Commit Graph
18314 Commits
Author SHA1 Message Date
Behdad Esfahbod e5a01efd07 [aat] Put back one set of compiled map in the plan
Use it if there's no user features. Reduces mallocs.
2025-03-13 02:47:20 -06:00
Behdad Esfahbod 7b48aa37a7 [aat] Reduce mallocs
Untested.
2025-03-13 01:07:35 -06:00
Behdad Esfahbod 792b9b9337 [bit-vector] Add a has() 2025-03-12 23:56:52 -06:00
Behdad Esfahbod 43d421d388 [aat] Speed up deleted-glyph removal 2025-03-12 23:36:09 -06:00
Behdad Esfahbod aa233ecedc [bit-vector] Add a hb_bit_vector_t type
Unused.
2025-03-12 23:05:04 -06:00
Behdad Esfahbod 402876d489 Merge pull request #5136 from harfbuzz/kerx-fix
Kerx fix
2025-03-12 22:48:59 -06:00
Behdad Esfahbod c38d518718 [aat/kerx] Fix initial-glyph collection logic
Test:
$ hb-shape GeezaPro.ttc --unicodes U+064A,U+064E,U+0651,U+0629
2025-03-12 22:39:33 -06:00
Behdad Esfahbod 2449eb0882 [aat] Adjust filtering to match collect_glyphs() 2025-03-12 21:03:24 -06:00
Behdad Esfahbod 6388ce2224 [kerx] We don't need right_set in machine kerning 2025-03-12 20:58:06 -06:00
Behdad Esfahbod b861b54178 [kerx] Cosmetic 2025-03-12 20:56:16 -06:00
Behdad Esfahbod 7146e5818f [kerx] Remove redundant check 2025-03-12 20:11:30 -06:00
Behdad Esfahbod 41b396c64b [kerx] Setup buffer glyphset 2025-03-12 20:02:58 -06:00
Behdad Esfahbod 83e0944f0f [kerx] Use a machine class cache
Like in morx.
2025-03-12 19:58:35 -06:00
Behdad Esfahbod 02733deded [kerx] Skip machine subtables that don't intersect the buffer
We do this extensively in morx. Do it here too.

Still no machine class cache in kerx.

Speeds up GeezaPro shaping Arabic text by 20%.
2025-03-12 19:44:07 -06:00
Behdad Esfahbod 45b2d28d04 Revert "[aat] Clear buffer glyph set"
This reverts commit b1a9219ac5.

Was unnecessary. collect_codepoints() clears the set.
2025-03-12 19:18:12 -06:00
Behdad Esfahbod d76a23a3f1 [aat] Micro-optimize 2025-03-12 19:16:59 -06:00
Behdad Esfahbod b1a9219ac5 [aat] Clear buffer glyph set
Should be clear already but no harm.
2025-03-12 19:13:14 -06:00
Behdad Esfahbod 54962b3fce [morx] Consolidate a check 2025-03-12 19:03:50 -06:00
Behdad Esfahbod 4807a021a1 [atomic] Kill hb_atomic_ptr_t<T>
Use hb_atomic_t<T *> instead.
2025-03-12 18:27:27 -06:00
Behdad Esfahbod 829d1eda16 [atomic] Clean up atomic_ptr_t 2025-03-12 18:19:04 -06:00
Behdad Esfahbod a0d76c5b37 [atomic] Templatize 2025-03-12 17:16:55 -06:00
Behdad Esfahbod bdee8658c6 [post] Fix a pointer type 2025-03-12 17:13:48 -06:00
Behdad Esfahbod 469502c99b [test-paint] Test fontations too
Luckily it passes!
2025-03-12 11:02:43 -06:00
Behdad Esfahbod fbb81e344b [fontations] Adjust alignment requirement 2025-03-12 00:52:12 -06:00
Behdad Esfahbod 8d300049f5 [fontations] Assert layout alignment 2025-03-12 00:43:10 -06:00
Behdad Esfahbod 05cfdb9105 [fontations] link_with instead of link_whole
Now that the hb-fontations symbol is called from libharfbuzz,
it gets pulled in properly.
2025-03-12 00:08:42 -06:00
Behdad Esfahbod 0b2a0bac47 Merge pull request #5130 from harfbuzz/malloc-rust
[fontations] Make fontations use hb_malloc et al
2025-03-11 23:04:11 -06:00
Behdad Esfahbod 5bf81c3758 [common] Make hb_malloc() et al public
New API:
+hb_malloc()
+hb_calloc()
+hb_realloc()
+hb_free()
2025-03-11 22:52:34 -06:00
Behdad Esfahbod a2c7dc0669 Merge pull request #5133 from harfbuzz/hb-coretext.cc
[coretext] Split common code into hb-coretext.cc
2025-03-11 20:43:32 -06:00
Behdad Esfahbod f0a67e9948 [coretext] Split common code into hb-coretext.cc 2025-03-11 20:37:30 -06:00
Behdad Esfahbod d014efd03d Merge pull request #5131 from harfbuzz/using2
Add hb_ft_face_create_from_blob_or_fail() et al
2025-03-11 13:37:48 -06:00
Behdad Esfahbod 759e1881e8 [coretext] Fix loading TTC files 2025-03-11 12:46:27 -06:00
Behdad Esfahbod 15fcfcb606 Merge pull request #5129 from harfbuzz/buffer-deserialize
Fix buffer deserialize
2025-03-11 12:01:20 -06:00
Behdad Esfahbod 5e6da54166 [test] Minor, accept absolute font paths 2025-03-11 12:00:18 -06:00
Behdad Esfahbod 8687f5c38e [face] Add tests for new constructors 2025-03-11 11:56:00 -06:00
Behdad Esfahbod 8e8a9f6f40 [face] Add hb_face_create_or_fail_using()
Fixes https://github.com/harfbuzz/harfbuzz/issues/5117

Untested.

New API:
+hb_face_create_or_fail_using
2025-03-11 11:56:00 -06:00
Behdad Esfahbod cbc205c2f0 [font/face] Refactor some code 2025-03-11 11:56:00 -06:00
Behdad Esfahbod b1c50eb938 [face] Rename a variable 2025-03-11 11:56:00 -06:00
Behdad Esfahbod f98c203e7a [coretext] Add hb_coretext_face_create_from_blob_or_fail()
Part of https://github.com/harfbuzz/harfbuzz/issues/5117

Untested.

New API:
+hb_coretext_face_create_from_blob_or_fail()
2025-03-11 11:56:00 -06:00
Behdad Esfahbod 8ca9fe7617 [face] Avoid edit-sanitizing in hb_face_count() 2025-03-11 11:56:00 -06:00
Behdad Esfahbod e4fe8bf95e [ft] Add hb_ft_face_create_from_blob_or_fail()
Part of https://github.com/harfbuzz/harfbuzz/issues/5117

Untested.

New API:
+hb_ft_face_create_from_blob_or_fail()
2025-03-11 11:56:00 -06:00
Behdad Esfahbod 3c30562dc7 Merge pull request #5132 from harfbuzz/cplusplus-more
[test-cplusplus] Add more headers
2025-03-11 11:55:46 -06:00
Behdad Esfahbod 756668d35c [directwrite] Drop dependency again as per review
https://github.com/harfbuzz/harfbuzz/pull/5132#issuecomment-2714186460
2025-03-11 11:27:02 -06:00
Behdad Esfahbod 6b2f859232 [test-c] Disable gobject include
It can't find its dependent headers since not installed.
2025-03-11 00:37:11 -06:00
Behdad Esfahbod 8ca892b0e2 [directwrite] Simplify headers 2025-03-11 00:24:11 -06:00
Behdad Esfahbod d247c116fb [test-draw] Test all font funcs 2025-03-11 00:15:40 -06:00
Behdad Esfahbod d2ccf595bd Directwrite is a C++ header? 2025-03-11 00:15:19 -06:00
Behdad Esfahbod 41dcc493e5 [test-cplusplus] Simplify 2025-03-10 23:49:05 -06:00
Behdad Esfahbod 57deae5fd4 [meson] Check correct dependency
Ouch!
2025-03-10 23:40:22 -06:00
Behdad Esfahbod dbad6cdfec [test-c(plusplus)] Add more headers 2025-03-10 23:23:16 -06:00