7ec1c41a55
[gsubgpos] Skippy-iter: Prefer correctness to performance
...
Prefer unsafe-to-concat correctness, over performance.
2022-11-23 21:12:41 -07:00
b6df347137
[perf] Add Hindi test to benchmark-shape
2022-11-23 20:15:16 -07:00
dce3502e10
[array] Add commented-out static asserts
...
They don't work.
2022-11-23 18:34:54 -07:00
a7fee43cef
[priority-queue] Minor micro-optimize
2022-11-23 17:46:32 -07:00
ff3cac0ccd
[subset] Fix unecessary trailing 0 bytes left by ContextFormat2 pruning.
...
Uneeded rules where beind removed from the count by the bytes for them was being left in the font.
2022-11-23 23:50:49 +00:00
f2851e4157
[test-map] Test has() getter with unique-ptr
2022-11-23 16:50:30 -07:00
5f3a780614
[font] Protect against div-by-zero
2022-11-23 16:45:46 -07:00
060ecac949
[font] Respect subfont slant setting in hb-draw
...
Fixes https://github.com/harfbuzz/harfbuzz/issues/3890
2022-11-23 16:31:37 -07:00
2e9b270a49
[hb-view] Fix cairo slanting condition
...
hb-draw already does slanting. If NOT hb-draw, we should slant
through cairo path. Donno why this was untested before.
This was double-slanting with hb-draw, and not slanting without it.
2022-11-23 16:18:39 -07:00
51028e63e6
[subset] Retain all glyphs in preprocessed face.
2022-11-23 22:51:16 +00:00
404cb99d86
[buffer-diff] Fix check for glyph flag equality
...
I'm not sure if the old behavior was intentional, but it was checking
that the glyph flags were a subset of the reference buffer's glyph
flags. I don't see why that is useful. Fix that.
Then make the buffer-verify code ignore flag differences when verifying
buffers, since our unsafe-to-concat flag at least, is conservative and
not guaranteed to be produced the same in fragments. See:
https://github.com/harfbuzz/harfbuzz/issues/3888
2022-11-23 15:37:03 -07:00
4c49daf7cd
[subset] actually use the preprocessed face in hb-subset.
...
Tests weren't actually using the preprocessed face due to this typo in util/hb-subset.
2022-11-23 22:33:57 +00:00
2c0afde737
[subset] add an inprogress accelerator to plan.
...
This allows subset code to cache information into the accelerator during preprocess subset. Previously the accelerator was created at the end of subsetting.
2022-11-23 13:41:52 -07:00
81640fdffe
[subset] fix leaked font in glyf::subset(...)
2022-11-23 13:41:52 -07:00
dcce53ddcb
[cff] Micro-optimize fetch_op
2022-11-23 13:20:48 -07:00
0bf7d9eb4d
[subset-cff] Micro-optimize encode_byte
2022-11-23 13:20:48 -07:00
a23f820427
[subset-cff] Micro-optimize array access
2022-11-23 13:20:48 -07:00
1e8f1ac677
[subset-glyf] Micro-optimize array access
2022-11-23 13:20:48 -07:00
356d135ed6
[subset-glyf] Reduce roundf calls
...
Saves 7% on MPLUS1-Variable/6000 benchmark.
2022-11-23 13:20:48 -07:00
463ae07e99
[subset] In the preprocess subset call always use long loca.
...
Long loca is needed so that we can store the trimmed glyph bytes to allow us to safely skip trimming in the later subset.
2022-11-23 12:52:55 -07:00
299ec902eb
[glyf] Move instanciation to serialize()
2022-11-22 16:03:02 -07:00
d8d881f22d
[subset-glyf] Don't create a second glyf accelerator
2022-11-22 15:24:16 -07:00
7560930070
[glyf] Add _create_font_for_instancing
2022-11-22 15:22:25 -07:00
40634ceeb0
[glyf] Adjust data types
2022-11-22 15:07:16 -07:00
ba0d28ea36
[glyf] Fix font error check
2022-11-22 15:01:48 -07:00
04c525019c
[glyf] Use a malloc instead of calloc
2022-11-22 14:54:55 -07:00
024aa81805
[glyf] Micro-optimize encode_coord
2022-11-22 14:51:42 -07:00
b6694597f9
[glyf] Micro-optimize encode_flag()
2022-11-22 14:49:01 -07:00
d47cfe7936
[glyf] Minor use operator ++
2022-11-22 14:45:04 -07:00
03e6bde790
[glyf] Minor adjustment to lastflag handling
...
No logic change.
2022-11-22 14:40:50 -07:00
0ca9fda889
[glyf] Remove misplaced comment
2022-11-22 14:39:10 -07:00
44be8ef4ce
[gvar] Skip degenerate all-untouched delta-sets
2022-11-22 14:21:25 -07:00
656bb223f1
[gvar] Micro-optimize unpack_points
2022-11-22 14:16:40 -07:00
a383027262
[gvar] Cosmetic
2022-11-22 14:15:17 -07:00
be89919a70
[gvar] is_valid() remove a check
...
I don't know why this check was there, but it doesn't make sense
because that function never returns 0 / false.
2022-11-22 13:56:20 -07:00
e8ddf107d0
[gvar] Optimize a loop
2022-11-22 13:29:32 -07:00
dd6fcec92c
[gvar] Remove a conditional
2022-11-22 13:27:40 -07:00
a02317238a
[gvar] Refactor deltas array access
2022-11-22 13:24:39 -07:00
c34c77698c
[gvar] Don't try IUP if all points are specified
2022-11-22 13:21:01 -07:00
27c4037e59
[gvar] Micro-optimize boundary-checking
2022-11-22 13:12:22 -07:00
ab8346fb6f
[gvar] Add an unlikely
2022-11-22 13:07:39 -07:00
1e8a342ea2
[gvar] Micro-optimize int types
2022-11-22 13:04:32 -07:00
4afcdf675b
More hb_memcpy
2022-11-22 12:56:48 -07:00
58a696d80e
More hb_memset
2022-11-22 12:56:05 -07:00
59c45f6deb
Use hb_memcpy instead of memcpy consistently
2022-11-22 12:54:50 -07:00
ac0efaf818
Use hb_memset instead of memset consistently
2022-11-22 12:50:36 -07:00
44a892a233
[shape] Use hb_memcmp instead of memcmp
2022-11-22 12:48:52 -07:00
c53c648127
[subset-cff] Another handrolled memcpy
2022-11-22 12:46:25 -07:00
ae578705c2
[array] Write hash as range for loop again
...
Now that our range loop is faster than our own iter.
2022-11-22 12:23:17 -07:00
13e1ca9eb5
[cff] Micro-optimize memcpy
2022-11-22 12:19:28 -07:00
Behdad Esfahbod
Garret Rieger