From: Nicholas Clark <nick@ccl4.org>
Date: Sun, 28 Dec 2008 13:06:42 +0000 (+0000)
Subject: Add a key flag HVhek_KEYCANONICAL for Perl_hv_common(), which signals that the
X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=commitdiff_plain;h=44b87b5061968d098b3906ca4b5e6f94469c8694;p=p5sagit%2Fp5-mst-13.2.git

Add a key flag HVhek_KEYCANONICAL for Perl_hv_common(), which signals that the
key is in canonical form - any key passed encoded in UTF-8 cannot be represented
as bytes, hence the downgrade check can be skipped. Use this internally for
shared hash key scalars, as they are always canonical.
---

diff --git a/hv.c b/hv.c
index b764c55..79f5973 100644
--- a/hv.c
+++ b/hv.c
@@ -388,8 +388,12 @@ Perl_hv_common(pTHX_ HV *hv, SV *keysv, const char *key, STRLEN klen,
 	if (flags & HVhek_FREEKEY)
 	    Safefree(key);
 	key = SvPV_const(keysv, klen);
-	flags = 0;
 	is_utf8 = (SvUTF8(keysv) != 0);
+	if (SvIsCOW_shared_hash(keysv)) {
+	    flags = HVhek_KEYCANONICAL | (is_utf8 ? HVhek_UTF8 : 0);
+	} else {
+	    flags = 0;
+	}
     } else {
 	is_utf8 = ((flags & HVhek_UTF8) ? TRUE : FALSE);
     }
@@ -595,7 +599,7 @@ Perl_hv_common(pTHX_ HV *hv, SV *keysv, const char *key, STRLEN klen,
 	}
     }
 
-    if (is_utf8) {
+    if (is_utf8 & !(flags & HVhek_KEYCANONICAL)) {
 	char * const keysave = (char *)key;
 	key = (char*)bytes_from_utf8((U8*)key, &klen, &is_utf8);
         if (is_utf8)
diff --git a/hv.h b/hv.h
index 2265326..9322377 100644
--- a/hv.h
+++ b/hv.h
@@ -367,6 +367,9 @@ C<SV*>.
 #define HVhek_FREEKEY	0x100 /* Internal flag to say key is malloc()ed.  */
 #define HVhek_PLACEHOLD	0x200 /* Internal flag to create placeholder.
                                * (may change, but Storable is a core module) */
+#define HVhek_KEYCANONICAL 0x400 /* Internal flag - key is in canonical form.
+				    If the string is UTF-8, it cannot be
+				    converted to bytes. */
 #define HVhek_MASK	0xFF
 
 /* Which flags enable HvHASKFLAGS? Somewhat a hack on a hack, as