|
4807a021
|
2025-03-12T18:27:27
|
|
[atomic] Kill hb_atomic_ptr_t<T>
Use hb_atomic_t<T *> instead.
|
|
|
b8327fe6
|
2025-02-06T20:01:12
|
|
Fix compiler errors
|
|
|
6091abcc
|
2025-02-06T02:40:24
|
|
[buffer] Avoid a copy in collect_codepoints()
|
|
|
e5e3bc40
|
2025-02-06T02:36:26
|
|
[buffer] Merge two functions into a template
|
|
|
c7378294
|
2025-02-06T01:42:25
|
|
[OT] Remove likely() from coverage check
Not that it matters, but because of the set-digest, this path
is not very likely() anymore.
|
|
|
3577992a
|
2025-02-06T01:30:59
|
|
[gsubgpos] Simplify a destruction
|
|
|
0a678a92
|
2025-02-06T01:02:44
|
|
Fix leak
|
|
|
4e75dfae
|
2025-02-05T23:57:37
|
|
Hide unused member
|
|
|
99043f20
|
2025-02-05T20:11:27
|
|
[gsubgpos] Tweak cache again
|
|
|
e535c683
|
2025-02-05T19:41:03
|
|
[Context] Simplify cache costing
|
|
|
726af2e4
|
2025-02-05T18:10:40
|
|
[PairPos] Use a class cache
10% speedup on Roboto-Regular. 5% on SF Pro.
|
|
|
acd122d0
|
2025-02-05T17:38:39
|
|
[gsubgpos] Prepare for per-lookup allocated caches
|
|
10cf348a
|
2024-11-11T10:32:01
|
|
[gsubgpos] Guard integer arithmetic against wrapping below zero
When either casting a signed integer to an unsigned one or subtracting
a value from an unsigned one, add assertions to verify we will not end
up below expected bounds.
|
|
|
7a890c2e
|
2024-09-22T13:19:02
|
|
Add hb_barrier() to switches of unions
https://github.com/harfbuzz/harfbuzz/pull/4864#issuecomment-2366923736
|
|
|
7aace3d3
|
2024-09-23T07:47:24
|
|
Ignore CGJ and Mongolian Variation Selectors during GPOS
Fixes https://github.com/harfbuzz/harfbuzz/issues/4869
|
|
|
e4e9f6a1
|
2024-09-22T11:17:48
|
|
[gsubgpos] Add a barrier
Fixes https://github.com/harfbuzz/harfbuzz/issues/4852
|
|
|
788b469a
|
2024-07-25T14:50:01
|
|
[ChainContext] Fix fast-path deviation from slow path
See comments. The input skippy was hitting the U+200D and NOT
seeing it skippable, whereas the context skippy would had.
Fixes https://github.com/harfbuzz/harfbuzz/issues/4813
|
|
|
0037ee28
|
2024-05-24T08:10:15
|
|
Merge pull request #4578 from harfbuzz/varc-table
`VARC` table
|
|
|
e2ab6c7b
|
2024-05-12T15:25:13
|
|
[kern/kerx] Add accelerator and set-digest filtering
|
|
|
ba1f194a
|
2024-05-02T12:18:11
|
|
[gsubgpos] Reduce stack use in recursion
|
|
|
ac411f26
|
2024-04-23T13:26:53
|
|
[Condition] Finish evaluation of ConditionValue
https://github.com/adobe-type-tools/opentype-spec-drafts/blob/main/condvalue_spec.md
|
|
|
f1ac867d
|
2024-03-14T13:30:21
|
|
[buffer] Add API for random state
Fixes https://github.com/harfbuzz/harfbuzz/issues/4620
New API:
+hb_buffer_set_random_state()
+hb_buffer_get_random_state()
|
|
|
1cb47eff
|
2024-02-06T14:49:31
|
|
Rename VariationStore to ItemVariationStore
|
|
|
7ee7e2e3
|
2023-11-04T15:02:28
|
|
[sanitize] Add remaining hb_barrier() annotations
I'm sure I've forgot some. But this is a good start.
|
|
aca55dbe
|
2023-11-03T10:51:37
|
|
[instancer] fix feature variations collect lookups
We already know which record to keep, so just collect lookups and no
need to check the substitutes_map
|
|
|
091e538d
|
2023-10-26T12:52:08
|
|
[instancer] support partial instancing for FeatureVariations
|
|
|
dcd3afca
|
2023-08-01T20:09:34
|
|
[skippy-iter] Remove unused num_items
|
|
|
39048099
|
2023-08-01T19:41:12
|
|
[skippy-iter] Remove early stop
The optimization in (Chain)RuleSet matching relies on
matching one, even if num_items is out of range.
Fixes https://github.com/harfbuzz/harfbuzz/issues/4358
|
|
|
2feac50b
|
2023-07-31T19:53:06
|
|
Revert "[gsubgpos] Keep another digest in the applicable_t"
This reverts commit fd79c7cecdf68fe4626943f29bd5edf1e603d2b2.
|
|
|
fd79c7ce
|
2023-07-31T19:52:15
|
|
[gsubgpos] Keep another digest in the applicable_t
The digest for all the remaining subtables combined.
The idea is to get out of the subtable look as soon as
no more can be applied.
Doesn't seem to speed up anything I tested. Going to revert.
|
|
|
e3fd69c8
|
2023-07-31T19:07:42
|
|
[layout] Inline another function
Code is smaller too.
|
|
|
60db1427
|
2023-07-24T20:49:32
|
|
[layout] Remove unused "max-size" cruft
|
|
|
c91899be
|
2023-07-21T11:43:47
|
|
[gdef] Use set-digest for mark-filterint-sets
Speeds up Noto Duployan-Regular.otf by 45% percent!
|
|
|
e8948a4e
|
2023-07-20T22:12:53
|
|
[gsubgpos] Fix optimization
Fixes https://github.com/harfbuzz/harfbuzz/issues/4336
|
|
|
e8de5cb0
|
2023-07-20T11:31:13
|
|
[gsubgpos] Use a couple variables
|
|
|
0ccd61a3
|
2023-07-19T16:42:30
|
|
[gsubgpos] Micro-optimize
|
|
|
b01c1650
|
2023-07-19T16:35:10
|
|
[gsubgpos] Minor error handling
|
|
|
ee5f200b
|
2023-07-19T12:26:50
|
|
[gsubgpos] Remove un unnecessary variable
|
|
|
42c6a3a1
|
2023-07-19T12:07:19
|
|
[gsubgpos] Remove unnecessary condition
Second is always set in this branch.
|
|
|
b5de54e1
|
2023-07-19T11:19:06
|
|
[gsubgpos] Also match second component in (Chain)RuleSet::apply()
Another 40% speedup in shaping Gulzar-Regular.
|
|
|
7a2f9dcf
|
2023-07-18T04:36:15
|
|
[matcher] Move initializers inline
|
|
|
d92a7a58
|
2023-07-17T11:27:38
|
|
[gsubgpos] Inline a couple functions
Produces smaller code.
|
|
|
4ea3737d
|
2023-07-17T09:50:28
|
|
[gsubgpos] Add fast-path for end-of-string in (Chain)RuleSet::apply()
|
|
|
49730531
|
2023-07-16T13:09:49
|
|
Config
|
|
|
645fabd1
|
2023-07-16T11:36:59
|
|
[gsubgpos] Vastly speed up ChainRuleSet / RuleSet matching
Match the first component in a fast loop.
Idea replicated from LigatureSet.
Speeds up Gulzar shaping by 22%!
37% in NotoNastaliqUrdu!
|
|
|
77080f86
|
2023-07-16T11:26:32
|
|
.
|
|
|
5c8f3b7f
|
2023-07-16T09:39:18
|
|
[gsubgpos] Conditionally cache backtrack again
|
|
|
52237eb1
|
2023-07-16T09:07:25
|
|
[ChainContext] More caching
|
|
|
62e10c9f
|
2023-07-16T08:43:18
|
|
Revert "[gsubgpos] Allocate iterators further up"
This reverts commit b9f364b8fcacf8ab32d0272190d8509f214e55f2.
|
|
|
b9f364b8
|
2023-07-16T08:42:36
|
|
[gsubgpos] Allocate iterators further up
To avoid calling set_match_func repeatedly.
Doesn't show speedup. Going to revert.
|
|
|
c2f454c7
|
2023-07-16T08:23:59
|
|
[gsubgpos] Fix residual from 5af80f349c4e040bfa853ee3f561ac16538b5988
|
|
|
5af80f34
|
2023-07-16T07:27:28
|
|
[ChainContext] Cache two class values
7% speedup shaping Gulzar.
|
|
|
8cde4fa9
|
2023-07-16T06:01:06
|
|
[gsubgpos] Inline a couple methods
|
|
|
d00b8873
|
2023-07-14T11:26:51
|
|
Revert "[subset/closure] Batch recursions in scheduled stages"
This reverts commit f2aaeeb3016e10bf91c251296391a381d5fc6385.
|
|
|
f2aaeeb3
|
2023-07-14T11:17:37
|
|
[subset/closure] Batch recursions in scheduled stages
Going to revert. Doesn't pass tests and savings are minor.
|
|
|
d84c5f29
|
2023-07-08T15:22:51
|
|
Revert "[gsubgpos] Make (Chain)Context funcs templatized"
This reverts commit 03ac08d267eb7ad2f1ac039115fc0030ea475fae.
|
|
|
e73223b9
|
2023-07-08T15:22:18
|
|
[gsubgpos] Make (Chain)Context funcs templatized
Unfortunately they still won't be inlined because the matcher_t
needs function pointers. So, no speed up. Going to revert.
|
|
|
1c8d54de
|
2023-07-08T14:28:07
|
|
[shape] Another always-inline
|
|
|
df45067f
|
2023-07-08T14:22:59
|
|
[shape] More always-inline
|
|
|
c3a28be5
|
2023-07-08T14:20:45
|
|
[shape] One more always-inline
|
|
|
9fea19eb
|
2023-07-08T13:56:24
|
|
[shape] Another always-inline
|
|
|
5b45c56c
|
2023-07-08T13:47:41
|
|
[shape] Always-inline more
Another 10% speedup in BM_Shape Gulzar-Regular.
|
|
|
bb9692e8
|
2023-07-08T13:45:48
|
|
[shape] Always-inline match_input
10% speedup in BM_Shape Gulzar-Regular.
|
|
|
6c451054
|
2023-07-08T12:55:34
|
|
[sanitize] Sprinkle a few unlikely's
|
|
|
4ecc62c2
|
2023-07-08T12:53:23
|
|
[sanitize] Simplify ChainRule::sanitize again
check_struct does the same now.
|
|
|
fb02f4a8
|
2023-07-08T12:29:18
|
|
[sanitize] Micro-optimize ChainRule::sanitize
|
|
|
403bc7ca
|
2023-07-07T20:17:32
|
|
[sanitize] Simplify Rule::sanitize()
|
|
|
385896ff
|
2023-07-07T19:56:19
|
|
[sanitize] Speed up ChainRule::sanitize()
10% speedup in sanitizing Gulzar-Regular.
|
|
|
2006d321
|
2023-07-07T17:07:11
|
|
[sanitize] Add "fastpath" for ArrayOfOffset16To<> objects with max size
Unfortunately this doesn't speed up NotoNastaliq or Gulzar as I was
hoping for. Their GSUB tables are not large enough for this to kick
in...
|
|
|
7a85663c
|
2023-07-07T19:21:18
|
|
Revert "[sanitize] Add "fastpath" for ArrayOfOffset16To<> objects with max size"
This reverts commit 10f8556c73f3cf231c6b5a900a6a1903f9516f90.
This was, unfortunately, wrong :(.
|
|
|
f2a3680f
|
2023-07-07T19:21:03
|
|
Revert "[gsubgpos] Limit (Chain)ContextFormat3 for consistency"
This reverts commit 90b48917dadbef51bc5e90904d0f81e81199b9e2.
|
|
|
90b48917
|
2023-07-07T18:50:01
|
|
[gsubgpos] Limit (Chain)ContextFormat3 for consistency
|
|
|
10f8556c
|
2023-07-07T17:07:11
|
|
[sanitize] Add "fastpath" for ArrayOfOffset16To<> objects with max size
|
|
|
68b78914
|
2023-07-07T17:07:11
|
|
[sanitize] Add "fastpath" for ArrayOfOffset16To<> objects with max size
Unfortunately this doesn't speed up NotoNastaliq or Gulzar as I was
hoping for. Their GSUB tables are not large enough for this to kick
in...
|
|
|
90752cd5
|
2023-07-07T15:47:28
|
|
[type] Add HeadlessArray16Of
|
|
|
2d6091fc
|
2023-06-30T11:48:56
|
|
[GPOS] Make AnchorMatrix sanitize lazy again
Was reverted in the previous commit, because it was incomplete.
|
|
|
690af7aa
|
2023-06-30T10:36:01
|
|
[GPOS] Sanitize Device tables lazily
This speeds up face loading for variable fonts by 80%!
Comparing before to after
Benchmark Time CPU Time Old Time New CPU Old CPU New
---------------------------------------------------------------------------------------------------------------------------------------------------------------
BM_Font/load_face_and_shape/Roboto-Regular.ttf/hb -0.0368 -0.0366 20 20 20 19
BM_Font/load_face_and_shape/RobotoFlex-Variable.ttf/hb -0.7149 -0.7162 77 22 77 22
BM_Font/load_face_and_shape/RobotoFlex-Variable.ttf/var/hb -0.7241 -0.7255 80 22 79 22
BM_Font/load_face_and_shape/SourceSansPro-Regular.otf/hb -0.1441 -0.1445 28 24 28 24
BM_Font/load_face_and_shape/AdobeVFPrototype.otf/hb -0.7893 -0.7910 66 14 66 14
BM_Font/load_face_and_shape/AdobeVFPrototype.otf/var/hb -0.7865 -0.7882 67 14 66 14
BM_Font/load_face_and_shape/SourceSerifVariable-Roman.ttf/hb -0.8895 -0.8900 227 25 226 25
BM_Font/load_face_and_shape/SourceSerifVariable-Roman.ttf/var/hb -0.8895 -0.8900 226 25 225 25
BM_Font/load_face_and_shape/Comfortaa-Regular-new.ttf/hb -0.5512 -0.5531 42 19 42 19
BM_Font/load_face_and_shape/NotoNastaliqUrdu-Regular.ttf/hb -0.1511 -0.1510 227 192 225 191
BM_Font/load_face_and_shape/NotoSerifMyanmar-Regular.otf/hb -0.1494 -0.1498 41 35 40 34
OVERALL_GEOMEAN -0.6443 -0.6456 0 0 0 0
|
|
|
9df07c3c
|
2023-06-08T11:46:06
|
|
[gsubgpos] Prevent a leak in closure
If the push wasn't successful we were constructing an
hb_set_t on the Crap data... At least that's my reading
of the code.
|
|
|
82741304
|
2023-06-04T09:41:41
|
|
[subset] start_embed never returns nullptr
Remove checks.
|
|
|
31949636
|
2023-05-04T16:29:48
|
|
Revert "[layout] Cache subtable coverages in hb_map_t"
This reverts commit 7a715d74e06720c17d28ba7b4c3da0b583f8d1d3.
|
|
|
7a715d74
|
2023-05-04T16:28:11
|
|
[layout] Cache subtable coverages in hb_map_t
Proof-of-concept. Going to revert. Memory consumption is more
than I like. It does speed up Roboto shaping another 15% though.
Perhaps if we could add logic to choose which subtables to
cache, this might be a useful approach.
|
|
|
975980d3
|
2023-05-04T15:56:06
|
|
[gsubgpos] Add apply_cached() as separate method
In prep for some other work.
|
|
|
3f2401e2
|
2023-05-03T16:56:52
|
|
[layout] Don't init iters successively multiple times
|
|
|
959f1634
|
2023-05-03T16:27:41
|
|
[gsubgpos] Minor reduce variable scope
|
|
|
fb795dc3
|
2023-05-03T15:17:20
|
|
[GPOS] Optimize iterator reset
Speeds up Gulzar shaping 5%.
|
|
|
8a8fc37c
|
2023-05-03T14:57:00
|
|
[GDEF] Use a cache for glyph classes
Shows 5% speedup for Roboto shaping.
|
|
|
49ac5e11
|
2023-05-02T12:12:26
|
|
[match_input] Micro-optimize
|
|
|
63afb4f2
|
2023-04-26T10:29:25
|
|
[syllabic] Better fix for previous issue
With previous fix the GPOS application was still reading the syllable()
member, which was already freed. This fix is more correct.
|
|
|
125450d2
|
2023-03-13T15:43:29
|
|
[subset] collect name_ids for FeratureParams
|
|
|
8708b9e0
|
2023-02-06T14:51:25
|
|
[GPOS] Avoid O(n^2) behavior in mark-attachment
Better implementation; avoids arbitrary limit on look-back.
|
|
|
661050b4
|
2023-02-06T12:38:17
|
|
Revert "[layout] Limit how far we skip when looking back"
This reverts commit 85be877925ddbf34f74a1229f3ca1716bb6170dc.
|
|
|
b29fbd16
|
2023-02-06T13:08:52
|
|
[gsubgpos] Refactor skippy_iter.match()
|
|
|
85be8779
|
2023-02-01T20:00:43
|
|
[layout] Limit how far we skip when looking back
See comments.
|
|
|
d18fd3f7
|
2023-01-31T16:28:49
|
|
[layout] Comment
|
|
|
7a4bd97e
|
2023-01-31T14:59:39
|
|
[layout] Build lookup accelerators lazily on-demand
Reduces memory consumption for large multi-script fonts
drastically.
|
|
|
83353f13
|
2023-01-31T14:32:14
|
|
[layout] Reduce memory use slightly
By using raw pointer instead of vector for subtable accelerator.
To be used for more memory saving by making subtable accelerators
lazy-loaded by shape-plans for large fonts.
|
|
|
49d75ef3
|
2023-01-27T15:37:11
|
|
[gsubgpos] Fix bug in cached ChainContextFormat2 application
|
|
|
d2279a20
|
2023-01-27T12:32:55
|
|
[gsubgpos] Avoid a copy into the vector
|
|
|
8d29be39
|
2023-01-26T11:41:58
|
|
[gsubgpos] Minor drop an unnecessary hb_iter
|
|
|
be8a87c4
|
2023-01-21T15:00:41
|
|
Move TRACE_DISPATCH after may_recurse()
Such that we don't get memory access issues if DEBUG_SANITIZE is
on and may_recurse() returns false.
|