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

p4raw-id: //depot/perl@14852

1 files changed, 6 insertions, 5 deletions

diff --git a/regexec.c b/regexec.c
index 900b491502..8c091a172d 100644
--- 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;