platform/upstream/libHarfBuzzSharp.git
5 years ago[atomic] Remove hb_atomic_int_impl_t now that it's always int
Behdad Esfahbod [Thu, 9 Aug 2018 05:51:35 +0000 (22:51 -0700)]
[atomic] Remove hb_atomic_int_impl_t now that it's always int

5 years ago[atomic] Unify typedef
Behdad Esfahbod [Thu, 9 Aug 2018 05:45:49 +0000 (22:45 -0700)]
[atomic] Unify typedef

Removes volatile from fallback implementation.  That was handwavy
anyway.

5 years ago[atomic] Change Mac atomic int type to int
Behdad Esfahbod [Thu, 9 Aug 2018 05:43:53 +0000 (22:43 -0700)]
[atomic] Change Mac atomic int type to int

If ever int is not 32bit, we'll fix this...

5 years ago[atomic] Change Windows atomic int type to int
Behdad Esfahbod [Thu, 9 Aug 2018 05:42:33 +0000 (22:42 -0700)]
[atomic] Change Windows atomic int type to int

5 years ago[atomic] Change Solaris atomic int to signed
Behdad Esfahbod [Thu, 9 Aug 2018 05:40:10 +0000 (22:40 -0700)]
[atomic] Change Solaris atomic int to signed

5 years ago1.8.7
Behdad Esfahbod [Thu, 9 Aug 2018 04:54:08 +0000 (21:54 -0700)]
1.8.7

5 years agoFix previous commit to use atomic operations
Behdad Esfahbod [Thu, 9 Aug 2018 05:01:04 +0000 (22:01 -0700)]
Fix previous commit to use atomic operations

5 years agoFix assertion failure
Behdad Esfahbod [Thu, 9 Aug 2018 04:42:18 +0000 (21:42 -0700)]
Fix assertion failure

https://bugzilla.mozilla.org/show_bug.cgi?id=1476334

5 years ago1.8.6
Behdad Esfahbod [Tue, 7 Aug 2018 16:55:42 +0000 (09:55 -0700)]
1.8.6

5 years agoAdd +hb_font_get_glyph_advances_for_direction
Behdad Esfahbod [Tue, 7 Aug 2018 16:52:06 +0000 (09:52 -0700)]
Add +hb_font_get_glyph_advances_for_direction

New API:
+hb_font_get_glyph_advances_for_direction

5 years agoActually add hb_font_get_glyph_[hv]_advances
Behdad Esfahbod [Tue, 7 Aug 2018 16:47:00 +0000 (09:47 -0700)]
Actually add hb_font_get_glyph_[hv]_advances

New API:
+hb_font_get_glyph_h_advances
+hb_font_get_glyph_v_advances

5 years agoMinor
Behdad Esfahbod [Tue, 7 Aug 2018 16:44:01 +0000 (09:44 -0700)]
Minor

5 years agoRemove get_advances_nil and alias it to get_advances_default
Behdad Esfahbod [Tue, 7 Aug 2018 16:37:52 +0000 (09:37 -0700)]
Remove get_advances_nil and alias it to get_advances_default

5 years ago[iter] Make operator bool explicit
Behdad Esfahbod [Mon, 6 Aug 2018 22:32:13 +0000 (15:32 -0700)]
[iter] Make operator bool explicit

We would need to write:

  for (Iter<T> it (...); bool (t); t++)

instead of:

  for (Iter<T> it (...); t; t++)

But I think it's an improvement in code readability in the long term.

Fixes https://github.com/harfbuzz/harfbuzz/issues/1127

5 years agoAdd batch advance width callback function
Behdad Esfahbod [Mon, 6 Aug 2018 16:45:17 +0000 (09:45 -0700)]
Add batch advance width callback function

New API:
+hb_font_funcs_set_glyph_h_advances_func
+hb_font_funcs_set_glyph_v_advances_func
+hb_font_get_glyph_h_advances
+hb_font_get_glyph_h_advances_func_t
+hb_font_get_glyph_v_advances
+hb_font_get_glyph_v_advances_func_t

5 years agobatchwidth
Koji Ishii [Wed, 1 Aug 2018 04:01:08 +0000 (13:01 +0900)]
batchwidth

5 years agoRemove most uses of direct comparison to Null objects
Behdad Esfahbod [Mon, 6 Aug 2018 13:30:12 +0000 (06:30 -0700)]
Remove most uses of direct comparison to Null objects

5 years agoPort _nil objects to Null() machinery
Behdad Esfahbod [Mon, 6 Aug 2018 13:17:48 +0000 (06:17 -0700)]
Port _nil objects to Null() machinery

Finally, unified!

5 years ago[BASE] Rename horzi/vert to h/v
Behdad Esfahbod [Mon, 6 Aug 2018 12:53:35 +0000 (05:53 -0700)]
[BASE] Rename horzi/vert to h/v

5 years ago[BASE] Misc fixes
Behdad Esfahbod [Mon, 6 Aug 2018 12:52:12 +0000 (05:52 -0700)]
[BASE] Misc fixes

The code was badly broken.  In better shape now, but still, needs
a full review before ever working.

5 years agoChange null-pool specialization to min_size again
Behdad Esfahbod [Mon, 6 Aug 2018 12:29:15 +0000 (05:29 -0700)]
Change null-pool specialization to min_size again

5 years agoMove Null system to hb-null.hh
Behdad Esfahbod [Mon, 6 Aug 2018 12:01:52 +0000 (05:01 -0700)]
Move Null system to hb-null.hh

5 years agoRemove unused macros
Behdad Esfahbod [Mon, 6 Aug 2018 11:58:34 +0000 (04:58 -0700)]
Remove unused macros

5 years agoMinor
Behdad Esfahbod [Mon, 6 Aug 2018 11:54:31 +0000 (04:54 -0700)]
Minor

5 years agoMinor
Behdad Esfahbod [Mon, 6 Aug 2018 11:51:38 +0000 (04:51 -0700)]
Minor

5 years agoMove hb_vector_t to hb-vector-private.hh
Behdad Esfahbod [Mon, 6 Aug 2018 11:42:46 +0000 (04:42 -0700)]
Move hb_vector_t to hb-vector-private.hh

5 years agoMove some more code around
Behdad Esfahbod [Mon, 6 Aug 2018 11:32:51 +0000 (04:32 -0700)]
Move some more code around

5 years agoMove null data definitions to hb-static.cc
Behdad Esfahbod [Mon, 6 Aug 2018 04:41:52 +0000 (21:41 -0700)]
Move null data definitions to hb-static.cc

Also remove "    " null data for Tag.  Just use zeroes.

5 years agoMinor
Behdad Esfahbod [Sat, 4 Aug 2018 02:55:09 +0000 (19:55 -0700)]
Minor

5 years agoMake lazy loader deal with OOM
Behdad Esfahbod [Sat, 4 Aug 2018 02:54:33 +0000 (19:54 -0700)]
Make lazy loader deal with OOM

5 years agoRename
Behdad Esfahbod [Sat, 4 Aug 2018 00:32:00 +0000 (17:32 -0700)]
Rename

5 years ago[ot] Comment
Behdad Esfahbod [Sat, 4 Aug 2018 00:06:42 +0000 (17:06 -0700)]
[ot] Comment

5 years ago[ot] Add misc tables to tables_t
Behdad Esfahbod [Sat, 4 Aug 2018 00:03:32 +0000 (17:03 -0700)]
[ot] Add misc tables to tables_t

Most unused.

5 years ago[ot] Move GDEF into tables_t
Behdad Esfahbod [Fri, 3 Aug 2018 23:57:40 +0000 (16:57 -0700)]
[ot] Move GDEF into tables_t

5 years ago[ot] Move GSUB/GPOS into tables_t
Behdad Esfahbod [Fri, 3 Aug 2018 23:45:57 +0000 (16:45 -0700)]
[ot] Move GSUB/GPOS into tables_t

5 years ago[ot] More shuffling
Behdad Esfahbod [Fri, 3 Aug 2018 23:34:25 +0000 (16:34 -0700)]
[ot] More shuffling

Misc. table loading now properly automated.

5 years ago[ot] Minor
Behdad Esfahbod [Fri, 3 Aug 2018 23:22:16 +0000 (16:22 -0700)]
[ot] Minor

5 years ago[ot] Move code around
Behdad Esfahbod [Fri, 3 Aug 2018 06:13:19 +0000 (23:13 -0700)]
[ot] Move code around

5 years ago[ot] Use switch for checking for blacklisted GDEF tables
Behdad Esfahbod [Fri, 3 Aug 2018 06:04:42 +0000 (23:04 -0700)]
[ot] Use switch for checking for blacklisted GDEF tables

Faster / more compact code.

5 years ago[ot] In GDEF blacklisting logic, remove the glyph index check for Times
Behdad Esfahbod [Fri, 3 Aug 2018 05:56:42 +0000 (22:56 -0700)]
[ot] In GDEF blacklisting logic, remove the glyph index check for Times

Just blacklist based on table sizes like we do for other fonts.

5 years agoMove code around
Behdad Esfahbod [Fri, 3 Aug 2018 03:04:37 +0000 (20:04 -0700)]
Move code around

5 years agoMinor
Behdad Esfahbod [Thu, 2 Aug 2018 09:22:38 +0000 (02:22 -0700)]
Minor

5 years agoFix for recent rename
Behdad Esfahbod [Thu, 2 Aug 2018 09:04:02 +0000 (02:04 -0700)]
Fix for recent rename

5 years ago[uniscribe] Update for recent changes
Behdad Esfahbod [Thu, 2 Aug 2018 09:03:13 +0000 (02:03 -0700)]
[uniscribe] Update for recent changes

5 years agoRemove unused table reference
Behdad Esfahbod [Thu, 2 Aug 2018 08:44:20 +0000 (01:44 -0700)]
Remove unused table reference

5 years agoMinor
Behdad Esfahbod [Thu, 2 Aug 2018 08:41:19 +0000 (01:41 -0700)]
Minor

5 years agoMake lazy-loader safe for nil objectification
Behdad Esfahbod [Thu, 2 Aug 2018 08:37:57 +0000 (01:37 -0700)]
Make lazy-loader safe for nil objectification

5 years agoMinor
Behdad Esfahbod [Thu, 2 Aug 2018 08:36:08 +0000 (01:36 -0700)]
Minor

5 years agoReduce storage by sharing face amongst lazy_loaders
Behdad Esfahbod [Thu, 2 Aug 2018 08:27:40 +0000 (01:27 -0700)]
Reduce storage by sharing face amongst lazy_loaders

5 years agoInternal templatization of lazy-loaders
Behdad Esfahbod [Thu, 2 Aug 2018 07:38:46 +0000 (00:38 -0700)]
Internal templatization of lazy-loaders

5 years agoRemove OT namespace from hb-machinery-private.hh
Behdad Esfahbod [Thu, 2 Aug 2018 06:59:09 +0000 (23:59 -0700)]
Remove OT namespace from hb-machinery-private.hh

5 years agoMinor
Behdad Esfahbod [Thu, 2 Aug 2018 05:50:45 +0000 (22:50 -0700)]
Minor

5 years ago[face] Sprinkle const in the API
Behdad Esfahbod [Thu, 2 Aug 2018 05:50:06 +0000 (22:50 -0700)]
[face] Sprinkle const in the API

5 years agoRename
Behdad Esfahbod [Thu, 2 Aug 2018 01:03:32 +0000 (18:03 -0700)]
Rename

5 years agoAdd comment re (our only) race condition
Behdad Esfahbod [Thu, 2 Aug 2018 00:53:30 +0000 (17:53 -0700)]
Add comment re (our only) race condition

5 years ago1.8.5
Behdad Esfahbod [Wed, 1 Aug 2018 21:51:51 +0000 (14:51 -0700)]
1.8.5

5 years ago[atomic] Fix Solaris ones to add proper barriers
Behdad Esfahbod [Wed, 1 Aug 2018 21:13:59 +0000 (14:13 -0700)]
[atomic] Fix Solaris ones to add proper barriers

5 years ago[atomic] Remove volatile from IBM impl signature
Behdad Esfahbod [Wed, 1 Aug 2018 21:02:39 +0000 (14:02 -0700)]
[atomic] Remove volatile from IBM impl signature

5 years ago[atomic] Oops
Behdad Esfahbod [Wed, 1 Aug 2018 21:00:46 +0000 (14:00 -0700)]
[atomic] Oops

5 years ago[atomic] Use read-barrier for get()
Behdad Esfahbod [Wed, 1 Aug 2018 20:59:59 +0000 (13:59 -0700)]
[atomic] Use read-barrier for get()

5 years ago[atomic] Some more minor tweaks
Behdad Esfahbod [Wed, 1 Aug 2018 20:59:31 +0000 (13:59 -0700)]
[atomic] Some more minor tweaks

5 years ago[ci] Fix Appveyor bot (#1123)
Ebrahim Byagowi [Wed, 1 Aug 2018 19:41:43 +0000 (00:11 +0430)]
[ci] Fix Appveyor bot (#1123)

5 years agoRevert "Add a new API, hb_ot_layout_get_feature_name_ids (#976)" (#1121)
Ebrahim Byagowi [Wed, 1 Aug 2018 07:45:44 +0000 (12:15 +0430)]
Revert "Add a new API, hb_ot_layout_get_feature_name_ids (#976)" (#1121)

This reverts commit 0c1b287b72e91e0898d75acb5d5acf1c6b9a7498.

5 years ago[atomic] Fix cast to fallback ptr_get()
Behdad Esfahbod [Wed, 1 Aug 2018 07:22:18 +0000 (00:22 -0700)]
[atomic] Fix cast to fallback ptr_get()

5 years ago[gobject] Fix copy/paste error
Behdad Esfahbod [Wed, 1 Aug 2018 07:19:25 +0000 (00:19 -0700)]
[gobject] Fix copy/paste error

5 years ago[atomic] On IBM, use light-weight sync for everything
Behdad Esfahbod [Wed, 1 Aug 2018 06:01:05 +0000 (23:01 -0700)]
[atomic] On IBM, use light-weight sync for everything

lwsync() is a full read/write-barrier.  That's all we need, never
need sync().  I'm not sure why an isync() was used in fetch_and_add,
but since that's a read-modify-write, I just changed it to have
lwsync() on both sides.

5 years ago[atomic] Add XXX items around Solaris ops
Behdad Esfahbod [Wed, 1 Aug 2018 06:00:15 +0000 (23:00 -0700)]
[atomic] Add XXX items around Solaris ops

Since add_int and cas are both read-modify-write, I wonder if we
also need a barrier after them.

5 years ago[atomic] Fix get() impl
Behdad Esfahbod [Wed, 1 Aug 2018 05:51:38 +0000 (22:51 -0700)]
[atomic] Fix get() impl

Originally, glib's atomic_get was implemented as "memory_barrier; load".
I copied this into cairo, fontconfig, and harfbuzz.  However, that's
wrong.  Correct way is "load; memory_barrier".  The details are long
and hard to fully grasp.  Best to read:

  https://www.kernel.org/doc/Documentation/memory-barriers.txt

Also see my report against GNOME:

  https://gitlab.gnome.org/GNOME/glib/issues/1449

Note that this is irrelevant if C++11-like atomic ops are available.

5 years ago[gobject] Hook up hb_map_t
Behdad Esfahbod [Wed, 1 Aug 2018 05:28:28 +0000 (22:28 -0700)]
[gobject] Hook up hb_map_t

5 years ago[atomic] Fix fallback impl
Behdad Esfahbod [Wed, 1 Aug 2018 05:21:21 +0000 (22:21 -0700)]
[atomic] Fix fallback impl

5 years ago[atomic] Add get_relaxed / set_relaxed
Behdad Esfahbod [Wed, 1 Aug 2018 04:05:51 +0000 (21:05 -0700)]
[atomic] Add get_relaxed / set_relaxed

To help TSan and be more "correct".

5 years ago[atomic] Use CONSUME, not ACQUIRE, memory-order for get()
Behdad Esfahbod [Wed, 1 Aug 2018 02:33:37 +0000 (19:33 -0700)]
[atomic] Use CONSUME, not ACQUIRE, memory-order for get()

Although, all implementations just elevate that to ACQUIRE.
But requirement for us is just CONSUME.

5 years agoRevert "[atomic] Make pointer get op relaxed instead of acquire"
Behdad Esfahbod [Wed, 1 Aug 2018 02:29:49 +0000 (19:29 -0700)]
Revert "[atomic] Make pointer get op relaxed instead of acquire"

This reverts commit b1e5650c67266dc158f22355fed206cd1c413f70.

After lots of head-scratching and finally finding the only truly
readable source to be the good old:

  https://www.kernel.org/doc/Documentation/memory-barriers.txt

I've convinced myself that we need consume memory-ordering on get().
The location of memory-barrier in a load should be after, not before
the load.  That needs fixing.  I'll do that separately.

5 years ago[subset] limit the max codepoint value to the unicode limit.
Garret Rieger [Wed, 1 Aug 2018 00:44:02 +0000 (17:44 -0700)]
[subset] limit the max codepoint value to the unicode limit.
When collecting all codepoints in the cmap avoid using large amount of memory for fonts that declare coverage over all 32 bit integers.

5 years ago[subset] Add hb_ot_layout_collect_features to harfbuzz-sections.txt.
Garret Rieger [Wed, 1 Aug 2018 00:59:19 +0000 (17:59 -0700)]
[subset] Add hb_ot_layout_collect_features to harfbuzz-sections.txt.
Add the fuzzer test case for feature collection timeout.

5 years ago[subset] use add_array to populate feature_indexes.
Garret Rieger [Tue, 31 Jul 2018 01:10:43 +0000 (18:10 -0700)]
[subset] use add_array to populate feature_indexes.
This is much faster then calling a bunch of individual add()'s.

5 years ago[subset] collect features first, then use those to collect lookups.
Garret Rieger [Tue, 31 Jul 2018 00:17:43 +0000 (17:17 -0700)]
[subset] collect features first, then use those to collect lookups.

5 years ago[indic] Flip default logic for double-halants in old-school
Behdad Esfahbod [Tue, 31 Jul 2018 22:27:29 +0000 (15:27 -0700)]
[indic] Flip default logic for double-halants in old-school

Oriya went down from 9 to 2.

BENGALI: 353725 out of 354188 tests passed. 463 failed (0.130722%)
DEVANAGARI: 707311 out of 707394 tests passed. 83 failed (0.0117332%)
GUJARATI: 366355 out of 366457 tests passed. 102 failed (0.0278341%)
GURMUKHI: 60729 out of 60747 tests passed. 18 failed (0.0296311%)
KANNADA: 951300 out of 951913 tests passed. 613 failed (0.0643966%)
MALAYALAM: 1048136 out of 1048334 tests passed. 198 failed (0.0188871%)
MYANMAR: 1115830 out of 1123883 tests passed. 8053 failed (0.716534%)
ORIYA: 42327 out of 42329 tests passed. 2 failed (0.00472489%)
SINHALA: 271596 out of 271847 tests passed. 251 failed (0.0923313%)
TAMIL: 1091754 out of 1091754 tests passed. 0 failed (0%)
TELUGU: 970555 out of 970573 tests passed. 18 failed (0.00185457%)

5 years ago[indic] Allow double-halant in old-spec Devanagari
Behdad Esfahbod [Tue, 31 Jul 2018 22:19:32 +0000 (15:19 -0700)]
[indic] Allow double-halant in old-spec Devanagari

Fixes https://github.com/harfbuzz/harfbuzz/issues/1071

5 years ago[test] Add Khmer tests, with NotoSansKhmer-Regular.ttf
Behdad Esfahbod [Tue, 31 Jul 2018 21:20:31 +0000 (14:20 -0700)]
[test] Add Khmer tests, with NotoSansKhmer-Regular.ttf

Note that there's minor positioning differences, and ONE reordering
difference between what we get for these and what Uniscribe gets.
Probably same as what's described in commit message for
1a96cc825dc9c8e3b6eef1403fe0864a1cfc0245

5 years ago[khmer] Typo
Behdad Esfahbod [Tue, 31 Jul 2018 21:10:31 +0000 (14:10 -0700)]
[khmer] Typo

5 years ago[test] Add Khmer test texts from recent bugs
Behdad Esfahbod [Tue, 31 Jul 2018 20:56:55 +0000 (13:56 -0700)]
[test] Add Khmer test texts from recent bugs

5 years ago[test] Move things around for shaper updates
Behdad Esfahbod [Tue, 31 Jul 2018 20:55:53 +0000 (13:55 -0700)]
[test] Move things around for shaper updates

5 years ago[khmer] Clear syllables before presentation features
Behdad Esfahbod [Tue, 31 Jul 2018 20:38:07 +0000 (13:38 -0700)]
[khmer] Clear syllables before presentation features

Probably not what Uniscribe does, but good idea?

5 years ago[khmer] Revert previous change, and remove pauses
Behdad Esfahbod [Tue, 31 Jul 2018 20:35:10 +0000 (13:35 -0700)]
[khmer] Revert previous change, and remove pauses

This makes test suite happy again (at 44) while fixing the sequences
we were fixing, which were the following with KhmerUI.ttf:

  U+1789,U+17BC
  U+1789,U+17D2,U+1789
  U+1789,U+17D2,U+1789,U+17BC

Fixes rest of https://github.com/harfbuzz/harfbuzz/issues/974

5 years ago[khmer] Apply ccmp after basic features
Behdad Esfahbod [Tue, 31 Jul 2018 19:49:02 +0000 (12:49 -0700)]
[khmer] Apply ccmp after basic features

Part of https://github.com/harfbuzz/harfbuzz/issues/974

5 years ago[khmer] Rewrite most of shaper to better follow spec
Behdad Esfahbod [Tue, 31 Jul 2018 18:45:32 +0000 (11:45 -0700)]
[khmer] Rewrite most of shaper to better follow spec

Khmer spec has only one reordering phase, and only simple prebase
matra and Coeng-Ro reordering.  Implement that.  Specifically,
this was done to address recognizing different orders of the matra
and Coeng-Ro sequence.  That said, some combinations are now
reordered differently from Uniscribe.  Not clear if that's intended
or a bug in Uniscribe.  The following two sequences render the same
in Uniscribe whereas we reorder them differently:

  U+17A0,U+17D2,U+179A,U+17C2
  U+17A0,U+17C2,U+17D2,U+179A

For that reason, our test suite numbers regressed slightly.  Used
to be at 34 for fails, now at:

KHMER: 299080 out of 299124 tests passed. 44 failed (0.0147096%)

But generally a good change, and removed lots of code.

Fixes https://github.com/harfbuzz/harfbuzz/issues/1026

5 years ago[shaper] Move code around
Behdad Esfahbod [Tue, 31 Jul 2018 08:27:04 +0000 (01:27 -0700)]
[shaper] Move code around

5 years ago[shaper] Remove a macro
Behdad Esfahbod [Tue, 31 Jul 2018 08:15:04 +0000 (01:15 -0700)]
[shaper] Remove a macro

5 years ago[atomic] Make pointer get op relaxed instead of acquire
Behdad Esfahbod [Tue, 31 Jul 2018 08:02:31 +0000 (01:02 -0700)]
[atomic] Make pointer get op relaxed instead of acquire

We only use it before cmpexch, so relaxed is fine and faster for
common case.

5 years ago[serialize] Increase stage count from 8 to 32
Behdad Esfahbod [Tue, 31 Jul 2018 00:03:06 +0000 (17:03 -0700)]
[serialize] Increase stage count from 8 to 32

Indic shaper uses many stages.  Now we are provably not limiting
functionality whereas the previous limit of 8 was assuming real-world
practices.

5 years ago[serialize] Remove unused truncate() method
Behdad Esfahbod [Mon, 30 Jul 2018 23:59:41 +0000 (16:59 -0700)]
[serialize] Remove unused truncate() method

6 years ago[subset] During lookup collection remember the features we've already processed.
Garret Rieger [Fri, 27 Jul 2018 00:42:02 +0000 (17:42 -0700)]
[subset] During lookup collection remember the features we've already processed.

6 years agoMinor
Behdad Esfahbod [Fri, 27 Jul 2018 21:55:29 +0000 (14:55 -0700)]
Minor

6 years ago[serialize] Allocate for markFilteringSet
Behdad Esfahbod [Thu, 26 Jul 2018 00:37:38 +0000 (17:37 -0700)]
[serialize] Allocate for markFilteringSet

6 years agoMinor
Behdad Esfahbod [Thu, 26 Jul 2018 00:22:05 +0000 (17:22 -0700)]
Minor

6 years agoFix unlikely leaks
prrace [Fri, 27 Jul 2018 20:58:27 +0000 (13:58 -0700)]
Fix unlikely leaks

6 years agoAdd hb-iter-private.hh
Behdad Esfahbod [Thu, 26 Jul 2018 00:07:17 +0000 (17:07 -0700)]
Add hb-iter-private.hh

Unused so far.

6 years agoMove some code from hb-open-type-private.hh to hb-machinery-private.hh
Behdad Esfahbod [Wed, 25 Jul 2018 23:58:47 +0000 (16:58 -0700)]
Move some code from hb-open-type-private.hh to hb-machinery-private.hh

6 years ago[closure] Allocate out set on the stack
Behdad Esfahbod [Wed, 25 Jul 2018 23:01:37 +0000 (16:01 -0700)]
[closure] Allocate out set on the stack