Tweak mark-base attachment

Apparently a base glyph can also become an attached component of a
ligature if the ligature-forming lookup used IgnoreBase.  This was
being confused with a non-first component of a MultipleSubst and
hence not matched for mark-attachment.  Tweak test to fix.

Fixes https://github.com/behdad/harfbuzz/issues/543

src/hb-ot-layout-gpos-table.hh

@@ -1087,7 +1087,9 @@ struct MarkBasePosFormat1
     do {
       if (!skippy_iter.prev ()) return_trace (false);
       /* We only want to attach to the first of a MultipleSubst sequence.  Reject others. */
-      if (0 == _hb_glyph_info_get_lig_comp (&buffer->info[skippy_iter.idx])) break;
+      if (!_hb_glyph_info_multiplied (&buffer->info[skippy_iter.idx]) ||
+	  0 == _hb_glyph_info_get_lig_comp (&buffer->info[skippy_iter.idx]))
+	break;
       skippy_iter.reject ();
     } while (1);
 

test/shaping/Makefile.am

@@ -62,6 +62,7 @@ TESTS = \
 	tests/indic-syllable.tests \
 	tests/language-tags.tests \
 	tests/ligature-id.tests \
+	tests/mark-attachment.tests \
 	tests/mark-filtering-sets.tests \
 	tests/mongolian-variation-selector.tests \
 	tests/spaces.tests \

test/shaping/tests/mark-attachment.tests

@@ -0,0 +1 @@
+fonts/sha1sum/98b7887cff91f722b92a8ff800120954606354f9.ttf::U+100F,U+103C,U+102F,U+1036:[uni103C102F=0+150|uni100F=0+550|uni1036=0@-150,0+0]