Change 24720: [PATCH] Re: [perl #36130] chr(-1) should probably return undef

[Rafael Garcia-Suarez]

Change 24720 by [EMAIL PROTECTED] on 2005/06/07 09:23:58

        Subject: [PATCH] Re: [perl #36130] chr(-1) should probably return undef
        From: Jarkko Hietaniemi <[EMAIL PROTECTED]>
        Date: Mon, 06 Jun 2005 20:28:35 +0300
        Message-ID: <[EMAIL PROTECTED]>

Affected files ...

... //depot/perl/pod/perlfunc.pod#472 edit
... //depot/perl/pp.c#459 edit
... //depot/perl/t/op/chr.t#2 edit

Differences ...

==== //depot/perl/pod/perlfunc.pod#472 (text) ====
Index: perl/pod/perlfunc.pod
--- perl/pod/perlfunc.pod#471~24691~    Fri Jun  3 02:29:36 2005
+++ perl/pod/perlfunc.pod       Tue Jun  7 02:23:58 2005
@@ -742,6 +742,10 @@
 to 255 (inclusive) are by default not encoded in UTF-8 Unicode for
 backward compatibility reasons (but see L<encoding>).
 
+Negative values give the Unicode replacement character (chr(0xfffd)),
+except under the L</bytes> pragma, where low eight bits of the value
+(truncated to an integer) are used.
+
 If NUMBER is omitted, uses C<$_>.
 
 For the reverse, use L</ord>.

==== //depot/perl/pp.c#459 (text) ====
Index: perl/pp.c
--- perl/pp.c#458~24717~        Tue Jun  7 01:18:59 2005
+++ perl/pp.c   Tue Jun  7 02:23:58 2005
@@ -3356,7 +3356,20 @@
 {
     dSP; dTARGET;
     char *tmps;
-    UV value = POPu;
+    UV value;
+
+    if (((SvIOK_notUV(TOPs) && SvIV(TOPs) < 0)
+        ||
+        (SvNOK(TOPs) && SvNV(TOPs) < 0.0))) {
+       if (IN_BYTES) {
+           value = POPu; /* chr(-1) eq chr(0xff), etc. */
+       } else {
+           (void) POPs; /* Ignore the argument value. */
+           value = UNICODE_REPLACEMENT;
+       }
+    } else {
+       value = POPu;
+    }
 
     SvUPGRADE(TARG,SVt_PV);
 

==== //depot/perl/t/op/chr.t#2 (text) ====
Index: perl/t/op/chr.t
--- perl/t/op/chr.t#1~24706~    Sat Jun  4 09:55:27 2005
+++ perl/t/op/chr.t     Tue Jun  7 02:23:58 2005
@@ -6,7 +6,7 @@
     require "test.pl";
 }
 
-plan tests => 26;
+plan tests => 34;
 
 # Note that t/op/ord.t already tests for chr() <-> ord() rountripping.
 
@@ -19,11 +19,24 @@
 is(chr(128), "\x80");
 is(chr(255), "\xFF");
 
-# is(chr(-1), undef); # Shouldn't it be?
+is(chr(-0.1), "\x{FFFD}"); # The U+FFFD Unicode replacement character.
+is(chr(-1  ), "\x{FFFD}");
+is(chr(-2  ), "\x{FFFD}");
+is(chr(-3.0), "\x{FFFD}");
+{
+    use bytes; # Backward compatibility.
+    is(chr(-0.1), "\x00");
+    is(chr(-1  ), "\xFF");
+    is(chr(-2  ), "\xFE");
+    is(chr(-3.0), "\xFD");
+}
 
 # Check UTF-8.
 
-sub hexes { join(" ",map{sprintf"%02x",$_}unpack("C*",chr($_[0]))) }
+sub hexes {
+    no warnings 'utf8'; # avoid surrogate and beyond Unicode warnings
+    join(" ",map{sprintf"%02x",$_}unpack("C*",chr($_[0])));
+}
 
 # The following code points are some interesting steps in UTF-8.
 is(hexes(   0x100), "c4 80");
End of Patch.