On Wed, 2007-12-26 at 14:01 +1100, Malcolm Tredinnick wrote: [...] > In general, the best way to provide something like this for the > community is to make it available as a third-party application. At the > moment, that isn't entirely simple, since the Django cache mechanism > isn't set up to allow third-party cache procedures. We could make > similar changes to what we did in the database backend to allow > third-party caches to be used without requiring core changes. > > I would be against including another caching mechanism in core at this > time, but making the changes to allow third-party caches that provide > the same API as Django's core caches would be a good change to make. > That provides the best of both worlds: room for experimentation with new > systems, without adding extra code to core for something that might not > be used too frequently (and which we would have to maintain forever, > regardless).
Sorry, that whole post came out horribly. I wrote one version, started editing and rewriting and then was distracted by a seagull or something and sent it in a form that only repeated most of what you already said. I suck. :-( The second paragraph above is my real point. What I also wanted to add was that including something like this would be overkill in many respects. We might as well just fix the existing file backend cache to handle resource limits if we were going to go down that path. I suspect it's not high priority since in most cases (1) generating the content once every now and again doesn't kill you (if it does, your serving hardware is a bit too feeble), so memcached hits the high-volume sweet spot and already does more than any hand-rolled version is going to do, (2) if caching gives you real benefits, it's because you are serving the same stuff over and over again, so storing one copy of everything you serve isn't too bad and a periodic cron job to blow away the cache files is the poor man's working solution if you're using the file-based cache. Yes, we could build in the second point to the file:/// backend, but it adds a little overhead to each request (since it's event based, rather than looping, you obviously have to check when requests occur) and if you really need that fine-granied control, memcached is probably going to be a better fit. Regards, Malcolm -- The hardness of butter is directly proportional to the softness of the bread. http://www.pointy-stick.com/blog/ --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Django developers" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/django-developers?hl=en -~----------~----~----~----~------~----~------~--~---
