The issue is somewhat theoretical, since in reasonbale fonts the indices
should always be valid, and in fact are fairly likely to be 0. On the other
hand, web fonts exist.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Aric Stewart <aric@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Like in GPOS_apply_ChainContextPos().
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Aric Stewart <aric@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
This preserves the current logic, but I wonder if we're not supposed to always
return "input_count". Returning 1 seems questionable if we matched more input
glyphs than that, and so does including the lookahead count.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Aric Stewart <aric@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
I.e., avoid accessing outside the "glyphs" array. If we're particularly
unlucky we may actually get a match against data outside the array.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Aric Stewart <aric@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
And avoid some endian conversions in the process.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Aric Stewart <aric@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Based on feedback from Henri, fixes regression from commit
63bcfa3354.
Signed-off-by: Jacek Caban <jacek@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Aric Stewart <aric@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
And cleanup a few style issues along the way.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Aric Stewart <aric@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
And cleanup a few style issues along the way.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Aric Stewart <aric@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
And cleanup some style issues along the way.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Aric Stewart <aric@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
And cleanup a style issue along the way.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Aric Stewart <aric@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
And cleanup some style issues along the way.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Aric Stewart <aric@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
E.g., the base form of 'ا' may be the same as the isolated form, in which case
the "isol" feature may not have a substitution for it. At the same time, while
the presentation form B glyph may look the same or similar as the base form,
it may have a different glyph ID. This can break e.g.
GSUB_LOOKUP_CONTEXT_CHAINED substitutions because of the unexpected glyph ID.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Aric Stewart <aric@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
In particular, if fLogicalOrder and fRTL are both set, the glyph order is
swapped compared to the character order. In all other cases it's the same.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Aric Stewart <aric@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
As opposed to just looking at the run direction.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Aric Stewart <aric@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
And standardize their type names.
Signed-off-by: Francois Gouget <fgouget@free.fr>
Signed-off-by: Aric Stewart <aric@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Henri Verbeet