Solved inconsistency in decoders inheritance hierarchy

[cdeler]

@tomchristie,

I tried to find a common ancestor in Decoders hierarchy (just wanted to make it clearer). Unfortunately there is a big gap between e.g. TextDecoder and BrotliDecoder (approach with Callable looks weak, but once Guido advised it python/typing#241 (comment))

If I did something strange or unnecessary, feel free to decline the PR

Beware: I introduced abc here.

[lovelydinosaur]

Okay, so I'm going to close this one off, since it seems like a technical bit of refactoring, in search of a problem that doesn't exist. Or to put it another way - I can't see what you're actually trying to solve here?

Please don't let that put you off tho' - the contributions are appreciated either way!

But let's not introduce extra hierarchy, ABCs, and other trickery unless there's a really clear end-user facing reason why we're doing so. 😀

[cdeler]

@tomchristie , I do agree that it might look like a refactoring without a target, but I wanted to solve this problem:

class DeflateDecoder(Decoder):
    # ...
    def __init__(self, request: "Request") -> None:
        self.request = request     # <-- here we do not call super type constructor, but just set the request property
        self.first_attempt = True

(it looks the same for GZipDecoder, BrotliDecoder, DeflateDecoder and MultiDecoder inhereted from Decoder with defined __init__ method). Having tried to solve that (adding super() call) I found that the decoders inheritance is strange and tried to fix it

Nevertheless, I accept the idea, that refactoring without a clear target is a root of evil (and just wanted to explain the cause of my PR)

[florimondmanca]

Yes, this rings a bell… :) #1095 (comment)