[asterisk-dev] Error msg: Asterisk is not Thread safe

[Russell Bryant]

On Dec 10, 2008, at 1:06 AM, Tilghman Lesher wrote:
> I'm not sure if removal was correct.  The code derives from issue  
> #1816
> and was a direct regression test for mutex behavior under certain BSD
> platforms.  While I agree that it needed fixing, I don't believe it  
> merited
> removal.  I think we need to reconsider that action.


I can see from the issue history that the patch did come from there.   
However, I do not see anywhere that indicates that the code actually  
pointed out a problem on some platform.  I think the goal was to just  
verify that the mutexes were getting initialized successfully.

If recursive mutexes were actually broken on a platform, Asterisk  
would go down in flames very quickly.  If you ran such a system with  
DEBUG_THREADS enabled, you would get ERROR messages all over the  
place.  To me, this is enough.

With that said, if you'd like to write a new valid test to verify that  
you can lock and unlock recursively without errors, then feel free.

--
Russell Bryant
Senior Software Engineer
Open Source Team Lead
Digium, Inc.