XML::Parser breaks on

r.joseph has asked for the wisdom of the Perl Monks concerning the following question:

Hello again everyone,

Been a long time since my last post, but I am really stuck this time, and can't figure why.

I am using XML::Parser to parse .RSS documents from Linux.com and Newsforge.net for a "live news feed" if you will. First, below is a snippet from Linux.com's RSS doc:

<item>
<title>Big software companies lose their minds!</title>
<link>http://linux.com/newsitem.phtml?sid=1&amp;aid=12492</link>
<description>Linux.com corresponent Mark Miller has some
views on big software companies.</description>
</item>
[download]

Now, if you look at the <link> tag, you see that there is a proper sequence, & to represent an ampersand. However, here is the problem. When XML::Parser encounters this chunk of data, it calls the Char handler, whatever you define it to be. Mine happens to be very simple, atleast right now (BTW, I am using the Subs style for the parser, but tha shouldn't matter):

sub found_char {
my ($ex, $str) = @_;
if ($ex->in_element('link') &&
      $ex->within_element('item')) {
 print "\t\tLink: $str\n";
}
}
[download]

So I should expect a simple string that has Link: and then the link, whatever that may be. However, it seems that XML::Parser instead, for some reason, splits on that escape sequence, so I get this output:

Link: http://linux.com/newsitem.phtml?sid=1
Link: &
Link: aid=12492
[download]

What I CANNOT figure out is why it seems to consider that string within the <link> element three strings!

Does anyone know how this can be fixed - I have seen this problem happen with other "non-element" data, and I just want it to grab all of the pertient data at one time.

Thanks a ton!

r. j o s e p h
"Violence is a last resort of the incompetent" - Salvor Hardin, Foundation by Issac Asimov

Comment on XML::Parser breaks on Select or Download Code

Replies are listed 'Best First'.
Re: XML::Parser breaks on by mirod (Canon) on Aug 16, 2001 at 12:09 UTC
This is a documented behaviour of XML::Parser. Actually XML::Parser documents the fact that this "can" happen. It actually happens for every entity, line break and expat input buffer boundary crossed. The review gives you a way to deal with it: basically you cannot use the data in the char handler, you just buffer it until you hit a tag (open _or_ close). By the way, did you try XML::RSS? Maybe it would make it easier for you to process your data.	[reply]
Re: XML::Parser breaks on by blakem (Monsignor) on Aug 16, 2001 at 12:14 UTC
I think you'll have to glue it all together. Here is a snippet that might help a bit. `sub xml_char { my ($xp, $txt) = @_; my $el = $xp->current_element(); $val{$el} .= $txt if $txt =~ /\S/; }` [download] Notice that %val is sort of a buffer area that will need to be cleaned up when you hit the end tag. -Blake	[reply] [d/l]
Re: Re: XML::Parser breaks on by clintp (Curate) on Aug 16, 2001 at 17:31 UTC
I've used something like this before, and like the general technique, but I question: `$val{$el}.=$txt if $txt =~ /\S/;` [download] Since the parser is actually allowed to break anywhere you could lose intra-word spaces or newlines (if they're signicficant).	[reply] [d/l]
Re: Re: Re: XML::Parser breaks on by blakem (Monsignor) on Aug 17, 2001 at 02:10 UTC
You're probably right and its a piece of code I haven't looked at in a long time. I do remember there being a reson for it, but can't remember it right now. Anyone looking at using this, should probably get rid of the `if` conditional. -Blake	[reply] [d/l]


The stupid question is the question not asked
	PerlMonks