Changeset 2811

Timestamp:

2001-10-31T19:41:49+13:00 (22 years ago)

Author:

sjboddie

Message:

* empty log message *

Location:

trunk/gsdl

Files:

• bin/script/mkcol.pl (modified) (1 diff)
• collect/demo/etc/collect.cfg (modified) ( previous)
• perllib/plugins/BasPlug.pm (modified) (3 diffs)
• perllib/plugins/ConvertToPlug.pm (modified) (2 diffs)
• perllib/plugins/WordPlug.pm (modified) (2 diffs)

trunk/gsdl/bin/script/mkcol.pl

@@ -164 +164 @@
     # load default plugins if none were on command line    
     if (!scalar(@plugin)) {
-    @plugin = (ZIPPlug,GMLPlug,TEXTPlug,HTMLPlug,EMAILPlug,
+    @plugin = (ZIPPlug,GAPlug,TEXTPlug,HTMLPlug,EMAILPlug,
            PDFPlug,RTFPlug,WordPlug,PSPlug,ArcPlug,RecPlug);
     }

trunk/gsdl/perllib/plugins/BasPlug.pm

@@ -272 +272 @@
     return undef;
     }
-    my $plugin_name = ref ($self);
     $file =~ s/^[\/\\]+//; # $file often begins with / so we'll tidy it up
 
-    my ($language, $encoding);
-    if ($self->{'input_encoding'} eq "auto") {
-    # use textcat to automatically work out the input encoding and language
-    ($language, $encoding) = $self->get_language_encoding ($filename);
-
-    } elsif ($self->{'extract_language'}) {
-    # use textcat to get language metadata
-    ($language, $extracted_encoding) = $self->get_language_encoding ($filename);
-    $encoding = $self->{'input_encoding'};
-
-    if ($extracted_encoding ne $encoding && $self->{'verbosity'}) {
-        print $outhandle "$plugin_name: WARNING: $file was read using $encoding encoding but ";
-        print $outhandle "appears to be encoded as $extracted_encoding.\n";
-    }
-
-    } else {
-    $language = $self->{'default_language'};
-    $encoding = $self->{'input_encoding'};
-    }
+    # Do encoding stuff
+    my ($language, $encoding) = $self->textcat_get_language_encoding ($filename);
 
     # create a new document
@@ -308 +290 @@
 
     if (!length ($text)) {
+    my $plugin_name = ref ($self);
     print $outhandle "$plugin_name: ERROR: $file contains no text\n" if $self->{'verbosity'};
 
@@ -384 +367 @@
 
     close FILE;
+}
+
+sub textcat_get_language_encoding {
+    my $self = shift (@_);
+    my ($filename) = @_;
+
+    my ($language, $encoding, $extracted_encoding);
+    if ($self->{'input_encoding'} eq "auto") {
+        # use textcat to automatically work out the input encoding and language
+        ($language, $encoding) = $self->get_language_encoding ($filename);
+    } elsif ($self->{'extract_language'}) {
+        # use textcat to get language metadata
+        ($language, $extracted_encoding) = $self->get_language_encoding ($filename);
+        $encoding = $self->{'input_encoding'};
+        if ($extracted_encoding ne $encoding && $self->{'verbosity'}) {
+        my $plugin_name = ref ($self);
+        my $outhandle = $self->{'outhandle'};
+            print $outhandle "$plugin_name: WARNING: $filename was read using $encoding encoding but ";
+            print $outhandle "appears to be encoded as $extracted_encoding.\n";
+        }
+    } else {
+        $language = $self->{'default_language'};
+        $encoding = $self->{'input_encoding'};
+    }
+    return ($language, $encoding);
 }
 

trunk/gsdl/perllib/plugins/ConvertToPlug.pm

@@ -243 +243 @@
         return undef;
     }
-    my $plugin_name = ref ($self);
     $file =~ s/^[\/\\]+//; # $file often begins with / so we'll tidy it up
 
@@ -257 +256 @@
 
     # Do encoding stuff
-    my ($language, $encoding);
-
-    # WordPlug's wvWare will always produce html files encoded as utf-8
-    if ($plugin_name eq "WordPlug") {
-    $self->{'input_encoding'} = "utf8";
-    $self->{'extract_language'} = 1;
-    }
-
-    if ($self->{'input_encoding'} eq "auto") {
-        # use textcat to automatically work out the input encoding and language
-        ($language, $encoding) = $self->get_language_encoding ($conv_filename);
-    } elsif ($self->{'extract_language'}) {
-        # use textcat to get language metadata
-
-    my ($extracted_encoding);
-        ($language, $extracted_encoding) = $self->get_language_encoding ($conv_filename);
-        $encoding = $self->{'input_encoding'};
-        if ($extracted_encoding ne $encoding && $self->{'verbosity'}) {
-            print $outhandle "$plugin_name: WARNING: $file was read using $encoding encoding but ";
-            print $outhandle "appears to be encoded as $extracted_encoding.\n";
-        }
-    } else {
-        $language = $self->{'default_language'};
-        $encoding = $self->{'input_encoding'};
-    }
+    my ($language, $encoding) = $self->textcat_get_language_encoding ($conv_filename);
 
     &BasPlug::read_file($self, $conv_filename, $encoding, $language, \$text);
     if (!length ($text)) {
+    my $plugin_name = ref ($self);
         print $outhandle "$plugin_name: ERROR: $file contains no text\n" if $self->{'verbosity'};
         return 0;

trunk/gsdl/perllib/plugins/WordPlug.pm

@@ -32 +32 @@
 }
 
+sub new {
+    my $class = shift (@_);
+
+    my $self = new ConvertToPlug ($class, @_);
+
+    # wvWare will always produce html files encoded as utf-8
+    if ($self->{'input_encoding'} eq "auto") {
+    $self->{'input_encoding'} = "utf8";
+    $self->{'extract_language'} = 1;
+    }
+
+    return bless $self, $class;
+}
+
 sub get_default_process_exp {
     my $self = shift (@_);
@@ -37 +51 @@
     return q^(?i)\.doc$^;
 }
-    
-
 
 # do plugin specific processing of doc_obj for HTML type