b0e7378fa9
Reroute Mandaic shaping through the Arabic shaper
...
We added Mandaic joining data to the Arabic shaper a while ago, but
were not actually using the Arabic shaper for Mandaic. Fixed.
2010-12-21 14:19:32 -05:00
88e7f37488
Annotate the Arabic joining table with block information
2010-12-21 14:18:24 -05:00
1482a39e56
Rename remaining metrics uses to extents
2010-12-17 20:15:05 -05:00
22c537657c
Rename TableDirectory to TableRecord as per OpenType 1.6
2010-12-14 23:51:29 -05:00
33e8b86197
Update 'head' table to OpenType 1.6
2010-12-13 15:54:56 -05:00
b768333514
Disallow ligature substitutions replacing one glyph
2010-12-13 14:22:35 -05:00
7eb875118d
Mozilla Bug 618592 - freeze on typekit
...
Fix apply_lookup() for zero-input broken fonts.
2010-12-13 14:13:35 -05:00
bbbbe80ec9
Rename original_direction to target_direction
2010-12-07 16:22:02 -05:00
ee8aaf976a
Fix arabic shaping of LTR text
...
We should ensure-direction before doing any complex work. The only
exception is mirroring that needs to see the original / final direction,
not the native. Handle that.
2010-12-07 16:20:42 -05:00
4f9e4a40bc
Fix failing checks
2010-11-22 11:30:32 -05:00
9da26d6669
Remove email address from Copyright headers
2010-11-22 11:03:18 -05:00
a45f000192
Clarify copyright notice
2010-11-18 13:59:10 -05:00
0884a8d9cd
Move generated table to its own file
2010-11-17 16:58:21 -05:00
14d784116b
Update Arabic joining table to include Mandaic
...
Mandaic was added to Unicode 6.0, but the joining data was not updated.
Draft ArabicShaping.txt from 6.1 includes the joining data for Mandaic.
Use that.
2010-11-17 16:52:58 -05:00
43bf2f7f1e
Add OpenType script tags for Unicode 5.2 and Unicode 6.0 scripts
...
Based on tags proposed by Microsoft for inclusion in OpenType.
2010-11-17 14:49:40 -05:00
afab01cf7c
Add Unicode 6.0 scripts
2010-11-17 14:35:34 -05:00
f234b68d18
Change HB_SCRIPT_MEITEI_MAYEK to HB_SCRIPT_MEETEI_MAYEK
...
The new name matches the Unicode name. Reported by Jonathan Kew.
2010-11-17 14:25:44 -05:00
300cb41417
Add XXX marks
2010-11-17 12:57:23 -05:00
3a852ae7fe
Save general category and combining class in the buffer for reuse
2010-11-03 16:37:24 -04:00
a5ab682b9b
More "unreached code" warning fixes
...
Ugly :(.
2010-11-03 16:10:12 -04:00
4e22c7e941
Add comment
2010-11-03 16:10:12 -04:00
0342034d1c
Pedantic
2010-11-03 16:10:12 -04:00
4a2d844c2f
Minor code shuffling
2010-11-03 15:28:56 -04:00
11e3ec444a
Fix a few more "unreachable code" warnings
2010-11-03 15:11:04 -04:00
2304856340
Remove another couple lines of dead code
2010-11-03 12:47:27 -04:00
b8783c85ac
Fix unreachable-code warning
2010-11-03 11:51:43 -04:00
c2709119c8
Move things around some more
2010-11-02 19:12:59 -04:00
194d4566ec
Move buffer var allocation local
2010-11-02 19:12:59 -04:00
1e7c1fcbc3
Move code around
2010-11-02 19:12:59 -04:00
bf94b3ad22
Move some more code around
2010-11-02 19:12:59 -04:00
6334658fe7
Simplify mark skipping logic
2010-11-02 19:12:59 -04:00
8c69e65abe
Rename lookup_flags to lookup_props since it's more than just flags
2010-11-02 19:12:59 -04:00
98370e89d1
WIP removing external synthesized GDEF support and implementing it internally
2010-11-02 19:12:58 -04:00
870e2d6eac
Remove unused function
2010-11-02 19:12:58 -04:00
1115890b90
More cleanup
2010-11-02 19:12:58 -04:00
dbf56b1d94
More lig-id cleanup
2010-11-02 19:12:58 -04:00
f6a23a0b91
More removal of lig-id code from buffer
2010-11-02 19:12:58 -04:00
dd2ffd282c
Minor renaming
2010-11-02 19:12:58 -04:00
fe263272a2
Move setting lig_id/component out of buffer and to the gsub code
2010-11-02 19:12:58 -04:00
2e2b2480c0
Always allocate new ligature id
...
No practical point in reusing ligature ids.
2010-11-02 19:12:58 -04:00
bf07d5a29c
Set component=0 for ligature glyph
2010-11-02 19:12:58 -04:00
37ab877149
Remove comment
2010-11-02 19:12:58 -04:00
88474c6fda
Get rid of the OpenType-specific internal buffer representation
...
Add variant integers to buffer item types. More cleanup coming.
2010-11-02 19:12:58 -04:00
6cb8c34930
Add hb_var_int_t
2010-11-02 19:12:57 -04:00
f22802431a
Remove trailing comma
2010-11-02 19:12:57 -04:00
13528d0c78
Supposedly implement vertical support in GPOS
...
Not tested at all.
2010-10-27 14:09:27 -04:00
9624de5b49
Clarify cursive_chain (and change its sign)
2010-10-27 13:44:59 -04:00
d6c9eadb88
Remove more pointless LONGTERMTODO items
2010-10-27 12:34:50 -04:00
e204674fe3
Rename hb_ot_layout_get_lig_carets() to hb_ot_layout_get_ligature_carets()
2010-10-27 12:32:02 -04:00
3357d145f8
Implement vertical support in get_lig_carets()
2010-10-27 12:30:46 -04:00
8eeed7eddc
Remove LONGTERMTODO item that I'll never fix
2010-10-27 12:07:49 -04:00
184a5279c6
Remove unused macro
2010-10-27 12:00:49 -04:00
af5d02a269
Rewrite Cursive joining to act more like other pair lookups
...
Look forward for next character instead of joining to the last character.
2010-10-27 11:54:26 -04:00
ea22c749c7
Fix Cursive positioning
...
Test case: "مرا" rendered using IranNastaliq.
2010-10-27 11:09:48 -04:00
aefdb64689
Fix segfault with Arabic combining marks
2010-10-27 10:40:39 -04:00
248e3c2ba4
Oops, remove extra mask setting that broke complex shaping
2010-10-27 01:23:14 -04:00
502f4cba3e
Divide get_metrics into get_advance and get_extents
...
Graphite module not updated.
Bump version to 0.3.
2010-10-27 01:13:56 -04:00
ec6c0e54d3
Fix blob leak
2010-10-26 11:28:14 -04:00
bd7378b2ef
Massage mask setting a bit more
...
Still finding the exact correct way the masks should be set.
2010-10-13 18:33:16 -04:00
961f9baa7b
Oops, actually set global mask
2010-10-13 17:17:00 -04:00
39dede9fff
Make sure boolean features always use value=1
...
Previously boolean features turned on the entire feature mask. This is
wrong if feature is Alternate and user has provided values bigger than one.
Though, I don't think other engines support such corner cases.
2010-10-13 15:54:06 -04:00
3506b2e78d
Return early if mask is 0
2010-10-13 15:38:52 -04:00
5c1c8c9c50
Make sure feature values don't leak out of their mask
2010-10-13 15:36:38 -04:00
852912fc2d
Fix applying default-value for features
...
Previously if a default global feature was overrided by a non-global
user feature, we were not setting any default mask for the feature,
essentially disabling the feature by default. Fix that.
2010-10-13 15:34:50 -04:00
2989be4919
Set user masks after complex masks
2010-10-13 15:18:29 -04:00
3111b8a0d5
Fix stupid bug, oops
2010-10-13 10:41:53 -04:00
758f68b860
Fix Arabic shaper
...
It's tested now. It works!
2010-10-12 17:37:44 -04:00
13403bc67a
Hookup Arabic shaper!
...
Not tested yet.
2010-10-12 17:23:54 -04:00
57ac0ecb78
Merge clearing masks and setting global masks
2010-10-12 17:07:02 -04:00
fc96596b7c
Form clusters before setting masks
2010-10-12 17:00:40 -04:00
1e80782244
Clear masks before setting them up, not after!
2010-10-12 17:00:07 -04:00
fd3d32d31c
Add hb_ot_shape_execute()
...
Not public yet.
2010-10-12 16:57:47 -04:00
49baa1f69e
Add hb_ot_complex_shaper_t stuff and start hooking Arabic shaper up
2010-10-12 16:50:36 -04:00
605ed468f3
Add hb_ot_shape_plan_t
...
This is the object that a separate plan/execute shaping API will use in
between the two stages.
2010-10-12 16:19:38 -04:00
895fb5d364
Refactor, in a different direction
2010-10-12 16:00:21 -04:00
d2ba016ca1
More refactoring
2010-10-12 15:35:45 -04:00
66e487dfbf
Update Arabic shaping table to Unicode 6.0.0.
2010-10-12 13:51:19 -04:00
f4792d99ee
Fix infinite loop!
...
Untested code is indeed buggy code.
Mozilla bug #603352 .
2010-10-12 12:32:18 -04:00
993d1e786a
Fix missing negation in unreachable code!
...
Mozilla bug #603346
2010-10-12 11:17:30 -04:00
52601275d5
More separation
2010-10-08 20:38:46 -04:00
6b7e675862
Minor
2010-10-08 20:30:04 -04:00
aa9c450bb2
Enable 'rtlm' mirroring
2010-10-08 20:27:38 -04:00
36925f695d
Rename hb_mask_allocator_t to hb_ot_map_t
...
hb_ot_plan_t may be a better name, donno.
2010-10-08 20:21:09 -04:00
f5dd3be46b
Improve checks
2010-10-08 20:16:23 -04:00
5a2b0b3878
Add hb-ot-map.cc
2010-10-08 20:14:57 -04:00
ecc4550ed7
Fix feature overriding
2010-10-08 20:05:29 -04:00
8f08c32750
Minor cleanup
2010-10-08 19:51:03 -04:00
a806762a31
Add hb-ot-shape-private.hh
2010-10-08 19:21:46 -04:00
4924affe0f
Add hb-ot-map-private.hh
2010-10-08 19:18:16 -04:00
a7c5046d6b
Add private hb_segment_properties_t
2010-10-08 18:47:47 -04:00
b897c607d9
Flip the OT bit-allocation vs gsub/gpos inside-out
...
We now build our entire attack plan before jumping in.
2010-10-08 18:41:57 -04:00
e89b7d2a61
Logically separate feature collection
2010-10-08 12:29:59 -04:00
5b88908f12
Minor
2010-10-08 12:23:01 -04:00
5360ce0c5c
Move some more code around
2010-10-07 21:21:11 -04:00
d9c7260788
Minor
2010-10-07 21:19:54 -04:00
efe0d682e8
Simplify compiling lookups
2010-10-07 21:12:46 -04:00
476c94218b
Rename
2010-10-07 17:47:33 -04:00
34db6f031d
Add XXX note
2010-10-07 01:21:19 -04:00
98aa3f6544
Call hb_ot_shape_setup_lookups_complex()
2010-10-06 00:23:36 -04:00
f1d07885dc
Rename setup_lookups()
2010-10-06 00:21:37 -04:00
affc5abac7
Move table_tag to hb_ot_shape_context_t
2010-10-06 00:18:16 -04:00
967240dd8b
Add internal hb_ot_shape_context_t
2010-10-05 23:00:05 -04:00
3eb936f153
Add Arabic/Syriac/N'ko shaping logic
...
Not hooked up just yet.
2010-10-05 19:22:06 -04:00
e81d7afe6e
Add hb_face_get_upem()
2010-10-04 18:18:48 -04:00
bd361945a8
Add API comments
2010-10-04 17:22:57 -04:00
645f6f265b
Add is_mutable() functions
...
Correspond to the make_mutable() ones.
2010-10-04 17:01:01 -04:00
19c0eab8cf
Add getters for all setter APIs
...
One in particular is not a straight getter: hb_font_unset_funcs() is
special because of the specific needs of the lifecycle management of
the user_data object.
2010-10-04 16:45:21 -04:00
f0feb084b0
Minor
2010-10-03 19:09:39 -04:00
ac0c1663fa
Avoid div-by-zero, validate upem
2010-10-01 19:09:23 -04:00
7f97d2cd90
Pedantic
2010-10-01 19:09:01 -04:00
2841436926
Don't zero glyph advances in MarkToBase and similar lookups
...
See email thread "Should MarkToBase attachment zero the mark advance?"
started by Jonathan Kew on 23 August 2010 for details.
2010-10-01 18:32:45 -04:00
36b3862009
One fewer cmp() implementation...
2010-09-29 12:10:24 -04:00
4e573715ae
Improve cmp function parameter namings and casts
...
No semantic change.
2010-09-28 16:28:57 -04:00
dca8aff246
Add comment re DejaVu Sans Mono having 'dflt' script
2010-09-28 16:25:45 -04:00
9dc45401c0
Fix stupid bug in bsearch cmp function!
2010-09-28 16:23:28 -04:00
f2a30bd605
Remove unimplemented method hb_font_get_funcs()
...
Got to add a suitable replacement.
2010-09-23 10:27:08 -04:00
7b9a38a112
Add test.c using public API
2010-09-22 17:42:59 -04:00
9ea7368fce
Fix hb_ot_layout leak
2010-09-22 17:38:44 -04:00
8e577acae2
Fix blob refcounting with insane SFNT table directories
2010-09-22 17:37:02 -04:00
5bd1e95236
Speedup Device table delta computation for common cases
2010-09-22 16:46:18 -04:00
ed4acbde9c
Fix NULL dereference
...
Reported by Jonathan Kew. Face table handling needs to be redone
anyway, but fix this for now.
2010-08-16 14:36:27 -04:00
258305c2a5
[GPOS] Fix div-by-zero
...
Patch by Jonathan Kew. Mozilla bug #465728 .
2010-08-13 14:10:02 -04:00
2422c4b96d
Add TODO
2010-08-13 14:00:34 -04:00
27f0b092a1
Logically separate feature allocation from application
2010-07-23 17:35:54 -04:00
6cf63132dc
Minor
2010-07-23 17:32:26 -04:00
4e4ef24e46
Towards separating bit allocation from shaping
2010-07-23 17:22:11 -04:00
acdba3f90b
Prefer C linkage
2010-07-23 15:39:27 -04:00
cc6d52279d
De-C++ where possible
...
Helps with avoiding many "extern C" declarations in source files.
2010-07-23 15:00:13 -04:00
0dd200d28f
Header dep cleanup
2010-07-23 14:56:29 -04:00
fb0d252462
dos2unix
2010-07-23 14:54:22 -04:00
81408cd51c
Don't use "operator ="
...
Some compilers don't like operator = defined on members of union members
either. Reported by Ginn Chen for Sun Studio compilers.
2010-07-23 14:46:57 -04:00
4f801bd658
Mozilla bug 580233 - check for zero-length record in hb sanitizer.
...
Patch / report by Jonathan Kew.
2010-07-21 16:37:01 -04:00
17e9ff938b
Fix 64bit issues with debug prints
2010-07-15 11:21:34 -07:00
fd56ae4963
Don't bother sorting, it's a safe font error if the array is not sorted
2010-07-08 00:53:40 -04:00
cc8a4abea6
Use bsearch where applicable
2010-07-08 00:40:04 -04:00
2f418f5709
Remove useless TODO
2010-07-07 22:07:40 -04:00
da8edbb622
Fix header
2010-06-09 07:15:39 -04:00
d05d13df02
Add a test for header preprocessor guards
2010-06-09 07:15:03 -04:00
0f0cd9d361
Fix header tags
2010-06-09 06:32:56 -04:00
f2a1b411b1
Followup fix for variation-selectors
...
Patch from Jonathan Kew
2010-06-03 11:37:51 -04:00
a224b4d502
Fix skipping variation-selectors
2010-06-02 22:31:41 -04:00
81a77b115d
Make feature sorting stable
2010-06-01 23:03:54 -04:00
fdc322a820
Minor
2010-05-28 20:55:52 -04:00
f062ec6bb2
Further simplify mask allocator
2010-05-28 20:54:43 -04:00
8af45fda47
Fix global feature handling
2010-05-28 20:41:20 -04:00
e04685ee7b
Simplify mask allocation
2010-05-28 20:37:06 -04:00
0e235d0fc9
Towards a mask allocator
2010-05-28 20:27:30 -04:00
81c5e8724b
Allow disabling default features
...
Patch from Jonathan Kew
2010-05-28 18:31:16 -04:00
2163afbf35
Add note about UTF-8 decoder
2010-05-27 14:04:15 -04:00
226faa58f4
Fix loop
2010-05-27 11:42:34 -04:00
56bef5680c
Fixes to Python and Graphite from Martin
2010-05-27 11:41:47 -04:00
b485da0b71
Disable Graphite as it crashes all over the place...
2010-05-27 11:39:19 -04:00
93ac709b1f
Cypriot is RTL
2010-05-26 16:22:00 -04:00
15c7379c16
Revert "Merge remote branch 'martin/master'"
...
This reverts commit 80af2812fbc442672ec2
2010-05-26 10:48:10 -04:00
80af2812fb
Merge remote branch 'martin/master'
2010-05-24 18:14:24 +01:00
c442672ec2
Fix struct initializers
2010-05-24 18:02:32 +01:00
f5ac9de259
Minor
2010-05-24 18:01:09 +01:00
f97bf4f812
Make sure we initialize all callbacks upon creation
...
Also fixes issue with unicode callbacks never being set really.
2010-05-24 17:50:19 +01:00
bf36a1074a
Move all callback functions in a vtable structs
2010-05-24 17:46:21 +01:00
3109a86add
hb-graphite now no longer has -ve advances within clusters. Fix infinite loop in tag_to_script(). python fixed to use tag_to_script and allow hbtestfont to be passed font files, where fontconfig knows about them.
2010-05-24 13:25:37 +01:00
72631c9d06
Merge branch 'master' of git://git.freedesktop.org/~behdad/harfbuzz-ng
2010-05-22 09:38:02 +01:00
1432ab15c1
Add graphical output to hbtestfont
2010-05-22 00:56:40 +01:00
bbc7a99d01
Move mirroring around a bit
2010-05-21 18:25:08 +01:00
bd0987386b
Add a few more standard features
2010-05-21 18:06:35 +01:00
75f3469ca6
Add note
2010-05-21 17:59:04 +01:00
1094a294f6
Add rtlm
2010-05-21 17:58:20 +01:00
074ea78749
Add ltra, ltrm, and rtla features
2010-05-21 17:53:10 +01:00
1ce7b87c4d
Cleanup bitmask allocation
2010-05-21 17:31:45 +01:00
dd22a8f7bf
Add note
2010-05-21 16:43:17 +01:00
cbd1d6a63a
Rename Grxxx to HbGrxxx
2010-05-21 15:16:43 +01:00
0375bdd202
Rename classes from Grxxx to HbGrxxx
2010-05-21 15:13:26 +01:00
a5a72e004b
Add hb-ot-shape.h, oops.
2010-05-21 15:12:52 +01:00
aa62a402a7
Fix Makefile to install hb-ot-shape.h
2010-05-21 14:42:11 +01:00
9722b8f005
Simple renames
2010-05-21 14:37:47 +01:00
02f28550d7
Graphite also forces us to link to libstdc++ currently
2010-05-21 14:37:28 +01:00
ca663bb23c
Move main shaper code into hb_ot_shape()
2010-05-21 14:34:23 +01:00
7acd232d36
Fix test failing
2010-05-21 14:20:48 +01:00
dd47924d88
Fix warnings
2010-05-21 14:18:08 +01:00
305ba86715
Import Graphite shaping backend by Martin Hosken
2010-05-21 14:15:39 +01:00
83f34677bc
Add hb_tag_from_string()
2010-05-21 13:43:49 +01:00
4a9a5c0b06
Don't allocate bits for features not available
2010-05-21 13:33:46 +01:00
e53d77142a
Speed up feature mask setting
...
Patch from Jonathan Kew.
2010-05-21 13:32:38 +01:00
4fa67f34ec
Add Unicode 5.2 scripts
2010-05-21 13:29:12 +01:00
ee1b322100
Fix feature mask setting
2010-05-21 12:55:57 +01:00
b490fa3433
Add hb_ot_tag_to_script()
2010-05-21 11:15:07 +01:00
280af1bddb
Let hb_face_get_table() return NULL if table not found
2010-05-20 18:33:35 +01:00
1cdbfd944e
Fix alternate, again
2010-05-20 17:47:28 +01:00
fdca3d5160
Fix alternate off-by-one
2010-05-20 17:46:20 +01:00
b634beb39e
Fix delta scale, again...
2010-05-20 17:44:52 +01:00
40335d4533
Remove unused operator
2010-05-20 17:35:14 +01:00
f7acd8df51
Do alternate glyph selection!
...
Kinda hand-wavy right now. Not tested.
2010-05-20 17:26:35 +01:00
750a229455
get_table() is allowed to return NULL. Use that to simplify code
2010-05-20 16:23:27 +01:00
99d9ef785f
Write hb_face_create_for_data() in terms of hb_face_create_for_tables()
...
Fixes lack of head_table initialization in create_for_tables() also.
2010-05-20 15:47:49 +01:00
6774463883
Apply user features to ranges!
2010-05-20 15:41:20 +01:00
9b60233385
Add _hb_bit_storage()
2010-05-20 15:31:12 +01:00
6b1b957f6d
Add lookup_map
2010-05-20 15:14:44 +01:00
60010a0c4d
Update always-apply mask from 0xFFFF to 1
...
We plan to use the first bit to be always on.
2010-05-20 14:05:02 +01:00
009aad5678
Invert the mask logic
...
Before, the mask in the buffer was inverted. That is, a 0 bit meant
feature should be applied and 1 meant not applied, whereas in the
lookups, the logic was positive.
Now both are in sync. When calling hb_buffer_add_glyph() manually,
the mask should be 1 instead of 0.
2010-05-20 14:00:57 +01:00
7f411dbfd9
Apply user features
...
No ranges yet
2010-05-20 13:40:18 +01:00
0db299ccad
Change hb_feature_t to keep tag/int instead of string
2010-05-20 13:30:09 +01:00
Behdad Esfahbod
Martin Hosken