Commit Graph

141 Commits

Author SHA1 Message Date
Behdad Esfahbod 882ebdadd6 [unsafe-to-break] Flag during cursive positioning 2017-08-10 20:23:17 -07:00
Behdad Esfahbod 2d158ca376 [unsafe-to-break] Flag during kerning 2017-08-10 20:20:30 -07:00
Volker H. Simonis ad273881c5 Minor fixes to pacify Coverity code scan
Fixes https://github.com/behdad/harfbuzz/pull/377
2016-12-18 00:14:06 -06:00
Behdad Esfahbod 59055b5494 [GX] Implement Feature Variations
Not hooked up to shaper yet.
2016-12-16 19:06:26 -06:00
Behdad Esfahbod cf3de4d8f7 [GX] Rename VarStore to VariationStore 2016-12-16 19:06:26 -06:00
Behdad Esfahbod 1124d2ece5 [GX] Fix thinko 2016-12-16 19:06:26 -06:00
Behdad Esfahbod 151d93de8a [GX] Hook up GPOS to Variation Store stored in GDEF
Untested.
2016-12-16 19:06:26 -06:00
Behdad Esfahbod 49cb87850c [GX] Pass apply-context down to get_anchor()
Needed to access GDEF for varStore.
2016-12-16 19:06:26 -06:00
Behdad Esfahbod dc9f229799 [GX] Port variation stuff to Variation Store design
Not hooked up to GDEF yet.
2016-12-16 19:06:26 -06:00
Behdad Esfahbod ca28670316 [GX] Towards anisotropic interpolation
Also hookup to ValueRecord and Anchors.
2016-12-16 19:06:26 -06:00
Behdad Esfahbod ebd7431f82 Partially revert 86c68c7a2c
That commit moved the advance adjustment for mark positioning to
be applied immediately, instead of doing late before.  This breaks
if mark advances are zeroed late, like in Arabic.  Also, easier to
hit it in RTL scripts since a single mark with non-zero advance is
enough to hit the bug, whereas in LTR, at least two marks are needed.

This reopens https://github.com/behdad/harfbuzz/issues/211
The cursive+mark interaction is broken again.  To be fixed in a
different way.
2016-02-24 15:53:40 +09:00
Behdad Esfahbod 86c68c7a2c [GPOS] Fix interaction of mark attachments and cursive chaining
Fixes https://github.com/behdad/harfbuzz/issues/211

What happens in that bug is that a mark is attached to base first,
then a second mark is cursive-chained to the first mark.  This only
"works" because it's in the Indic shaper where mark advances are
not zeroed.

Before, we didn't allow cursive to run on marks at all.  Fix that.
We also where updating mark major offsets at the end of GPOS, such
that changes in advance of base will not change the mark attachment
position.  That was superior to the alternative (which is what Uniscribe
does BTW), but made it hard to apply cursive to the mark after it
was positioned.  We could track major-direction offset changes and
apply that to cursive in the post process, but that's a much trickier
thing to do than the fix here, which is to immediately apply the
major-direction advance-width offsets...  Ie.:

https://github.com/behdad/harfbuzz/issues/211#issuecomment-183194739

If this breaks any fonts, the font should be fixed to do mark attachment
after all the advances are set up first (kerning, etc).

Finally, this, still doesn't make us match Uniscribe, for I explained
in that bug.  Looks like Uniscribe applies minor-direction cursive
adjustment immediate as well.  We don't, and we like it our way, at
least for now.  Eg. the sequence in the test case does this:

- The first subscript attaches with mark-to-base, moving in x only,
- The second subscript attaches with cursive attachment to first subscript
  moving in x only,
- A final context rule moves the first subscript up by 104 units.

The way we do, the final shift-up, also shifts up the second subscript
mark because it's cursively-attached.  Uniscribe doesn't.  We get:

[ttaorya=0+1307|casubscriptorya=0@-242,104+-231|casubscriptnarroworya=0@20,104+507]

while Uniscribe gets:

[ttaorya=0+1307|casubscriptorya=0@-242,104+-211|casubscriptnarroworya=0+487]

note the different y-offset of the last glyph.  In our view, after cursive,
things move together, period.
2016-02-16 16:07:20 +07:00
Behdad Esfahbod 6ab920224c [GPOS] Minor
No effect.
2016-02-11 16:57:52 +07:00
Behdad Esfahbod cbc3a76c5a [GPOS] Merge fixing of offsets for cursive and mark attachments
Part of fixing https://github.com/behdad/harfbuzz/issues/211
2016-02-11 16:48:13 +07:00
Behdad Esfahbod 7d8d58ac81 [GPOS] Divide position_finish() into two phases, for advances and offsets
Right now the position_finish_advances() is empty.  To be used for
spacing attachments proposal later.
2016-02-11 16:34:28 +07:00
Behdad Esfahbod 8474231567 [ot] Minor shuffling code around 2016-02-11 16:27:41 +07:00
Behdad Esfahbod b0b11614e9 [GPOS] Add harmless recursion in fix_mark_attachment()
Will do nothing.  Just useful for merging two functions.
2016-02-11 15:28:55 +07:00
Behdad Esfahbod 686567baab [GPOS] Merge attach_chain() and cursive_chain()
Differentiate, using new attach_type().
2016-02-11 15:25:28 +07:00
Behdad Esfahbod 806ad8dc65 [GPOS] Minor shuffling 2016-02-11 14:53:11 +07:00
Behdad Esfahbod 0f6278d1fb [GPOS] Negate sign of attach_lookback(), and rename it to attach_chain()
No behavior change.  Preparing to unify how cursive and mark attachments
work.
2016-02-11 14:49:10 +07:00
Behdad Esfahbod fc38e6034c [layout] Collect coverage glyphs in PairPos
Apparently class=0 is used for ClassDef1.  See:
https://github.com/adobe-type-tools/afdko/issues/90
2015-12-23 14:50:53 +01:00
Behdad Esfahbod 550bd14d2b Actually commit changes for previous commit 2015-11-04 22:58:58 -08:00
Behdad Esfahbod ff31b3c1b8 Skip GPOS cursive / attachment adjustmnent if none happened
I'm starting to really like how free these new scratch_flags are.
2015-11-04 22:47:30 -08:00
Behdad Esfahbod f966649747 Fix another memory access issue discovered by libFuzzer
Fixes https://github.com/behdad/harfbuzz/issues/139#issuecomment-146984679
2015-10-13 00:30:50 -04:00
Behdad Esfahbod f396fbb000 Fix return value of sanitize when subformat is not readable
This is a fix on top of the previous issue fixed in
c917965b9e.

This was caught by "libFuzzer" testing.
2015-10-09 12:25:55 -04:00
Behdad Esfahbod b47159011c Define return_trace()
Not functional change (expected!).
2015-09-29 14:57:02 +01:00
Behdad Esfahbod c917965b9e Add missing returns
Fixes possible invalid read of two bytes.

Reported by Behzad Najjarpour Jabbari, Secunia Research.
2015-09-29 14:45:35 +01:00
Behdad Esfahbod 6578575cc8 [GPOS] Fix cursive connection with mix of RTL and non-RTL lookups
See thread "Issue with cursive attachment" started by Khaled.
Turned out fixing this wasn't as bad as I had assumed.  I like the
new code better; we now have a theoretical model of cursive
connections that is easier to reason about.
2015-08-25 20:29:36 +01:00
Behdad Esfahbod 58f2a73fb9 [GPOS] Rewrite cursive attachment slightly differently
In anticipation for upcoming fix for bug reported by
Khaled in thread "Issue with cursive attachment".
2015-08-25 18:58:24 +01:00
Behdad Esfahbod 2c8b3b2e53 [debug] Print lookup index in debug-apply output 2015-08-18 14:36:43 +01:00
Behdad Esfahbod fc3c59a1d7 Fix unused var warnings 2015-04-08 13:04:29 -07:00
Behdad Esfahbod bd047d3b7f [layout] Minor 2015-02-25 15:43:31 -08:00
Behdad Esfahbod 1a2322134a [layout] Don't check glyph props against lookup flags when recursing
Shouldn't be needed.  I have a hard time imagining this breaking any
legitimate use case.
2015-02-25 15:43:30 -08:00
Behdad Esfahbod 095a1257cc [layout] Port sanitize() to use dispatch()
Needed some rework of Extension table.  Hopefully I got it right, and
the new template usage doesn't break any compilers...
2015-02-25 15:43:30 -08:00
Behdad Esfahbod 40c58923cb [layout] Refactor Lookup::dispatch() 2015-02-25 15:43:30 -08:00
Behdad Esfahbod 70366f5d19 [layout] Refactor get_subtable() 2015-02-25 15:43:29 -08:00
Behdad Esfahbod f72f326aea Minor 2015-02-25 15:43:29 -08:00
Behdad Esfahbod 8e36ccfd4f [layout] Use dispatch() for add_coverage() 2015-02-25 15:43:29 -08:00
Behdad Esfahbod 50b8dc79da [layout] Add may_dispatch()
No functional change right now.
2015-02-25 15:43:29 -08:00
Behdad Esfahbod de2118ed7a Make sanitize() a const method
This makes a lot of code safer.  We only try modifying the object in one
place, after making sure it's safe to do so.  So, do a const_cast<> in
that one place...
2015-02-25 15:43:28 -08:00
Behdad Esfahbod 6759ed95a3 Minor 2015-02-25 15:43:28 -08:00
Behdad Esfahbod 6b599dac1f Remove unnecessary check in sanitize 2015-02-25 15:43:28 -08:00
Behdad Esfahbod 365576d246 [layout] Allocate iters in the context
Can be further optimized, but I think I didn't break anything.

Saves another 3% off Roboto shaping.
2015-02-25 15:43:28 -08:00
Behdad Esfahbod 514564f544 [layout] Move skippy_iter setup from constructor into init() 2015-02-25 15:43:27 -08:00
Behdad Esfahbod b051be542a [lookup] Add skippy_iter.reset()
Towards reducing the cost of initializing skippy_iter()
2015-02-25 15:43:27 -08:00
Behdad Esfahbod 696266981d [layout] Merge forward and backward iterators 2015-02-25 15:43:26 -08:00
Behdad Esfahbod 37d13acd8d [layout] Remove some unnecessary checks in skippy 2015-02-25 15:43:26 -08:00
Behdad Esfahbod baa14e1814 [lookup] Don't initialize skippy if coverage match fails
Currently:

  - Initializing skippy is very expensive,

  - Our lookup accelerator (using set-digests) can be very ineffecite,

As such, we end up many times initializing skippy but then failing
coverage check.  Reordering fixes that.

When, later, we fix our accelerator to have truly small false-positive
rate (for example by using the frozen-sets), then we might want to
reorder these checks such that we wouldn't calculate coverage number
if skippy is going to fail.

This shows a 5% speedup with Roboto already.
2015-02-25 15:43:26 -08:00
Behdad Esfahbod 7788993bc1 [layout] Use setter method to set c->lookup_props 2015-02-25 15:43:25 -08:00
Behdad Esfahbod 67dfa8c7c2 When matching second glyph of kerning pairs, use bsearch
Roboto has glyphs (like 'F') that have 200 kerning pairs.
Add a handcoded bsearch instead of previous linear search.

This doesn't show much speedup though, apparently we spend the
bulk of the time somewhere before here.
2015-01-19 17:00:31 -08:00