Move essentials of magpie's cache into the database

[sdcosta]

Moving Magpie's cache into the database

* The best way of implementing this would probably be to subclass RSS_cache and make it work with the database. * If a feed returns a 304 not modified, we do not need to see if there are new items (as we currently do) * All the feed content is already in the database. We store an md5sum in the database and magpie's cache becomes redundant. We only need to save the Etags and the last accessed times. #137

[Darth_Sebulba05@…]

This will work just as well..

Until then, if I add new feeds, I'll just have to remember to chmod it manually.

[james@…]

The magpie cache is not needed at all. I turned it off with:

: define('MAGPIE_CACHE_ON', false);

and haven't looked back.

[Darth_Sebulba05@…]

That means you have to fetch the whole feed every time, doesn't it?

[sdcosta]

It would seem quite selfish if someone turned magpie's cache off unless they had a very good reason to. Of course, if you have protected your installation so that updating is private then it should not matter. Having a caching sytem for fetching feeds is essential to being a good netizen. Websites are known to ban IP's and even user-agents if feeds are fetched too often. If you would like to know why the cache is important please see

http://magpierss.sourceforge.net/faq.php#why-is-it-important

Point 1 (however) is not applicable to gregarius because we wanted to allow multiple users to share the same cache. We are going to change that behavior in this ticket.

[sdcosta]

Fixed in [1155]