On Sat, Apr 07, 2001 at 08:39:48PM -0400, Tavin Cole wrote:
> On Sat, Apr 07, 2001 at 07:20:01PM -0500, Brandon wrote:
> > There's some bug in the new Client restructuring as it has broken by
> > code. It's also possible that it was broken before and changing my code
> > has just revealed its brokenness. However, since the stack trace
> > penetrates into FNPClient and I haven't changed
> > SimplifiedClient.dumpBuckets() lately, I think it's most likely a problem
> > with FNPClient. At the very least it should return a more informative
> > exception.
> > 
> > java.lang.NullPointerException
> >         at 
> > java.io.SequenceInputStream.nextStream(SequenceInputStream.java:98)
> >         at java.io.SequenceInputStream.<init>(SequenceInputStream.java:80)
> >         at 
> > Freenet.client.FNPClient$BInsertInstance.<init>(FNPClient.java:389)
> >         at Freenet.client.FNPClient.obtainClient(FNPClient.java:44)
> >         at 
> > Freenet.client.SimplifiedClient.dumpBuckets(SimplifiedClient.java:356)
> >         at 
> > Freenet.client.SimplifiedClient.dumpBuckets(SimplifiedClient.java:321)
> >         at Freenet.client.SimplifiedClient.put(SimplifiedClient.java:152)
> >         at Freenet.client.SimplifiedClient.put(SimplifiedClient.java:137)
> >         at 
> > Freenet.contrib.xmlrpc.LocalStreamingClient.endInsert(LocalStreamingClient.java:130)
> >         at 
> > Freenet.contrib.xmlrpc.LocalStreamingClient$Pump.run(LocalStreamingClient.java:224)
> 
> Actually, this was unrelated to the restructuring.  It's a result of my fixing
> Fred to compute the encryption key for CHKs from the whole data stream instead
> of the non-metadata part only.  You'll get this if you attempt to pass a
Just like libfreenet does now, right? Does that mean that you can move the
metadata boundary around and get the same CHK, as you can in libfreenet?
> NullBucket in for the metadata Bucket, which it seems like you're doing in
> LocalStreamingClient.  The best way to fix this, I think, is to fix NullBucket
> so that it does return an input stream instead of null.  This input stream
> would of course be length 0, giving an EOF on any read.
> 
> The work around is to just create a new empty FileBucket and pass that in.
>     
-- 
Always hardwire the explosives
        -- Fiona Dexter quoting Monkey, J. Gregory Keyes, Dark Genesis

_______________________________________________
Devl mailing list
Devl at freenetproject.org

Reply via email to