On 1/25/06, hugo <[EMAIL PROTECTED]> wrote: > You might have written this as a simple template loader - and used the > basic template loaders as starting point, just overloading the > get_template functionality. That way you would be able to use the > standard render_to_xxx functions - you only would have to change the > TEMPLATE_LOADERS setting. And that way integration with the Django core > is quite simple, as it is just another bunch of template loaders the > user can just activate.
I'm not sure that would work, because I suspect much of the performance savings Wojtek is getting is from the fact that his patch caches *compiled* templates, not the raw template code. (He has a template that loads another template within a loop.) I don't see how a template loader -- in the TEMPLATE_LOADERS framework -- would be able to cache compiled templates. I was thinking we could add a CACHE_COMPILED_TEMPLATES setting and change get_template() to give it a "cache" keyword argument, which would default to the value of CACHE_COMPILED_TEMPLATES. Any quick thoughts before I implement this? Adrian -- Adrian Holovaty holovaty.com | djangoproject.com | chicagocrime.org
