Add flake8 to tests or black to CI

[mrocklin]

Currently it looks like PRs can pass tests without passing flake8 tests. Do we want to start enforcing style?

[jhamman]

Do we want to start enforcing style?

I would like this. Is there a standard flake8 config and way to enforce this for dask sub projects?

[mrocklin]

Nothing fully standard. We could grab the setup.cfg from dask/dask or dask/distributed

…

On Fri, Jul 27, 2018 at 7:09 AM, Joe Hamman ***@***.***> wrote: Do we want to start enforcing style? I would like this. Is there a standard flake8 config and way to enforce this for dask sub projects? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#106 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AASszPiUqjXNdf4Rw-PmLiwH4buVVrpPks5uKx8LgaJpZM4Viyd0> .

[jhamman]

See #109. As far as I can tell, we've been running flake8 all along. Maybe @lesteve or @guillaumeeb has some idea why things weren't being raised in #78.

[guillaumeeb]

I remember asking myself if the none.sh job_script function would show something in Travis if flake checks failed.

The thing that troubled me is that we are running flake8 and py.test in one Travis operation. Maybe we should add a && between the two calls to take into account both return codes?

[mrocklin]

In dask/distributed we run flake8 in a separate travis-ci run. It seems to work pretty well.

…

On Sat, Jul 28, 2018 at 6:51 AM, Guillaume EB ***@***.***> wrote: I remember asking myself if the none.sh job_script function would show something in Travis if flake checks failed. The thing that troubled me is that we are running flake8 and py.test in one Travis operation. Maybe we should add a && between the two calls to take into account both return codes? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#106 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AASszEPdw22n5GMXGHVIaXSC_QqVZcM5ks5uLGxRgaJpZM4Viyd0> .

[guillaumeeb]

Closed with #110.