52b6e0baa0
When serializing cmap14 order the offsets from smallest to largest.
...
Current versions of OTS fail fonts with cmap 14's who's last offset does not point to the a block at the end of the table.
2020-02-26 11:09:54 -08:00
a99134c5be
add oss-fuzz 20886 test file
2020-02-26 09:58:03 -08:00
d0aaba5c50
fixes oss-fuzz 20886
...
hb_set_t::resize () is needed after compact()
2020-02-26 09:35:32 -08:00
05a25c1a5b
[cff] minor, remove unused fields
2020-02-26 19:35:27 +03:30
9fe0dc3464
[draw] Pass draw_helper_t itself around instead recreating it
...
Specially helpful if we want to change the design
2020-02-26 17:40:46 +03:30
1b8b863898
minor
2020-02-26 16:36:48 +03:30
4cdaa9d1f4
[glyf] Simplify contour end logic
...
So no need for infinite loop here
2020-02-26 16:29:14 +03:30
132fcfbc47
[fuzz] minor don't abort main.cc when the file was empty or not found
2020-02-26 16:15:17 +03:30
84163c83d3
[draw] Skip commands and paths not contributing anything
...
They aren't contributing to rendering and making issue for stroking, let's skip them
ourselves as Skia does also https://skia-review.googlesource.com/c/skia/+/268166
They are useful for extracting extents and so which that functionality won't be effected by this change.
2020-02-26 16:09:28 +03:30
073d4954e0
[draw] Port glyf path extract to draw_helper_t
2020-02-26 15:40:40 +03:30
0ebf3a4e62
[draw] Move common CFF path building logic to draw_helper_t
2020-02-26 15:15:22 +03:30
c400cb8863
Re-implement hb_set_t::del_range ( #2194 )
...
* optimize hb_set_del_range()
fix issue #2193
* fixed bug & added tests
* coding & comment tweaks
2020-02-25 16:06:03 -08:00
c21eb86bfd
Merge pull request #2163 from harfbuzz/absolute-link
...
added add_link_abs()
2020-02-26 02:54:38 +03:30
4081439d2a
tweak reflecting review & add test cases
2020-02-25 15:03:12 -08:00
152000d9c7
[fuzz] Practice variations on font object
2020-02-25 21:16:57 +03:30
036d868913
[draw] Add a fuzzer
...
Specially checks correctness of the API semantics:
* no move happens when a path is already opened with move-to.
* no path will be left open and close-path will happen at the end of opened paths.
* no path opens with a move-to and will be closed with no length.
* paths start and ending points matches.
* no line/quadratic/cubic command will be issued when no path is started.
2020-02-25 19:09:44 +03:30
de896278f7
coding & comment tweaks
2020-02-25 07:12:20 -08:00
2f97aa65e5
[cff] Make sure previous is ended on processing a seac
2020-02-25 18:07:44 +03:30
acc2d4738e
[cff] Make path extract easier to read and more defensive
2020-02-25 18:05:48 +03:30
b59eb54f70
[glyf] Refactor, move get_points inside Glyph
2020-02-25 12:56:32 +03:30
173b745da8
fixed bug & added tests
2020-02-24 22:56:57 -08:00
a5012e97c4
optimize hb_set_del_range()
...
fix issue #2193
2020-02-24 17:09:48 -08:00
96b71e802f
[fuzz] make the custom loader to handle multiple files
...
Actually this was the way it used to work :)
2020-02-24 23:01:02 +03:30
1f5a54c768
[gvar] fix infinite loop introduced by 11f3fca
...
The attempt on removing end_points had made the code unreadable
and has intrdouced infinite, fixed by making the code clear what
it tries to achieve.
2020-02-24 14:30:07 +03:30
f00eb4ebfa
[gvar] Don't compare against Null address
2020-02-24 09:25:19 +03:30
19b78d56cd
[gvar] Accept coord_count even if is higher than gvar's axisCount
2020-02-24 09:25:18 +03:30
197e2e929b
[gvar] Return gracefully even if gvar wasn't used at all
2020-02-24 08:55:35 +03:30
346ab9884b
[gvar] Return gracefully even if coords_count != gvar's axis count
...
It shouldn't be that common to have the situation, even so, no need to
reject entire a glyph even if gvar isn't used.
2020-02-24 08:48:54 +03:30
5f23a76e99
[cff] Don't remove coords const qualification
...
We should find some way to audit all the const removals, perhaps by
enabling -Wold-style-cast warning on the project and turning all the
implicit const removals to const_cast.
2020-02-24 08:23:25 +03:30
ff5223ba60
remove invalid glyphs from closure ( #2188 )
...
fix issue #2186
2020-02-23 15:53:21 -08:00
e17fd0d91c
[tools] More on py3 compatibility
2020-02-24 00:10:11 +03:30
0cf050a7b1
[draw] Merge consequent move-to commands of CFF/CFF2
2020-02-23 23:23:50 +03:30
86bd5a0ba1
[draw] End CFF paths with a line-to
...
Issue a line-to command when start and end point of a path isn't same, matches freetype also.
2020-02-23 23:23:50 +03:30
a46ba770ab
[number] minor
2020-02-22 17:15:06 +03:30
4aa354beba
chain second OffsetTo::serialize_copy() to first
2020-02-21 10:17:26 -08:00
6120f50c8f
fix Head with link.bias
2020-02-21 10:17:26 -08:00
e65bff0943
workaround whence_t as bit-field
2020-02-21 10:17:26 -08:00
1aed6d223c
fix whence=Head; position as 28 bits
2020-02-21 10:17:25 -08:00
4ca8e0d989
re-implement Tail; rewrote name table with it
2020-02-21 10:17:25 -08:00
c1313e4be2
minor: undid unintended change
2020-02-21 10:17:25 -08:00
9c65680b48
rename enums to avoid conflict on win
2020-02-21 10:17:25 -08:00
555f6f1daa
merge add_link_abs() into add_link()
2020-02-21 10:17:25 -08:00
017cd945bf
added add_link_abs()
2020-02-21 10:17:25 -08:00
07504569d5
[gsub] minor format
2020-02-21 13:05:44 +03:30
411225426f
Fix glyph closure for alternate substitution.
...
It was not filtering on the glyphs to be retained.
2020-02-21 13:01:23 +03:30
da37880e17
[draw] Add a private bit field instead null-checking quad-to callback
...
This doesn't imply the the design is good or stable just is better
to not fail due to null quad-to callback so as our quad-cubic curves
translation is currently table limited which can be revisited anytime.
2020-02-20 19:04:20 +03:30
79b2b8a91f
minor
2020-02-19 16:41:13 +03:30
8c652f72fc
Minor, switch to https links where possible
2020-02-19 16:32:44 +03:30
bbcbcafc25
[tool] Minor, move input files link
2020-02-19 16:21:47 +03:30
83db1e117e
[ci] Compatibility with Python 3
...
Disabled CMake test runners as it goes for Python 2 and updated the bots
Install Python 3 on macOS 10.12 bot
2020-02-19 16:17:45 +03:30
Garret Rieger
ariza
Ebrahim Byagowi