[Dtdparse-discuss] Patches for XML::DTDParse::Catalog

A bug with dtdparse is that it does not handle whitespace properly
in system identifiers.  For example, say the following entity
declaration exists:

  <!ENTITY % parm.ent PUBLIC "-//Earl Hood//ENTITIES
			      A newline in the PubID//EN">

The newline is kept when parsed, but when dtdparse attempts to
resolve the pubid, it cannot since the newline/whitespace is
not compressed when doing the lookup in the catalog.

An immediate fix to the problem is the following patch:

<patch>
--- Catalog.pm.org	Sat Feb 15 19:21:53 2003
+++ Catalog.pm	Sat Feb 15 19:27:33 2003
@@ -84,6 +84,7 @@ sub system_map {
 sub public_map {
     my($self, $pubid) = @_;
 
+    $pubid =~ s/\s+/ /g;
     return $self->_find('PUBID', $pubid);
 }
 
@@ -111,6 +112,7 @@ sub reverse_public_map {
 sub declaration {
     my($self, $pubid) = @_;
 
+    $pubid =~ s/\s+/ /g;
     foreach my $dir (@{$self->{'DIRECTIVE'}}) {
 	my %hash = %{$dir};
 
</patch>

Years ago, I wrote a complete(?) SGML Open Catalog parser module.
If there is any interest in it, just let me know.

--ewh