XML::Twig - resolving parameter entity decalarations

jxh has asked for the wisdom of the Perl Monks concerning the following question:

Consider :

#!/usr/bin/perl -w

use strict;

use XML::Twig;

my $twig;
my $xml;

$twig=XML::Twig->new(
        pretty_print => 'record_c',
        NoLWP => 1,
        keep_spaces => 1,
        load_DTD => 1,
        expand_external_ents => 1,
        keep_encoding => 1,
);

{
        local $/;
        $xml = <DATA>;
}

$twig->parse("$xml");
$twig->print


__DATA__
<?xml version="1.0"?>
<!DOCTYPE garden SYSTEM "./garden.dtd">
<garden>
        <plant>&flower;</plant>
</garden>
[download]

My external dtd, garden.dtd looks like this :

<!ENTITY flower "%flower_ref;">
<!ENTITY % flower_ref "daffodil">
[download]

When I run my code, I get : illegal parameter entity reference at line 2, column 17, byte 36 at /usr/local/lib/perl5/site_perl/5.6.1/sun4-solaris/XML/Parser.pm line 185
Why is this ?
I am looking to have all referaances resolved when printing...
Many Thanks.

Comment on XML::Twig - resolving parameter entity decalarations Select or Download Code

Replies are listed 'Best First'.
Re: XML::Twig - resolving parameter entity decalarations by mirod (Canon) on May 03, 2006 at 17:07 UTC
The XML might not well-formed. I can't tell you why off the top of my head, I haven't used parameter entities in a while, but if you put the xml in a separate file and if you run `xmllint` (the XML checker that comes with libxml2) it complains that the entity flower is not defined. That said, when I add the proper element declarations and when I reverse the order of the 2 entity declarations, xmllint is quite happy, so it looks indeed like XML::Twig does not deal properly with parameter entities, I have to check, I haven't touched that corner of the code in years. Actually, after looking a little deeper, it looks like it's an XML::Parser problem: when loading the external DTD, XML::Twig parses a fake document that's just the DTD and a dummy element, in order to get the entity values, otherwise XML::Parser does not read the external DTD at all. But when I parse the document created from the DTD, either with XML::Parser or just with `xmlwf` (the xml checker that comes with expat), I get an "illegal parameter entity reference" error. The document is: `<!DOCTYPE dummy [ <!ENTITY % flower_ref "daffodil"> <!ENTITY flower "%flower_ref;"> ]> <dummy/>` [download] So this looks a bit tricky to fix. If you have any idea, I'll take it ;--(	[reply] [d/l]

Replies are listed 'Best First'.

Re: XML::Twig - resolving parameter entity decalarations
by mirod (Canon) on May 03, 2006 at 17:07 UTC

The XML might not well-formed.

I can't tell you why off the top of my head, I haven't used parameter entities in a while, but if you put the xml in a separate file and if you run xmllint (the XML checker that comes with libxml2) it complains that the entity flower is not defined.

That said, when I add the proper element declarations and when I reverse the order of the 2 entity declarations, xmllint is quite happy, so it looks indeed like XML::Twig does not deal properly with parameter entities, I have to check, I haven't touched that corner of the code in years.

Actually, after looking a little deeper, it looks like it's an XML::Parser problem: when loading the external DTD, XML::Twig parses a fake document that's just the DTD and a dummy element, in order to get the entity values, otherwise XML::Parser does not read the external DTD at all.

But when I parse the document created from the DTD, either with XML::Parser or just with xmlwf (the xml checker that comes with expat), I get an "illegal parameter entity reference" error.

The document is:

<!DOCTYPE dummy [
<!ENTITY % flower_ref "daffodil">
<!ENTITY flower "%flower_ref;">
]>
<dummy/>
[download]

So this looks a bit tricky to fix. If you have any idea, I'll take it ;--(

[reply]
[d/l]