Re: draft concept for acl caching

On Tue, 2003-01-28 at 20:06, Henrik Nordstrom wrote:

> I do not think refcounting is really needed. acls are only created
> once and never destroyed.. A reconfigure only reloads the ACL data,
> it does not destroy the actual acl.

I hadn't dug in that much. My recollection though is that a new acl is
created and the old one cbdatafreed(). There are checks for that
condition in acl.cc.

> Having the cache operate without change for the lifespan of the
> request even during/after a reconfigure is probably preferable in my
> opinion.

Yes, that is the goal:
Extant requests operate on the old data/results. New requests only
operate on current at the time acl's.

Rob

-- 
GPG key available at: <http://users.bigpond.net.au/robertc/keys.txt>.