userdb cache

[Havoc Pennington]

Hi,

Tim Dijkstra wrote:
> Already some time ago I prepared a patch to clean up the code regarding
> userdb caching a bit and to add a configure switch to bypass it
> entirely. (The latter is off by default, but IMHO it should be on and
> people should rely on nscd).

IIRC I was willing to agree we need to turn it on, but only after 
gathering some data on performance (with and without nscd), which is the 
purpose of the configure switch. I may remember incorrectly of course ;-)

> Back then, Havoc said it was OK to commit. In the mean time I got cvs
> access, but because this is my first commit to the dbus-repo I attach
> the patch once again and am asking for another heads-up.  

Go for it, maybe only on HEAD (not the 1.0 branch).

Havoc