Quoting Toad <[EMAIL PROTECTED]>:
> On Tue, Jun 10, 2003 at 03:01:48PM -0700, [EMAIL PROTECTED] wrote:
> > Hopefully, we set our goal to be that every node gives to Freenet as much
> > as it can, and takes from Freenet all that it can. Surely optimizing for
> > some perceived set of conditions i
On Tue, Jun 10, 2003 at 02:01:30PM -0700, Ian Clarke wrote:
> > maxNodeConnections - Raise this. A _lot_. That's what nio is all about ;-)
>
> It is worth noting that having more references in the DS may not always
> be a good thing - as if a node has too many references then each
> individual
On Wed, Jun 11, 2003 at 02:54:41PM +0100, Kevin Steen wrote:
> At 10/06/2003 20:44, you wrote:
> >More recent changes to the code complete implementation of bandwidth
> >limiting on both input and output (testers would be appreciated on this
> >particular issue). One issue here is that the overall
On Tue, Jun 10, 2003 at 03:01:48PM -0700, [EMAIL PROTECTED] wrote:
> Quoting Toad <[EMAIL PROTECTED]>:
>
> > 56k modems do not contribute anything to the network, except perhaps
> > actual content. I don't want to make freenet particularly hostile to
> > transient nodes on 56k links, but it's not
I'm not saying to optimize for 56k. I'nm saying, don't optimize for something if it's going to be at the expense of something else, in this situation.
I don't belive I get what you're saying, but this might explain my viewpoint. If you optimize for one thing at the expense of another, in this situation, that is bad.
He didn't exactly say that 56k modems contribute nothing to the network. In
reality, 56k modems do not contribute to the overall health and functioning
of Freenet in general. Reasons for this should be obvious.
Since Freenet has more pressing issues that affect *all* users (and not just
users
Interesting log. Getting NullPointerExceptions on what should be valid
InetAddress objects. Looks like a libgcj bug. Any ideas?
11-Jun-03 11:43:48 AM (freenet.transport.tcpAddress, QThread-2, DEBUG):
getHostAddress() took 0 ms
11-Jun-03 11:43:48 AM (freenet.transport.tcpConnection, QThread-2, DE
Alright, an update, I've got fproxy to work. You need to copy
'(CVS)/freenet/node/http/templates' to '/freenet/node/http/templates' in the
root directory of your filesystem. Quick hack I know, and it may only work on
Unix(es) but it seems to work. No bugs in any of the servlets, that I can
see.
Haven't got it to work perfectly yet. There's the HTML template issue
preventing fproxy or other HTTP servlets from working at all, and it seems to
dump zombie processes all over the place. Also, there's a strange problem
with Frost - the node seems to automatically return a blank/corrupted file
At 10/06/2003 20:44, you wrote:
>More recent changes to the code complete implementation of bandwidth
>limiting on both input and output (testers would be appreciated on this
>particular issue). One issue here is that the overall bandwidthLimit and
>averageBandwidthLimit settings are no longer supp
> However, there is still a problem with the HTML templates. Fred tries to
> access the templates which are stored by default in
> "/freenet/node/http/templates". These are contained in freenet.jar, but
> I'm not sure how to implement this in gcj, with its binary executables.
Do it anyway you feel
> I had to remove SimpleContext.java and InsertServlet.java from the
> Makefile since they no longer seem to exist in the source tree.
> And I added MultipleFileInfolet.java, RTInfolet.java,
> ConfigUpdateListener.java, InvertPrivateKeyRequest.java,
> InverPrivateKey.java, NewInvertPrivateKey.java
Hi,
On Tue, Jun 10, 2003 at 05:24:19PM -0500, Tiernan Hubble wrote:
> On Tuesday 10 June 2003 15:49, Mark Wielaard wrote:
> > Could you send the patch?
> >
> Attaching, hope it works. (I also downgraded -O2 to -O since that causes
> problems with some files, and added -Wno-deprecated to disable an
14 matches
Mail list logo