Changeset 24060 for main/trunk/greenstone2/perllib/plugins/MetadataXMLPlugin.pm

Timestamp:

2011-05-19T17:51:07+12:00 (13 years ago)

Author:

ak19

Message:

Dr Bainbridge fixed the unicode issue that appeared after the 2.84 release, when the parser instantiation got changed (by removing its ProtocolEncoding argument, which used to set this encoding to Latin-1).

File:

• main/trunk/greenstone2/perllib/plugins/MetadataXMLPlugin.pm (modified) (6 diffs)

main/trunk/greenstone2/perllib/plugins/MetadataXMLPlugin.pm

@@ -147 +147 @@
 
     # create XML::Parser object for parsing metadata.xml files
-    my $parser = new XML::Parser('Style' => 'Stream',
+    my $parser = new XML::Parser('Style' => 'Stream',                             
+                                  'Pkg' => 'MetadataXMLPlugin',
                                   'PluginObj' => $self,
-                                  'Pkg' => 'MetadataXMLPlugin',
-                  'Handlers' => {'Char' => \&Char,
+                    'Handlers' => {'Char' => \&Char,
                          'Doctype' => \&Doctype
                          });
@@ -246 +246 @@
     my $filename = &util::filename_cat($base_dir, $file);
     return undef unless $self->can_process_this_file($filename);    
-
+    
     $self->{'metadata-file'} = $file;
     $self->{'metadata-filename'} = $filename;
@@ -266 +266 @@
 
     if ($@) {
+    print STDERR "**** Error is: $@\n";
     my $plugin_name = ref ($self);
     my $failhandle = $self->{'failhandle'};
@@ -301 +302 @@
     my $self = shift(@_);
     my ($expat, $element) = @_;
-
+    
     if ($element eq "FileSet") {
     $self->{'saved_targets'} = [];
@@ -356 +357 @@
     }
     elsif ($element eq "Metadata") {
-    my $metadata_name = decode("utf8",$self->{'metadata_name'});
-    my $metadata_value = decode("utf8",$self->{'metadata_value'});
+    # text read in by XML::Parser is in Perl's binary byte value
+    # form ... need to explicitly make it UTF-8
+    
+    my $metadata_name = decode("utf-8",$self->{'metadata_name'});
+    my $metadata_value = decode("utf-8",$self->{'metadata_value'});
     
     &metadatautil::store_saved_metadata($self,
@@ -384 +388 @@
 # things down significantly in some cases.
 sub Char {
-    if ($]<5.008) {
-    use bytes;  # Necessary to prevent encoding issues with XML::Parser 2.31+ and Perl 5.6
-    }
+    use bytes;  # Necessary to prevent encoding issues with XML::Parser 2.31+ 
+
+#    if ($]<5.008) {
+#   use bytes;  # Necessary to prevent encoding issues with XML::Parser 2.31+ and Perl 5.6
+#    }
     $_[0]->{'Text'} .= $_[1];
     return undef;