Commit Graph

14195 Commits

Author SHA1 Message Date
Behdad Esfahbod 60c6b7786d Disable -Wcast-function-type-strict
https://github.com/harfbuzz/harfbuzz/pull/3859#issuecomment-1295409126
2022-10-28 14:19:39 -06:00
Martin Storsjö d88269c827 freetype: Fix function signatures to match without casts
Clang 16 has got a new stricter warning for casts of function types
(see 1aad641c79).

This new warning gets included as part of the existing error
diagnostic setting of -Wcast-function-type.

This fixes errors like these:

../src/hb-ft.cc:1011:34: error: cast from 'void (*)(FT_Face)' (aka 'void (*)(FT_FaceRec_ *)') to 'FT_Generic_Finalizer' (aka 'void (*)(void *)') converts to incompatible function type [-Werror,-Wcast-function-type-strict]
    ft_face->generic.finalizer = (FT_Generic_Finalizer) hb_ft_face_finalize;
                                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2022-10-28 13:49:41 -06:00
Qunxin Liu 8a5524833c [instance] update OS2/.usWeightClass and .usWidthClass when
no-prune-unicode-ranges option is enabled
2022-10-27 11:59:31 -06:00
خالد حسني (Khaled Hosny) fddeba26e4
Merge pull request #3857 from 2xsaiko/outgoing/cmake-abs-path
Pass through absolute paths to cmake config directly
2022-10-26 22:34:02 +02:00
Marco Rebhan db292f6f02
Pass through absolute paths to cmake config directly
The previous code concatenates includedir to _harfbuzz_prefix verbatim,
which results in a wrong final include path in case includedir is an absolute
path. Instead, we can let meson determine the absolute include and lib paths
in advance and save them in the cmake config.

This is an issue in nixpkgs, where includedir is set to the final (absolute)
path of the built library in the Nix store, which causes CMake projects
depending on harfbuzz to not configure.

See https://github.com/NixOS/nixpkgs/issues/180054.
2022-10-26 22:13:44 +02:00
Behdad Esfahbod e854739b2d [fuzzing] Add test font for previous commit 2022-10-26 13:12:56 -06:00
Behdad Esfahbod 477d71724c [glyf] Limit points in a glyf to 10000 roughly
Only enforced when components are being expanded.

Fixes https://github.com/harfbuzz/harfbuzz/issues/3838
2022-10-26 13:11:47 -06:00
Behdad Esfahbod 9aad3dba8f [SingleSubst] Fix degenerate-lookup test
Part of https://github.com/harfbuzz/harfbuzz/issues/3853
2022-10-26 13:04:02 -06:00
Garret Rieger 83769b9cb1 [subset] add comment for why we retain empty lookups. 2022-10-21 18:29:23 -06:00
Garret Rieger 70e2de2bd4 [subset] Always output Lookup's even if they are empty.
The rest of layout subsetting depends on lookup indices being consistent with those computed during planning. So if an empty lookup is discarded during the subset phase it will invalidate all subsequent lookup indices. Generally we shouldn't end up with an empty lookup as we pre-prune them during the planning phase, but it can happen in rare cases such as when a subtable is considered degenerate (eg. #3853)
2022-10-21 18:29:23 -06:00
Joel Auterson 3377ddf69e Add missing underscore 2022-10-21 11:09:35 -06:00
Joel Auterson 28cb1b3954 Fix working-directory 2022-10-21 11:09:35 -06:00
Joel Auterson dbbbf02d77 Fix GHA workflow 2022-10-21 11:09:35 -06:00
Joel Auterson 756b500018 Add ARM CI workflow 2022-10-21 11:09:35 -06:00
Joel Auterson c813f84235 Make build work for arm-none-eabi 2022-10-21 11:09:35 -06:00
Behdad Esfahbod b0b7a65388 [subset] Fix check-symbols failure
Fixes https://github.com/harfbuzz/harfbuzz/issues/3850
2022-10-20 17:13:26 -06:00
Khaled Hosny 970321db7b 5.3.1 2022-10-19 22:06:46 +02:00
Khaled Hosny 7c8be866c9 [doc] Hide another experimental symbol 2022-10-19 22:04:57 +02:00
Garret Rieger d1bf6c0c26 [subset] only preprocess in benchmark when experimental api is enabled. 2022-10-17 20:14:02 +00:00
Garret Rieger db22bfb3cc [subset] Remove Franklin from the tests which is not an open source font. 2022-10-17 13:15:47 -06:00
Behdad Esfahbod a73137d101 [face] Fix annotation 2022-10-17 12:18:24 -06:00
Behdad Esfahbod d65af60213 [shape] Adjust Grapheme clusters for Katakana voiced sound marks
Fixes https://github.com/harfbuzz/harfbuzz/issues/3844
2022-10-15 14:26:02 -06:00
Behdad Esfahbod bda42fc34a
Merge pull request #3842 from harfbuzz/patch_mode
[subset] Begin implementing a subset accelerator
2022-10-14 15:08:22 -06:00
Garret Rieger f53ebf5584 [subset] Add hb_subset_preprocess to experimental symbol list for check-symbols. 2022-10-14 19:38:19 +00:00
Garret Rieger fdb98ed88e [subset] add missing HB_EXTERN. 2022-10-14 18:30:39 +00:00
Garret Rieger 515863e57c [subset] Remove add accelerator flag, replace with new api method.
Adds hb_subset_preprocess() which preprocesses the face and attaches accelerator data.
2022-10-13 23:42:00 +00:00
Garret Rieger 573640c99f [subset] Add hb-subset-accelerator.hh to Make soure list. 2022-10-13 23:21:35 +00:00
Behdad Esfahbod a756bd1944 [glyf] Use component phantom points after transformation 2022-10-13 17:18:19 -06:00
Behdad Esfahbod 82d19c08fc Revert "[glyf/Composite] Remove phantom points when not needed anymore"
This reverts commit 527e63a3bd.
2022-10-13 17:17:03 -06:00
Behdad Esfahbod 527e63a3bd [glyf/Composite] Remove phantom points when not needed anymore 2022-10-13 17:13:40 -06:00
Garret Rieger 3394ec7048 [subset] use subset accelerator in tests.
This ensures it produces equivalent subsets as without the accelerator.
2022-10-13 23:02:54 +00:00
Garret Rieger f4903defc4 [subset] use the accelerator in the subsetting benchmark. 2022-10-13 21:38:54 +00:00
Garret Rieger 01481db582 [subset] use accelerator unicode to gid map if available. 2022-10-13 21:12:22 +00:00
Garret Rieger 4ec5eb955f [subset] add a subset accelerator.
Can be optionally attached to the face during subsetting. Contains data which can accelerate future subsets.
2022-10-13 19:40:31 +00:00
Garret Rieger f105c28749 [subset] Suggested flags to enable more performant subset production for use in incxfer. 2022-10-13 18:53:41 +00:00
Behdad Esfahbod 5769d422c5 [type] Add HBFixed template 2022-10-13 12:12:24 -06:00
Behdad Esfahbod 8c29dcaee4 s/HBFixed/F16DOT16/g 2022-10-13 12:06:15 -06:00
Behdad Esfahbod 294b1c9f6e Use snprintf instead of sprintf 2022-10-13 11:30:02 -06:00
Garret Rieger 9559d3c1c1 [repacker] fix incorrect coverage table size estimation.
During splitting of PairPosFormat2 the code was assuming the maximum size of the generated coverage table would be equal too the current size. This is incorrect size the new coverage table may not preserve the ranges found in the original coverage table (since we are splitting based on class, not coverage) and in the worst case may convert from format2 to format1. So use the size of a format1 table as the max size.
2022-10-11 14:06:04 -06:00
Khaled Hosny dcd8fe2c3b Revert "[ci] Mark automatic releases as draft"
This reverts commit 9a28df411e.

It creates a new draft release for each upload, which is not what we are
after.
2022-10-08 19:33:38 +02:00
Khaled Hosny 3ce4b8f5c9 5.3.0 2022-10-08 19:10:07 +02:00
Khaled Hosny 90eee512a0 [doc] Add symbols to harfbuzz-sections.txt 2022-10-08 19:02:28 +02:00
Khaled Hosny 9a28df411e [ci] Mark automatic releases as draft
Fixes https://github.com/harfbuzz/harfbuzz/issues/2780
2022-10-08 18:22:13 +02:00
Garret Rieger 99f4668e19 [repacker] use mutable copies of Coverage/ClassDef in MarkBasePos shrink operation.
Also make mutable copies (when needed) of the top level subtables during a split operation.
2022-09-30 10:58:50 -06:00
Garret Rieger 8f1bf23cc9 [subset] optimize glyf subsetting w/ retain gids.
When retain gids is enabled the subset plan may require the output of many empty glyphs. This change optimizes the glyf subsetting code when the number of retained glyphs << number of output glyphs. Unnessecary lookups to the glyph map are reduced by iterating through the glyph map instead of the output glyph set.
2022-09-29 15:44:49 -06:00
Qunxin Liu e94fe2adf3 [instance] minor optimizations to glyf instancing 2022-09-29 12:14:43 -06:00
Qunxin Liu d5fc4a73c0 [instance] add tests for featureVariations
Also updated the script that is used to generate tests.With fonttools,
we now do instancing first and then subsetting.

With different order of subsetting and instancing operations on the same
VF file, fonttools seems to generate 2 different font files with different
glyph set.
1. do subsetting and then instancing: this seems result in a larger glyph
   set in the font file. Lookups are collected from both retained features
   and all possible alternate featurevariations, this leads to a larger
   glyph set after glyph closurei. And instancer doesn't redo glyph
   closure, it does lookups pruning only.

2. do instancing and then subsetting: lookups are collected from
   features that are replaced already and possible alternate feature
   variations
2022-09-29 12:14:43 -06:00
Qunxin Liu b90ce34a25 [instance] fix for non-empty .notdef glyph metrics update
we need pass in source glyph's outline data to calculate boundaries, and
then drop the outline data if notdef-outline is specified.

Also disable shifting points for instancing in get_points () API
2022-09-29 12:14:43 -06:00
Qunxin Liu b706c6f77e [instance] update FeatureList with variations 2022-09-29 12:14:43 -06:00
Qunxin Liu 8f3a7017c3 [instance] support FeatureVariations table 2022-09-29 12:14:43 -06:00