On Thu, 16 Nov 2000, you wrote:
> On Thu, 16 Nov 2000, Dave Bryson wrote:
> I probably jotted that down at some point quickly.
> I was thinking to access the cache at a single point
> and catch just the general Exception so that all
> the Runtime had to know about was the GlobalCache
> and nothing of it's internals if possible.
Ahhh... ok.
> Right now that caching code isn't being used
> inside the Runtime. I was thinking of trying
> to make a general cache: permanent objects,
> expirable object, refreshable objects. The permanent
> objects would be managed externally like the template
> cache where the Runtime figures out whether the
> template needs to be reloaded. They wouldn't expired or
> be refreshed automatically.
Could "permanent objects" possibly be future compiled templates?
If we want to be able to cache all these different types, it may be
worth while then for us to also consider the ability to set and upper
bound on the size of cache. Maybe adding and LRU check or something
along those lines. Configurable of course.
> I was thinking that of making a single cache that could
> deal with all cases and then optimize the snot
> out of it, if at all possible.
Sounds cool. I'll spend some time thinking about this a little more.
Probably the first step is to simply get it working in Velocity as is (
not much work really ), then build on this other stuff.
Anything new and cool we do here would probably be benefical in Turbine as
well.
BTW, does anyone know of a free profiler on Linux?
--
Dave Bryson
[EMAIL PROTECTED]
----------------------
