Commit Graph

4000 Commits

Author SHA1 Message Date
Behdad Esfahbod 81e2b9b8a9 Never skip over CGJ
We might want to tweak this some more.  For now, never skipping over
it is better behavior than always skipping.

Part of https://github.com/behdad/harfbuzz/issues/554
2017-10-15 10:49:06 +02:00
Behdad Esfahbod 74f1b88e37 [ucdn] Fix compiler warnings 2017-10-13 11:01:55 +02:00
Behdad Esfahbod c2cf68dbdf [coretext] Fix build 2017-10-13 10:30:19 +02:00
Behdad Esfahbod 95883fc5d4 [coretext] Pass correct font size to CoreText
CoreText "point"s are not typographic points, but CSS pixels.  Ie.
they are 96 per inch, not 72 per inch.
2017-10-13 10:22:20 +02:00
Behdad Esfahbod 4f9a83ec21 Merge remote-tracking branch 'origin/master' 2017-10-12 14:08:58 +02:00
Behdad Esfahbod fa48ccbe12 [indic] Special-case Kannada Ra,H,ZWJ sequence
Fixes https://github.com/behdad/harfbuzz/issues/435
2017-10-12 14:07:37 +02:00
n8willis 40b05d7b77 Replacing 'cluster' with 'syllable' when comments describe certain Indic shaping behavior (#563)
* Clarified wording of syllable/cluster behavior in Uniscribe in final_reordering; changed one other probable typo.

* Additional syllable/cluster swap in comments for final reordering and for initial-reordering matra decomposition.
2017-10-12 13:48:48 +02:00
Behdad Esfahbod 8923033eb2 [coretext] Use fabs() to silence compiler warning 2017-10-12 12:02:04 +02:00
Dominik Röttsches dd4b321b4a [coretext] Activate tracking for system fonts
Another attempt at fully fixing
https://github.com/behdad/harfbuzz/issues/360
2017-10-12 11:54:17 +02:00
Behdad Esfahbod 6a2cbc6ef5 [coretext] Use fabsf() instead of abs()
Fixes bots.

Link libharfbuzz.so with -lm.  Surprising that we survived without
it so far!
2017-10-12 10:46:09 +02:00
Behdad Esfahbod 4e4781319b [ft] Add hb_ft_font_changed()
When the font size or variations settings on underlying FT_Face change,
one can call hb_ft_font_changed() and continue using hb_font created using
hb_ft_font_create().

Fixes https://github.com/behdad/harfbuzz/issues/559

New API:
hb_ft_font_changed()
2017-10-12 10:33:16 +02:00
Behdad Esfahbod 94b3cafc3a Add hb_face_get_table_tags()
New API:
hb_face_get_table_tags()

Fixes https://github.com/behdad/harfbuzz/issues/560
2017-10-11 17:22:44 +02:00
Behdad Esfahbod e1b6d92302 Remove cast of functions to (hb_destroy_func_t)
Fixes https://github.com/behdad/harfbuzz/issues/474
2017-10-11 15:51:31 +02:00
Behdad Esfahbod 717fcb51dd [coretext] Minor 2017-10-11 15:38:21 +02:00
Behdad Esfahbod 374bb48d90 Change ptem API to make 0 mean unset 2017-10-11 15:37:50 +02:00
Behdad Esfahbod 06c14225b2 [coretext] Minor 2017-10-11 15:29:53 +02:00
Behdad Esfahbod a4b46212c5 [coretext] Adjust font size check for 0 2017-10-11 15:29:22 +02:00
Behdad Esfahbod 269cf1ca82 Correctly initialize ptem 2017-10-11 15:24:22 +02:00
Behdad Esfahbod 84686bf4c7 [coretext-aat] Also pass through CoreText if font has kerx table
SFNSText has kerx table which apparently is applied.
2017-10-11 15:02:48 +02:00
Behdad Esfahbod 296d0134c9 [coretext] Change default point size to 12 2017-10-11 14:09:30 +02:00
Dominik Röttsches a5ebe1d4ae [coretext] Recreate CTFont if pt size changed
Attempt at fixing #360
2017-10-11 13:33:05 +02:00
Dominik Röttsches db7a73ce07 [coretext] Fix build 2017-10-11 13:24:48 +02:00
Behdad Esfahbod f334130756 [coretext] Another try 2017-10-11 13:17:46 +02:00
Behdad Esfahbod a8e466c3c1 [coretext] Move CT_Font to font data
Towards implementing optical sizing.  Untested; won't compile.

https://github.com/behdad/harfbuzz/issues/360
2017-10-11 13:05:59 +02:00
Behdad Esfahbod f9b4c6570a [coretext] Move font size to a macro 2017-10-11 12:51:25 +02:00
Behdad Esfahbod 8d450dd188 Infrastructure for creating CTFont with correct font size
https://github.com/behdad/harfbuzz/issues/360
2017-10-11 12:43:46 +02:00
Behdad Esfahbod 16d02a58cf [coretext] Change default font size from 36 to 18 2017-10-11 12:28:06 +02:00
Behdad Esfahbod c0c2dbc871 Remove dirty tracking
Turns out I don't need this to resolve CoreText optical sizing
issue after all.

https://github.com/behdad/harfbuzz/issues/360
2017-10-11 12:23:35 +02:00
Behdad Esfahbod b57f18da70 Add hb_font_[sg]et_ptem() to set/get point size on font
New API:
hb_font_set_ptem()
hb_font_get_ptem()

Needed for hb-coretext optical sizing:
https://github.com/behdad/harfbuzz/issues/360
2017-10-11 11:47:47 +02:00
Behdad Esfahbod 3f9370d9e5 Fix TODO item 2017-10-11 11:34:53 +02:00
Behdad Esfahbod 5f50f8837f [arabic] Fix cluster merging 2017-10-07 13:17:33 +02:00
Behdad Esfahbod c2545b9215 Add check for HB_EXTERNs
https://github.com/behdad/harfbuzz/pull/555
2017-10-07 12:57:30 +02:00
fanc999 48a9406839 Fix build of HarfBuzz tools and HarfBuzz-GObject on Visual Studio (#555)
* hb-buffer.h: Mark hb_buffer_diff() for export

This will fix the tools builds on Visual Studio, as the symbol is used
by the tools.

* build: Adapt NMake Makefiles for GLib 2.53.4 or later

glib-mkenums was ported from a PERL script to a Python script, so we
need to update how we generate the enum sources for HarfBuzz-GObject in
the NMake builds.  Let this be known in the build documentation for MSVC
builds.

One of the problems with the underlying cmd.exe that the NMake Makefiles
run on is that shebang lines are not recognized, so we need to to test
run the script with Python and see whether it succeeded by outputing a
source file that is larger than 0 in file size (since running the PERL
version of the script will clearly fail and cause an empty file to be
created).

If it succeeds, we then run a small Python utility script that makes the
necessary string replacements, and we are done.  If that fails, then we
run the glib-mkenums script with PERL, and do the replacements with the
PERL one-liners as we did before.

We need to make replace.py use latin-1 encoding when using Python 3.x to
cope with the copyright sign that is in the generated enum sources.
2017-10-07 12:57:14 +02:00
Behdad Esfahbod c9e2cf6f55 [ot] Remove 'mal' and 'gle' lang tags
Fixes https://github.com/behdad/harfbuzz/issues/477
2017-10-04 16:59:22 +02:00
Behdad Esfahbod 621c49cb86 Improve performance of ligature component matching
This O(N^2) was introduced in 8b2c94c43f.
Make it O(N).
2017-10-04 15:06:48 +02:00
Behdad Esfahbod ab8d70ec70 [arabic] Implement Unicode Arabic Mark Ordering Algorithm UTR#53
Fixes https://github.com/behdad/harfbuzz/issues/509
2017-10-04 14:47:10 +02:00
Behdad Esfahbod b6fe0ab636 Add info_cc() convenience macro 2017-10-04 13:37:08 +02:00
Behdad Esfahbod 7f9e7f8689 Adjust normalizer for out-of-order marks
We are going to implement Unicode Arabic Mark Ordering Algorithm:

  http://www.unicode.org/reports/tr53/tr53-1.pdf

which will reorder marks out of their sorted ccc order.  Adjust
normalizer to stop combining as soon as dangerous ordering is
detected.
2017-10-04 13:20:33 +02:00
Behdad Esfahbod a252ad61f0 Minor 2017-10-04 13:07:08 +02:00
Behdad Esfahbod 4c05a405ac Revert "Treat HAMZA ABOVE similar to SHADD for sorting purposes"
This reverts commit 5a33057576.

Proper fix coming soon.
2017-10-04 13:06:51 +02:00
Behdad Esfahbod d695cacbf4 Shorthand format controls should not be ignored
Fixes https://github.com/behdad/harfbuzz/issues/503
2017-10-04 11:33:47 +02:00
Behdad Esfahbod 50e95229d7 Add Chinantec language tags
Fixes https://github.com/behdad/harfbuzz/issues/516
2017-10-04 11:28:04 +02:00
Behdad Esfahbod 9786fcd8fd Fix GPOS v_origin ordering
This should affect mark attachment in vertical text.  I have no font to test,
but this sounds the right order.

Noticed while debugging https://github.com/behdad/harfbuzz/issues/532
2017-10-03 17:22:43 +02:00
Behdad Esfahbod 63db692fa9 [indic] Fix shaping of U+0AFB GUJARATI SIGN SHADDA
Fixes https://github.com/behdad/harfbuzz/issues/552
2017-10-03 16:00:18 +02:00
Behdad Esfahbod f559c63307 [indic] Implement Consonent_With_Stacker
Fixes https://github.com/behdad/harfbuzz/issues/528
2017-10-03 15:20:07 +02:00
Behdad Esfahbod e07669fc43 [use] Implement Consonant_With_Stacker
Related to https://github.com/behdad/harfbuzz/issues/528
2017-10-03 14:57:14 +02:00
Behdad Esfahbod 5d98de1f38 Tweak mark-base attachment
Apparently a base glyph can also become an attached component of a
ligature if the ligature-forming lookup used IgnoreBase.  This was
being confused with a non-first component of a MultipleSubst and
hence not matched for mark-attachment.  Tweak test to fix.

Fixes https://github.com/behdad/harfbuzz/issues/543
2017-10-03 13:49:40 +02:00
Behdad Esfahbod c44657a05d Tweak input matching some more
Followup to 8b2c94c43f

Allow matching sequences of marks attached to different ligatures,
as supposedly the base of the subsequent marks were already jumped
over.
2017-10-03 13:14:17 +02:00
Behdad Esfahbod 8b2c94c43f Tweak ligature component matching for ligature formation
If two marks want to ligate and they belong to different components of the
same ligature glyph, and said ligature glyph is to be ignored according to
mark-filtering rules, then allow.

Example Burmese senquence:

  U+1004,U+103A,U+1039,U+101B,U+103D,U+102D

Test font provided by Norbert Lindenberg.

Fixes https://github.com/behdad/harfbuzz/issues/545
2017-10-02 20:03:35 +02:00
Behdad Esfahbod 71c0a1429d [indic] Fix shaping of U+0A51
Mark it as matra below to allow the sequence U+0A15, U+0A51, U+0A47.
Oh well...

Fixes https://github.com/behdad/harfbuzz/issues/524
2017-10-02 18:57:03 +02:00
Behdad Esfahbod 6eb8950e62 [ucdn] Oops! 2017-10-02 18:30:05 +02:00
Behdad Esfahbod cbec0cd65e Remove some really old cruft 2017-10-02 18:26:52 +02:00
Behdad Esfahbod b3dff7720c [ucdn] Update README 2017-10-02 18:22:51 +02:00
Behdad Esfahbod 57c55ef834 [ot] Improve shaper selection heuristic 2017-10-02 18:21:27 +02:00
Behdad Esfahbod 5680ef884c [use] Hack to fix shaping of U+1A7F
Fixes https://github.com/behdad/harfbuzz/issues/525
2017-10-02 18:20:51 +02:00
Behdad Esfahbod 0faa16a253 [ucdn] Update to Unicode 10
Update to commit c000ebf79c095a7d58cf90090bde5715592c4834
plus this bug-fix: https://github.com/grigorig/ucdn/issues/18
2017-10-02 17:15:46 +02:00
Behdad Esfahbod ea535a1dfa [use] Update to Unicode 10 2017-10-02 17:02:39 +02:00
Behdad Esfahbod 29c244aff6 Minor 2017-10-02 16:36:21 +02:00
Behdad Esfahbod 1535f8c672 Add Unicode 10 scripts 2017-10-02 16:12:18 +02:00
Behdad Esfahbod bdbe974f52 [indic] Update table to Unicode 10 2017-10-02 10:00:31 -04:00
Behdad Esfahbod 819cc36049 [indic] Move manual code out of generated table 2017-10-02 09:03:16 -04:00
Behdad Esfahbod da4866f717 [arabic] Update table to Unicode 10 data 2017-10-02 08:57:06 -04:00
Behdad Esfahbod 1a0a356a0f [indic] Categorize U+0980 BENGALI ANJI as placeholder
Fixes https://github.com/behdad/harfbuzz/issues/538
2017-10-02 08:49:23 -04:00
jfkthame 19e77e01bc [shape-plan] Don't look for a cached plan if variation coords are specified. Fixes #549. (#550)
This is a temporary hack just to avoid incorrect shaping results, pending proper support
for caching shape-plans with variation coordinates.
2017-10-02 08:08:49 -04:00
mhosken 66128d3cfb Fix missing xscale in rtl graphite unpacking (#541)
* Fix missing xscale in rtl graphite unpacking

* Oops didn't need to mess with yscale
2017-09-26 14:29:45 -04:00
Behdad Esfahbod 3ca69c8c32 Use strtod_l() to correctly parse decimal numbers in French & other locales
Test with, eg.:
$ LC_ALL=fr_FR.utf-8 ./hb-view NotoSansArabic-VF.ttf بهداد --variations wght=1.2
2017-09-14 20:50:35 -04:00
Behdad Esfahbod d03f11f246 Fix buffer_diff for empty buffers
If buffers are empty, content type should be ignored.

This fixes last of the failing tests: fuzzed.tests.  Green again!
2017-09-04 20:14:13 -07:00
Behdad Esfahbod 7cc348041d [unsafe-to-break] Fix unsafe-to-break for cluster-level=1
Fixes tests/shaping/tests/cluster.tests
2017-09-04 20:06:21 -07:00
Behdad Esfahbod 06cb162cd7 [indic] Treat Consonant_With_Stacker as consonant
Fixes https://github.com/behdad/harfbuzz/issues/528
"Kannada JIHVAMULIYA and UPADHMANIYA insert dotted circles"
2017-09-01 10:34:56 -07:00
Behdad Esfahbod c449d2d8c9 [unsafe-to-break] Mark during fraction-slash formation
Fixes tests/automatic-fractions.tests
2017-08-30 17:28:22 -07:00
Behdad Esfahbod 099472e08b hb_buffer_diff() tweak
I like to have a mode where CONTAINS_NOTDEF and CONTAINS_DOTTEDCIRCLE are not
returned.  Abused a value of -1 for that.  hb-shape now uses it.  Fixes two
of the six tests failing with --verify in test/shaping/run-tests.sh.
2017-08-30 16:45:06 -07:00
Behdad Esfahbod 224d20e4e1 [docs] Minor 2017-08-23 14:49:10 -07:00
Behdad Esfahbod b2dd0c1c32 Add hb_glyph_info_get_glyph_flags()
New API.
2017-08-23 13:12:54 -07:00
Behdad Esfahbod 3e8f4f1452 Rename HB_BUFFER_DIFF_FLAG_MASK_MISMATCH to HB_BUFFER_DIFF_FLAG_GLYPH_FLAGS_MISMATCH 2017-08-22 17:56:25 -07:00
Behdad Esfahbod 8820ba29df Fix warning about "may be used uninitialized" 2017-08-15 17:12:21 -07:00
Emil A Eklund (eae) 91770e1c56 Fix signed/unsigned warning (#522)
Change hb_buffer_diff to explicitly cast result of abs to unsigned when
comparing with position_fuzz to avoid unsafe signed/unsigned comparions
warnings on windows.
2017-08-15 16:25:18 -07:00
Behdad Esfahbod 587f15ece3 Minor 2017-08-14 15:23:10 -07:00
Behdad Esfahbod 338e61ad1b Fix typo 2017-08-14 12:36:38 -07:00
Jonathan Kew 331d66c7fe Add function to compare two buffers
Based on patch from Jonathan Kew.

Needs more cleaning up and documentation.

New API:
hb_buffer_diff_flags_t
hb_buffer_diff()
2017-08-14 11:44:12 -07:00
Behdad Esfahbod 219af509ef [graphite2] Fix warning 2017-08-13 15:10:26 -07:00
Behdad Esfahbod 239119a631 [unsafe-to-break] Mark all positions as unsafe in alternative shapers 2017-08-13 15:08:34 -07:00
Behdad Esfahbod 05fabbd03e [unsafe-to-break] Towards verifying unsafe-to-break in --verify
We break and shape fragments and reconstruct shape result from them.
Remains to compare to original buffer.  Going to add some buffer
comparison API and use here, instead of open-coding.
2017-08-11 20:00:55 -07:00
Behdad Esfahbod 6ce25f57c6 Fix hb_buffer_append()
Ouch!
2017-08-11 19:31:05 -07:00
Behdad Esfahbod 1c17c2bde5 [unsafe-to-break] Copy flag to all glyphs in a cluster
Makes consumption easier.
2017-08-11 19:06:07 -07:00
Behdad Esfahbod ec104e5912 [unsafe-to-break] Mark unsafe is cluster merging is disabled
We were relying on cluster merges not requiring unsafe flagging because
they get merged.  If cluster level requests no merging, then we flag
unsafe when merge would have happened.
2017-08-11 18:24:27 -07:00
Behdad Esfahbod 39a9749430 New API: hb_buffer_append() 2017-08-11 15:52:06 -07:00
Behdad Esfahbod 14a639ea59 [unsafe-to-break] Fix Use-of-uninitialized-value in unsafe_to_break_set_mask
Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=3011
2017-08-11 11:30:39 -07:00
Behdad Esfahbod e43aad5552 [unsafe-to-break] Flag in Thai PUA shaping 2017-08-10 20:54:15 -07:00
Behdad Esfahbod eae009602b [unsafe-to-break] Flag in Hangul shaper 2017-08-10 20:50:48 -07:00
Behdad Esfahbod b65aa1cd47 [unsafe-to-break] Flag during mark attachment 2017-08-10 20:32:02 -07:00
Behdad Esfahbod 882ebdadd6 [unsafe-to-break] Flag during cursive positioning 2017-08-10 20:23:17 -07:00
Behdad Esfahbod 2d158ca376 [unsafe-to-break] Flag during kerning 2017-08-10 20:20:30 -07:00
Behdad Esfahbod a7c4972cce [unsafe-to-break] Flag during fallback positioning 2017-08-10 20:16:15 -07:00
Behdad Esfahbod e2a2b5b229 [unsafe-to-break] Be careful with flag propagation when merging clusters 2017-08-10 20:11:35 -07:00
Behdad Esfahbod f2868c2008 Set mask to 0, instead of 1, by default
This shouldn't matter.
2017-08-10 20:10:53 -07:00
Behdad Esfahbod 9e005c5d86 [unsafe-to-break] Mark Indic-like clusters as unsafe-to-break 2017-08-10 20:10:53 -07:00
Behdad Esfahbod 5287ccc935 [unsafe-to-break] Allocate public bits from 0 2017-08-10 15:02:34 -07:00
Behdad Esfahbod 91ce0c45d6 [safe-to-break] Fix logic when there is no out-buffer 2017-08-10 15:02:34 -07:00
Behdad Esfahbod a043c99f77 [unsafe-to-break] Simplify logic
Always use the algorithm for non-monotone case.  It's more robust.
2017-08-10 15:02:34 -07:00
Behdad Esfahbod 40bd7e9a1c [unsafe-to-break] Add UNSAFE_TO_BREAK flag
Not all shapers code is updated to set this properly.
GSUB and Arabic shaper are updated.
GPOS and other shapers are NOT.

Fixes https://github.com/behdad/harfbuzz/issues/224
2017-08-10 15:02:28 -07:00
Behdad Esfahbod ac8c4e56d8 [ft] Fix theoretical leak 2017-08-09 22:05:08 -07:00
Behdad Esfahbod 5a33057576 Treat HAMZA ABOVE similar to SHADD for sorting purposes
Part of https://github.com/behdad/harfbuzz/issues/509
2017-08-09 17:23:28 -07:00
Behdad Esfahbod 68af14d5cc Protect against div-by-zero in CBDT extent code
Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=1452#c5

CC https://github.com/behdad/harfbuzz/issues/139
2017-08-09 17:09:21 -07:00
Behdad Esfahbod 5dc30451b8 Two fixes to avar mapping
1. Handle segment with two entries correctly,

2. Fix rounding math.  Ouch!

Fixes https://github.com/behdad/harfbuzz/issues/521
2017-08-07 21:37:07 -07:00
Behdad Esfahbod dc2c418e39 [check-defs/symbols.sh] Drop empty-symbol lines
Fixes https://github.com/behdad/harfbuzz/issues/510
2017-08-03 12:34:15 +01:00
Ebrahim Byagowi 65f64d1400 Unbreak arm-none-eabi build again (#514) 2017-07-19 02:42:53 +04:30
Behdad Esfahbod c1432bce3c [arabic] Adjust feature order again
Fixes https://github.com/behdad/harfbuzz/issues/505
2017-07-14 17:35:17 +01:00
Cosimo Lupo 9813be3d12 [coretext] Allow to disable kern (#508)
* Minor

* [coretext] Fix leak

* [coretext] Do not reset num_features

* [coretext] allow to disable kern; re-enabling doesn't seem to be working
2017-07-14 17:11:46 +01:00
Behdad Esfahbod 9dd29c681e [use] Allow up to two medial-below letters
Fixes https://github.com/behdad/harfbuzz/issues/376
2017-07-14 17:01:27 +01:00
Behdad Esfahbod 216b003c91 [use] Fix shaping of U+AA29 CHAM VOWEL SIGN AA
Part of https://github.com/behdad/harfbuzz/issues/376
Also see https://github.com/roozbehp/unicode-data/issues/6

Test added, using NotoSansCham built from Noto Phase III sources.
2017-07-14 16:38:51 +01:00
Behdad Esfahbod f1cd7ca893 [indic] Add github URL 2017-07-14 15:59:40 +01:00
Behdad Esfahbod 3cc84f45b9 [indic] Fix https://github.com/behdad/harfbuzz/issues/478 2017-07-14 15:50:22 +01:00
Behdad Esfahbod e359a4b8f5 [indic] Disable automatic ZWNJ handling for Indic features
Fixes https://github.com/behdad/harfbuzz/issues/294

Also fixes a bunch of other Indic issues.  Test results after:

BENGALI: 353725 out of 354188 tests passed. 463 failed (0.130722%)
DEVANAGARI: 707307 out of 707394 tests passed. 87 failed (0.0122987%)
GUJARATI: 366355 out of 366457 tests passed. 102 failed (0.0278341%)
GURMUKHI: 60732 out of 60747 tests passed. 15 failed (0.0246926%)
KANNADA: 951201 out of 951913 tests passed. 712 failed (0.0747968%)
KHMER: 299071 out of 299124 tests passed. 53 failed (0.0177184%)
MALAYALAM: 1048136 out of 1048334 tests passed. 198 failed (0.0188871%)
ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%)
SINHALA: 271662 out of 271847 tests passed. 185 failed (0.068053%)
TAMIL: 1091754 out of 1091754 tests passed. 0 failed (0%)
TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%)

Before:

BENGALI: 353725 out of 354188 tests passed. 463 failed (0.130722%)
DEVANAGARI: 707307 out of 707394 tests passed. 87 failed (0.0122987%)
GUJARATI: 366349 out of 366457 tests passed. 108 failed (0.0294714%)
GURMUKHI: 60732 out of 60747 tests passed. 15 failed (0.0246926%)
KANNADA: 951190 out of 951913 tests passed. 723 failed (0.0759523%)
KHMER: 299070 out of 299124 tests passed. 54 failed (0.0180527%)
MALAYALAM: 1048136 out of 1048334 tests passed. 198 failed (0.0188871%)
ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%)
SINHALA: 271662 out of 271847 tests passed. 185 failed (0.068053%)
TAMIL: 1091753 out of 1091754 tests passed. 1 failed (9.15957e-05%)
TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%)
2017-07-14 14:22:52 +01:00
Behdad Esfahbod cdf1fd0627 [indic] Add infrastructure to disable ZWNJ-skipping in context-matching
Not used yet.
2017-07-14 12:43:34 +01:00
Dominik Schlösser 4e21ec5469 Fix for reordering of Tibetan vowel u (#443)
* Undone change for Tibetan vowel u

* removed comment on reordering that became invalid with roll-back

* Support for Dzongkha contractions with multiple vowel-signs

* Removed non-functional and unnecessary defines for HB_MODIFIED_COMBINING_CLASS_CCC138,140
2017-07-14 12:14:23 +01:00
Ebrahim Byagowi 3b0e47ca00 Fix arm-none-eabi build (fixes #451) (#496) 2017-06-19 14:47:09 +04:30
Ebrahim Byagowi 76c4873e8c Support branch prediction helpers on clang compiles (#491) 2017-06-02 13:23:10 -04:00
Behdad Esfahbod 7dba30639a Handle allocation failure in hb-language code 2017-06-01 11:44:42 -04:00
Sebastian Rasmussen 92e2c4baaf Avoid using strdup inside library. (#488)
If an application provides a malloc replacement through
hb_malloc_impl() it is important that it is used to allocate
everything, but the use of strdup() circumvents this and
causes system malloc() to be called instead. This pairs
badly with the custom hb_free_impl() being called later.
2017-05-29 10:53:30 -07:00
Khaled Hosny 06cfe3f736 Do not skip TAG characters in glyph substitution (#487)
Hide them like Mongolian Free Variation Selectors instead.

Fixes https://github.com/behdad/harfbuzz/issues/463
2017-05-17 11:32:47 -07:00
Behdad Esfahbod 1817221620 Minor 2017-05-16 14:26:28 -07:00
mhosken 74b99ef224 Fix graphite2 rtl conversion (#475) 2017-04-20 11:13:22 -07:00
Chris Peterson aacca37590 Fix clang -Wcomma warnings (#471) (#472)
clang's new -Wcomma compiler option warns about possible misuse of the
comma operator such as between two statements.

hb-common.cc:190:9 [-Wcomma] possible misuse of comma operator here
hb-ot-layout-gsubgpos-private.hh:345:30 [-Wcomma] possible misuse of
comma operator here
hb-shape-plan.cc:438:26 [-Wcomma] possible misuse of comma operator here
2017-04-17 23:25:24 -07:00
William Hua 5ecf96e3a2 Use absolute paths of ragel generated headers (#467)
https://github.com/behdad/harfbuzz/issues/455
2017-04-16 22:33:42 -07:00
jfkthame c04c1fe86e Blacklist GDEF table in additional Tahoma versions. (#459)
There are more broken versions of Tahoma out there on various Windows releases,
so we need to add them to our blacklist to avoid broken rendering.
See https://bugzilla.mozilla.org/show_bug.cgi?id=1279925 for details.
2017-04-11 14:29:13 -07:00
ebraminio adfd4ae1cf [cmake] Improve third party libraries support (#461) 2017-04-11 23:18:18 +04:30
Behdad Esfahbod a4471d0c2c Move list of ragel sources to Makefile.sources as well 2017-04-05 15:42:11 +02:00
ebraminio d2acaf6d72 Split ragel generated files lists and remove hardcoded rl files lists (#453) 2017-04-05 11:21:23 +02:00
jfkthame 740fdbcd0e avoid UBSan warning in get_stage_lookups (#450)
See https://bugzilla.mozilla.org/show_bug.cgi?id=1336600
2017-04-03 13:22:39 +02:00
Behdad Esfahbod edcf6344bc Blacklist more versions of Padauk
Patch from Phil Race.
2017-03-24 10:24:52 -07:00
Khaled Hosny 91570a1eeb Just always use strtod here 2017-03-22 23:07:20 +02:00
Chun-wei Fan 539571c1a9 src/hb-common.cc: Fix build on older Visual Studio
Visual Studio only supported strtof() from Visual Studio 2013 onwards, so
use strtod() instead to do the operation, which should do the same thing,
sans going to a double, not a float.
2017-03-22 23:03:49 +02:00
Behdad Esfahbod 47e7a1800f Revert "Fix Context lookup application when moving back after a glyph delete"
This reverts commit b9b005f3a4.

This introduced invalid access cases. Revert until I fix correctly.
2017-03-10 13:23:02 -08:00
Behdad Esfahbod 3c080a7a0a Fix buffer serialize of empty buffer
Ouch!
2017-03-07 18:13:28 -08:00
Behdad Esfahbod 9ac9af725c Add TODO item 2017-03-05 13:51:01 -08:00
Behdad Esfahbod 5aec2fb8d0 Remove TODO item that is not going to happen 2017-03-02 11:46:33 -08:00
Behdad Esfahbod b9b005f3a4 Fix Context lookup application when moving back after a glyph delete
This was broken forever, since days that we did not allow moving
tape backwards. Works now. Reported by Doug Felt.
2017-03-01 14:27:23 -08:00
Behdad Esfahbod 85630996b8 Fix buffer-overrun with Bengali reph positioning code
This has no security implications whatsoever since we always keep
and extra element at the end of buffer, just in case.

Discovered by oss-fuzz
CC https://github.com/behdad/harfbuzz/issues/139
Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=660
2017-02-25 13:32:20 -08:00
Dominik Röttsches a657f23ca3 Blacklist another instance of Padauk (#419)
In https://crbug.com/681813 another instance of Padauk was identified
triggering collapsed glyphs. Blacklist this version by patching
hb-ot-layout.cc to print out gdef, gsub, and gpos table length, then
adding those to the list of blacklisted versions.
2017-02-20 13:52:07 -06:00
jfkthame 44f7d6ecde Guard against underflow when adjusting length (#421)
* Guard against underflow when adjusting length

With the fuzz-testcase in mozilla bug 1295299, we end up with a recursed lookup that removes 3 items, when `match_positions[idx]` is 0, which results in (unsigned) `end` wrapping to a huge value.

Making `end` a signed int is probably the simplest route to a fix.

Fixes https://bugzilla.mozilla.org/show_bug.cgi?id=1295299.

* Add testcase for #421.
2017-02-16 19:03:24 -08:00
jfkthame 45766b673f [indic] Add support for Grantha marks that may be used in Tamil to th… (#401)
* [indic] Add support for Grantha marks that may be used in Tamil to the Indic table.

See https://bugzilla.mozilla.org/show_bug.cgi?id=1331339.

Testcase: U+0BA4,U+0BC6,U+1133c,U+0BAA,U+1133c,U+0BC6,U+1133c

* [indic] Add test for Grantha nukta that is allowed in Tamil by ScriptExtensions.txt
2017-02-16 09:40:21 -08:00
Behdad Esfahbod d4bb52b9fd Unbreak hb-coretext build 2017-02-09 14:13:25 -08:00
Behdad Esfahbod c8dfed8e1e Merge pull request #357 from khaledhosny/graphite-scale
[graphite] Fix shaping with varying font sizes
2017-02-08 14:36:18 -08:00
Behdad Esfahbod 7c47474f20 Set LC_ALL instead of LANG when creating harfbuzz.def
Hopefully fixes Travis faliure?
https://github.com/behdad/harfbuzz/pull/413#issuecomment-278481882
https://travis-ci.org/khaledhosny/harfbuzz/jobs/199785332#L1065
2017-02-08 14:33:28 -08:00
Philip Withnall ffde3c9f9e hb-font: Fix a potentially undefined use of memcmp() (#413)
While it’s fine to call memcmp(x, 0, 0) in practice, the C99 standard
explicitly says that this is not allowed: even if the length is zero,
the pointer arguments must be valid.

http://stackoverflow.com/a/16363034

Coverity ID: 141178

Signed-off-by: Philip Withnall <withnall@endlessm.com>
2017-02-08 13:56:57 -08:00
Behdad Esfahbod 73c6dcbb45 Silence Coverity warning
Fixes https://github.com/behdad/harfbuzz/issues/408
2017-02-07 16:53:21 -08:00
Behdad Esfahbod 466b3e58bd Shuffle things around a bit 2017-02-03 16:57:15 -08:00
Behdad Esfahbod fc8189b639 Minor 2017-02-03 16:17:08 -08:00
Behdad Esfahbod d3d3691806 Add dirty-state tracking to hb_face_t 2017-02-03 15:42:03 -08:00
Behdad Esfahbod 2171f48b4b Add dirty-state tracking to hb_font_t 2017-02-03 10:58:09 -08:00
Behdad Esfahbod 95808bad2c Add new API hb_font_set_face()
For completeness.
2017-02-03 10:39:40 -08:00
jfkthame 4ec19319ab Add Win10 Anniversary Update version of Tahoma to GDEF blacklist. (#412) 2017-02-02 10:48:23 -08:00
Behdad Esfahbod 1dd630a7f2 Minor
Fixes https://github.com/behdad/harfbuzz/issues/411
2017-02-01 11:57:21 -08:00
Behdad Esfahbod e888f642db Route Adlam through Arabic shaper
Fixes joined Adlam rendering.

Fixes https://github.com/googlei18n/noto-fonts/issues/828
2017-01-26 14:50:14 -08:00
jfkthame 72c754873c Add Win7 version of himalaya.ttf to the GDEF table blacklist. (#407) 2017-01-26 14:35:41 -08:00
Behdad Esfahbod 22af28a3f1 [var] Implement MVAR table
Unhooked. Untested.
2017-01-23 23:42:53 -08:00
Behdad Esfahbod 67a191164d [var] Whitespace 2017-01-23 23:21:30 -08:00
Behdad Esfahbod 79e8e27ffd [var] Hook up advance variations to hb-ot-font
Yay, works!
2017-01-23 17:57:27 -08:00
Behdad Esfahbod 6074340371 [var] Fix bug in HVAR/VVAR impl 2017-01-23 17:55:16 -08:00
Behdad Esfahbod bd3b11d926 [var] Prepare for hooking up advance variations to hb-ot-font 2017-01-23 17:37:19 -08:00
Behdad Esfahbod a4fca9f005 [var] Implement DeltaSetIndexMap and advance delta fetching
Not hooked up to hb-ot-font yet.
2017-01-23 12:25:40 -08:00
Behdad Esfahbod 7b399f73ef [var] Start adding HVAR/VVAR 2017-01-23 11:41:43 -08:00
Behdad Esfahbod b732c53eb5 Add LArrayOf<> 2017-01-22 20:31:53 -08:00
Behdad Esfahbod 5e156fa5ed Add LOffsetTo<> 2017-01-22 20:28:56 -08:00
Behdad Esfahbod d6bdbbca30 [var] Remove HB_OT_TAG_fvar/avar
We're not going to define public macros for tags for tables.
There's little use to those and they are implementation detail.
2017-01-22 20:16:51 -08:00
Behdad Esfahbod a11d9a0466 Minor 2017-01-22 20:11:57 -08:00
Behdad Esfahbod 7a860f656e [var] Add macros for standard axes 2017-01-22 19:52:09 -08:00
Behdad Esfahbod a484e237b1 [var] Implement 'avar' table mapping
Untested!
2017-01-22 19:41:33 -08:00
Behdad Esfahbod 5ec96d30ca [var] Adjust API in prep for 'avar' implementation
The 'avar' table does not allow random access to axis maps,
so change API to avoid quadratic-time implementation.

Removed -hb_ot_var_normalize_axis_value(), added
+hb_ot_var_normalize_variations() and
+hb_ot_var_normalize_coords() instead.
2017-01-22 18:52:00 -08:00
Behdad Esfahbod 8a577aaa0d [var] Stub implementation of avar table 2017-01-22 18:22:40 -08:00
Behdad Esfahbod 7647a05a0b Minor 2017-01-22 17:47:49 -08:00
Behdad Esfahbod d2f249e745 Stub out doc comments for new symbols 2017-01-22 17:42:33 -08:00
Behdad Esfahbod 5dfd341f51 Minor 2017-01-22 16:55:40 -08:00
Behdad Esfahbod 825e40407d [hb-ft] Remove use of variable-length array
Hopefully also fixes build failure on msvc.
2017-01-22 16:41:45 -08:00
Behdad Esfahbod 47ee34e847 [var] Hook up variations to FreeType face
hb-view correctly renders variations with ft font-funcs now.
hb-ot-font needs HVAR implementation.
2017-01-21 18:10:08 -08:00
Behdad Esfahbod bb1e19268f [var] Rename var_coord to variation
Looks much better.
2017-01-21 17:42:50 -08:00
Behdad Esfahbod 64fe92bf2d [var] Remove use of variable-length arrays 2017-01-21 15:36:35 -08:00
Behdad Esfahbod 2491134b38 [var] Add hb_font_set_var_coords() 2017-01-21 15:21:50 -08:00
Behdad Esfahbod 113393efec Remove a few unused includes 2017-01-21 15:12:03 -08:00
Behdad Esfahbod 2d40923ca9 Minor 2017-01-21 15:06:15 -08:00
Behdad Esfahbod 8b2a580470 [var] Add hb_font_set_var_coords_design() 2017-01-21 15:05:00 -08:00
Behdad Esfahbod 0dcc7b49a8 [var] Minor lifecycle adjustment 2017-01-21 14:50:22 -08:00
Behdad Esfahbod bf0d3a665b [var] Add hb_font_get_var_coords_normalized() 2017-01-21 14:48:46 -08:00
Behdad Esfahbod b3c0714b4b [var] Add hb_var_coord_t and parsing routines 2017-01-20 20:30:03 -08:00
Behdad Esfahbod 72364103bc Move code around 2017-01-20 20:16:53 -08:00
Behdad Esfahbod 785982bf83 [var] Flesh out some more 2017-01-20 20:05:45 -08:00
Behdad Esfahbod 422c0c36c8 [var] Flesh out a bit 2017-01-20 19:14:54 -08:00
Behdad Esfahbod b8376b1090 Minor 2017-01-20 18:19:28 -08:00
Behdad Esfahbod f2e73d37e7 [var] Remove over-optimization 2017-01-20 17:42:18 -08:00
Behdad Esfahbod 469926cc45 [var] Change double to float in API 2017-01-20 17:40:46 -08:00
Behdad Esfahbod 55d42fd667 Start adding hb-ot-var.h and implementation
Supports enumerating variation axes, normalizing values, etc.
2017-01-19 19:35:48 -08:00
Behdad Esfahbod 272b511532 Minor 2017-01-19 19:33:07 -08:00
Khaled Hosny 2452543fdd [ot] Fix automatic fraction for RTL scripts (#405)
The numbers for right-to-left scripts are processed also from right to
left, so the order of applying “numr” and “dnom” features should be
reversed in such case.

Fixes https://github.com/behdad/harfbuzz/issues/395
2017-01-18 12:48:13 -08:00
Sascha Brawer 1337428e4f Update language tags to OpenType 1.8.1 (#403)
Resolves https://github.com/behdad/harfbuzz/issues/324
2017-01-18 04:51:02 -08:00
Behdad Esfahbod 784d1180de Fix build with clang 2017-01-11 11:58:13 -08:00
Behdad Esfahbod 1f810daf16 Port math table to hb_lazy_table_loader_t 2017-01-09 23:50:56 -08:00
Behdad Esfahbod ebbcc11122 Add hb_lazy_table_loader_t 2017-01-09 22:47:47 -08:00
Behdad Esfahbod 1af9d924a4 Move lazy-loader to more generic place 2017-01-09 22:22:37 -08:00
Behdad Esfahbod ac274331c2 In lazy-loading, remember allocation failure 2017-01-09 22:20:38 -08:00
Behdad Esfahbod 29fb0cb727 Minor 2017-01-09 21:18:55 -08:00