On Thu, Mar 17, 2011 at 22:10, C. Michael Pilato <cmpil...@collab.net> wrote: > On 03/17/2011 03:43 PM, Greg Stein wrote: >> Dude. If you want to delete *one* prop from the dav_cache, then just torch >> them all (ie. set to null). HTTPv2 does not use them. HTTPv1 only uses the >> vsn URL. And it is a cache. >> >> So don't be tricky. Blast the entire dav_cache rather than tweak it. > > Yes, the more I examined real-world usage, the more I decided that full > cache invalidation would be fine. But sometimes, it's good exercise to > force yourself to dive into an API you've never looked at before and try > something new. I managed to successfully implement the functionality I set > out to implement, and that's intellectually satisfying. I'm sure you know > and can appreciate this feeling.
You bet! > I've attached the patch here in case someone else wants to model similar > functionality on it later. (Philip has a similar patch for another bit of > unrelated functionality.) But, yeah, it's a full dav_cache purge that I'll > be committing. Cool. At a minimum, the demonstration of a custom function could be useful. And yeah... maybe we'll do something nifty with prop skels in the future. Cheers, -g
