVEL,
HB_MAX_GRAPH_EDGE_COUNT) but handed out ColorLine stops uncharged, so a
single 65535-stop ColorLine shared across up to 16384 gradient edges
could deliver ~10^9 stops from one hb_font_paint_glyph() call, keeping
the rendering thread busy for minutes per glyph in every backend.

Give hb_paint_context_t the session's live budget via
funcs->get_budget_remaining_ptr(), with the same local HB_BUDGET_GLYPH
fallback hb_draw_session_t uses, charge HB_BUDGET_16 per stop delivered
in ColorLine::get_color_stops() (the one place every consumer fetches
stops, so vector, raster, cairo and client paint functions are all
covered), and make recurse() skip further paints once the budget is
negative, as it already does for an oversized graph.  The backend's one
budget is only spent, never re-armed, so this does not reintroduce the
cross-subsystem budget multiplication removed in #6147.  hb-ft-colr.hh
has the same shape and gets the identical change.

With the default 2^24 budget about a million stops per session are
delivered before the rest of the walk is skipped best-effort.  A
separate quadratic cost inside cairo's own gradient pattern building is
not addressed here.  Adds a regression test that fails without the fix.

Discovered by Anthropic's OSS Scanner, as vulnerability ANT-2026-F2NWHRVP.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-10-30 16:14:51 -06:00
Behdad Esfahbod c34cf564b1 [ft] Allow font queries from color paint callbacks (#6490)
Release the face lock around solid, gradient, and custom palette
callbacks. Keep lazy FreeType color-stop reads locked and give each
getter call its own iterator, so callbacks can query the same font
without deadlocking or disturbing another stop query.

Snapshot the selected palette once per COLRv1 traversal, shared by
the bounds and paint passes. Nested paints can select a different
palette on the face; they must not change the outer paint's colors.
Return failure if the palette copy cannot be allocated. Stops remain
lazy and no cache policy or limits change.

Add a bounded callback regression with the existing COLRv1 fixture.
It queries font extents, performs nested paints with another palette,
and checks repeated and paged stop reads against the OT backend.
The original code times out; unlocking without copying the palette
fails the color comparison. The full Meson suite passes: 274 tests, no failures.

Reported by Anthropic's OSS Scanner as ANT-2026-YPLDSFHA.

Assisted-by: OpenAI Codex
2026-10-30 14:52:35 -06:00
Behdad Esfahbod 472c3e0fd3 [draw] Add rendering work-budget API
Expose budget policy and remaining-work callbacks on draw and paint function
objects.  The public setters clamp invalid negative caps, distinguish finite
defaults from explicit unlimited operation, and report whether a backend
participates in accounting.

Let draw sessions borrow the live counter owned by draw_data.  OpenType glyf
and CFF traversal then debit the same counter in place, while unsupported
backends retain a finite per-glyph fallback.  Forward budget access through
the parent-font draw adaptor.

Document the new API and cover nil, default, unlimited, clamped, concrete,
and live-depletion behavior.

Tests: meson test -C build (272/272 passed)

Assisted-by: OpenAI Codex <codex@openai.com>
2026-08-26 17:11:21 -06:00
Nussu06 9de7e9d563 [paint] pass color_glyph its own user_data
The dispatch wrapper handed the callback user_data->push_clip_glyph
instead of user_data->color_glyph, so a client registering both
callbacks with different contexts got the wrong pointer on every
COLR glyph.
2026-07-17 11:43:34 -06:00
Khaled Hosny b4ed968183 [test/paint] Test the fill_path callback 2026-07-12 19:36:15 +03:00
Behdad Esfahbod 788706d568 [test] Normalize linear paint transcripts
Reduce linear gradient anchors and normalize color lines before
printing them in test-paint. This makes the transcript reflect the
renderer-facing gradient geometry instead of backend-specific raw
COLRv1 anchors and stop ranges.

Also normalize signed zero in transform output so the fixtures stay
stable across backends.

Testing:
CCACHE_DISABLE=1 meson test -C build test-paint --print-errorlogs

Assisted-by: OpenAI Codex
2026-05-27 12:40:44 +02:00
Behdad EsfahbodandClaude Opus 4.6 f3c36ef3d6 [test] Add API tests for push_clip_path_start/_end
Folded into the existing test-paint binary.  Two tests,
both font-independent (registered ahead of the glyph-based
flavor setup so they run even on builds without the test
COLRv1 font):

* test_push_clip_path_nil_returns_null: an unset paint-funcs
  table returns NULL from push_clip_path_start and clears
  the out-param draw_data, and push_clip_path_end is safe
  to call on the nil stub.

* test_push_clip_path_round_trip: registered start/end fire
  in order, the returned draw-funcs + draw-data round-trip
  through the public entry points, and every hb_draw_*()
  call reaches the sink accumulator.  Asserts the count
  includes hb_draw_close_path's synthetic back-to-start
  line_to so the contract is documented by the test.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-16 16:18:20 -06:00
Behdad Esfahbod 469502c99b [test-paint] Test fontations too
Luckily it passes!
2025-03-12 11:02:43 -06:00
Matthias Clasen 0ce90f60ab [test-paint] Make it easier to regenerate test output
This command will regenerate the expected output for all the
paint tests:

GENERATE_DATA=1 G_TEST_SRCDIR=./test/api ./build/test/api/test-paint
2025-03-06 20:10:13 -05:00
Behdad Esfahbod 9d149c5284 [test-paint] Redirect generated output to stderr 2025-02-26 15:07:46 -07:00
Behdad Esfahbod 753093d60e [test-colrv1] Add test files from https://roettsch.es/var_colrv1.html
These are not currently used for any testing. To be used later.
2025-02-26 14:41:04 -07:00
Behdad Esfahbod 6fe1441782 [test-paint] Update comment 2025-02-26 14:40:08 -07:00
Behdad Esfahbod 755929c48d Fix more compiler warnings
In case FreeType is not available.
2024-10-12 20:08:32 -06:00
Behdad Esfahbod 22e1a5a785 [test-paint] Fix warnings 2024-08-06 12:20:34 -06:00
Behdad Esfahbod 7d7541e35d [test-paint] Print paint-color-glyph 2023-08-14 17:55:09 -06:00
Behdad Esfahbod 58effbcc4f . 2023-08-14 16:47:36 -06:00
Behdad Esfahbod 474b99d122 [test-paint] Fix build without FreeType
Fixes https://github.com/harfbuzz/harfbuzz/issues/4103
2023-02-04 10:16:11 -07:00
Behdad Esfahbod e377888990 [ft-colr] Conditionalize on (unreleased0 FreeType 2.13.0
That's the version that the color API is called stable, and
includes changes that we rely on.
2023-01-26 13:27:55 -07:00
Khaled Hosny c08e5d094a [hb-draw] Fix MSVC warning
warning C4305: 'initializing': truncation from 'double' to 'float'
2023-01-21 15:27:56 +02:00
Behdad Esfahbod c86bab3fb4 [test-paint] Use G_GNUC_PRINTF 2023-01-17 06:29:17 -07:00
Behdad Esfahbod f02c4ebb40 [test-paint] Adjust condition for enabling ft backend 2023-01-16 14:17:44 -07:00
Behdad Esfahbod 60a4f2e6f5 [test-paint] Detect FreeType COLRv1 from struct size 2023-01-12 19:12:46 -07:00
Behdad Esfahbod b783967242 [test] Minor save glyph_count 2023-01-12 19:11:10 -07:00
Matthias Clasen cca0ed9bcd test-paint: Add variable font tests
These needs freetype master to work.
2023-01-12 21:09:50 -05:00
Matthias Clasen 1cb5a87feb test-pain: Print out extend for color lines
Update all expected results to include this value.
2023-01-10 22:13:44 -05:00
Matthias Clasen 6648e6e3e5 Add one more paint test
Add a test for the clip_box_top_left_glyph,
since we've seen broken rendering with it.
2023-01-10 22:07:44 -05:00
Matthias Clasen b20871322f test-paint: Fix use of g_test_skip 2023-01-09 16:30:42 -05:00
Matthias Clasen eb0f0279d2 test-paint: Avoid g_test_skip_printf
This is relatively recent api we don't need.
2023-01-09 15:33:34 -05:00
Matthias Clasen e61c2be41c test-paint: Skip tests if ft COLRv1 is missing 2023-01-09 15:25:29 -05:00
Matthias Clasen 1d662632d9 test-paint: More output for failures 2023-01-09 15:01:59 -05:00
Matthias Clasen d0108d31a2 Add more paint tests
These tests compare the output of the ft and ot
implementations for all the glyphs in the test_glyphs
font.
2023-01-09 14:42:30 -05:00
Matthias Clasen 5f976d86a7 test-paint: Use %.3g for results
This produces more readable output.
2023-01-09 14:24:32 -05:00
Matthias Clasen a085efa699 test-paint: Fix font setup for ft
We must call hb_ft_font_set_funcs after
setting the font scale.
2023-01-09 13:11:16 -05:00
Matthias Clasen 26f713a7f5 paint-test: Use larger scales
A scale of 20 is noise, so use 1000.
2023-01-09 12:54:06 -05:00
Matthias Clasen 42047070dd test-paint: More helpful output
No need to print ASCII chars as hex.
2023-01-08 16:08:16 -05:00
Behdad Esfahbod 3fd6c0d97a [test-paint] Remove unused FT_Library 2023-01-08 12:30:26 -07:00
Behdad Esfahbod 28be4f8805 [test-paint] Actually run against hb-ft 2023-01-08 12:19:41 -07:00
Matthias Clasen ed7d0234e1 test-paint: Limit the precision of clip boxes
Update expected test results.
2023-01-08 11:57:43 -07:00
Matthias Clasen f3ce137420 test-paint: Be more flexible for expected results
When generating the expected output with GENERATE_DATA=1,
Glib's test framework puts out some comments at the top
of the file. Ignore them when comparing the expected
output. This makes it possible to directly use the output
of

GENERATE_DATA=1 ./test-paint -p TESTCASE

as expected result for TESTCASE.
2023-01-08 09:01:31 -05:00
Matthias Clasen 2a51567925 [paint] Add a test for recursion 2022-12-26 17:00:20 -07:00
Matthias Clasen 1eb4d002f2 Try to fix the build with msvc 2022-12-26 11:29:26 -07:00
Behdad Esfahbod 6ccbfabd4f [paint] Return bool from paint_image()
Fixes https://github.com/harfbuzz/harfbuzz/issues/3974
2022-12-24 10:44:25 -07:00
Matthias Clasen 3478728edb Fix test-paint build without freetype 2022-12-24 09:16:29 -05:00
Matthias Clasen 9b9d7c7b8e Plug a memory lek in paint tests 2022-12-24 05:34:47 -05:00
Matthias Clasen 55b7af6b62 Tweak paint-tests 2022-12-23 18:34:09 -05:00
Matthias Clasen 44b48845b7 Add tests for hb_color_line_t
Test a few things that were broken with the
ft implementation before.
2022-12-23 16:28:53 -05:00
Behdad Esfahbod a4a86c0ec2 [test-paint] g_test_message 2022-12-23 13:52:22 -07:00
Behdad Esfahbod dcab567988 [test-paint] Don't use g_test_fail_print() for older glib 2022-12-23 13:47:37 -07:00