Re: [Bayonne-devel]libexec timeout provokes segfault

[David S. <dy...@os...>]

There are three things I am doing which might improve this situation.  The
First, I believe, is that the default mask (as gathered from getTrapDefault)
should be set to 0x7 rather than 0x3 so that it covers timeouts rather than
ignores them if no ^timeout trap is in your active script.

Second, I am moving the pika audio buffers from the audio play/record thread
objects which only exist during an active audio session into the trunk object
so that they are only allocated once, and so the pika thread never is at risk
of having a buffer to an object that may already have been deleted.

Finally, I am doing a new Common C++ (and ccaudio) release, probably later
today, which fixes some issues with delete[] vs delete.  These changes are in
cvs currently.

David

On Thu, 21 Sep 2000, Matt Benjamin wrote:
> We have found that bayonne (.4.3a and .5.1) can be made to
> segfault by calling libexec with a timeout smaller than the execution time
> of the script.  
> 
> Script fragment and backtrace available, on request.
> 
> 
> Matt
> 
>  
> Matt Benjamin President/CTO
>  
> The Linux Box 
> 206 South Fifth Ave. Suite 150 
> Ann Arbor, MI  48104
>  
> tel. 734-761-4689 
> fax. 734-769-8938
> pgr. 734-431-0118    
> 
> 
> _______________________________________________
> Bayonne-devel mailing list
> Bay...@li...
> http://lists.sourceforge.net/mailman/listinfo/bayonne-devel
-- 
OST - Your "source" for free telephony software