Closed blueyed closed 5 years ago
did this practically happen?
Yes, with py34-xdist on Travis, see the original ref.
Yikes, thanks
Rebased.
Pushed a fixup.
I also wonder if caching the ImportError wouldn't make sense?!
Can you please check the travis failure @blueyed?
I think it is related to pytest itself failing with the test. Will have to investigate.
IIRC AppVeyor skips the test.
Any clues about the failures on Travis?
Still the same as with https://github.com/pytest-dev/py/pull/207#issuecomment-464888877 I guess. No energy to look into this now.
OK. Another option is to drop support for those old pytest versions.
With the "importlib" option getting used in pytest, we will need to require a higher version of py
anyway.
Fixed the new test to cd back (using monkeypatch).
Merging #207 into master will increase coverage by
0.03%
. The diff coverage is100%
.
@@ Coverage Diff @@
## master #207 +/- ##
==========================================
+ Coverage 82.04% 82.07% +0.03%
==========================================
Files 55 55
Lines 10136 10153 +17
Branches 1141 1141
==========================================
+ Hits 8316 8333 +17
Misses 1556 1556
Partials 264 264
Please add a CHANGELOG then it is good to go I think
Pushed with [ci skip]
- hopefully that will still trigger it on the merge.
Python 3.4 might raise FileNotFoundError due to
os.getcwd()
failing on a non-existing cwd. This is fixed in Python 3.5.Ref: https://github.com/pytest-dev/pytest/pull/4787#issuecomment-463341251