Commit Graph

1913 Commits

Author SHA1 Message Date
Behdad Esfahbod 7e160dc347 [gpos] Fix unsafe-to-break of mark-attachment
This was undetected because most of the time marks were in same
cluster of their base already.

Fixes https://github.com/harfbuzz/harfbuzz/issues/3093
2022-01-19 09:27:55 -07:00
Behdad Esfahbod c8195dbc01
Merge pull request #3365 from harfbuzz/gdef-fix
GDEF fix
2022-01-19 06:56:05 -08:00
Simon Cozens 282642311f Test for #2140 2022-01-19 16:05:09 +02:00
Simon Cozens fc5354c5e5 Add test for #2516 2022-01-19 16:04:58 +02:00
luz paz e2e305066a Fix various typos
Found via `codespell -q 3 -S ./perf/texts -L actualy,ba,beng,fo,gir,inout,nd,ot,pres,ro,te,teh,timne`
2022-01-16 05:39:03 -08:00
Garret Rieger 87496bf63e [subset] fix fuzzer timeout if visisted_paint goes into error. 2022-01-13 11:08:24 -08:00
Garret Rieger 067f90a820 [subset] Fix for fuzzer timeout.
Fixes https://oss-fuzz.com/testcase-detail/5549945449480192

In prune_langsys: move LangSys visited check up before any work is done for a LangSys. In this particular case the compare() method is responsible for the majority of the time spent and wasn't being guarded with a visisted check.
2021-12-14 17:34:21 -07:00
Garret Rieger c4573c2ec7 [repacker] don't infinite loop if visited or roots is in error.
Fixes https://oss-fuzz.com/testcase-detail/5205038086094848
2021-12-14 15:57:48 -07:00
Behdad Esfahbod 23159084b4 [morx] Reverse graphemese, not whole buffer
Fixes https://github.com/harfbuzz/harfbuzz/issues/3314
Supersedes https://github.com/harfbuzz/harfbuzz/pull/3315
2021-12-10 15:02:57 -07:00
Qunxin Liu 706014f69d [subset] (Chain)ContextSubst glyph_closure fix
- When pos_glyphs is empty, use current full glyphs set as input for
subsequent recursive closure process
- Also increase max_lookup_visit_count to 35000 cause a real font file hit
previous limit 20000 and some lookups are dropped unexpectedly
2021-12-10 11:08:36 -07:00
Qunxin Liu 51655a078e [subset] COLR : only include glyphs after COLR closure 2021-12-06 13:29:21 -07:00
Qunxin Liu 74b46b29e7 [subset] MATH: don't serialize coverage table when iterator is empty
when iterator is empty, just set coverage offset to 0.
serialize() in coverage will at lease write out a 16-bit format header.
2021-12-03 08:00:33 -07:00
Khaled Hosny 6c81cd9543 [tests] Add tests for platform shapers
Fixes https://github.com/harfbuzz/harfbuzz/issues/3115
2021-12-01 19:55:48 +02:00
Garret Rieger 9121ed0cec [subset] Improve sharing of Ligature subtables.
Ligature subtables use virtual links to enforce an ordering constraint between the subtables and the coverage table. Unfortunately this has the sideeffect of prevent the subtables from being shared by another Ligature with a different coverage table since object equality compares all links real and virtual. This change makes virtual links stored separately from real links and updates the equality check to only check real links. If an object is de-duped any virtual links it has are merged into the object that replaces it.
2021-12-01 08:37:04 -07:00
Garret Rieger 95329081c2 [subset] further optimize cmap4 packing. 2021-11-28 07:47:49 -07:00
Garret Rieger 599143824c [subset] Don't pad glyphs when using long loca. 2021-11-26 18:01:04 -07:00
Garret Rieger d9660fd58a [subset] Make cmap4 packing more optimal.
The current CMAP4 implementation uses whatever the current codepoint ranges are and then encodes them as indivudal glyph ids or as a delta if possible. However, it's often possible to save bytes by splitting up existing ranges and encoding parts of them using deltas where the cost of splitting the range is less than encoding each glyph individual.
2021-11-26 13:21:50 -07:00
Behdad Esfahbod 720ab0883b [util] Add --single-par
Use it in aots tests

Fixes https://github.com/harfbuzz/harfbuzz/issues/3129
Related https://github.com/harfbuzz/harfbuzz/issues/3298
2021-11-25 11:49:16 -07:00
Qunxin Liu 903a6baece [subset] layout_features filtering fix
we should not use get_size (), which returns length * item_size
2021-11-25 11:24:41 -07:00
Khaled Hosny 69d8f27c69 [meson] Require 0.55.0
We implicitly require it for building ragel subproject. This new version
requirement should satisfied in both Fedora 33 and Debian bullseye, and
not be too cutting edge for us.
2021-11-22 03:11:36 +02:00
Qunxin Liu 84dc4e85e8 [subset] avoid writing out duplicate extra glyph names in post table
Add check for possible duplicate with other name index
2021-11-19 17:22:35 -07:00
Qunxin Liu ca418cac74 [subset] keep features that have FeatureParams and the tag is "size" 2021-11-19 13:31:25 -07:00
Qunxin Liu e88fc41ef3 [subset] inputSequence could be empty, change the sanity check 2021-11-19 12:57:43 -07:00
Garret Rieger ace98cc65f [subset] Only sanitize recursion depth in COLR. 2021-11-10 10:34:46 -07:00
Garret Rieger f51b48c8e7 [subset] Fix fuzzer found memory leak.
Happens because an insert into a map with an invalid key reports successful, but this causes the set being inserted to be lost.
2021-11-02 17:00:07 -07:00
Qunxin Liu 540f19b6fe [subset] fix bug in (Chain)ContextFormat2
Only keep rulesets for glyphs class numbers that survived in coverage
2021-11-02 16:05:55 -07:00
Garret Rieger e260eeb9be [subset] Update test goldens for gdef.glyphset. 2021-11-02 16:56:52 -06:00
Qunxin Liu 60e203644b [subset] use glyphset_gsub instead of glyphset for GDEF 2021-11-02 16:56:52 -06:00
Garret Rieger e39647c0f1 [subset] update subset goldens to fix tests. 2021-11-02 16:26:02 -06:00
Qunxin Liu 0a7563a53f [subset] fuzzer fix: https://oss-fuzz.com/testcase?key=6254792024915968
Make sure input is valid, each gid has a corresponding offset value in
the map
2021-11-02 15:25:18 -07:00
Qunxin Liu 364b6b3989 [subset] Don't assume FeatureList is sorted
Though the spec said FeatureRecords are sorted alphabetically by feature
tag, there're font files with unsorted FeatureList. And harfbuzz is not
able to subset these files correctly because we use binary search in
finding featureRecords when collecting lookups. Also
find_duplicate_features needs to be updated to handle this.
2021-11-02 15:15:02 -07:00
Garret Rieger 49c9392412 [subset] During LigatureSubstFormat1 subsetting always place Coverage last.
In Windows 7 on Chrome if the coverage table comes before any of the LigatureSet or Ligature subtables the font won't load. This changes the packing order to always place the Coverage table last. Virtual links are used to ensure the repacker maintains the desired ordering.

Coincidentally fontTools also does the same thing (a3f988fbf6/Lib/fontTools/ttLib/tables/otTables.py (L1137)) to reduce overflows during packing.
2021-11-02 12:19:13 -06:00
Matthias Clasen 8dc86d0703 Add a test font scale rounding behavior
This test verifies the fix for
https://github.com/harfbuzz/harfbuzz/issues/3274
2021-11-01 17:59:48 -07:00
Qunxin Liu 85deddb16e [subset] fuzzer fix: https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=40406 2021-10-27 16:16:15 -07:00
Behdad Esfahbod 4262636926
Merge pull request #3248 from googlefonts/connected_components
[repacker] Keep connected subgraphs in the same space.
2021-10-23 13:18:22 -07:00
Qunxin Liu 8c583db9b5 [subset] COLRv1: update subset() method for new strutc ClipList and VarIdxMap
Also fix issues in struct PaintTransform definition
2021-10-23 11:22:39 -07:00
David Corbett 8eaee2f887 [USE] Only skip default ignorables within CCSes 2021-10-08 10:15:08 -06:00
Garret Rieger 79937d24b7 [repacker] Add repacker test that requires space splitting. 2021-10-05 12:53:21 -07:00
Khaled Hosny bb57df7a4f [test] Add test for https://github.com/harfbuzz/harfbuzz/issues/2757#issuecomment-933531978 2021-10-05 16:04:10 +02:00
Khaled Hosny 1624550377 [test] Add test for https://github.com/harfbuzz/harfbuzz/issues/3244 2021-10-05 16:04:10 +02:00
Qunxin Liu 794b00db4b [subset] fuzzer fix: https://oss-fuzz.com/testcase-detail/6616166961905664 2021-09-29 10:29:06 -06:00
Garret Rieger d5beb96e06 [repacker] add repacker isolation test on real font. 2021-09-22 18:18:50 -06:00
Qunxin Liu ca7b9daef0 [subset] subset MATH table 2021-09-22 18:08:30 -04:00
Garret Rieger c2cc566c9d [subset] Fix subset_offset_array adding unused space to serializer.
ArrayOf.serialize_append allocates space for the new item, but ArrayOf.pop() does not recover the allocated space. So in the case where the revert path was entered the extra space added by serialize_append gets left in the serialization buffer. This moves the snapshot to before ArrayOf.serialize_append is called so that revert cleans up the buffer extend.
2021-09-22 16:07:18 -06:00
Qunxin Liu 8c5c81746d [subset] fix find_duplicate_features
We should check each feature against all other features with the same
tag for duplicates.
2021-09-22 15:35:38 -06:00
Garret Rieger 74f96d9d4b [repacker] fix heap use after free in repacker.
Don't store a reference to the link in overflow records as the link object may be freed if the sorted graph vector is resized.
2021-09-19 09:06:17 -06:00
Garret Rieger d700646c33 [subset] In preperation for 3.0.0 release remove legacy subsetting api methods. 2021-09-16 15:31:29 -06:00
David Corbett a0101a31b5 [Unicode 14] Add tests 2021-09-14 18:38:50 -04:00
Khaled Hosny 4d9b7de0b6 [style] Prepare for release 2021-09-13 10:10:10 -06:00
Behdad Esfahbod 201cbe3c66 [style] Rename HB_STYLE_TAG_SLANT to HB_STYLE_TAG_SLANT_ANGLE 2021-09-10 04:59:16 -06:00