[khmer] Fix Coeng vs Halant confusion
Test suite results are unchanged (34).
diff --git a/src/hb-ot-shape-complex-khmer.cc b/src/hb-ot-shape-complex-khmer.cc
index 7c2cd79..18e3c94 100644
--- a/src/hb-ot-shape-complex-khmer.cc
+++ b/src/hb-ot-shape-complex-khmer.cc
@@ -317,7 +317,7 @@
if ((FLAG_UNSAFE (info[i].khmer_category()) & (JOINER_FLAGS | FLAG (OT_N) | FLAG (OT_RS) | MEDIAL_FLAGS | FLAG (OT_Coeng))))
{
info[i].khmer_position() = last_pos;
- if (unlikely (info[i].khmer_category() == OT_H &&
+ if (unlikely (info[i].khmer_category() == OT_Coeng &&
info[i].khmer_position() == POS_PRE_M))
{
/*
@@ -538,7 +538,7 @@
* and possibly multiple substitutions happened prior to this
* phase, and that might have messed up our properties. Recover
* from a particular case of that where we're fairly sure that a
- * class of OT_H is desired but has been lost. */
+ * class of OT_Coeng is desired but has been lost. */
if (khmer_plan->virama_glyph)
{
unsigned int virama_glyph = khmer_plan->virama_glyph;
@@ -548,7 +548,7 @@
_hb_glyph_info_multiplied (&info[i]))
{
/* This will make sure that this glyph passes is_coeng() test. */
- info[i].khmer_category() = OT_H;
+ info[i].khmer_category() = OT_Coeng;
_hb_glyph_info_clear_ligated_and_multiplied (&info[i]);
}
}
