Fix RFC3339 representation

[marcelstoer]

https://now.httpbin.org/ returns an invalid RFC3339 representation. It currently shows 2-digit milliseconds. I hope I tracked it correctly back to Maya core.py.

RFC3339 5.6 allows at most 1 digit for milliseconds: https://tools.ietf.org/html/rfc3339#section-5.6.

time-secfrac    = "." 1*DIGIT
partial-time    = time-hour ":" time-minute ":" time-second [time-secfrac]

I'm not 100% happy with the fix as .%f apparently isn't guaranteed to always return 6 digits (platform dependent). https://stackoverflow.com/a/35643540/131929 shows a more stable version based on splitting and reformatting the original string.

[timofurrer]

Thanks for the contribution and sorry for the long wait.

The change looks good. Would you be up for writing a test to ensure that the RFC3339 is correctly implemented?

🎉

[marcelstoer]

Sure, I try not to forget it.

[timofurrer]

Great! Ready when you are 🎉

[marcelstoer]

I get 10 failed tests on macOS before I even start (also on master). This is regardless of whether I do pipenv run pytest tests/ or simply run make. Are the tests supposed to pass on macOS at all?

[timofurrer]

Are the tests supposed to pass on macOS at all?

Yes, they should.

Can you open a new issue with some more information about the test failures you get?

[timofurrer]

Thanks for the contribution 🎉

[timofurrer]

I've released v0.4.3 containing this fix!

[marcelstoer]

To fix the originally reported problem someone would have to deploy that latest version for https://now.httpbin.org/

[timofurrer]

That "someone" would be Kenneth. I see what I can do.

[marcelstoer]

@kennethreitz - ping, how about upgrading httpbin.org?