For 5.12: saner behaviour for `length`

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

diff --git a/pp.c b/pp.c
index b4e81d6..6110b4c 100644 (file)
--- a/pp.c
+++ b/pp.c
@@ -2625,7 +2625,7 @@ PP(pp_i_modulo_1)
      /* This is the i_modulo with the workaround for the _moddi3 bug
       * in (at least) glibc 2.2.5 (the PERL_ABS() the workaround).
       * See below for pp_i_modulo. */
-     dVAR; dVAR; dSP; dATARGET; tryAMAGICbin(modulo,opASSIGN);
+     dVAR; dSP; dATARGET; tryAMAGICbin(modulo,opASSIGN);
      {
          dPOPTOPiirl;
          if (!right)
@@ -3018,25 +3018,35 @@ PP(pp_length)
     dVAR; dSP; dTARGET;
     SV * const sv = TOPs;
 
-    if (SvAMAGIC(sv)) {
-       /* For an overloaded scalar, we can't know in advance if it's going to
-          be UTF-8 or not. Also, we can't call sv_len_utf8 as it likes to
-          cache the length. Maybe that should be a documented feature of it.
+    if (!SvOK(sv) && !SvGMAGICAL(sv)) {
+       /* FIXME - this doesn't allow GMAGIC to return undef for consistency.
+        */
+       SETs(&PL_sv_undef);
+    } else if (SvGAMAGIC(sv)) {
+       /* For an overloaded or magic scalar, we can't know in advance if
+          it's going to be UTF-8 or not. Also, we can't call sv_len_utf8 as
+          it likes to cache the length. Maybe that should be a documented
+          feature of it.
        */
        STRLEN len;
-       const char *const p = SvPV_const(sv, len);
+       const char *const p
+           = sv_2pv_flags(sv, &len,
+                          SV_UNDEF_RETURNS_NULL|SV_CONST_RETURN|SV_GMAGIC);
 
-       if (DO_UTF8(sv)) {
+       if (!p)
+           SETs(&PL_sv_undef);
+       else if (DO_UTF8(sv)) {
            SETi(utf8_length((U8*)p, (U8*)p + len));
        }
        else
            SETi(len);
-
+    } else {
+       /* Neither magic nor overloaded.  */
+       if (DO_UTF8(sv))
+           SETi(sv_len_utf8(sv));
+       else
+           SETi(sv_len(sv));
     }
-    else if (DO_UTF8(sv))
-       SETi(sv_len_utf8(sv));
-    else
-       SETi(sv_len(sv));
     RETURN;
 }
 
@@ -4837,7 +4847,7 @@ PP(pp_split)
             }
         }
     }
-    else if (do_utf8 == ((RX_EXTFLAGS(rx) & RXf_UTF8) != 0) &&
+    else if (do_utf8 == (RX_UTF8(rx) != 0) &&
             (RX_EXTFLAGS(rx) & RXf_USE_INTUIT) && !RX_NPARENS(rx)
             && (RX_EXTFLAGS(rx) & RXf_CHECK_ALL)
             && !(RX_EXTFLAGS(rx) & RXf_ANCH)) {
@@ -4845,7 +4855,7 @@ PP(pp_split)
        SV * const csv = CALLREG_INTUIT_STRING(rx);
 
        len = RX_MINLENRET(rx);
-       if (len == 1 && !(RX_EXTFLAGS(rx) & RXf_UTF8) && !tail) {
+       if (len == 1 && !RX_UTF8(rx) && !tail) {
            const char c = *SvPV_nolen_const(csv);
            while (--limit) {
                for (m = s; m < strend && *m != c; m++)