I don't get it. If we were missing writes, the logical thing would be to reenable fixKeys, or even try to write to all registered keys.
Might be responsible for the extremely long send times we see now... but
reports from users suggest that 5018+ has serious problems that previous
builds didn't, so maybe it is a recent phenomenon. I suspect a bug in
the queueing code.
On Sat, Aug 02, 2003 at 04:34:09PM -0700, Zlatin Balevsky wrote:
> Update of /cvsroot/freenet/freenet/src/freenet/transport
> In directory sc8-pr-cvs1:/tmp/cvs-serv5728/src/freenet/transport
>
> Modified Files:
> WriteSelectorLoop.java
> Log Message:
> hmm. perhaps we _were_ missing writes..
>
> Index: WriteSelectorLoop.java
> ===================================================================
> RCS file: /cvsroot/freenet/freenet/src/freenet/transport/WriteSelectorLoop.java,v
> retrieving revision 1.37
> retrieving revision 1.38
> diff -u -r1.37 -r1.38
> --- WriteSelectorLoop.java 2 Aug 2003 21:42:49 -0000 1.37
> +++ WriteSelectorLoop.java 2 Aug 2003 23:34:07 -0000 1.38
> @@ -526,7 +526,7 @@
> sent = ((SocketChannel)curKey.channel()).
> write(currentJob.data);
> //this should be here, not in finally
> - if (sent==0) currentSet.remove(curKey);
> + //if (sent==0) currentSet.remove(curKey);
> } finally {
> if(currentJob.data.limit() != oldLimit)
> currentJob.data.limit(oldLimit);
>
> _______________________________________________
> cvs mailing list
> [EMAIL PROTECTED]
> http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/cvs
--
Matthew J Toseland - [EMAIL PROTECTED]
Freenet Project Official Codemonkey - http://freenetproject.org/
ICTHUS - Nothing is impossible. Our Boss says so.
pgp00000.pgp
Description: PGP signature
