From: Karl <khw@karl.(none)>
Date: Wed, 14 Jan 2009 00:51:53 +0000 (-0700)
Subject: Add warning about octal > 377 in some regexes
X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=commitdiff_plain;h=609122bd053b77225626a0b7630039ddc39620ae;p=p5sagit%2Fp5-mst-13.2.git

Add warning about octal > 377 in some regexes

(Tweaked by rgs)

Message-ID: <496D3F02.6020204@khwilliamson.com>
---

diff --git a/pod/perldiag.pod b/pod/perldiag.pod
index e06dd5e..cb82633 100644
--- a/pod/perldiag.pod
+++ b/pod/perldiag.pod
@@ -4777,6 +4777,15 @@ In code that currently says C<use AutoLoader; @ISA = qw(AutoLoader);>
 you should remove AutoLoader from @ISA and change C<use AutoLoader;> to
 C<use AutoLoader 'AUTOLOAD';>.
 
+=item Use of octal value above 377 is deprecated
+
+(D deprecated, W regexp) There is a constant in the regular expression whose
+value is interpeted by Perl as octal and larger than 377 (255 decimal, 0xFF
+hex).  Perl may take this to mean different things depending on the rest of
+the regular expression.  If you meant such an octal value, convert it to
+hexadecimal and use C<\xHH> or C<\x{HH}> instead.  If you meant to have
+part of it mean a backreference, use C<\g> for that.  See L<perlre>.
+
 =item Use of %s in printf format not supported
 
 (F) You attempted to use a feature of printf that is accessible from
diff --git a/regcomp.c b/regcomp.c
index 3ad1f3b..e061528 100644
--- a/regcomp.c
+++ b/regcomp.c
@@ -7430,6 +7430,19 @@ tryagain:
                             I32 flags = 0;
 			    STRLEN numlen = 3;
 			    ender = grok_oct(p, &numlen, &flags, NULL);
+
+			    /* An octal above 0xff is interpreted differently
+			     * depending on if the re is in utf8 or not.  If it
+			     * is in utf8, the value will be itself, otherwise
+			     * it is interpreted as modulo 0x100.  It has been
+			     * decided to discourage the use of octal above the
+			     * single-byte range.  For now, warn only when
+			     * it ends up modulo */
+			    if (SIZE_ONLY && ender >= 0x100
+				    && ! UTF && ! PL_encoding
+				    && ckWARN2(WARN_DEPRECATED, WARN_REGEXP)) {
+				vWARNdep(p, "Use of octal value above 377 is deprecated");
+			    }
 			    p += numlen;
 			}
 			else {
diff --git a/t/op/pat.t b/t/op/pat.t
index 586b317..eb1034a 100755
--- a/t/op/pat.t
+++ b/t/op/pat.t
@@ -13,7 +13,7 @@ sub run_tests;
 
 $| = 1;
 
-my $EXPECTED_TESTS = 3961;  # Update this when adding/deleting tests.
+my $EXPECTED_TESTS = 3962;  # Update this when adding/deleting tests.
 
 BEGIN {
     chdir 't' if -d 't';
@@ -3978,6 +3978,13 @@ sub run_tests {
         iseq $te [0], '../';
     }
 
+	# This currently has to come before any "use encoding" in this file.
+    {
+        local $Message;
+        local $BugId   = '59342';
+        must_warn 'qr/\400/', '^Use of octal value above 377';
+    }
+
 
     SKIP: {
         # XXX: This set of tests is essentially broken, POSIX character classes