MacOSX-dev Digest, Vol 48, Issue 3

[Alastair Houghton]

On 6 Dec 2007, at 03:39, Chad Armstrong wrote:

> That's an interesting point, Scott.  I'll look into this further.  I  
> noticed that Panther's kern.maxprocperuid is set to only 100,  
> whereas Leopard is 266.  I'll need to also check other versions of  
> Mac OS X and see how they might different.
>
> The only problem with your suggestion is that each task is called  
> sequentially.  One is called, and the next one will not start until  
> the previous one has completed.  The PID continues to rise with each  
> time the NSTask is launched, but there are not a bunch of tasks  
> being called all at once.

Does your code call -waitUntilExit on the NSTasks?  At the underlying  
UNIX layer, if you don't check the exit status, you will end up with  
zombie processes that still take up process table entries.

If you want to check for this problem, you can do so by looking at the  
output of the "ps x" command in Terminal.  The STAT column will show  
"Z" for zombie processes.

Kind regards,

Alastair.

--
http://alastairs-place.net