replaced m/^G.../g in while loop with s/...// for utf-8 checking.

is easier to understand, and faster.

1 files changed, 8 insertions, 9 deletions

diff --git a/httpd/cgi-bin/check b/httpd/cgi-bin/check
index 5982b9a..151c786 100755
--- a/httpd/cgi-bin/check
+++ b/httpd/cgi-bin/check
@@ -9,7 +9,7 @@
 # This source code is available under the license at:
 #     http://www.w3.org/Consortium/Legal/copyright-software
 #
-# $Id: check,v 1.154 2001-07-20 06:23:03 link Exp $
+# $Id: check,v 1.155 2001-07-21 11:32:06 duerst Exp $
 
 #
 # We need Perl 5.004.
@@ -80,9 +80,9 @@ my $element_ref = 'http://www.htmlhelp.com/reference/html40/';
 
 #
 # Strings
-$VERSION    =  q$Revision: 1.154 $;
+$VERSION    =  q$Revision: 1.155 $;
 $VERSION    =~ s/Revision: ([\d\.]+) /$1/;
-$DATE       =  q$Date: 2001-07-20 06:23:03 $;
+$DATE       =  q$Date: 2001-07-21 11:32:06 $;
 $MAINTAINER =  'gerald@w3.org';
 $NOTICE     =  ''; # "<p><strong>Note: This service will be ...</strong>";
 
@@ -437,10 +437,10 @@ EOHD
     my $p;
     for (@{$File->{Content}}) {
       $line++;
-      # while loop needed for very long lines (>32K),
+      # substitution needed for very long lines (>32K),
       # to avoid backtrack stack overflow
-      $p = 0;        # make sure this works for empty lines
-      while (m/\G(   # \G: start where we left in previous loop
+      my $l = $_;
+      $l =~ s/
           [\x00-\x7F]                           # ASCII
         | [\xC2-\xDF]        [\x80-\xBF]        # non-overlong 2-byte sequences
         |  \xE0[\xA0-\xBF]   [\x80-\xBF]        # excluding overlongs
@@ -449,9 +449,8 @@ EOHD
         |  \xF0[\x90-\xBF]   [\x80-\xBF]{2}     # planes 1-3
         | [\xF1-\xF3]        [\x80-\xBF]{3}     # planes 4-15
         |  \xF4[\x80-\x8F][\x80-\xBF]{2}        # plane 16
-        )
-        /xg) { $p = pos; } # assignement needed because pos lost outside loop
-      push @lines, $line unless ($p == length);
+        //xg;
+      push @lines, $line if (length $l);
     }
   }
   if(@lines) {