Conversation
|
I think this is an incorrect use of The idea is that the xfails may start reporting XPASS at some point if fixed accidentally. Here's some of @pganssle's explanations on doing it properly: https://blog.ganssle.io/articles/2021/11/pytest-xfail.html / https://ganssle.io/talks/#xfail-and-skip / https://ganssle.io/talks/#xfail-lightning. |
|
This was broken by pytest-dev/pytest#11233. And it was reported upstream in pytest-dev/pytest#12231, and fixed in pytest-dev/pytest#12280. So we can either wait till a pytest release contains this fix, or temporarily turn them off.
That's why this change was bad; xfails are expected to fail, so printing an exception traceback explaining the known failure pollutes the logs and pretty much forces these to be turned off. It will is listed in the summary at the end (which is what the |
|
@henryiii yes, so I'd suggest setting |
Signed-off-by: Henry Schreiner <[email protected]>
c36a32a to
bdfb0bd
Compare
|
I just added a todo, don't think we need anything too fancy, and we don't have to drop it the second pytest releases. ;) Otherwise followed your suggestion and put that sequence of letters in. |
This reverts commit 8d86d31.
These xfails are being printed as tracebacks on every test run. Since they are known to not work, let's hide them for now.