On Thursday 04 November 2010 23:03:06 Robert Hailey wrote: > > On 2010/11/04 (Nov), at 11:06 AM, Matthew Toseland wrote: > > > On Wednesday 03 November 2010 20:00:48 Robert Hailey wrote: > >> > >> Example patch attached. Small, but untested! > > > > Could you please explain what it is that you think is going on, and > > what the patch will do to fix it? > > I would still like to discuss/theorize on the recent topics, but I've > observed some more unexpected behavior and think I found a bigger and > more directly-responsible bug: > > Suppose our node is mostly-idle > That is, none are in a opennet grace period (which would keep them > connected) > We lose one peer... > A short time later, the node starts an announcer to "top off" our peer > connections (+1 is easy, right?). No... > Twelve announcement replies come back and displace all of our veteran > peers
No. We don't announce when we are 1 peer from the maximum. We only announce when we have very few peers. See Announcer.enoughPeers. > > (patch attached) No. Refusing to ever replace a connected peer with an announced peer is unacceptable IMHO. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part. URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20101105/2f2a8765/attachment.pgp>
