Precomputing the hash value for a string representable in bytes, but passed in

in UTF-8, would result in storing the wrong hash value in the hash, and hence
failing lookups. I guess not that much XS code precomputes hash values.

diff --git a/ext/XS/APItest/APItest.xs b/ext/XS/APItest/APItest.xs
index 78ee526..e63d804 100644 (file)
--- a/ext/XS/APItest/APItest.xs
+++ b/ext/XS/APItest/APItest.xs
@@ -448,10 +448,14 @@ common(params)
        if ((svp = hv_fetchs(params, "action", 0)))
            action = SvIV(*svp);
        if ((svp = hv_fetchs(params, "val", 0)))
-           val = *svp;
+           val = newSVsv(*svp);
        if ((svp = hv_fetchs(params, "hash", 0)))
            hash = SvUV(*svp);
 
+       if ((svp = hv_fetchs(params, "hash_pv", 0))) {
+           PERL_HASH(hash, key, klen);
+       }
+
        result = (HE *)hv_common(hv, keysv, key, klen, flags, action, val, hash);
        if (!result) {
            XSRETURN_EMPTY;

diff --git a/ext/XS/APItest/t/hash.t b/ext/XS/APItest/t/hash.t
index 1ef99ed..47d4437 100644 (file)
--- a/ext/XS/APItest/t/hash.t
+++ b/ext/XS/APItest/t/hash.t
@@ -150,6 +150,27 @@ if ($] > 5.009) {
     }
 }
 
+{
+    my $as_utf8 = "\241" . chr 256;
+    chop $as_utf8;
+    my $as_bytes = "\243";
+    foreach my $key ('N', $as_bytes, $as_utf8, "\x{2623}") {
+       my $ord = ord $key;
+       foreach my $hash_pv (0, 1) {
+           my %hash;
+           is (XS::APItest::Hash::common({hv => \%hash, keypv => $key,
+                                          val => $ord, hash_pv => $hash_pv,
+                                          action =>
+                                          XS::APItest::HV_FETCH_ISSTORE}),
+               $ord, "store $ord \$hash_pv = $hash_pv");
+           is_deeply ([each %hash], [$key, $ord], "First key read is good");
+           is_deeply ([each %hash], [], "No second key good");
+
+           is ($hash{$key}, $ord, "Direct hash read finds $ord");
+       }
+    }
+}
+
 exit;
 
 ################################   The End   ################################

diff --git a/hv.c b/hv.c
index 78a1097..b764c55 100644 (file)
--- a/hv.c
+++ b/hv.c
@@ -606,6 +606,11 @@ Perl_hv_common(pTHX_ HV *hv, SV *keysv, const char *key, STRLEN klen,
            if (flags & HVhek_FREEKEY)
                Safefree(keysave);
             flags |= HVhek_WASUTF8 | HVhek_FREEKEY;
+           /* If the caller calculated a hash, it was on the sequence of
+              octets that are the UTF-8 form. We've now changed the sequence
+              of octets stored to that of the equivalent byte representation,
+              so the hash we need is different.  */
+           hash = 0;
        }
     }