|
0fceaef0
|
2023-06-30T08:50:49
|
|
[benchmark-subset] Report in microseconds, not milli
We're in that range now. :)
|
|
|
547dc1a4
|
2023-06-30T08:48:43
|
|
[benchmark-font] Add load_face_and_shape benchmark
To measure face-loading performance
|
|
|
fcf70af1
|
2023-06-30T08:43:10
|
|
[perf] Fix break
|
|
|
794fac32
|
2023-06-29T20:55:52
|
|
[glyf] Minor another range for loop
|
|
|
0ae16766
|
2023-06-29T20:38:42
|
|
[CompositeGlyph] Minor use a range for loop
|
|
|
05c01ad5
|
2023-06-29T16:49:20
|
|
[SimpleGlyph] Minor use a range for loop
|
|
|
4123e0c7
|
2023-06-29T16:27:35
|
|
[path-builder] Speed up mid-point calculation
|
|
|
73376b24
|
2023-06-29T16:14:45
|
|
[subset/cff1] More error handling
|
|
|
cc44b3bc
|
2023-06-29T16:12:10
|
|
[subset/cff1] Handle an error condition
Fixes https://oss-fuzz.com/testcase-detail/5191907895279616
|
|
|
fc38c01a
|
2023-06-29T15:29:15
|
|
Minor inline a function
|
|
|
a520177e
|
2023-06-29T14:56:05
|
|
[gvar] Remove unnecessary initialization
|
|
|
ca44c38c
|
2023-06-29T14:41:21
|
|
Add HB_ALWAYS_INLINE
With MSVC implementation as well.
|
|
|
0a00dc0c
|
2023-06-29T14:17:47
|
|
[draw] Add a few unlikely's
|
|
|
2d9c3da0
|
2023-06-29T13:31:11
|
|
[draw] Inline more functions
|
|
|
17f29c81
|
2023-06-29T13:27:09
|
|
[benchmark_font/draw_glyphs] Implement quadratic_to
Not interested in the fallback implementation here.
|
|
|
d26b3ade
|
2023-06-29T13:25:52
|
|
[draw-session] Inline small functions
|
|
|
7aba1e6d
|
2023-06-29T12:44:34
|
|
[cache] Minor use a range for loop
|
|
|
afae537c
|
2023-06-29T12:42:44
|
|
[cache] Minor remove .init()
Use constructor.
|
|
|
c49ca371
|
2023-06-29T12:38:45
|
|
[Glyph] Minor micro-optimize away a function call
For when coords are not set.
|
|
|
5fb7b02a
|
2023-06-29T12:28:43
|
|
[CompositeGlyph] Minor use range loop
|
|
|
20b32b04
|
2023-06-29T12:08:11
|
|
[SimpleGlyph] Micro-optimize phantom_only
|
|
|
d1660eaf
|
2023-06-29T11:59:03
|
|
[subset/cff] Minor use HB_OPTIMIZE_SIZE_VAL
|
|
|
04809ede
|
2023-06-29T11:58:53
|
|
[CompositeGlyph] Micro-optimize translate()
|
|
|
229db973
|
2023-06-29T11:54:10
|
|
[CompositeGlyph] Minor use vector::push()
|
|
|
75101802
|
2023-06-29T10:54:49
|
|
[CompositeGlyph] Micro-optimize
|
|
|
e9d74d6b
|
2023-06-29T10:53:52
|
|
[SimpleGlyph] Minor add an unlikely to error condition
|
|
|
6abca413
|
2023-06-29T10:39:49
|
|
[path-builder] Micro-optimize
No need to initialize these when has_data=false.
|
|
|
5703c1c4
|
2023-06-29T10:35:23
|
|
[path-builder] Add a couple of unlikely's
|
|
|
f5eead04
|
2023-06-28T15:07:15
|
|
[glyf/gvar] Optimize getting variable phantom points
Used when there's no HVAR table and get_h_advance().
I see some 10% speedup with a SourceSerifVariable-Roman with
the HVAR table removed.
|
|
|
e42d6df5
|
2023-06-28T14:25:11
|
|
[glyf] Minor refactor a couple of lines and micro-optimize
|
|
|
c31471d1
|
2023-06-28T13:51:54
|
|
[glyf] Micro-optimize
|
|
|
83d75d5e
|
2023-06-28T13:35:24
|
|
[glyf] always_inline consume_point
7% speedup in:
BM_Font/draw_glyph/Roboto-Regular.ttf/hb
|
|
|
8795cced
|
2023-06-28T13:29:29
|
|
Revert "[SimpleGlyph] Avoid branches in read_points"
This reverts commit ba062c713e469f91f57f3c85990f721789ec7c2a.
|
|
|
ba062c71
|
2023-06-28T13:24:37
|
|
[SimpleGlyph] Avoid branches in read_points
Calculate total bytes in read_flags and bounds-check once.
This slows things down apparently, so going to revert.
|
|
|
62f5ed46
|
2023-06-28T12:02:52
|
|
[subset/cff] Fix an infinite loop
Fixes https://oss-fuzz.com/testcase-detail/5419002026131456
|
|
87b57361
|
2023-06-28T15:38:40
|
|
README.python.md: Add some notes for Windows
Note that the DLLs for HarfBuzz and its deps must be found %PATH% on Windows so
that pygobject is able to import and use HarfBuzz in Python scripts.
Also note that for Visual Studio builds, it is recommended that Visual Studio
2019 or later is used for the build, possibly in regards with how the
preprocessor handles C++ code when running g-ir-scanner.
|
|
10b9ca54
|
2023-06-28T07:46:33
|
|
Fix typo: subtitution -> substitution
|
|
|
73a82af1
|
2023-06-27T17:05:22
|
|
[subset/cff] Remove stale wrong code
With my recent rewrite of serialize_header, this seems to
work now.
|
|
|
158eba66
|
2023-06-27T16:22:54
|
|
[subset/cff] Fix compiler error
About variable with internal linkage with no definition.
|
|
|
b0472199
|
2023-06-27T16:07:54
|
|
[subset/cff1] Micro-optimize code
Let compiler see whether optional argument is present.
|
|
|
1cabb65b
|
2023-06-27T16:01:17
|
|
[subset/cff] Micro-optimize CFFIndex::serialize_header
Unfortunately hb_array_t::begin/end are faster than its _begin/_end.
As such, a range loop on array itself is faster than range loop on
pipeline starting with array. Rework code to loop on the array always.
|
|
|
dfaf9cd5
|
2023-06-27T14:56:32
|
|
[subset/cff1] Minor optimize
|
|
|
41f85d01
|
2023-06-27T12:50:04
|
|
[benchmark-font] Add RobotoFlex
We didn't have a variable TTF before!
|
|
|
585c8426
|
2023-06-27T11:28:38
|
|
[subset/cff1] Minor micro-optimize
|
|
|
00904503
|
2023-06-27T08:18:18
|
|
[config] Graduate avar2 from boring-expansion
|
|
|
3edd6cdc
|
2023-06-27T06:56:00
|
|
[vector] Minor micro-optimize shrink_vector
The compiler seems to understand this pattern better.
|
|
|
aed21563
|
2023-06-26T22:51:50
|
|
[subset/cff1] Micro-optimize
|
|
|
40a1c08f
|
2023-06-26T19:03:04
|
|
[subset/cff1] Micro-optimize
|
|
|
bd3e7877
|
2023-06-26T18:55:39
|
|
[subset/cff1] Reuse a function
|
|
|
ffd23d3a
|
2023-06-26T18:54:06
|
|
[subset/cff1] Micro-optimize Charset serialization
|
|
|
452557cd
|
2023-06-26T18:40:54
|
|
Replace a free with hb_free
Ouch.
|
|
|
1fc128f6
|
2023-06-26T18:34:27
|
|
Replace a few memset/memcpy's with hb_ equivalents
|
|
|
2c359635
|
2023-06-26T18:28:11
|
|
[face-builder] Don't zero allocation
We overwrite it.
|
|
|
347b9448
|
2023-06-26T18:09:40
|
|
[null] Fix getting Crap(hb_bytes_t)
Fixes https://oss-fuzz.com/testcase-detail/6187272924692480
|
|
|
49c52fa9
|
2023-06-26T17:55:29
|
|
[cmap] Don't zero a few allocations unnecessarily
|
|
|
e4b2d9c3
|
2023-06-26T17:52:37
|
|
[pool] Don't clear unused memory
|
|
|
6129702e
|
2023-06-26T17:34:54
|
|
[subset/cff] Simplify some allocation embedding
|
|
|
99db06b4
|
2023-06-26T17:30:31
|
|
[var] Simplify a couple of copy operations
|
|
|
73bcd3f2
|
2023-06-26T17:24:40
|
|
[var] Adjust a few allocations to not clean space
Since we immediately copy.
|
|
|
b1e7e8ba
|
2023-06-26T17:19:45
|
|
[algs] Fix return of hb_memset()
|
|
|
c2bab073
|
2023-06-26T16:34:53
|
|
[subset/cff] Minor optimization
|
|
|
570fb4df
|
2023-06-26T16:26:12
|
|
[subset/cff] Add a vector pre-alloc
|
|
|
e447d394
|
2023-06-26T16:02:41
|
|
[subset/cff1] Micro-optimize
|
|
|
50499e9e
|
2023-06-26T15:32:27
|
|
[subset/cff1] Minor use rvalues in a few vector push() places
|
|
|
e5b3d4b1
|
2023-06-26T14:24:38
|
|
[subset/cff1] Minor use ?:
|
|
|
81e2db7c
|
2023-06-26T14:17:44
|
|
[subset/cff1] Speed up plan_subset_charset
|
|
|
d3f90a8c
|
2023-06-26T14:02:26
|
|
[subset/cff] Use a typedef for glyph_to_sid_map_t
|
|
|
6783701b
|
2023-06-26T12:31:22
|
|
[subset] Handle an error condition
Fixes https://oss-fuzz.com/testcase-detail/6306810588692480
|
|
|
7d3d157d
|
2023-06-26T12:02:21
|
|
[subset/cff] Speed up hb_plan_subset_cff_fdselect
|
|
|
f79d961a
|
2023-06-26T11:32:37
|
|
[vector] Speedup push()
|
|
|
b4b80bca
|
2023-06-26T11:29:32
|
|
[subset/cff1] Speed up plan_subset_charset
|
|
|
7305ec47
|
2023-06-26T12:35:03
|
|
Merge pull request #4300 from pnacht/scorecard-action
Add Scorecard action
|
|
4a628b23
|
2023-06-26T15:19:29
|
|
[meson] Remove incorrect option for ICU subproject
Fixes #4298 .
|
|
3bdb8639
|
2023-06-26T11:25:10
|
|
Add Scorecard badge to README
|
|
|
27ef6c08
|
2023-06-26T11:22:18
|
|
Create scorecard.yml
|
|
|
6c4f975d
|
2023-06-25T23:37:08
|
|
Add a fuzzer font
|
|
|
87f1b80d
|
2023-06-25T23:14:01
|
|
[bimap] Write with a map & vector combo
Faster for getting keys.
Speeds up 10% in:
BM_subset/subset_glyphs/RobotoFlex-Variable.ttf/retaingids/10
|
|
|
e7ad017c
|
2023-06-25T22:25:24
|
|
[subset/cff] Allocate more memory for retain-gids
To avoid serializing twice.
20% speedup in:
BM_subset/subset_glyphs/SourceSansPro-Regular.otf/retaingids/10
|
|
|
30f1ab86
|
2023-06-25T19:29:03
|
|
[subset/cff1] Fix typo
|
|
|
7c7db0fc
|
2023-06-25T19:07:40
|
|
[subset/cff1] Minor shuffle code around
|
|
|
1e09ebeb
|
2023-06-25T18:24:42
|
|
[subset/cff1] Minor use vector length instead of map population
|
|
|
23109dde
|
2023-06-25T18:21:57
|
|
[subset/cff1] Minor use an exact allocation
|
|
|
393f0f9f
|
2023-06-25T18:14:56
|
|
[map] Rename resize() to alloc()
Better matches the functionality, and hb_vector_t.
|
|
|
793f663b
|
2023-06-25T17:33:57
|
|
[vector] Speed up hb_vector_t<hb_array_t<U>>
|
|
|
289bad82
|
2023-06-25T17:26:40
|
|
[subset/cff1] Minor use a dagger
|
|
|
31d971d1
|
2023-06-25T17:23:52
|
|
[vector] Sprinkle std::addressof
|
|
|
c03c0a9d
|
2023-06-25T17:13:59
|
|
[subset/cff1] Remove unnecessary remap_sid_t::reset
The object is initialized automatically.
|
|
|
6ed0d04b
|
2023-06-25T17:10:36
|
|
[subset/cff1] In remap_sid_t::reset(), reset everything
Previous code wasn't resetting next=0; tests were passing.
It's always called when sidmap is empty. So, redundant,
but keeping as is.
|
|
|
7e97233c
|
2023-06-25T17:06:02
|
|
[subset/cff1] Comment
|
|
|
86a550dc
|
2023-06-25T17:01:40
|
|
[subset/cff1] Speed up string writing
|
|
|
dc1b1724
|
2023-06-25T16:35:36
|
|
[subset/cff1] Minor reuse a returned value
|
|
|
95341bcc
|
2023-06-25T16:31:54
|
|
[subset/cff1] Minor use {} initialization syntax
|
|
|
be0cec28
|
2023-06-25T16:01:13
|
|
[map] Micro-optimize iteration
Make is_real() faster (removes a shift). is_used() gets inlined
so is not slowed down by this change.
|
|
|
9dbea607
|
2023-06-25T14:28:02
|
|
[subset/cff1] Fix resource leak in error case
|
|
e310473c
|
2023-06-24T21:50:58
|
|
[use] Allow multiple CMAbv glyphs on subjoined
|
|
|
d4bbe3f4
|
2023-06-23T18:13:30
|
|
[subset/cff] Reuse a calculate index total data size
|
|
|
ec0fbf8f
|
2023-06-23T16:13:13
|
|
[subset/ClassDef] Micro-optimize use_class_zero calc
Probably never matters.
|
|
|
b557a841
|
2023-06-23T15:31:54
|
|
[algs] Speed up fasthash for aligned uint64_t
|
|
|
fc80d20c
|
2023-06-23T15:21:30
|
|
[serialize] Only hash at most 128 bytes for object_t
Optimization. Shouldn't in reality bring down the hash performance.
Byte objects differ in their early bytes anyway.
|
|
|
19eb5e3b
|
2023-06-23T15:12:13
|
|
[subset/cff1] Micro-optimize Charset::serialize
|