[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: kded died
From:       Steffen Hansen <stefh () mip ! sdu ! dk>
Date:       1999-06-22 16:59:41
[Download RAW message or body]

On Tue, 22 Jun 1999, Stephan Kulow wrote:

> Hi!
> 
> when I start kded (tried with kword before, but it
> dies also alone :) I get after a lot of other debugs:
[...]
> autoregistering service SpellTool
> ASSERT: "u == TC_KService" in
> /home/kulow/prod/kdelibs/kio/kregfactories.cc (205)
> 
> Here is the resulting stack trace (not too much information for me :)
> 
> #0  0xef1945b4 in XInternAtom () from /usr/openwin/lib/libX11.so.4
> #1  0x1aeb4 in remove_property ()
>     at /home/kulow/prod/kdelibs/corba/kded/kded-server.cc:50
> #2  0x1b11c in sighandler ()
>     at /home/kulow/prod/kdelibs/corba/kded/kded-server.cc:69
> #3  <signal handler called>
> #4  0xef1945b4 in XInternAtom () from /usr/openwin/lib/libX11.so.4
> #5  0x1aeb4 in remove_property ()
>     at /home/kulow/prod/kdelibs/corba/kded/kded-server.cc:50
> #6  0x1b11c in sighandler ()
>     at /home/kulow/prod/kdelibs/corba/kded/kded-server.cc:69
> #7  <signal handler called>
> #8  0xee587fe4 in _libc_kill () from /usr/lib/libc.so.1
> #9  0xee53a4cc in abort () from /usr/lib/libc.so.1
> #10 0xee9c6614 in __default_terminate () from
> /opt/local/lib/libmico2.2.7.so
> #11 0xee9c6648 in __terminate () from /opt/local/lib/libmico2.2.7.so
> #12 0xee9c7100 in __throw () from /opt/local/lib/libmico2.2.7.so
> #13 0xee8ea740 in CORBA::Exception::_terminate_handler ()
>    from /opt/local/lib/libmico2.2.7.so
> #14 0xee9c6648 in __terminate () from /opt/local/lib/libmico2.2.7.so
> #15 0xee9c7100 in __throw () from /opt/local/lib/libmico2.2.7.so
> #16 0x1b7f8 in main (argc=1, argv=0xeffff694)
>     at /home/kulow/prod/kdelibs/corba/kded/kded-server.cc:149

The trace doesn't really tell where it went wrong. Hmm.

> The interesting part happens when I try it again. It tells me then, 
> that there is already an instance running - so it died before the
> remove_property was successfull :(

Kind of weird. remove_property() is called for in the signal handler
that should handle everything catchable. So the property should be gone
after kded dies. I'll have to look at it when i get home.

> I noticed, that the property lists a PID. Isn't this a bit dangerous
> in X? I access the X server from the client and from the server I'm
> compiling on on a regular base, so it would be quite dangerous to
> assume the one I'm starting kword from is the same as the one I'm
> started kded from.

The PID isn't used for any communication. It's only used to minimize the
chances to get a race between two kdeds started at the same time. The
current code might barf if you manage to start 2 kdeds on two different
machines but with the same PID. I think thats highly unlikely...

greetings,
-- 
Steffen Hansen                              |
email: stefh@mip.sdu.dk, stefh@imada.sdu.dk,|  
       hansen@kde.org                       | ABC...VWXKZ :)
URL:   http://www.mip.sdu.dk/~stefh         | 

[prev in list] [next in list] [prev in thread] [next in thread]