Because the db caching doesn't support atomic operations, it was unsafe to store a list of 'keys' in a single key. So, rather than store the list, I just append each key with a specific tag, and then filter for it later. This means I don't need to worry too much about atomic usage with lists (i.e. queued requests).
However - I still can think of many instances where I would need atomic usage, so I will probably implement this later on down the line. Hopefully, the atomic modifications will be accepted by the core devs.
This also contains threaded cache cleaning, which means you no longer need to rely on requests to clean the cache (which would have potentially slowed the user query down), and will remove any cache entries past their expiry date every 3 minutes.
Enjoy!
Cal
- filter
- cache
- db
- backend
- like
manything need to do with RequestContext, but it's too tedious.
use
`render_to_response("/my.html", {'key':value,},request)`
instead of
`render_to_response("/my.html", {'key':value,},new RequestContext(request)) `
and you can also use
`render_to_response("/my.html", {'key':value,},new RequestContext(request))`