Misapplied regex optimizations when \C is present.

Fixes 20001230.002.

What still remains broken is that the submatches that
have \C in them get their UTF8 flag on because their
parent SV has it on.  This will result in malformed
UTF8 if a \C happened to match a non-ASCII byte.

p4raw-id: //depot/perl@8836

1 files changed, 3 insertions, 0 deletions

diff --git a/regcomp.c b/regcomp.c
index 547f756a30..69d114e9e9 100644
--- a/regcomp.c
+++ b/regcomp.c
@@ -1930,6 +1930,8 @@ Perl_pregcomp(pTHX_ char *exp, char *xend, PMOP *pm)
 	r->reganch |= ROPT_LOOKBEHIND_SEEN;
     if (RExC_seen & REG_SEEN_EVAL)
 	r->reganch |= ROPT_EVAL_SEEN;
+    if (RExC_seen & REG_SEEN_SANY)
+	r->reganch |= ROPT_SANY_SEEN;
     Newz(1002, r->startp, RExC_npar, I32);
     Newz(1002, r->endp, RExC_npar, I32);
     PL_regdata = r->data; /* for regprop() */
@@ -2638,6 +2640,7 @@ tryagain:
 	    break;
 	case 'C':
 	    ret = reg_node(pRExC_state, SANY);
+	    RExC_seen |= REG_SEEN_SANY;
 	    *flagp |= HASWIDTH|SIMPLE;
 	    nextchar(pRExC_state);
 	    break;