XML::LibXML, svg, default namespace: Why so slow?

7stud has asked for the wisdom of the Perl Monks concerning the following question:

Dear Monks,

libxml is supposed to be fast, but I am trying to alter a small XML file, and XML::LibXML is intolerably slow:

use strict;
use warnings;
use 5.016;
use XML::LibXML;


my $dom = XML::LibXML->load_xml(string => <<'END_OF_XML');
<?xml version="1.0" encoding="UTF-8" standalone="no" ?>
<!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 1.1//EN" "http://www.w3.org/Grap
+hics/SVG/1.1/DTD/svg11.dtd">
<svg xmlns="http://www.w3.org/2000/svg" 
     xmlns:xlink="http://www.w3.org/1999/xlink" 
     version="1.1">    
  <tag> 
    <text>
      <tspan>Hello</tspan>
    </text>
  </tag>
</svg>
END_OF_XML

#Deal with the namespace declared in the svg tag:

my $xpc = XML::LibXML::XPathContext->new($dom); 
$xpc->registerNs('ns', 'http://www.w3.org/2000/svg');
my ($tspan) = $xpc->findnodes('//ns:text/ns:tspan');

#Change the tspan's text:

$tspan->removeChildNodes();
$tspan->appendTextNode("Goodbye");

#Prove that the tspan's text changed:

say $tspan; 
say $dom->toString;
[download]

Am I doing something wrong?

Comment on XML::LibXML, svg, default namespace: Why so slow? Download Code

Replies are listed 'Best First'.
Re: XML::LibXML, svg, default namespace: Why so slow? by choroba (Cardinal) on Jan 03, 2015 at 22:10 UTC
When I ran your script, it seemed tolerably fast (0.056s). But, when I removed the DOCTYPE declaration, it got even faster: 0.043s. When I put it back but changed the URL to a non-existent one, the script started to crawl at 0.447s. So, maybe, check your connection speed? لսႽ† ᥲᥒ⚪⟊Ⴙᘓᖇ Ꮅᘓᖇ⎱ Ⴙᥲ𝇋ƙᘓᖇ	[reply]
Re: XML::LibXML, svg, default namespace: Why so slow? by 7stud (Deacon) on Jan 04, 2015 at 00:00 UTC
Thanks. I wasn't timing my program because it took approximately 5 seconds to run. I thought I tried removing the DOCTYPE to see if that was the problem, but now when I remove it, the results appear instantaneously, so you were right: the problem was my connection.	[reply]
Re: XML::LibXML, svg, default namespace: Why so slow? by locked_user sundialsvc4 (Abbot) on Jan 04, 2015 at 02:49 UTC
Hmmm ... when presented with a `DOCTYPE` (which of course references a URL at `w3.org` as its means of defining the type), does anything attempt to reach-out to the Internet to retrieve that DTD? If it did, and failed, a timeout could occur. I quite-honestly don’t recall ... But, if it is “noticeably slow,” then it obviously can’t be the job of XML-parsing. It must be “something else.” Like this, maybe.