DragonFly bugs List (threaded) for 2004-04
Re: troubles with caps
:> - Cache results locally so it does not have to perform an IPC for
:> every request (CAPS has or will have a mechanism whereby the
:> service will be able to notify clients of changes to have clients
:> invalidate their caches).
:I would skip that part for the moment. This one should be evaluated wether
:it is worth the effort. E.g. for uid <=> name mappings most portable apps
:already cache them internally.
I think it's fairly important for libc to do some caching locally.
While it is true that programs like ls use auxillary caching procedures
such as user_from_uid(), I do not consider this desireable because
such auxillary caching mechanisms are not tied into the invalidation
mechanism that CAPS has (and in the future the invalidation mechanism
may be up-call based). Tight integration is needed for the caching
to work properly.