EBCDIC tweak (native instead of uni) and comment update.

p4raw-id: //depot/perl@14852

diff --git a/regexec.c b/regexec.c
index 900b491..8c091a1 100644 (file)
--- a/regexec.c
+++ b/regexec.c
@@ -959,8 +959,8 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, char *strend, char *sta
                to_utf8_lower((U8*)m, tmpbuf1, &ulen1);
                to_utf8_upper((U8*)m, tmpbuf2, &ulen2);
 
-               c1 = utf8_to_uvuni(tmpbuf1, 0);
-               c2 = utf8_to_uvuni(tmpbuf2, 0);
+               c1 = utf8_to_uvchr(tmpbuf1, 0);
+               c2 = utf8_to_uvchr(tmpbuf2, 0);
            }
            else {
                c1 = *(U8*)m;
@@ -984,9 +984,10 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, char *strend, char *sta
             * text of the node.  The c1 and c2 are the first
             * characters (though in Unicode it gets a bit
             * more complicated because there are more cases
-            * than just upper and lower: one is really supposed
-            * to use the so-called folding case for case-insensitive
-            * matching (called "loose matching" in Unicode).  */
+            * than just upper and lower: one needs to use
+            * the so-called folding case for case-insensitive
+            * matching (called "loose matching" in Unicode).
+            * ibcmp_utf8() will do just that. */
 
            if (do_utf8) {
                UV c, f;