Discourage users from using unittest support somewhat

[ionelmc]

After some discussion @nedbat on IRC - he simply don't believe TestCase are bad 😜

[nicoddemus]

Thanks @ionelmc!

I like the change, but I'm curious about what the discussion entailed. Also, the title talks about discouraging users from using unittest, while your comment indicates that @nedbat actually thinks it is OK to keep TestCase subclasses around.

Would love to understand better what was the discussion about. 😁

[ionelmc]

Well, @nedbat, what do you think of it now?

[RonnyPfannschmidt]

personally i think we should put it even more harsh and ourtight say that some of us would be happy to fix those details as part of a consulting gig, but its certainly neither easy nor fun work

[ionelmc]

@nicoddemus the discussion was basically me saying mean things about @nedbat's testcases and bragging about converting around 40k sloc worth of testcase code to fixtures. It only took a week - what's the big deal 😁

[nicoddemus]

@ionelmc thanks for the summary! 👍

Could you please take a look at the failure on the docs env?

[coveralls]

Coverage remained the same at 92.832% when pulling d98d122 on ionelmc-patch-1 into beb77c1 on master.

[coveralls]

Coverage remained the same at 92.832% when pulling c8032a9 on ionelmc-patch-1 into beb77c1 on master.

[nicoddemus]

Thanks @ionelmc!

(Just a small heads up, prefer to open PRs from your own fork, even if you have write access to the repository like you do here, otherwise CI services will run twice: once for your branch and another for the PR).