fix connectionpool.ConnectionPool.close()

[tharvik]

See #873 for reference.

[Lukasa]

Hrrrm, those test failures seem to be unrelated to this change: something is going on with install. I'll investigate.

[Lukasa]

Cool, we bumped into this. No point piling on, the reports are already there, so let's just work around.

[Lukasa]

Fab, so I've proposed a fix that should resolve the busted tests in #875. When that (or something like it) gets merged, you'll need to rebase on top of it to solve the test failures.

Sorry about that inconvenience!

[haikuginger]

While we're in here, should we switch from pass to a NotImplemented exception akin to RequestMethods.urlopen()?

[tharvik]

While we're in here, should we switch from pass to a NotImplemented exception akin to RequestMethods.urlopen()?

That would be even nicer, I'll do it when #875 is merged.

[Lukasa]

@haikuginger @tharvik I don't actually think we want to do that. I think we shouldn't punish people who unconditionally call super() in close(), and it gives us the opportunity to be a bit more flexible later.

[haikuginger]

I'll defer to @Lukasa's judgement on that, then; I'm 👍 on this change.

[tharvik]

Okay, I won't add anything to the PR then.

[Lukasa]

All better, mergy mergy mergy. Thanks @tharvik! ✨

[sigmavirus24]

@Lukasa you did the merge, don't forget the release notes ;)

[haikuginger]

Thanks very much @tharvik! 🍰