Changeset 13550 for trunk/gsdl/perllib/classify/GenericList.pm

Timestamp:

2007-01-10T14:00:02+13:00 (17 years ago)

Author:

mdewsnip

Message:

Improvements to Unicode handling in preparation for adding the Unicode collation code. Now all metadata values are converted from UTF-8 strings to Unicode strings, and this means that the normal Perl length() and substr() functions work.

File:

• trunk/gsdl/perllib/classify/GenericList.pm (modified) (7 diffs)

trunk/gsdl/perllib/classify/GenericList.pm

@@ -233 +233 @@
             $metadata_value =~ s/^\s*//;
             $metadata_value =~ s/\s*$//;
-            push(@{$self->{$metadata_group . ".list"}->{$section_OID}}, $metadata_value);
+
+            # Convert the metadata value from a UTF-8 string to a Unicode string
+            # This means that length() and substr() work properly
+            # We need to be careful to convert classifier node title values back to UTF-8, however
+            my $metadata_value_unicode_string = &convert_utf8_string_to_unicode_string($metadata_value);
+
+            # Add the metadata value into the list for this combination of metadata group and section
+            push(@{$self->{$metadata_group . ".list"}->{$section_OID}}, $metadata_value_unicode_string);
         }
         last if (@metadata_values > 0);
@@ -305 +312 @@
     my %metavaluetoOIDsubhash = ();
 
-    my $lastpartition = &unicode::substr($sortedmetavalues[0], 0, 1);
+    my $lastpartition = substr($sortedmetavalues[0], 0, 1);
     foreach my $metavalue (@sortedmetavalues) {
-        my $metavaluepartition = &unicode::substr($metavalue, 0, 1);
+        my $metavaluepartition = substr($metavalue, 0, 1);
 
         # Is this the start of a new partition?
@@ -371 +378 @@
 
 
-sub unicode_length
-{
-    my $utf8string = shift(@_);
-
-    my @unicodestring = @{&unicode::utf82unicode($utf8string)};
-    return scalar(@unicodestring);
+sub convert_utf8_string_to_unicode_string
+{
+    my $utf8_string = shift(@_);
+
+    my $unicode_string = "";
+    foreach my $unicode_value (@{&unicode::utf82unicode($utf8_string)}) {
+    $unicode_string .= chr($unicode_value);
+    }
+    return $unicode_string;
+}
+
+
+sub convert_unicode_string_to_utf8_string
+{
+    my $unicode_string = shift(@_);
+
+    my @unicode_array;
+    for (my $i = 0; $i < length($unicode_string); $i++) {
+    push(@unicode_array, ord(substr($unicode_string, $i, 1)));
+    }
+    return &unicode::unicode2utf8(\@unicode_array);
 }
 
@@ -385 +407 @@
     my $lastpartitionend = shift(@_);
 
-    my $partitionstart = &unicode::substr($metavalue, 0, 1);
+    my $partitionstart = substr($metavalue, 0, 1);
     if ($partitionstart le $lastpartitionend) {
-    $partitionstart = &unicode::substr($metavalue, 0, 2);
+    $partitionstart = substr($metavalue, 0, 2);
     # Give up after three characters
     if ($partitionstart le $lastpartitionend) {
-        $partitionstart = &unicode::substr($metavalue, 0, 3);
+        $partitionstart = substr($metavalue, 0, 3);
     }
     }
@@ -403 +425 @@
     my $partitionstart = shift(@_);
 
-    my $partitionend = &unicode::substr($metavalue, 0, &unicode_length($partitionstart));
+    my $partitionend = substr($metavalue, 0, length($partitionstart));
     if ($partitionend gt $partitionstart) {
-    $partitionend = &unicode::substr($metavalue, 0, 1);
+    $partitionend = substr($metavalue, 0, 1);
     if ($partitionend le $partitionstart) {
-        $partitionend = &unicode::substr($metavalue, 0, 2);
+        $partitionend = substr($metavalue, 0, 2);
         # Give up after three characters
         if ($partitionend le $partitionstart) {
-        $partitionend = &unicode::substr($metavalue, 0, 3);
+        $partitionend = substr($metavalue, 0, 3);
         }
     }
@@ -428 +450 @@
 
     # Create an hlist partition
-    my %child_classifier_node = ( 'Title' => $partitionname,
+    my %child_classifier_node = ( 'Title' => &convert_unicode_string_to_utf8_string($partitionname),
                   'childtype' => "VList",
                   'contains' => [] );
@@ -471 +493 @@
     # Otherwise create a sublist (bookshelf) for the metadata value
     else {
-        my %child_classifier_node = ( 'Title' => $metavalue,
+        my %child_classifier_node = ( 'Title' => &convert_unicode_string_to_utf8_string($metavalue),
                       'childtype' => "VList",
                       'contains' => [] );