| Subject: | Parameter entity references in internal subset break parser? |
Hi,
the first occurrence of a parameter reference in the internal subset
fires the default handler. After that the default handler is called for
everything instead of the dedicated handlers specified in the
constructor to XML::Parser.
I wouldn't bet my life on this not being an error in my XML instead of a
bug in XML::Parser but all other parsers I have tested (xmllint, Xerces)
do not report any well-formedness errors and process everything as expected.
The following script (also attached) illustrates the problem. Compare
the behavior of the original version to the one where "%common;" removed.
It actually doesn't matter whether the file "common.txt" exists. I also
didn't see any difference, when NoExpand was set to false.
Regards,
Guido
use strict;
use XML::Parser;
use Data::Dumper;
my $xml = <<EOF;
<!DOCTYPE mytype [
<!ENTITY % common SYSTEM "common.txt">
%common; <!-- Remove this line for comparison! -->
<!ATTLIST mytype foo CDATA "bar">
]>
<mytype foo="bar"/>
EOF
sub default_handler { shift; print "Default: ", Dumper \@_ }
sub catchall_handler { shift; print "Catch all: ", Dumper \@_ }
my $p = XML::Parser->new (NoExpand => 1,
Handlers => {
Default => \&default_handler,
Doctype => \&catchall_handler,
DoctypeFin => \&catchall_handler,
Attlist => \&catchall_handler,
});
$p->parse($xml);
| Subject: | xml-parser.pl |
use strict;
use XML::Parser;
use Data::Dumper;
my $xml = <<EOF;
<!DOCTYPE mytype [
<!ENTITY % common SYSTEM "common.txt">
%common;
<!ATTLIST mytype foo CDATA "bar">
]>
<mytype foo="bar"/>
EOF
sub default_handler { shift; print "Default: ", Dumper \@_ }
sub catchall_handler { shift; print "Catch all: ", Dumper \@_ }
my $p = XML::Parser->new (NoExpand => 1,
Handlers => {
Default => \&default_handler,
Doctype => \&catchall_handler,
DoctypeFin => \&catchall_handler,
Attlist => \&catchall_handler,
});
$p->parse($xml);