Matthew Dapena-Tretter

This is great data; thank you. Can to do me a favor and add one more thing to your benchmarks? A version of `get_hash` that's the same as the original...

Sorry, @Alp-Phone! I've been extremely busy. (Plus GitHub doesn't notify when you update comments.) I'll try to look at it soon. What might be helpful is to run some benchmarks...

Sorry, I don't quite follow your previous benchmarks. It looks like your custom `get_hash` is still pickling the processor when we were trying to compare it to something that just...

@pccompassion It's likely that you're just using a cache strategy that generates the image (or verifies it's existence) on access. Check out the optimization section of the docs.

I'm really sorry @NewVadim but I can't really understand. If you're worried about too-frequent disk access, you can read [these tips](http://django-imagekit.readthedocs.org/en/latest/caching.html#optimizing) for optimization. On the other hand, if you're getting...

@NewVadim I think I understand. You're saying that the image doesn't exist, but you don't want it to raise an error?

Hm, in that case, I'm leaning towards closing this as a #wontfix. AFAIK, this is how Django handles missing files in templates (e.g. ImageFields, etc.) everywhere, so it makes sense...

AFAIK IK behaves just like Django here. If you want to include an image that may not exist, you can do truthiness checks, but accessing an attribute of a non-existent...

If you're getting an error checking truthiness, that's a bug.

afk right now but there's a section on this in [the docs](http://django-imagekit.readthedocs.io/en/latest/advanced_usage.html).