Changeset 21958 for main/trunk

Timestamp:

2010-04-27T11:18:22+12:00 (14 years ago)

Author:

kjdon

Message:

ppthtml and xslhtml don't seem to output utf8, so remove the input_encoding option to html/txt secondary plugins so that it will use textcat to work it out itself.

Location:

main/trunk/greenstone2/perllib/plugins

Files:

• ExcelPlugin.pm (modified) (1 diff)
• PowerPointPlugin.pm (modified) (3 diffs)

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

@@ -79 +79 @@
     my $text_options = $secondary_plugin_options->{'TextPlugin'};
 
-    push(@$html_options, "-input_encoding", "utf8");
+    # xslhtml doesn't output utf8, let Greenstone work out the encoding
+    #push(@$html_options, "-input_encoding", "utf8");
     push(@$html_options,"-extract_language") if $self->{'extract_language'};
     push(@$html_options, "-file_rename_method", "none");
     push(@$html_options, "-processing_tmp_files");
 
-    push(@$text_options, "-input_encoding", "utf8");
+    #push(@$text_options, "-input_encoding", "utf8");
     push(@$text_options,"-extract_language") if $self->{'extract_language'};
     push(@$text_options, "-file_rename_method", "none");

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

@@ -101 +101 @@
     $self->{'file_type'} = "PPT";
 
-    # ppthtml outputs utf-8 already.
     #these are passed through to gsConvert.pl by ConvertBinaryFile.pm
     $self->{'convert_options'} = "-windows_scripting" if $self->{'windows_scripting'};
@@ -117 +116 @@
  
     if (defined $html_options){
-    push(@$html_options,"-input_encoding", "utf8");
+    # ppthtml doesn't output utf-8 necessarily - let Greenstone determine the encoding
+    #push(@$html_options,"-input_encoding", "utf8");
     push(@$html_options,"-extract_language") if $self->{'extract_language'};
     push(@$html_options,"-file_rename_method", "none");
@@ -128 +128 @@
     } 
     if (defined $text_options){
-    push(@$text_options,"-input_encoding", "utf8");
+    #push(@$text_options,"-input_encoding", "utf8");
     push(@$text_options,"-extract_language") if $self->{'extract_language'};
     push(@$text_options,"-file_rename_method", "none");
     }
     if (defined $pageimg_options){
+    #is this true??
     push(@$pageimg_options,"-input_encoding", "utf8");
     push(@$pageimg_options,"-extract_language") if $self->{'extract_language'};