CI: pip>=10.0 may cause breakages in testing?

[EricCousineau-TRI]

From this issue pypa/pip#5240 (comment)
(@peteflorence had run into this, which was resolved in RussTedrake/underactuated#116)

I've noticed what might look like a similar failure in AppVeyor results:
https://ci.appveyor.com/project/wjakob/pybind11/build/1.0.2770/job/42dem9qbijtwx6er#L46
(From PR #1367)

Should .appveyor.yml not upgrade its pip version, or perhaps constrain pip to upgrade to <10.0?

[peteflorence]

Hey Eric, if you scroll through other comments in that issue there seem to be other solutions for example using pip2 instead of pip. Explicitly specifying 9.0.3 worked for our use case but you might want a different solution.

…

On Fri, Apr 20, 2018 at 12:07 PM Eric Cousineau ***@***.***> wrote: From this issue pypa/pip#5240 (comment) <pypa/pip#5240 (comment)> ***@***.*** <https://github.com/peteflorence> had run into this, which was resolved in RussTedrake/underactuated#116 <RussTedrake/underactuated#116>) I've noticed what might look like a similar failure in AppVeyor results: https://ci.appveyor.com/project/wjakob/pybind11/build/1.0.2770/job/42dem9qbijtwx6er#L46 (From PR #1367 <#1367>) Should .appveyor.yml not upgrade its pip version, or perhaps constrain pip to upgrade to <10.0? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#1372>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFYQqHfv46g-wQZjZGxVaJHYutgDL25qks5tqgfOgaJpZM4Tdt5W> .

[henryiii]

See the changes in the reference PR. It's still using pip10. You really should be only using python -m pip, not pip the command. 😠 🙄

[wjakob]

This is fixed now AFAIK.