We're storing the same 40 bytes long storage_id hash several million times

[smileychris]

We're storing the same 40 bytes long storage_id hash several million times

[SmileyChris]

Yeah, it's probably not the smartest of my ideas :P

In fact, I'm not really sold on the whole pickle-hash idea, since it seems that the django-storages s3 backend can't handle being pickled anyway. My reasoning is that if you were using a storage backend which has initialisable arguments as part of the class, you want to be able to differentiate between two unique versions of those storage classes.

But that's me rambling. I think I'll get around to just try/excepting the pickle and putting a hash of the full class path if pickling fails.

Back on topic, if you want to refactor the code to use a separate model for the storage then I'd be happy to pull it in. You'll need to make sure that the current storage hash lookup is cached - don't need another database hit...

[SmileyChris]

Done (haven't pushed a new version to pypi yet, but it's in the repo)

[SmileyChris]

(I've actually pushed this to a branch for now... i'm not quite happy with the migrations)

[SmileyChris]

(ok, i'm happy enough now, merged to master and deleted branch)