Mauricio Fernandez <[EMAIL PROTECTED]> writes: > On Thu, Sep 28, 2006 at 10:56:56PM -0700, Suraj N. Kurapati wrote: >> >> Both problems should now be fixed in r70. Please try that. >> > >> > My problem still exists: wmiirc freezes when retagging. >> >> I'm pretty confident that the concurrency bugs have been squashed in >> ruby-ixp, so it should be quite stable now. So instead, I will >> inspect the ruby-wmii branch that you are using, with the help of >> the logs you posted, to find the bug. > > I reproduced the problem with branch-ruby-ixp + Ruby-IXP r70 + > wmii-3.1 + wmii-3.1_libixp_overload_fix.patch. > > wmiirc would freeze when an IXP::Client was reused in a child process; I fixed > it by reopening the connection after a fork(). > > I have just pushed a couple patches to ruby-wmii (/branch-ruby-ixp) to > * sync against Ruby-IXP's svn (r70) > * fix the retagging/freezing bug > > Let's see how stable it is :) (and if any events are lost in practice).
Okay, I've been running the same instance of ruby-wmii for about 50 hours now and it feels quite stable. There have been no visible or feelable failures neither in wmiirc nor in wmii. However, it looks to me like there's a memory leak somewhere: ,---- | athene:~# ps aux | USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND | [...] | olli 7531 2.4 23.2 162384 150524 ? Sl Sep30 74:06 ruby /home/olli/.wmii-3/wmiirc | [...] `---- A freshly started wmiirc uses about 2.8% of the memory: ,---- | olli 13956 6.1 2.8 28624 18608 ? Ssl 14:34 0:09 ruby /home/olli/.wmii-3/wmiirc `---- I put the log file on my website, as it is quite big and I didn't want to attach it to this list: http://www.sopos.org/olli/wmiirc.log olli -- GnuPG-Key: http://www.sopos.org/pgp/olli.asc GnuPG-Fingerprint: F27A BA8C 1CFB B905 65A8 2544 0F07 B675 9A00 D827 NP: Muff Potter - "Unkaputtbar" _______________________________________________ wmii@wmii.de mailing list http://wmii.de/cgi-bin/mailman/listinfo/wmii