New approach to fix this:
69f9fbc420
Previous approach was reverted as it was too broad. See context:
https://github.com/behdad/harfbuzz/issues/347#issuecomment-267838368
With U+05E9,U+05B8,U+05C1,U+05DC and Arial Unicode, we now (correctly) disable
GDEF and GPOS, so we get results very close to Uniscribe, but slightly different
since our fallback position logic is not exactly the same:
Before: [gid1166=3+991|gid1142=0+737|gid5798=0+1434]
After: [gid1166=3+991|gid1142=0@402,-26+0|gid5798=0+1434]
Uniscribe: [gid1166=3+991|gid1142=0@348,0+0|gid5798=0+1434]
Back in the old days, we used to apply 'calt' and 'cswh' in Arabic shaper,
with a pause in between. Then we disabled the 'cswh' because Microsoft
disabled it, but forgot to remove the unnecessary pause. Do that now.
This has the benefit that it fixes shaping with monbaiti from Windows 10.
In that version of that font, the lookups from 'calt' are duplicated in
'rclt', and Mongolian was changed to go through Universal Shaping Engine.
We still use the Arabic shaper for Mongolian. With a pause after 'calt',
we were applying the duplicate lookups from 'calt' and 'rclt' twice. It
happened to be the case that these lookups were NOT idempotent. So we
were getting wrong shaping. See thread "Windows 10 monbaiti.ttf upgrade
(5.01 -> 5.51) caused loss of diacritical marks when shaped with harfbuz"
on the mailing list. This fixes that.
This aims to make Syriac Abbr Mark sizing more accurate when repeating segments are used, by adding an extra repeat and tightening up the spacing slightly rather than leaving a shortfall corresponding to a partial repeat-width.
The feature is enabled for any character in the Arabic shaper.
We should experiment with using it for Arabic subtending marks.
Though, that has a directionality problem as well, since those
are used with digits...
Fixes https://github.com/behdad/harfbuzz/issues/141
Previously, we expected users to provide BOT/EOT flags when the
text *segment* was at paragraph boundaries. This meant that for
clients that provide full paragraph to HarfBuzz (eg. Pango), they
had code like this:
hb_buffer_set_flags (hb_buffer,
(item_offset == 0 ? HB_BUFFER_FLAG_BOT : 0) |
(item_offset + item_length == paragraph_length ?
HB_BUFFER_FLAG_EOT : 0));
hb_buffer_add_utf8 (hb_buffer,
paragraph_text, paragraph_length,
item_offset, item_length);
After this change such clients can simply say:
hb_buffer_set_flags (hb_buffer,
HB_BUFFER_FLAG_BOT | HB_BUFFER_FLAG_EOT);
hb_buffer_add_utf8 (hb_buffer,
paragraph_text, paragraph_length,
item_offset, item_length);
Ie, HarfBuzz itself checks whether the segment is at the beginning/end
of the paragraph. Clients that only pass item-at-a-time to HarfBuzz
continue not setting any flags whatsoever.
Another way to put it is: if there's pre-context text in the buffer,
HarfBuzz ignores the BOT flag. If there's post-context, it ignores
EOT flag.
Follows the order of the Arabic/Syriac specs. Also don't stop
between rlig and calt in non-Arabic scripts.
Micro-tests for Arabic and Mongolian added for the latter.
This reverts commit d5bd0590ae.
The reasoning behind that logic was flawed and made under
a misunderstanding of the original problem, and caused
regressions as reported by Jonathan Kew in thread titled
"tibetan marks" in Oct 2013. Apparently I have had fixed
the original problem with this commit:
7e08f1258d
So, revert the faulty commit and everything seems to be in good
shape.
Unicode 6.2.0 Section 16.2 / Figure 16.3 says:
"For backward compatibility, between Arabic characters a ZWJ acts just
like the sequence <ZWJ, ZWNJ, ZWJ>, preventing a ligature from forming
instead of requesting the use of a ligature that would not normally be
used. As a result, there is no plain text mechanism for requesting the
use of a ligature in Arabic text."
As such, we flip internal zwj to zwnj flags for GSUB matching, which
means it will block ligation in all features, unless the font
explicitly matches U+200D glyph. This doesn't affect joining behavior.
Testing shows that this is closer to what Uniscribe does.
Reported by Khaled Hosny:
"""
commit 568000274c
...
This commit is causing a regression with Amiri, the string “هَٰذ” with
Uniscribe and HarfBuzz before this commit, gives:
[uni0630.fina=3+965|uni0670.medi=0+600|uni064E=0@-256,0+0|uni0647.init=0+926]
But now it gives:
[uni0630.fina=3+965|uni0670.medi=0+0|uni064E=0@-256,0+0|uni0647.init=0+926]
i.e. uni0670.medi is zeroed though it has a base glyph GDEF class.
"""
The test case is U+0647,U+064E,U+0670,U+0630 with Amiri.
Before, we were zeroing advance width of attached marks for
non-Indic scripts, and not doing it for Indic.
We have now three different behaviors, which seem to better
reflect what Uniscribe is doing:
- For Indic, no explicit zeroing happens whatsoever, which
is the same as before,
- For Myanmar, zero advance width of glyphs marked as marks
*in GDEF*, and do that *before* applying GPOS. This seems
to be what the new Win8 Myanmar shaper does,
- For everything else, zero advance width of glyphs that are
from General_Category=Mn Unicode characters, and do so
before applying GPOS. This seems to be what Uniscribe does
for Latin at least.
With these changes, positioning of all tests matches for Myanmar,
except for the glitch in Uniscribe not applying 'mark'. See preivous
commit.
Behdad Esfahbod
Ebrahim Byagowi
jfkthame
Chun-wei Fan