Firstly, jira seems to be down so I'm replying by email again, I'll
update jira when it's back up.

I see your point and I see the code for it (339-350 right?) but I'm
still getting the error listed below, which disapears when I apply the
patch. I guess the problem must be in the adding of new Writables to
the map.

WARN crawl.MapWritable: Unable to load meta data entry, ignoring.. :
java.io.IOException: unable to load class for id: 36

On 7/25/07, Doğacan Güney <[EMAIL PROTECTED]> wrote:
> On 7/25/07, Robert Young <[EMAIL PROTECTED]> wrote:
> > The message which was appearing in the logs is pasted below.
> >
> > Basically, in org.apache.nutch.crawl.MapWritable#getKeyValueEntry the
> > Writable is instantiated. It's class is determined by a two byte code
> > (which is written to crawldb I guess), if there is no entry for the
> > class it fails to create it, regardless of if it's a Writable. You're
> > right in that it can potentially handle any Writabl object, but only
> > if it has a maping for it's class.
>
> If you add a writable that does not have a mapping, MapWritable
> automatically creates it and then stores the mapping internally. When
> a MapWritable is written, any new mapping is also written (as a byte
> and the corresponding class name). So, when you read a MapWritable it
> first reads all the mappings (that are not already statically defined)
> then proceeds to reading the <Writable,Writable> map. So I think your
> problem is caused by something else (perhaps, there is a bug in
> MapWritable's implementation but that is what that code is trying to
> do).
>
> Also, replying to JIRA generated emails does not add comments to
> issues (despite what the email is saying). So please use JIRA to
> reply.
>
> >
> > Cheers
> > Rob
> >
> > 07/07/25 11:52:00 WARN crawl.MapWritable: Unable to load meta data
> > entry, ignoring.. : java.io.IOException: unable to load class for id:
> > 36
> >
> > On 7/25/07, Doğacan Güney (JIRA) <[EMAIL PROTECTED]> wrote:
> > >
> > >     [ 
> > > https://issues.apache.org/jira/browse/NUTCH-527?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12515283
> > >  ]
> > >
> > > Doğacan Güney commented on NUTCH-527:
> > > -------------------------------------
> > >
> > > What was the error you were having? MapWritable supports reading and 
> > > writing *all* writables. The ones defined at the top of the file are an 
> > > optimization and shouldn't affect correctness (basically mapwritable 
> > > first writes a byte and the associated classname, then writes that byte 
> > > to indicate classname everywhere else. For commonly used types we 
> > > statically define the association so that "first write the byte then the 
> > > classname" phase is not necessary).
> > >
> > > > MapWritable doesn't support all hadoops writable types
> > > > ------------------------------------------------------
> > > >
> > > >                 Key: NUTCH-527
> > > >                 URL: https://issues.apache.org/jira/browse/NUTCH-527
> > > >             Project: Nutch
> > > >          Issue Type: Bug
> > > >    Affects Versions: 0.9.0
> > > >         Environment: Tested on Solaris and Windows with Java 1.5
> > > >            Reporter: Rob Young
> > > >         Attachments: mapwritable.patch
> > > >
> > > >
> > > > The map of classes which implement org.apache.hadoop.io.Writable is not 
> > > > complete. It does not, for example, include 
> > > > org.apache.hadoop.io.BooleanWritable. I would happily provide a patch 
> > > > if someone would explain what the Byte parameter is.
> > >
> > > --
> > > This message is automatically generated by JIRA.
> > > -
> > > You can reply to this email to add a comment to the issue online.
> > >
> > >
> >
>
>
> --
> Doğacan Güney
>
-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
Nutch-developers mailing list
Nutch-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nutch-developers

Reply via email to