repeat operator (x) doesn't preserve utf8-ness

[p5sagit/p5-mst-13.2.git] / pp.c

diff --git a/pp.c b/pp.c
index 316c793..03ced37 100644 (file)
--- a/pp.c
+++ b/pp.c
@@ -1076,10 +1076,10 @@ PP(pp_repeat)
            SP -= items;
     }
     else {     /* Note: mark already snarfed by pp_list */
-       SV *tmpstr;
+       SV *tmpstr = POPs;
        STRLEN len;
+       bool isutf = SvUTF8(tmpstr) ? TRUE : FALSE;
 
-       tmpstr = POPs;
        SvSetSV(TARG, tmpstr);
        SvPV_force(TARG, len);
        if (count != 1) {
@@ -1092,7 +1092,10 @@ PP(pp_repeat)
            }
            *SvEND(TARG) = '\0';
        }
-       (void)SvPOK_only(TARG);
+       if (isutf)
+           (void)SvPOK_only_UTF8(TARG);
+       else
+           (void)SvPOK_only(TARG);
        PUSHTARG;
     }
     RETURN;
@@ -1199,15 +1202,8 @@ PP(pp_ncmp)
     {
       dPOPTOPnnrl;
       I32 value;
-#ifdef __osf__ /* XXX Configure probe for isnan and isnanl needed XXX */
-#if defined(USE_LONG_DOUBLE) && defined(HAS_LONG_DOUBLE)
-#define Perl_isnan isnanl
-#else
-#define Perl_isnan isnan
-#endif
-#endif
 
-#ifdef __osf__ /* XXX fix in 5.6.1 --jhi */
+#ifdef Perl_isnan
       if (Perl_isnan(left) || Perl_isnan(right)) {
          SETs(&PL_sv_undef);
          RETURN;
@@ -1892,6 +1888,7 @@ PP(pp_hex)
     STRLEN n_a;
 
     tmps = POPpx;
+    argtype = 1;               /* allow underscores */
     XPUSHn(scan_hex(tmps, 99, &argtype));
     RETURN;
 }
@@ -1909,6 +1906,7 @@ PP(pp_oct)
        tmps++;
     if (*tmps == '0')
        tmps++;
+    argtype = 1;               /* allow underscores */
     if (*tmps == 'x')
        value = scan_hex(++tmps, 99, &argtype);
     else if (*tmps == 'b')
@@ -2013,12 +2011,12 @@ PP(pp_substr)
        RETPUSHUNDEF;
     }
     else {
-        if (utfcurlen) {
+       if (utfcurlen)
            sv_pos_u2b(sv, &pos, &rem);
-           SvUTF8_on(TARG);
-       }
        tmps += pos;
        sv_setpvn(TARG, tmps, rem);
+       if (utfcurlen)
+           SvUTF8_on(TARG);
        if (repl)
            sv_insert(sv, pos, rem, repl, repl_len);
        else if (lvalue) {              /* it's an lvalue! */
@@ -2031,7 +2029,7 @@ PP(pp_substr)
                                "Attempt to use reference as lvalue in substr");
                }
                if (SvOK(sv))           /* is it defined ? */
-                   (void)SvPOK_only(sv);
+                   (void)SvPOK_only_UTF8(sv);
                else
                    sv_setpvn(sv,"",0); /* avoid lexical reincarnation */
            }
@@ -4445,7 +4443,8 @@ PP(pp_pack)
            if ((*pat != 'a' && *pat != 'A' && *pat != 'Z') || pat[1] != '*')
                DIE(aTHX_ "/ must be followed by a*, A* or Z*");
            lengthcode = sv_2mortal(newSViv(sv_len(items > 0
-                                                  ? *MARK : &PL_sv_no)));
+                                                  ? *MARK : &PL_sv_no)
+                                            + (*pat == 'Z' ? 1 : 0)));
        }
        switch(datumtype) {
        default: