nutch-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
Subject Re: Huge Problem trying to develop plugin for Nutch
Date Sat, 26 Mar 2005 21:49:18 GMT
Hi Stefan,

 Thanks for the tip. Yeah, my plugin.xml file looks just like yours:

<?xml version="1.0" encoding="UTF-8" ?> 
<plugin id="parse-rss" name="RSS Parse Plug-in" version="1.0.0" provider-name="edu.usc.cs599">
  <extension-point id="net.nutch.parse.Parser" name="Nutch Content Parser" /> 
<library name="parse-rss.jar">
  <export name="*" /> 
  <library name="jdom.jar" /> 
  <library name="jaxen-full.jar" /> 
  <library name="commons-feedparser-0.5-beta.jar" /> 
  <library name="saxpath.jar" /> 
  <library name="commons-httpclient-3.0-beta1.jar" /> 
  <library name="log4j-1.2.6.jar" /> 
 <extension id="net.nutch.parse.rss" name="RssParse" point="net.nutch.parse.Parser">
  <implementation id="net.nutch.parse.rss.RSSParser" class="net.nutch.parse.rss.RSSParser"
contentType="text/xml" pathSuffix="" /> 

The strange this is, if you look at my program that I sent to the list, I made
it do like the same thing as the PluginClassLoader class in the net/nutch/plugin package,
except I used the URLClassLoader (which the PluginClassLoader is based on). I dynamically
was able to load all the necessary jar files in my program and then run the feedparser.
The problem is when I try to do the same thing in nutch, with the exact same location of the
jar files on my system, with the exact same code, I get the error about it not being able
to find org/jdom/Document. If you look at the log file that I sent, the nutch crawler's PluginRepository
was able to parse my plugin's plugin.xml file, get all the jar file pointers out of it's "runtime"
element, and then load them via the PluginClassLoader. Those jar files included the jdom.jar
file, which I tested (via the jar tvf command) to ensure that it has the org/jdom/Document
class (and of course I know it has that class in !
 there cuz it works from my program). So yeah, the weird thing is, for whatever
reason, the feedparser library can't find the jdom class when it's run from inside my parse-rss

Any help on this would be greatly appreciated.

Thanks in advance.

  Chris Mattmann

----- Original Message -----
From: Stefan Groschupf <>
Date: Saturday, March 26, 2005 6:17 am
Subject: Re: Huge Problem trying to develop plugin for Nutch

> Do you have the jars configured in your plugin xml?
> Like:
>  <runtime>
>       <library name="parse-msword.jar">
>          <export name="*"/>
>       </library>
>       <library name="poi-2.1-20040508.jar"/>
>       <library name="poi-scratchpad-2.1-20040508.jar"/>
>    </runtime>
> You can have internal jars and jars you export.
> Internal jars are only available to the plugin classloader of your 
> plugin but exported jars are available for other plugin 
> classloaders as 
> well.
> See:
> Does this help?
> Stefan

View raw message