From: Nicholas Clark <nick@ccl4.org>
Date: Thu, 22 Oct 2009 12:30:03 +0000 (+0100)
Subject: Perl_utf16_to_utf8() should treat "\0" like any every other odd-length input.
X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=commitdiff_plain;h=26cc780bb8911dcc1e2c3e971db95322db3e2d9f;p=p5sagit%2Fp5-mst-13.2.git

Perl_utf16_to_utf8() should treat "\0" like any every other odd-length input.

The "be understanding" bodge to not panic, introduced in 1de9afcdf18cf98b, is
no longer needed now that c28d61051c446453 fixes the underlying problem.
---

diff --git a/ext/XS-APItest/t/utf16_to_utf8.t b/ext/XS-APItest/t/utf16_to_utf8.t
index 592d0b1..40a6288 100644
--- a/ext/XS-APItest/t/utf16_to_utf8.t
+++ b/ext/XS-APItest/t/utf16_to_utf8.t
@@ -23,15 +23,13 @@ for my $ord (0, 10, 13, 78, 255, 256, 0xD7FF, 0xE000, 0xFFFD,
     }
 }
 
-# Currently this is special-cased to work. Should it?
-
-is(utf16_to_utf8("\0"), "\0", 'Short string to utf16_to_utf8');
-
-# But anything else is fatal
-
-my $got = eval {utf16_to_utf8('N')};
-like($@, qr/^panic: utf16_to_utf8: odd bytelen 1 at/, 'Odd byte length panics');
-is($got, undef, 'hence eval returns undef');
+foreach ("\0", 'N', 'Perl rules!') {
+    my $length = length $_;
+    my $got = eval {utf16_to_utf8($_)};
+    like($@, qr/^panic: utf16_to_utf8: odd bytelen $length at/,
+	 "Odd byte length panics for '$_'");
+    is($got, undef, 'hence eval returns undef');
+}
 
 for (["\xD8\0\0\0", 'NULs'],
      ["\xD8\0\xD8\0", '2 Lows'],
@@ -40,7 +38,7 @@ for (["\xD8\0\0\0", 'NULs'],
      ["\xDC\0\xDC\0", 'High High'],
     ) {
     my ($malformed, $name) = @$_;
-    $got = eval {utf16_to_utf8($malformed)};
+    my $got = eval {utf16_to_utf8($malformed)};
     like($@, qr/^Malformed UTF-16 surrogate at/,
 	 "Malformed surrogate $name croaks for utf16_to_utf8");
     is($got, undef, 'hence eval returns undef');
@@ -53,7 +51,7 @@ for (["\xD8\0\0\0", 'NULs'],
 }
 
 my $in = "NA";
-$got = eval {utf16_to_utf8_reversed($in, 1)};
+my $got = eval {utf16_to_utf8_reversed($in, 1)};
 like($@, qr/^panic: utf16_to_utf8_reversed: odd bytelen 1 at/,
      'Odd byte length panics');
 is($got, undef, 'hence eval returns undef');
diff --git a/utf8.c b/utf8.c
index dc12df8..c504891 100644
--- a/utf8.c
+++ b/utf8.c
@@ -958,12 +958,6 @@ Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
 
     PERL_ARGS_ASSERT_UTF16_TO_UTF8;
 
-    if (bytelen == 1 && p[0] == 0) { /* Be understanding. */
-	 d[0] = 0;
-	 *newlen = 1;
-	 return d + 1;
-    }
-
     if (bytelen & 1)
 	Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen %"UVuf, (UV)bytelen);