|
2bcbfd35
|
2025-02-07T17:58:13
|
|
[hvgl] Indent
|
|
|
fc0966d9
|
2025-02-07T17:53:49
|
|
[hvgl] Fix the complex-number transform code path
Tested by U+9DAD at wght=900 of PingFangUI.
|
|
|
32a6940d
|
2025-02-07T16:15:53
|
|
[hvgl] Minor const
|
|
|
9685d393
|
2025-02-07T16:57:19
|
|
[hvgl] Another clamp
Huh. Errors down to just 1!
|
|
|
3c84aa52
|
2025-02-07T13:52:32
|
|
[hvgl] Move transforms around
This doesn't seem to quite match the spec. But produces fewer wrong
shapes. Down to 834 out of 34485 characters.
|
|
|
3e1e2ab3
|
2025-02-07T13:45:18
|
|
[hvgl] Another transform fix
|
|
|
d4511648
|
2025-02-07T17:35:26
|
|
[draw] Simplify quadratic-to-cubic conversion
Hopefully introduces less error.
|
|
|
40b9930c
|
2025-02-07T16:46:49
|
|
[hvgl] Minor move condition
Previous code was safe but was taking address of an out-of-range
array member if end was 0. Was never dereferenced though.
|
|
|
7e84cf81
|
2025-02-07T13:23:01
|
|
[hvgl] Simplify transform
|
|
|
a71ad886
|
2025-02-07T16:42:42
|
|
[hvgl] Rename a variable
|
|
|
d5174c41
|
2025-02-07T13:13:02
|
|
[hvgl] Fix transform order
|
|
|
76dbae15
|
2025-02-07T11:34:33
|
|
[hvgl] Minor
|
|
|
453aea05
|
2025-03-02T16:59:19
|
|
[hvgl] Use double instead of float for coords & transforms
As "spec" requires.
|
|
|
00bc3ecf
|
2025-03-02T16:58:37
|
|
[geometry] Use int literals instead of float ones
In anticipation of templatizing on float type.
|
|
|
9f85e974
|
2025-01-30T17:58:28
|
|
[hvgl] Remove a resize(0) that was not doing any good
In hb_vector_t<double> we use hb_realloc() for resizing, which
doesn't care what the vector current length is.
|
|
|
dbef4e32
|
2025-01-30T14:54:41
|
|
[hvgl] Name a value
|
|
|
da3fca46
|
2025-01-30T08:41:31
|
|
[hvgl] Reduce another malloc
|
|
|
44780db5
|
2025-01-30T08:31:48
|
|
[hvgl] Remove unused parameter
|
|
|
9579cee8
|
2025-01-30T08:26:16
|
|
[hvgl] Add table version check
|
|
|
d4b29817
|
2025-01-30T08:01:56
|
|
[hvgl] Remove use of a transient variable
|
|
|
9b361172
|
2025-01-30T07:58:21
|
|
[hvgl] Micro-optimize vector allocation
|
|
|
a2fc9523
|
2025-01-30T07:51:10
|
|
[hvgl] Use a common "scratch" vector
To avoid reallocation.
|
|
|
de0e1bdd
|
2025-01-30T07:30:51
|
|
[hvgl] Minor simplify sanitize
|
|
|
4e56a0b8
|
2025-01-30T06:54:59
|
|
[hvgl] Use smarter algorithm for cycle detection
Great speedup as we don't use hb_set_t. Yay!
|
|
|
cd73ca3c
|
2025-01-30T06:43:51
|
|
[hvgl] Possibly save a malloc
|
|
|
7a171dcc
|
2025-01-30T06:42:35
|
|
[hvgl] Simplify resize call
|
|
|
60562fce
|
2025-01-30T05:09:35
|
|
[hvgl] Remove an unnecessary cast
|
|
|
26fc2d7c
|
2025-01-30T05:08:24
|
|
[hvgl] Minor var type declaration
|
|
|
fc62af31
|
2025-01-30T05:03:34
|
|
[hvgl] Bail on error
Easier than handling a vector too short.
|
|
|
2a3477c8
|
2025-01-30T04:59:49
|
|
[hvgl] Micro-optimize
|
|
|
d9d58958
|
2025-01-30T04:46:08
|
|
[geometry] Minor avoid a default initialization
Compiler probably doesn't care though.
|
|
|
7d3b340d
|
2025-01-30T04:26:31
|
|
[hvgl] Grammar
|
|
|
241304b2
|
2025-01-30T04:25:54
|
|
[hvgl] Micro-optimize
|
|
|
460f0f19
|
2025-01-30T04:18:36
|
|
[hvgl] Cosmetic
|
|
|
b191d084
|
2025-01-30T04:16:56
|
|
[hvgl] Micro-optimize
|
|
|
500290bb
|
2025-01-30T03:54:54
|
|
[hvgl] Fix typo
|
|
|
3bdae43a
|
2025-01-30T01:23:21
|
|
[hvgl] Add a barrier
Just in case.
|
|
|
c52c98ab
|
2025-01-28T03:59:39
|
|
[hvgl] Remove a duplicate variable
|
|
|
0a25471b
|
2025-01-28T03:49:59
|
|
[hvgl] Fix vector resize
I meant to resize to exact, but the false wasn't doing that.
|
|
|
4bced206
|
2025-01-28T03:06:00
|
|
[hvgl] Rename a method
|
|
|
ea50fa28
|
2025-01-28T02:11:31
|
|
[hvgl] Remove commented out leftover
|
|
|
8183ab13
|
2025-01-22T05:35:09
|
|
[hvgl] Apply shape deltas
|
|
|
de87cea1
|
2025-01-22T05:34:55
|
|
[hvgl] Minor bound checking
|
|
|
5fa2c628
|
2025-01-28T01:07:40
|
|
[hvgl] Avoid a div-by-zero
|
|
|
37a8e4af
|
2025-01-22T05:34:30
|
|
[algs] Minor tweak to HBFloat
|
|
|
c6bbc4ed
|
2025-01-22T05:33:33
|
|
[open-type] Fix NumType
|
|
|
1b42d811
|
2025-01-28T00:36:34
|
|
[hvgl] Micro-optimize segment access
|
|
|
9f78bae8
|
2025-01-22T04:38:14
|
|
[hvgl] Draw PartShape
With no variations so far.
|
|
|
caf04058
|
2025-01-22T04:37:27
|
|
[hvgl] Fix NumType float conversion
|
|
|
3a6dc3d2
|
2025-01-28T00:27:03
|
|
[hvgl] Micro-optimize
|
|
|
469c59e7
|
2025-01-21T21:43:47
|
|
[hvgl] Start processing PartShape outline
|
|
|
fe964860
|
2025-01-21T21:41:51
|
|
[hvgl] Fix Align<>::get_size()
|
|
|
6e92267c
|
2025-01-28T00:06:48
|
|
[hvgl] Only pass the needed coords around
|
|
|
df890f69
|
2025-01-21T20:06:07
|
|
[hvgl] Use UnsizedArrayOf for a couple of types
|
|
|
e9c531dc
|
2025-01-20T18:02:38
|
|
[hvgl] Simplify Align<>
|
|
|
80a8c5fa
|
2025-01-28T00:00:00
|
|
[hvgl] Whitespace
|
|
|
a6ba9603
|
2025-01-20T18:00:23
|
|
[hvgl] Move Align to proper place
|
|
|
14a75335
|
2025-01-20T17:46:44
|
|
[hvgl] Traverse down the tree
Still no path fetching.
|
|
|
4fc69418
|
2025-01-27T23:43:39
|
|
[hvgl] Rename "be" var to "le"
|
|
|
386ea1ba
|
2025-01-20T17:23:51
|
|
[hvgl] Fix sanitize
|
|
|
8ff4cfbc
|
2025-01-19T16:55:25
|
|
[hvgl] Fix null size of Part
|
|
|
303cd634
|
2025-01-27T23:14:09
|
|
[hvgl] Disable unnecessary code
|
|
|
9c788d12
|
2025-01-27T21:17:38
|
|
[hvgl] Comment
|
|
|
3d0fae3f
|
2025-01-27T19:29:47
|
|
[hvgl] Optimize initial getting of coordinates
|
|
|
f31fa0ae
|
2025-01-27T14:52:19
|
|
[hvgl] Optimize extremum transforms some more
Passed the 2x threshold, faster than CoreText.
|
|
|
e7a05b1b
|
2025-01-27T14:48:40
|
|
[hvgl] Micro-optimize
|
|
|
4ba06925
|
2025-01-27T14:42:17
|
|
[hvgl] Minor use a variable
|
|
|
727b647e
|
2025-01-27T14:37:36
|
|
[hvgl] Speed up transform variations
|
|
|
be30bf9d
|
2025-01-27T14:22:31
|
|
[hvgl] Simplify master transforms a bit more
|
|
|
93ddf8f8
|
2025-01-27T14:01:04
|
|
[hvgl] Move variables around
|
|
|
49fd66eb
|
2025-01-27T13:58:59
|
|
[hvgl] Split apply_transforms into two loops
The paper says walk them together, but the master can be applied
separately before the extremums.
|
|
|
19114e0c
|
2025-01-27T13:56:43
|
|
[hvgl] Micro-optimize transforms
Measurable speedup.
|
|
|
b7a9f15d
|
2025-01-27T13:52:12
|
|
[hvgl] Simplify transforms
Speeds up a bit too.
|
|
|
3a8d6d60
|
2025-01-27T13:41:07
|
|
[hvgl] Iterate all four iterators together
Instead of trying every row, sparsely process non-trivial rows.
|
|
|
a1fba4eb
|
2025-01-27T12:52:03
|
|
[hvgl] Micro-optimize
|
|
|
8f505a70
|
2025-01-27T11:40:09
|
|
[hvgl] Speed up simd path
Only process coord!=0 columns.
For some reason I don't see any speedup :-(.
|
|
|
c111f088
|
2025-01-26T23:27:14
|
|
[hvgl] Simplify sanitize more
|
|
|
d4600d8d
|
2025-01-26T23:24:50
|
|
[hvgl] Simplify more sanitize
|
|
|
41ee17e8
|
2025-01-26T23:19:43
|
|
[hvgl] Simplify sanitize
|
|
4f3c3ee0
|
2025-01-26T22:58:49
|
|
[hvgl/simd] Only enable if little-endian
|
|
|
8e7eeb5f
|
2025-01-26T22:54:52
|
|
Revert "[hvgl] Implement AVX"
This reverts commit 82ac479cbb49a88946025be2b84315bfd6091d24.
|
|
|
e7ecc8e5
|
2025-01-26T22:52:29
|
|
[hvgl] Implement AVX
Shows like only 2% / 3% speedup. Going to revert.
|
|
|
8bd24679
|
2025-01-26T22:15:16
|
|
[hvgl] Minor macro def
|
|
|
c0f88772
|
2025-01-26T21:11:37
|
|
[hvgl] Micro-optimize
|
|
|
50137437
|
2025-01-26T20:29:48
|
|
[hvgl] Micro-optimize
|
|
|
1e8c858e
|
2025-01-26T20:22:39
|
|
[hvgl] Rewrite a loop
No functionality change.
|
|
|
07a97adb
|
2025-01-26T20:19:07
|
|
[hvgl] Remove an unnecessary check
It's a rare case; no need to check it first.
|
|
|
6da4fb10
|
2025-01-26T19:13:43
|
|
[config] Add HB_NO_SIMD
|
|
|
2ba20cfe
|
2025-01-26T19:05:19
|
|
[hvgl] Micro-optimization leftover
|
|
|
76017d56
|
2025-01-26T18:03:04
|
|
[hvgl/simd] Remove unaligned SIMD codepath
The generic path is faster it seems.
|
|
|
c5e3fb5d
|
2025-01-26T17:57:06
|
|
[hvgl/simd] Reduce alignment requirement
Speeds up a bit.
|
|
|
587b1677
|
2025-01-26T17:28:08
|
|
[hvgl] Implement Apple SIMD 4x4 matrix usage
Shows around 7% improvement only.
|
|
|
6252d435
|
2025-01-26T14:32:24
|
|
[hvgl] Move code
|
|
|
640d778b
|
2025-01-25T19:56:04
|
|
[hvgl] Minor micro-optimize
|
|
|
93f0dd97
|
2025-01-25T12:44:42
|
|
[hvgl] Optimize away complex math
Most of the time rotation is zero. It's the same test.
|
|
|
25ed2c19
|
2025-01-25T12:33:45
|
|
[hvgl] Micro-optimize some more
|
|
|
8e210127
|
2025-01-25T12:29:17
|
|
[hvgl] Micro-optimize segment access
|
|
|
575a831b
|
2025-01-25T12:16:19
|
|
[hvgl] Micro micro optimize
|
|
|
9c99c292
|
2025-01-25T12:05:25
|
|
[hvgl] Micro-optimize apply_transforms()
|
|
|
7b724814
|
2025-01-25T12:01:53
|
|
[hvgl] Speed up hot loop
|