Changeset 1891 for trunk/gsdl

Timestamp:

2001-02-01T15:26:44+13:00 (23 years ago)

Author:

paynter

Message:

Named characters like é and ì are translated
to UTF8 internally.

File:

• trunk/gsdl/perllib/plugins/HTMLPlug.pm (modified) (3 diffs)

trunk/gsdl/perllib/plugins/HTMLPlug.pm

@@ -40 +40 @@
 use BasPlug;
 use ghtml;
+use unicode;
 use util;
 use parsargv;
@@ -165 +166 @@
     return 1;
 }
+
+
+
 
 sub replace_images {
@@ -520 +524 @@
 }
 
+
+# Extend the BasPlug read_file so that strings like é are
+# converted to UTF8 internally.  
+#
+# We don't convert < or > or & or " in case
+# they interfere with the GML files
+
+sub read_file {
+    my ($self, $filename, $encoding, $textref) = @_;
+
+    &BasPlug::read_file($self, $filename, $encoding, $textref);
+    
+    # Convert things like é to their UTF8 equivalents
+    $$textref =~ s/&(lt|gt|amp|quot);/&z$1;/go;
+    $$textref =~ s/&([^;]+);/&unicode::ascii2utf8(\&ghtml::getcharequiv($1,1))/gseo;
+    $$textref =~ s/&z(lt|gt|amp|quot);/&$1;/go;
+}
+
 1;