Test on Python 3.8, drop 3.3 and 3.4

[asvetlov]

Fixes #129

@Tinche please review

[michael-k]

Can you add the Python 3.8 classifier to setup.py?

Would be nice to see a release with Python 3.8 support.

[asvetlov]

Done

[L1ghtn1ng]

@Tinche Can you please merge this in so it can support py3.8? Also when is a release going to be done(tagged) with py3.8 support?

[blueyed]

👍
Missed this when creating #134 - which still makes up for a cleaner .travis.yml though. Please consider adopting it, or we can otherwise do it (#134) later.

Created #135 to revisit coverage reporting also (which is useful with PRs like this). Can come afterwards though.

[blueyed]

I think it would be nicer to have this rebased on master, so that e.g. 0f86f79 was not necessary (a fixup after master was merged into it).

[Tinche]

I plan on pulling this in before the next release, so soon! Could I ask that merge conflicts are fixed please @asvetlov

[fulaphex]

@Tinche there are no more conflicts, could you merge it to unblock the release?

[Tinche]

@Tinche there are no more conflicts, could you merge it to unblock the release?

There are still conflicts :)

[fulaphex]

I might be missing something, but here is how it looks on my end:

That's why I pinged you

[simonfagerholm]

@fulaphex I'm guessing that was the state at the time of the last push, not current state

[L1ghtn1ng]

Even for me it is saying there are no conflicts

[Tinche]

Huh, weird.

[michael-k]

I tried to rebase this branch on master and ended up in rebase hell, but squash merging it worked fine. I've created a PR #152 which is effectively a rebase-and-squash of this PR (including proper attribution).

Hope this helps to unblock the next release.

[Tinche]

Superceded by #152.