Commit Graph

11509 Commits

Author SHA1 Message Date
Khaled Hosny 03538e872a 2.8.0 2021-03-16 19:28:35 +02:00
Khaled Hosny 4c34478b28
Merge pull request #2899 from harfbuzz/ci-fix
[ci] Don’t install meson from its master branch
2021-03-16 10:37:49 +02:00
Behdad Esfahbod daeb277346
Merge pull request #2900 from harfbuzz/nodiscard-buffer
Mark buffer modification methods that can fail with nodiscard
2021-03-15 18:50:02 -07:00
Khaled Hosny f2d08578e7 [tests] Increase shape-fuzzer timeout 2021-03-16 01:15:40 +02:00
Khaled Hosny cad753e20f [ci] Use known working Ubuntu version
GitHub Actions seems to be moving ubuntu-latest from ubuntu-18.04 to
ubuntu-20.04, but gcovr is broken for us in the new version.
2021-03-16 01:11:12 +02:00
Khaled Hosny ff86c72e09 [ci] Avoid duplicate builds on pull requests
Don’t run GitHub Actions on pushing to branches other than master. This
was already the case for the linux-ci workflow.
2021-03-16 01:11:12 +02:00
Behdad Esfahbod f5356bcb61 [ci] Switch back to c++2a
Okay I guess we can be bothered to fix things up for next version of C++.

Reverts meat of c0b3b9b258
2021-03-15 16:52:39 -06:00
Behdad Esfahbod 9e397ff2fb [buffer] Fix order of HB_INTERNAL HB_DISCARD
Sucks that has to be specified in this order. But that is what it is for now.

Was only exhibiting problem on C++>=17 since that's when the [[nodiscard]]
was introduced.
2021-03-15 16:51:17 -06:00
Behdad Esfahbod 9582640eed [ci] Remove more cruft
All were commented out already.
2021-03-15 16:48:16 -06:00
Behdad Esfahbod c0b3b9b258 [ci] Don't test C++2a build
One of the bots is unhappy when HB_NODISCARD comes after HB_INTERNAL.
No idea why. But, again, we're testing HarfBuzz, not C++, not clang. Ugh.

In file included from src/harfbuzz.cc:1:
In file included from src/hb-aat-layout.cc:30:
In file included from src/hb-aat-layout.hh:32:
In file included from src/hb-ot-shape.hh:32:
In file included from src/hb-ot-map.hh:32:
src/hb-buffer.hh:335:15: error: an attribute list cannot appear here
  HB_INTERNAL HB_NODISCARD bool move_to (unsigned int i); /* i is output-buffer index. */
              ^~~~~~~~~~~~
https://app.circleci.com/pipelines/github/harfbuzz/harfbuzz/1693/workflows/77459205-a189-45d3-bc58-52a8fd952c3f/jobs/155912/parallel-runs/0/steps/0-110?invite=true
2021-03-15 16:48:07 -06:00
Behdad Esfahbod 15f3ef0136 [test] Fix output format 2021-03-15 16:01:35 -06:00
Behdad Esfahbod 8450f43ae1 [buffer] HB_NODISCARD next_glyph() 2021-03-15 16:01:35 -06:00
Behdad Esfahbod f4bc7673db [buffer] Implement copy_glyph() in terms of output_info() 2021-03-15 16:01:35 -06:00
Behdad Esfahbod f73982a699 [buffer] Implement replace_glyph() in terms of replace_glyphs(1,1)
I get exact same binary size with this, suggesting that compiler is
optimizing these as needed.
2021-03-15 16:01:35 -06:00
Behdad Esfahbod 862f913489 [buffer] Implement output_glyph() in terms of replace_glyphs(0,1)
To my surprise, saves ~20kb in my build (non-size-optimized) build.
The output_glyph() method is never used in the fast paths, so doesn't
matter if is not fully optimized for the special case it is.
2021-03-15 16:01:35 -06:00
Behdad Esfahbod 34a1204f10 [buffer] HB_NODISCARD output_glyph()
Also, generalize and use replace_glyphs() in morx where output_glyph() was used
in a loop.
2021-03-15 16:01:35 -06:00
Khaled Hosny 07315d9c83 [ci] Don’t install meson from its master branch
We are testing Harfbuzz not meson!
2021-03-15 22:45:25 +02:00
Behdad Esfahbod e6be9eb4fb [buffer] HB_NODISCARD output_info() 2021-03-15 14:12:10 -06:00
Behdad Esfahbod 2a0dbb3ee5 [buffer] HB_NODISCARD copy_glyph() 2021-03-15 14:10:39 -06:00
Behdad Esfahbod b05e5d9a79 [buffer] HB_NODISCARD next_glyphs() 2021-03-15 14:08:08 -06:00
Behdad Esfahbod 4ae8aab83b [buffer] HB_NODISCARD has_separate_output() 2021-03-15 14:01:52 -06:00
Behdad Esfahbod 8d3701f507 [buffer] HB_NODISCARD in_error() 2021-03-15 14:01:32 -06:00
Behdad Esfahbod 41e05479b6 [buffer] HB_NODISCARD shift_forward() 2021-03-15 14:00:00 -06:00
Behdad Esfahbod 83b3784d1d [buffer] HB_NODISCARD make_room_for() 2021-03-15 13:59:14 -06:00
Behdad Esfahbod c355508a89 [buffer] HB_NODISCARD ensure_glyphs() / ensure_unicode() 2021-03-15 13:58:30 -06:00
Behdad Esfahbod bc22305b6a [buffer] HB_NODISCARD ensure_inplace() 2021-03-15 13:57:18 -06:00
Behdad Esfahbod cac6c86d2f [buffer] HB_NODISCARD move_to() 2021-03-15 13:56:46 -06:00
Behdad Esfahbod 05d2d37f9a [buffer] HB_NODISCARD ensure() 2021-03-15 13:56:46 -06:00
Behdad Esfahbod d8028a0762 [buffer] HB_NODISCARD enlarge() 2021-03-15 13:34:36 -06:00
Behdad Esfahbod 3f1998a065 [buffer] HB_NODISCARD replace_glyph() 2021-03-15 13:33:44 -06:00
Behdad Esfahbod 607979d12f [buffer] HB_NODISCARD replace_glyphs() 2021-03-15 13:23:48 -06:00
Behdad Esfahbod 906c9928bb [buffer] Return success status from buffer ops that can fail
Previous error-handling philosophy was that user doesn't need to
immediately know whether operation failed. But as can be seen after
we added malloc-failing fuzzing, there's just so many places in the
code that a failure of these operations needs to be mitigated before
further operations. So I'm moving towards returning success here,
and possibly making it nodiscard.
2021-03-15 13:13:45 -06:00
Behdad Esfahbod a5b8e7db4d [hangul] Improve error handling
I did a review; changed some "return"s to "break"s, which should be identical.
Removed one check just before "continue" because not necessary.
The added error check is the actual fix.

Should fix https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=31755
2021-03-15 12:46:58 -06:00
Behdad Esfahbod 99767f9386 [hangul] Whitespace 2021-03-15 12:36:59 -06:00
Behdad Esfahbod 3622120fab [subset] Make ClassDef format2 .intersects() return false if range value 0
We treat Class0 as "doesn't intersect".  That's the only meaningful
interpretation. If one allos Class0 to mean "intersects", then the
intersects() result should be true iff glyphset is non-empty.

Related to https://github.com/harfbuzz/harfbuzz/issues/2703
2021-03-10 12:21:43 -07:00
David Corbett e19de65eae
Update hb-ot-tag-table.hh (#2890) 2021-03-08 10:12:47 -08:00
Khaled Hosny 7686ff854b
[ot] Keep substituted Default_Ignorables (#2886)
Don’t replace Default_Ignorables with zero-width space if they are
substituted or multiplied, not just when ligated.

After this change, HarfBuzz output matches that of Uniscribe and
CoreText for the new tests.

Fixes https://github.com/harfbuzz/harfbuzz/issues/2883
2021-03-04 13:09:32 -08:00
Behdad Esfahbod 5efa04c890 [Makefile] Rebuild .def files if config changed
I was getting check-symbols failure because my previous build was
without CoreText, and after reconfiguring with CoreText, the old
harfbuzz.defs file was not being regenerated.
2021-03-02 16:30:13 -07:00
Behdad Esfahbod d351bbf0fb [Makefile] Remove unused variable HBNODISTHEADERS
Not sure what it was used for before.
2021-03-02 16:30:13 -07:00
Behdad Esfahbod fd489433a8 [indic] Fix cluster-merging logic with cluster-level=1
Was producing non-monotonic cluster numbers because our faulty logic
was not merging clusters if something from before base and after base
had switched positions.

Fixes https://github.com/harfbuzz/harfbuzz/issues/2272
2021-03-02 16:30:09 -07:00
Behdad Esfahbod 2902529b92 [subset] Fix HB_TINY build
Fixes https://github.com/harfbuzz/harfbuzzjs/issues/34#issuecomment-789247723
2021-03-02 15:05:39 -07:00
Behdad Esfahbod 7cb22ba7eb
Include C headers with their C++ names (#2882)
Remove unnecessary includes.

Fixes build with some known broken SDKs (Nintendo Switch?)

https://en.cppreference.com/w/cpp/header

Fixes https://github.com/harfbuzz/harfbuzz/pull/2881
2021-03-01 11:44:06 -08:00
Behdad Esfahbod dbcf2f417f [cmake] Fix syntax 2021-03-01 12:43:32 -07:00
Behdad Esfahbod 93745f9f49
Merge pull request #2875 from harfbuzz/constexpr-inttype
[constexpr] Add constructors to IntType
2021-02-23 14:54:45 -07:00
Behdad Esfahbod 486da35cc0 m Add comments to IntType cast out operator
Okay, bots seem to be happy. Merging.
2021-02-23 13:58:35 -07:00
Behdad Esfahbod 83b66bfb66 Another try to fix narrowing error
../src/hb-ot-layout-gsubgpos.hh: In instantiation of ‘void OT::ChainRule::serialize_array(hb_serialize_context_t*, OT::HBUINT16, Iterator) const [with Iterator = hb_map_iter_t<hb_array_t<const OT::IntType<short unsigned int> >, const hb_map_t*&, (hb_function_sortedness_t)0, 0>; typename hb_enable_if<hb_is_iterator_of<Lhs, typename Lhs::item_t>::value>::type* <anonymous> = 0; OT::HBUINT16 = OT::IntType<short unsigned int>]’:
../src/hb-ot-layout-gsubgpos.hh:2341:30:   required from here
../src/hb-ot-layout-gsubgpos.hh:2326:15: error: narrowing conversion of ‘(unsigned int)g’ from ‘unsigned int’ to ‘short unsigned int’ inside { } [-Werror=narrowing]
       c->copy (HBUINT16 {g});
       ~~~~~~~~^~~~~~~~~~~~~~

https://github.com/harfbuzz/harfbuzz/pull/2875
2021-02-23 13:04:38 -07:00
Behdad Esfahbod 6c4bb60829 Fix narrowing errors with recent changes 2021-02-22 23:05:55 -07:00
Behdad Esfahbod d6bd00a488 Revert back IntType out cast to signed/unsigned
Previous commit didn't fix the bots. Putting it back now that I
understand why I initially did the "Wide" casts.  But only doing
it for out-cast this time.  This causes "narrowing" warnings
whenever we are converting signed/unsigned to smaller HBUINT16
etc.  But those are valuable warnings.  We should address those
separately instead of ignoring.

Maybe we should start using uint16_t more liberally in the
internal subsetter function signatures then.
2021-02-22 22:42:50 -07:00
Behdad Esfahbod 0983601399 Add back wider cast to IntType
My local clang12 is fine, but many bots are not:

../src/hb-ot-cff1-table.hh: In instantiation of ‘bool CFF::Charset1_2<TYPE>::sanitize(hb_sanitize_context_t*, unsigned int) const [with TYPE = OT::IntType<unsigned char>]’:
../src/hb-ot-cff1-table.hh:554:13:   required from here
../src/hb-ot-cff1-table.hh:377:60: error: comparison between signed and unsigned integer expressions [-Werror=sign-compare]
       if (unlikely (!ranges[i].sanitize (c) || (num_glyphs < ranges[i].nLeft + 1)))
                                                ~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~

Enabling the extra cast operator mentioned in previous commit to see if
that fixes this case.

Again, I'd be happy to say "use 1u instead of 1" if this was universally
erred on.  But since some compilers happily compile this while others
err, it would be a huge headache. Let's see...

https://github.com/harfbuzz/harfbuzz/pull/2875
2021-02-22 22:33:17 -07:00
Behdad Esfahbod 567cedcc5f Narrow down cast operators on IntType
Say for USHORT, we were implementing casts from and to unsigned.
With this change, we cast from and to uint16_t only.  This allows
compiler more opportunities to catch possible narrowing issues in
the code.

It needed a couple of fixes in the codebase though, because
previously, if a USHORT was participating in arithmetic with signed
numbers, eg. "u + 1", the result would have been unsigned.  With
this change, it would be signed.  The correct fix is to update the
code to read "u + 1u".

That said, I think about conditionally adding back the cast
out to signed/unsigned, to facilitate better type deduction.
But I couldn't think of a real situation where that would help
with anything.  So I didn't add.  Here's what it was:

  template <typename Type2 = hb_conditional<hb_is_signed (Type), signed, unsigned>,
           hb_enable_if (sizeof (Type) < sizeof (Type2))>
  operator hb_type_identity_t<Type2> () const { return v; }

https://github.com/harfbuzz/harfbuzz/pull/2875
2021-02-22 22:32:42 -07:00