Version: 0.6.1a (using KDE KDE 3.1) Installed from: RedHat RPMs OS: Linux Had a rapidly growing .xsession-errors file on RH 8 system yesterday. Noticed Kopete was using lots of CPU (about 40% on Athlon 1.33 Ghz). I killed Kopete, and .xsession-errors errors stopped. Error message was: select: Bad file descriptor I was only running AIM. Prior to yesterday, no problems...installed this version when it first came out. I've reviewed dates on my RPMs, and I did upgrade some KDE components via apt on 3/11. At that time, I had some dependency problem with Kopete, so I uninstalled the RPM, and reinstalled it. Only change yesterday was my install of Samba, which occurred about the same time as the Kopete problem. samba-2.2.7-2 Sat 15 Mar 2003 01:47:53 PM PST openssl-devel-0.9.6b-31 Tue 11 Mar 2003 06:29:18 PM PST openssl-0.9.6b-31 Tue 11 Mar 2003 06:29:12 PM PST file-3.39-9 Tue 11 Mar 2003 06:29:10 PM PST kernel-2.4.18-26.8.0 Tue 11 Mar 2003 06:29:04 PM PST kopete-0.6.1a-4akc Tue 11 Mar 2003 05:14:45 PM PST xmms-skins-1.2.7-13 Tue 11 Mar 2003 05:12:11 PM PST xmms-mpg123-1.2.7-13 Tue 11 Mar 2003 05:12:10 PM PST xmms-1.2.7-13 Tue 11 Mar 2003 05:12:08 PM PST kdenetwork-3.1-3.1.8.0 Tue 11 Mar 2003 03:44:30 PM PST kdenetwork-devel-3.1-3.1.8.0 Tue 11 Mar 2003 03:44:26 PM PST redhat-artwork-0.68-1.0.8.0 Tue 11 Mar 2003 03:44:21 PM PST kdepim-3.1-3.1.8.0 Tue 11 Mar 2003 03:44:13 PM PST kdepim-devel-3.1-3.1.8.0 Tue 11 Mar 2003 03:44:08 PM PST kdegames-3.1-1.3.8.0 Tue 11 Mar 2003 03:44:01 PM PST kde-3.1-4.0.8.0 Tue 11 Mar 2003 03:43:51 PM PST qt-3.1.1-6.3.8.0 Tue 11 Mar 2003 03:43:47 PM PST qt-devel-3.1.1-6.3.8.0 Tue 11 Mar 2003 03:43:34 PM PST qt-designer-3.1.1-6.3.8.0 Tue 11 Mar 2003 03:43:24 PM PST
I have this error as well, with the only three differences being that i'm not using redhat, it uses 100% CPU time and i'm running a CVS version (HEAD before it became unstable)
Subject: Re: [Kopete-devel] Errors filling up .xsession-errors Is this related or comparable to the following bug reports? http://bugs.kde.org/show_bug.cgi?id=54520 http://bugs.kde.org/show_bug.cgi?id=54955
Subject: Errors filling up .xsession-errors > ------- Additional Comments From klingens@kde.org 2003-03-16 16:59 ------- > Subject: Re: [Kopete-devel] Errors filling up .xsession-errors > > Is this related or comparable to the following bug reports? > > http://bugs.kde.org/show_bug.cgi?id=54520 > http://bugs.kde.org/show_bug.cgi?id=54955 The problem that i'm having is more similar to 54955 than 54520 in that i can connect and use Kopete (and not have the problem) and some random amount of time after disconnecting (it's doesn't happen right after i disconnected) I start with select(): bad file descriptor errors I'm also running glibc 2.2.5 instead of 2.3.1 and the plugins that i was using when this happened were the AOL, ICQ, MSN, and Y! plugins (disabled the Y! plugin and it still did it)
*** Bug 56033 has been marked as a duplicate of this bug. ***
Subject: Re: Errors filling up .xsession-errors I've figured out how to reproduce this bug. I clicked 'Disconnect All' (to disconnect obviously). I had to click it three times to get it to disconnect all 4 protocols. The first click only disconnects Yahoo and ICQ (and if you let it go for too long, they reconnect again), the second click disconnects OSCAR, and the third one disconnects MSN, and starts the spewing of the select errors to .xsession-errors
Subject: Re: Errors filling up .xsession-errors > I've figured out how to reproduce this bug. I clicked 'Disconnect All' (to > disconnect obviously). I had to click it three times to get it to > disconnect all 4 protocols. The first click only disconnects Yahoo and ICQ > (and if you let it go for too long, they reconnect again), the second click > disconnects OSCAR, and the third one disconnects MSN, and starts the > spewing of the select errors to .xsession-errors oh yeah, this is with CVS head from yesterday.
I'm also experiencing this bug--my .xsession-errors fills up to many GB, causing me to run out of free disk space. Until there is a fix I've taken to deleting it every 5 minutes using cron, which seems to be working.
Workaround: link your .xsession-errors to /dev/null until the bug is fixed.
I'm using Jabber, Oscar, and ICQ with 0.6.1a (KDE 3.1.90 CVS >= 20030308), and I see this pretty regularly. When it starts giving the "select: Bad file descriptor " errors, and I quit it gives the following backtrace: (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)...[New Thread 16384 (LWP 31795)] [New Thread 32769 (LWP 31796)] (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...0x413d2b89 in wait4 () from /lib/libc.so.6 #0 0x413d2b89 in wait4 () from /lib/libc.so.6 #1 0x41446000 in sys_sigabbrev () from /lib/libc.so.6 #2 0x4121a061 in waitpid () from /lib/libpthread.so.0 #3 0x40a19793 in KCrash::defaultCrashHandler(int) () from /usr/lib/libkdecore.so.4 #4 0x4135f9d8 in sigaction () from /lib/libc.so.6 #5 0x40053d58 in KopeteContact::displayName() const () from /usr/lib/libkopete.so.1 #6 0x4006faf1 in KopeteMessageManagerFactory::cleanSessions(KopeteProtocol*) () from /usr/lib/libkopete.so.1 #7 0x4004ff56 in KopeteProtocol::unload() () from /usr/lib/libkopete.so.1 #8 0x4190a5b2 in OscarProtocol::unload() () from /usr/lib/kde3/kopete_oscar.so #9 0x40058b42 in LibraryLoader::remove(KopetePlugin*) () from /usr/lib/libkopete.so.1 #10 0x40056a0a in LibraryLoader::~LibraryLoader() () from /usr/lib/libkopete.so.1 #11 0x4005b134 in KStaticDeleter<LibraryLoader>::destructObject() () from /usr/lib/libkopete.so.1 #12 0x40a23e3d in KGlobal::deleteStaticDeleters() () from /usr/lib/libkdecore.so.4 #13 0x409b60a3 in KApplication::~KApplication() () from /usr/lib/libkdecore.so.4 #14 0x40a3118b in KUniqueApplication::~KUniqueApplication() () from /usr/lib/libkdecore.so.4 #15 0x080605cb in endl(kdbgstream&) () #16 0x0805e02c in KIconLoader::unknown() () #17 0x4134ea51 in __libc_start_main () from /lib/libc.so.6
>Workaround: link your .xsession-errors to /dev/null until the bug is fixed. Normally, i would do this, but i'm doing debugging on too many other apps to not have my .xsession-errors file.
Other workarounds: - Run Kopete from a console so output doesn't go to .xsession-errors but to stdout - Disable the debug area that causes the spam. Kopete uses some areas in the 14xxx range. In this particular case that might not work though, because the select() problem is more likely to be generated in kdelibs instead. Martijn
Subject: Re: Errors filling up .xsession-errors On Thursday 03 April 2003 03:18 pm, you wrote: > ------- You are receiving this mail because: ------- > You are on the CC list for the bug, or are watching someone who is. > You are a voter for the bug, or are watching someone who is. > > http://bugs.kde.org/show_bug.cgi?id=56028 > > > > > ------- Additional Comments From klingens@kde.org 2003-04-03 23:18 ------- > Other workarounds: > > - Run Kopete from a console so output doesn't go to .xsession-errors but to > stdout - Disable the debug area that causes the spam. Kopete uses some > areas in the 14xxx range. In this particular case that might not work > though, because the select() problem is more likely to be generated in > kdelibs instead. > > Martijn Actually, I think I'll just look into fixing it :-) But thanks for the other workarounds, i'm sure they'll help other people too.
I can't reproduce this bug anymore with either Kopete 0.6.2 or Kopete CVS HEAD, so i'll close it. Please feel free to reopen it if you still have this bug after upgrading to Kopete 0.6.2.
I still see it in 0.6.2 on Mandrake 9.1, after signing off from all IM services but leaving the program running for a while.
do you know how long you left it running? Before, this would happen almost immediately after signing off. (at least for me)
It was probably an hour or so after signing off; it's hard to say because I was away from the computer. It used to happen very soon after signing off for me too, but it definitely still happens.
as long as nobody has a some clear evidence for the protocol causing this the bug won't be fixed.
inference says that it's the OSCAR plugin, which could be why I don't see it anywith with Kopete HEAD. I'm in the process of rebuilding Kopete right now and will run tests soon.
As the OSCAR plugin in HEAD has changed quite a bit (and I've added checking for a connection in some places) this might be gone now. If it appears with CVS HEAD then please make sure you find out when it does so (connection lost != connection disconnected by user!).
this bug appears to be fixed in Kopete CVS HEAD. Will test tomorrow with Kopete 0.6.2.
Can't reproduce this bug with 0.6.2 at the moment, will wait overnight (approx. 8-10 hours from now to see if this problem still exists). If anyone else can reproduce, what steps do you take?
This problem has been fixed in Kopete CVS.