Changeset 16753

Timestamp:

2008-08-13T13:10:23+12:00 (16 years ago)

Author:

ak19

Message:

get_language_encoding for HTMLFiles strips out the comments before trying to match on html tags

File:

• gsdl/trunk/perllib/plugins/ReadTextFile.pm (modified) (2 diffs)

gsdl/trunk/perllib/plugins/ReadTextFile.pm

@@ -340 +340 @@
     (exists $self->{'converted_to'} && $self->{'converted_to'} eq 'HTML')){
 
+    # remove comments, including multiline ones, so that we don't match on 
+    # inactive tags (those that are nested inside comments)
+    $text =~ s/<!--.*?-->//sg;
+
     # remove <title>stuff</title> -- as titles tend often to be in English
     # for foreign language documents
@@ -349 +353 @@
     }
     # check the meta http-equiv charset tag unless it is commented out
-    elsif (($text !~ /<!--[^<>]?<meta http-equiv/i) && ($text =~ /<meta http-equiv.*content-type.*charset=(.+?)\"/i)) {            
+    elsif ($text =~ m/<meta http-equiv.*content-type.*charset=(.+?)\"/i) {             
         $best_encoding = $1;
 #       print STDERR "**** meta tag found, encoding is: $best_encoding\n";