Update grapheme break algorithm to Unicode 5.1.0 (based on UAX #29)

diff --git a/lib/unicode.c b/lib/unicode.c
index 55390eb1a52e33b4579fa4a2811003f5599d1633..16c593282cb8f13232ada5066e80026d3fe1b485 100644 (file)
--- a/lib/unicode.c
+++ b/lib/unicode.c
@@ -603,6 +603,12 @@ int utf32_iterator_grapheme_boundary(utf32_iterator it) {
   /* GB9 */
   if(gbafter == unicode_Grapheme_Break_Extend)
     return 0;
   /* GB9 */
   if(gbafter == unicode_Grapheme_Break_Extend)
     return 0;

+  /* GB9a */
+  if(gbafter == unicode_Grapheme_Break_SpacingMark)
+    return 0;
+  /* GB9b */
+  if(gbbefore == unicode_Grapheme_Break_Prepend)
+    return 0;

   /* GB10 */
   return 1;
 
   /* GB10 */
   return 1;