[vdr] Re: Segment faults 43

[Rainer Zocholl <UseNet-Posting-Nospam-74308-@zocki.toppoint.de>]

rahrenbe@cc.hut.fi(Rolf Ahrenberg)  04.12.04 00:31

>On Fri, 3 Dec 2004, Klaus Schmidinger wrote:

>>  virtual void Activate(bool On) {}
>>               // to do things like starting/stopping a thread.

>Just a little notice: osdteletext plugins doesn't use this method at
>all, 

But VDR did not warn?

In a multitasing environmentit's a lottery to assume that
something will run on it's own in the way/sequence it is programmed...


>but my latest patch for it does. You (not Klaus :) might want to
>check it out if it cures the crashing - I cannot do that by myself,
>because everything is working nicely on my setup ;)

>vdr-osdteletext-0.4.1-fixed-a-little-bit-more.diff.gz

I'll try ASAP. Thanks.


Another point which is usually a clear indication of (timing)
problems with thread management:
When the application is run under the debugger no
faults are issued and all seems to run stable!
Attaching gdb to VDR makes it run stable!
That makes it so hard to find the cause, 
as only "post mortem" debugging is possible.

Third point:
Maybe the problem occurs only on faster CPUs?
I'm using here an AMD Duron 1800 with reduced FSB of 100 
giving a final CPU clock of 1300MHz on an Asrock K7S8XE+ (SiS748).
Too "athcool" is used.




Rainer