Cannot dynamically mark a test as xfail

[gbgbgbg]

From inside a test, the test cannot be marked as xfail (or xpass)

This is useful when particular test combinations are invalid.

I believe the test xtest_dynamic_xfail_set_during_setup covers this case, but by starting with xtest, it is not being run. I have not found any documented reason for its exclusion, nor a ticket that addresses.

[pytestbot]

GitMate.io thinks the contributor most likely able to help you is @RonnyPfannschmidt.

[RonnyPfannschmidt]

test_dynamic_xfail_no_run shows how it currently works in a kind of messy way, i beleive that mechanism can be used to do what i think you asked for

please report if my guess was correct

[gbgbgbg]

I want to be able to mark a test as xfail, and then let it run, and have a possibility of generating a XPASS result

[gbgbgbg]

ah, it looks like run=True is the trick!

This will work for me, but the xtest_dynamic_xfail_set_during_setup case is more natural for me

Why is xtest_dynamic_xfail_set_during_setup disabled?

[RonnyPfannschmidt]

according to the history 8 years ago when applymarker got introduced the test had gotten "disabled"

[RonnyPfannschmidt]

i'm closing as the actual question seems to be answered, in case something is missing please call out and thanks

[gbgbgbg]

is there a ticket for re-enabling xtest_dynamic_xfail_set_during_setup ?

[RonnyPfannschmidt]

no, alos that test uses a mechanism which instead of modifying the collected test, it modifies the function of the test (which may be used in different tests

as part of the upcoming mark cleanup we are going to remove that mechanism eventually (with some deprecation grace periods

[sharan-cs2000]

this is due to
assert 0, "Could this be represented as an exemption" --> comparing against integer value, upon failure you will above issues, correcting this to:
assert "Could this be represented as an exemption", "Could this be represented as an exemption" --> correct values

XFAIL This is stdout

XFAIL This is stderr

As work-around we can use it accordingly, need to look assert() against sys.stdout.write/sys.stderr.write usage..

[rth]

test_dynamic_xfail_no_run shows how it currently works in a kind of messy way, i beleive that mechanism can be used to do what i think you asked for

For other users who end up here, the following appears to work,

def test_something(request):
    if some_dynamic_condition:
        request.applymarker(pytest.mark.xfail(run=False))

    # continue running the code below in any case
    some_other_code()

It might be helpful to have this as part of the docs.