Add test environment using pluggy from master branch#2744
Conversation
pluggy>=0.5 already warns about those
For some reason, the previous approach brakes 'coveralls' because pip still tries to install the 'pluggy' master requirement (git+https://...)
goodboy
left a comment
goodboy
left a comment
There was a problem hiding this comment.
I think you've got one API mismatch and maybe a slight factoring you can do in the tox.ini other wise looks good.
| rep = outcome.get_result() | ||
| rep.headerlines += ["header1"] | ||
| return rep | ||
| outcome.set_result(rep) |
There was a problem hiding this comment.
@nicoddemus that method doesn't exist I don't think?
It's force_result().
I do, however, prefer this name better and think we should move to it.
It fits with Future.set_result() which may be something that's handy if we ever want to offer an async API.
There was a problem hiding this comment.
i think the current name is much better explaining whats going on as a futures set_result will raise a exception on second use
There was a problem hiding this comment.
@RonnyPfannschmidt fair enough.
Either way I'm confused how this test is passing if the method name is wrong?
There was a problem hiding this comment.
Hmm I will have to take a closer look later, indeed this is weird.
There was a problem hiding this comment.
Ahh of course, it raises an error at the rep.headerlines += ["header1"] line. Plus it is an "xfail" test so it is not really running.
There was a problem hiding this comment.
I'm updating to force_result anyway to avoid future confusion.
| [testenv:py35-pluggymaster] | ||
| passenv={[testenv]passenv} | ||
| commands={[testenv]commands} | ||
| setenv= |
There was a problem hiding this comment.
can these two be pulled from the py27 job above?
Even though the test is not running at the moment (xfail), at least we avoid future confusion
|
Done requested changes. |
|
@nicoddemus please rebase |
| commands={[testenv:py27-pluggymaster]commands} | ||
| setenv= | ||
| _PYTEST_SETUP_SKIP_PLUGGY_DEP=1 | ||
| deps = |
There was a problem hiding this comment.
@nicoddemus heh I was actually thinking to derive setenv and deps just because they're common to these two envs.
Not a big deal though :)
goodboy
left a comment
goodboy
left a comment
There was a problem hiding this comment.
As long as that call to force_result() actually made a difference I think this is good 👍
|
Fixed the merge conflict. 😓 |
|
After this is merged, we should re-trigger pluggy's master to ensure it is fixed. |
4 participants
Fix #2737
After merging this we can conclude pytest-dev/pluggy#79.