Adam Hopkins
Adam Hopkins
No worries. It really should only be the domain, but that doesn't mean because of historical reasons we shouldn't be supporting the scheme too. `//` is legit. For this PR...
If that is the the decision we need to do some validation. As it stands there is certainly a use when you would need an empty scheme and we are...
Yup. Consistency is the pain point here. I am hesitant to add yet another minor fix, and will try and take a crack in the coming weeks at revamping this....
@StuartFarmer Can you put together a PR so we can review it?
> @ahopkins If I have some more time, I can file a PR. It will make Sanic a lot more usable in more complex applications where you need more control...
> @ahopkins are you saying that this is fixed in 3.7? I meant to say that the example @mikepurvis shows above [has some 3.7+ only methods](https://docs.python.org/3/library/asyncio-eventloop.html#asyncio.Server.serve_forever).
> All those methods are just sugar to make using asyncio more ergonomic, though. If Sanic needs to support 3.6, they could be easily reimplemented, or even just directly backported/monkey-patched,...
@StuartFarmer any luck?
@StuartFarmer Did you ever do any work on this?
The latest release added a `shutdown` for tasks that are named. I am thinking that maybe we could add to this functionality and simply add all tasks and that should...