Re: XML::Twig - Memory usage problem

mirod wrote:
> howard chen wrote:
> 
>> On Wed, Nov 5, 2008 at 1:07 AM, mirod <[email protected]> wrote:
>>> If your XML includes id attributes, you can either use the id => undef option
>>> when you create the twig, or use the updated version on my site.
>> Are you sure problem only exist with "id attributes" ? I don't have
>> the "id" attribute in my example XML provided. (In fact, no attributes
>> used at my XML),
>> However, I do have a node called "company_id". Don't know if it is related.
> 
> It is not related. It's really only with an id attribute.
> 
> When there are id attributes in the XML (given by the id option or just 'id' by 
> default), then XL::Twig keeps a hash { id => element } . This way you can get 
> efficiently the element from the id, using elt_id (or getElementById) on the 
> twig. The value, the reference to the element, is weakened, so this doesn't 
> prevent the element to be destroyed by a purge, but the entry in the hash 
> remained, which can add up to quite a bit of wasted memory after a while.
> 
> If you don't have id's, then I still haven't been able to reproduce the leak. 
> Back to the drawing board I guess.

I am afraid I can't reproduce the leak. The code I am testing is the one you 
sent in your first email, but I see no leak with it.

Could you maybe give me some more details on what you do in the company sub? And 
the result of perl -v? The system I test on gives 'This is perl, v5.8.8 built 
for x86_64-linux-thread-multi'.

Thanks


#!/usr/bin/perl

use strict;
use warnings;

use XML::Twig;

my $doc= shift @ARGV;

print "testing more complex handler on $doc ";
END { print "- size: ", memory_size(), "\n"; }

# display memory size
sub memory_size
{ open( my $status, '<', "/proc/$$/status") or die "cannot find status info in 
/proc/$$/status: $!";
   my( $size)= map { m{^VmSize:\s+(\d+\s+\w+)} } <$status>;
   $size=~ s{(\d\d\d )}{ $1};
   return $size;
}

my $tx= TextXML->new;
$tx->parse( $doc);

package TextXML;

sub new { return bless {}; }

sub parse {
     my( $self, $doc)= @_;
     my $t = XML::Twig->new(
             twig_handlers => {
                     company => sub {
                             $self->company( $_[0], $_[1] );
                             $_[0]->purge;
                     },
             }
     );

     $t->parsefile( $doc );
     $t->purge;
}

sub company {
     my ( $self, $t, $elt ) = @_;
     my @data= map { $_->tag => $_->text } $elt->children;
     $elt->first_child( 'company_id')->subs_text( qr{^c(\d+)}, '$elt( c => $1)');
}
_______________________________________________
Perl-XML mailing list
[email protected]
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs