The issue described in this pr is the cause for long-running Samba  
smbd processes to grow to the user's memory limit and then die.

I have applied the patches from the pr on two NetBSD 2.0 machines and  
the size of the smbd processes stays constant. I am about to do the  
same on two more 3.0 systems.

Is there any technical reason why these patches have not been  
committed? Is there a better way to solve the problem?

-- 
     aew