Michael

About this httplib replacement. As I had a look into this lately, I'd say replacing it wouldn't be too complicated. In the end, it's only three classes in there. As...

Another wishlist thing. Not sure how you see this, but imho `unittest` is a major pain in the ass. Clunky syntax, no way to properly parametrize tests, inflexible, unpythonic, ......

Agreed, but it's a major change and piece of work. The main question is: Is it desired at all by the core devs?

Ok, fair enough. In that case, I may actually give this a shot somewhen. I was mainly afraid, the whole idea may be rejected, due to adding further dependencies or...

Well, I'm exactly one of these (few) unlucky users who, as you said, use a blocking pool to limit the number of parallel connections to remote hosts and as well...

Hmm, I guess it's not avoidable to track, if the data was read completely. As you said, it's hard to safely determine the state otherwise. And if an error occured...

Exhaust-reading on acquire may not be a good idea. Imagine a huge download, that was interrupted, and no one cares about the remaining unread data anymore. Better save the time...

Our initial grief were the slow and incomplete ufuncs and we'd love to see a faster version integrated into numpy as well. I agree, our main focus was getting a...

Yes, I agree. Which backend (we referred to them as 'implementation' in npg, anyways) is used should be transparent for the user, except explicitly asked for one. We had implemented...

I'd vote for better error handling in general. It's not only `POSKeyError`s, but also all sorts of `zlib.error`s or `UnpicklingError`s, that might creep into the database over the years and...