Change 14143 by jhi@alpha on 2002/01/09 01:32:05
SANY (. with /s) wasn't happy on Unicode.
Affected files ...
.... //depot/perl/regexec.c#245 edit
.... //depot/perl/t/op/pat.t#147 edit
Differences ...
==== //depot/perl/regexec.c#245 (text) ====
Index: perl/regexec.c
--- perl/regexec.c.~1~ Tue Jan 8 18:45:06 2002
+++ perl/regexec.c Tue Jan 8 18:45:06 2002
@@ -3877,7 +3877,15 @@
}
break;
case SANY:
- scan = loceol;
+ if (do_utf8) {
+ loceol = PL_regeol;
+ while (scan < loceol) {
+ scan += UTF8SKIP(scan);
+ hardcount++;
+ }
+ }
+ else
+ scan = loceol;
break;
case CANY:
scan = loceol;
==== //depot/perl/t/op/pat.t#147 (xtext) ====
Index: perl/t/op/pat.t
--- perl/t/op/pat.t.~1~ Tue Jan 8 18:45:06 2002
+++ perl/t/op/pat.t Tue Jan 8 18:45:06 2002
@@ -6,7 +6,7 @@
$| = 1;
-print "1..843\n";
+print "1..845\n";
BEGIN {
chdir 't' if -d 't';
@@ -2620,3 +2620,21 @@
print "<\x{2029}>" =~ /<\s>/ ? "ok 843\n" : "not ok 843\n";
}
+{
+ print "# . with /s should work on characters, not bytes\n";
+
+ my $s = "\x{e4}\x{100}";
+
+ # This is not expected to match: the point is that
+ # neither should we get "Malformed UTF-8" warnings.
+ print $s =~ /\G(.+?)\n/gcs ?
+ "not ok 844\n" : "ok 844\n";
+
+ my @c;
+
+ while ($s =~ /\G(.)/gs) {
+ push @c, $1;
+ }
+
+ print join("", @c) eq $s ? "ok 845\n" : "not ok 845\n";
+}
End of Patch.
