TST: Windows tests now xfail where needed#2027
Conversation
richardjgowers
left a comment
richardjgowers
left a comment
There was a problem hiding this comment.
Looks good, and is failing, which is the idea right?
|
Yes, 2 fail + 21 xfail was expected for Appveyor as I see here; once the linked PR has been properly reviewed / merged this should probably be rebased to confirm removal of the failure. |
* all unit tests that currently fail or error on Windows (Appveyor) will now strictly xfail * Appveyor test suite failure is no longer allowed in CI--this prevents regressions in Windows compatibility progress; the objective now is to reduce xfails instead of actual failures
richardjgowers
force-pushed
the
windows_xfail
branch
from
1c74e78 to
d5b8f3e
Compare
Codecov Report
@@ Coverage Diff @@
## develop #2027 +/- ##
========================================
Coverage 88.59% 88.59%
========================================
Files 143 143
Lines 17305 17305
Branches 2649 2649
========================================
Hits 15331 15331
Misses 1376 1376
Partials 598 598Continue to review full report at Codecov.
|
|
I've rebased this, and updated the repo settings on github so that the appveyor must pass before we can merge (ie windows tests are equal citizens to linux/osx) |
|
21 xfails and no errors for the branch & merge Appveyor tests looks right |
* all unit tests that currently fail or error on Windows (Appveyor) will now strictly xfail * Appveyor test suite failure is no longer allowed in CI--this prevents regressions in Windows compatibility progress; the objective now is to reduce xfails instead of actual failures
2 participants
all unit tests that currently fail or error on
Windows (Appveyor) will now strictly xfail
Appveyor test suite failure is no longer allowed
in CI--this prevents regressions in Windows compatibility
progress; the objective now is to reduce xfails instead
of actual failures (and remove the
xfailtag as appropriatealong the way).
There should be two actual failures until this PR branch can be rebased on top of a merge from #2024.
From the current master this will mean 13 fails + 9 errors [and 1 previous xfail] --> 2 actual failures fixed in #2024 + 21 xfails (eventual objective: reduce back to 1 xfail).