`match` should not be an iterator

[tooxie]

I think it makes it more complicated to work with the function because in order to subscript it one has to force it into a list first. If each result had to be fetched from the server then it would make a lot of sense to yield them one by one, but since the response is already downloaded and parsed, returning all results at once makes no difference.

[sampsyo]

Hi! Thanks for the suggestion. I generally like iterator-based APIs, though, and they're quite common in modern Python. Even dict.items(), for example, is an iterator these days. I agree there's not a huge performance problem to be solved here, but I think the difference is small enough that it's probably not worth the effort to change.