On Tue, 31 Jul 2001, Perrin Harkins wrote:

> > The idea of a volatile cache for
> > certain included files would be very useful.
> 
> Would it?  Your data is cached by your objects, right?  It seems like this
> won't actually help you unless you move the caching out of the objects and
> into the templates, which probably isn't what you want if you reuse these
> objects on multiple templates.

Currently we cache on two levels: 

 1.  raw html generated from the templates
 2.  serialized objects

The logic for caching/fetching both of these is done in our mod_perl
handlers (I actually wanted our objects responsible for caching the data,
but I lost an argument with a certain Sr. Engineer, but that is another
story). The overhead of processing templates is enough that it pays to
cache the raw output when we can.  That said, the speed improvements that
Doug is working on based on Andy's Stash implemented in C, might mean that
caching raw output isn't needed.


Regarding the need for cached templates, I can easily see a requirement
for "personalized" content on certain pages which is the reason I start
hungering for a good template caching solution.  I am thinking of a page
where most of the data is cached for a certain ttl, and an include which
contains the calls to personalized data is labeled "volatile" so it is the
only template that has anything processed on each request. 

I don't know if this makes much sense in the context of what y'all talked
about at the BOF, so let me know if I am off on an unrelated course...

Craig






Reply via email to