784fe9ac67
[cmap] Simplify caching
2023-01-29 09:26:52 -07:00
a451aa5465
Add back a null check
...
This was accidentally dropped in the previous commit.
2023-01-29 11:25:28 -05:00
318aa10708
[ot-font] Use the cmap cache more
...
Use the cmap cache for get_nominal_glyph and
get_variation_glyph as well. The first of these
is used a lot in pango.
2023-01-29 09:17:17 -05:00
f8a744d9d5
[ot-font] Add a cmap cache
...
Speeds up Roboto shaping by 7%, for 1kb per face.
2023-01-28 13:38:07 -07:00
a34a204bf7
[subset-plan] Simplify unicodes allocation
2023-01-11 12:27:19 -07:00
bd4b040e7f
[shape-plan] Simplify glyphs_requested allocation
2023-01-11 11:23:48 -07:00
023f595dec
[cmap] Speed up DefaultUVS::copy even more
...
Another 14% on SourceHanSerifVF/10 benchmark.
2022-12-03 11:18:05 -07:00
4ca6105108
[cmap] Remove double-min
2022-12-03 11:15:06 -07:00
cd29147e30
[cmap] Minor cast
2022-12-03 10:41:42 -07:00
4cdb5cc69b
[cmap] Minor change iterator
2022-12-03 10:40:24 -07:00
a2d33779e1
Fix arm bot build
2022-12-03 09:49:00 -07:00
dabbf13d40
[cmap] Speed up DefaultUVS::copy
2022-12-03 09:46:11 -07:00
2658370f00
[subset] make the cmap cache in accelerator const.
2022-11-29 22:04:10 -07:00
7a004a7ac2
[subset] Cache per subtable cmap unicode mappings.
2022-11-28 19:48:15 -07:00
c503cf003e
[cmap] Store offset, not pointer, in cmap cache
2022-11-28 15:53:35 -07:00
a15ad778fe
[arabic-fallback] Generate PUA table from data
...
Uses packtab for more compact arrays.
2022-06-19 12:29:10 -06:00
8c27c51c27
[arabic-pua] Rename symbols
2022-06-19 10:47:38 -06:00
7698962911
[arabic-fallback] Disable PUA shaping under HB_NO_OT_SHAPER_ARABIC_FALLBACK
2022-06-19 10:42:06 -06:00
55350377b0
[cmap/ft] Only map 0xF000 range if font_page is NONE
2022-06-19 10:42:06 -06:00
c3f590bb1e
[arabic] Support legacy PUA shaping
...
Support legacy pre-OpenType Windows 3.1-era fonts, by remapping PUA code
points in cmap table and letting our fallback shaper build the GSUB
table.
Uniscribe applies also mset-like substitution, but our fallback mark
positioning gives better results, so this is not implemented.
2022-06-19 10:42:05 -06:00
5dc12d7d8d
[cmap] Rewrite set_for() slightly
2022-06-03 01:37:02 -06:00
9552955e08
Add an unlikely
2022-06-03 01:33:01 -06:00
a7a688616a
[cmap] Convert another map use to unique_ptr
2022-06-02 18:59:15 -06:00
f82ee17a75
[map] Pre-size map in constructor if we can
2022-05-18 12:19:49 -06:00
8f9f0c494b
[subset] Enforce cmap12 group ordering constraints in collect_mapping.
...
Fixes fuzzer issue: https://oss-fuzz.com/testcase-detail/6365271012540416
2022-05-10 12:15:09 -06:00
f10ddb8dd8
[cmap] Use -1 as Unicode sentinel, not U+FFFF in Format12 serialize
2022-05-05 11:21:24 -06:00
8a19968c8b
[cmap] Use iterator bool operator
2022-05-05 11:17:23 -06:00
052812b6ba
Merge pull request #3561 from googlefonts/cmap_opt
...
[subset] Further cmap subsetting speed optimizations
2022-05-04 15:38:30 -06:00
f0c04114bc
[subset] Embed unicode to gid list vector in subset plan.
2022-05-03 22:02:59 +00:00
3fff2e9182
[perf/benchmark-font] Cosmetic
2022-05-02 16:42:10 -06:00
307d2d8bb6
[cmap] Sprinkle some 'unlikely's
2022-05-02 16:30:22 -06:00
088133d939
[subset] cache cp to new gid list in subset plan.
...
This avoids having to recompute the ordered list multiple times during cmap generation.
2022-05-02 21:29:16 +00:00
6922a2561f
[subset] Change serialize_rangeoffset_glyid back to using iterator.
2022-04-29 23:30:32 +00:00
c66fd50c26
[subset] in cmap4 serialization save cp to gid iter to memory.
...
Iterator accesses are slow and it's iterated multiple times.
2022-04-29 23:18:53 +00:00
17b98563dc
[subset] In cmap4 serialization reduce unnessecary calls into the iterator.
...
Gives ~20% speedup for large subsets.
2022-04-29 22:49:02 +00:00
5e241094bf
[subset] In unicodes cache cleanup if set insert fails.
2022-04-29 22:45:16 +00:00
a424a92ce5
[subset] s/void */intptr_t.
2022-04-29 22:14:03 +00:00
aad67f5629
[subset] cache results of collect_unicodes.
2022-04-29 22:05:34 +00:00
b4236b7de6
[subset] Optimize Cmap4 collect_unicodes.
...
Use set add_range() instead of individual add() calls.
2022-04-29 19:22:00 +00:00
f41945e313
[cmap] In collect_unicodes() of format 12/13, limit to max Unicode
...
Fixes fuzzer timeout:
https://oss-fuzz.com/testcase-detail/5062368881672192
2022-03-21 18:24:30 -06:00
ac1bb3e39e
[machinery] Move accelerators to constructor/destructor
2022-01-20 12:10:05 -07:00
e062376ef1
[machinery] Make accelerator lazy-loader call Xinit/Xfini
...
Instead of init/fini. To isolate those functions. To be turned into
constructor/destructors, ideally one per commit (after some SFINAE
foo.)
2022-01-19 17:09:34 -07:00
8a69e00639
[meta] Use std::addressof() instead of hb_addressof()
2022-01-13 16:17:34 -07:00
1d9ef3a75a
[subset] Actually fix end_cp unitialized warning.
2021-12-01 11:53:10 -07:00
d8635dfe5a
[subset] Fix warning about uninitialized use of end_cp.
2021-12-01 11:17:46 -07:00
95329081c2
[subset] further optimize cmap4 packing.
2021-11-28 07:47:49 -07:00
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
c852b86841
Rename HBGlyphID to HBGlyphID16
2021-09-19 16:30:12 -04:00
2bd911b8b4
[subset] handle cmap4 overflows.
...
If a cmap4 subtable overflows during serialization drop it and the corresponding EncodingRecord. Don't drop the corresponding cmap12 table if it would have otherwise been removed.
2021-09-02 14:43:17 -06:00
b9a176e268
[subset] speedup cmap4 subsetting for large codepoint counts. ( #3178 )
...
glyphIdArray generation implementation was O(n^2). Refactored to use a hashmap to reduce complexity. After the change subset time for a 22k codepoint subset went from 7s to 0.7s.
2021-08-29 10:33:12 -06:00
Behdad Esfahbod
Matthias Clasen
Garret Rieger
Khaled Hosny