Version: 1.0.1+ #3210 svn-rev: 689476 (using KDE KDE 3.5.7) Installed from: Gentoo Packages Compiler: gcc-Version 4.1.2 (Gentoo 4.1.2) CXXFLAGS=-march=nocona -O2 -pipe -g OS: Linux My bouncer is on a normal DSL line with a 24h disconnect. When I connect to the bouncer after such a reconnect has happened, the backlog is replayed until the disconnect, most of the chans vanish and are not "recreated". Only the "last" chan in the list stays there and is not lost. Eg i have this list of chans: #wesnoth, #wesnoth-de, #wesnoth-dev and #konversation active on my bouncer, when connecting to it in the morning shortly all open, then alls but #konversation are closed again, in #konversation the complete backlog of the night is available, when looking at the plain logfiles I see that logging for all the other chans stopped when reaching the disconnect in the backlog. The chan #konversation is just one example in this case, when not connected to #konversation, too, then #wesnoth-dev is the only one "active" after the reconnects. This problem did already exist in 1.0.1, too, there it sometimes came together with a crash (cf bug #145614) but that crash seems to have been solved in svn. There is no real problem when being logged in into the bouncer when the reconnect happens, only when the backlog is replayed. Some system/bouncer data: I am running konversation from svn rev 689476 (that is version 1.0.1+ #3210). My bouncer is miau, version 0.6.4 running on a slug/nslu2 (version from the unslug package manager ipkg: miau 0.6.4-1). Do you need any more information?
Okay, today I had 5 instead of only 4 chans open and all beside the servertab were lost. So it seems to not be possible to tell exactly which chans might be lost...
what's chandiscon set to in your miaurc? If it's set to part try setting it to something else.
It was set to the default value, I will now try to set it to "nothing", I hope this works nicely. If it does I will tell the miau dev to change to comment to not only mention mirc but also konversation in that list. (Of course I will report back in here if it does work nicely then)
No, this change did not help, still exactly the same problem even though I did change chandiscon from 'part' to 'nothing'.
Created attachment 21220 [details] configuration file for miau (the bouncer in question) This bouncer config might help to reproduce the problems. I have the impression that quite some traffic has to take place after a disconnect of the bouncer so that the chans are lost when attaching to the bouncer. I did change all nickname/passwddata, the rest is identical to my config.
Created attachment 21348 [details] A ralog where I forced to create such a problem Okay, I did now "force" the problem to occur. This is what I have done: *Detach the bouncer *Force a reconnect to the server but "resetting" my dsl line which results in a serverdisconnect because of a changed IP *Create some traffic in the chans I was in (some chat was taking place in #wesnoth-dev, some in #konversation) *After about 1:30 after the server reconnect: reattach the bouncer in konversation with active rawlog and copy the rawlog to a file that is now attached in here In the file I only changed the passwords, the rest is an unchanged rawlog.
Forgot to mention: The last line of log konversation has in it's own "normal" log files is that one, everything afterwards is not in there: [15:45:19] >> :Ivanovic!n=ivanovic@pD9EC3CD9.dip0.t-ipconnect.de QUIT :[14:27:26] Read error: 110 (Connection timed out)
Created attachment 21363 [details] Rawlog of Konversation having the problem I will attach two corresponding logs, one of Konversation while replaying the backlog (and have the tabs vanish) and one from miau including all the traffic it had that night.
Created attachment 21364 [details] corresponding debug log from miau, including all traffic (in and out)
Is this still an issue for you with current versions of miau and Konversation?
Nope, the issue is fixed by now.
Thanks for getting back to us!