On Thu, Mar 28, 2024 at 09:19:08AM -0700, Junio C Hamano wrote:

Jeff King [off-list ref] writes:

quoted

quoted

With the following band-aid, we can skip the test and the output
from "sh t4126-*.sh -i -v -x" might give us a clue that explains how
such a failure happens.  Unfortunately GitHub CI's win test does not
give us insight into a test that did not fail, so I did not get
anything useful from the "ls -l" down there (I already knew that
sample patches are empty files).

We package up the failed test output and trash directories for each run.
You can find the one for this case here:

  https://github.com/git/git/actions/runs/8458842054/artifacts/1364695605

What I meant was that with the band-aid that (1) sets prerequisite
so that Windows would not fail and (2) has some diagnostic in the
code that sets prerequisite, because the overall test does not fail,
we do not package up that diagnostic output.

Right, I meant that we could look at the run without the band-aid (which
is what the link points to). But I guess maybe you realized already that
it would not be helpful because of the "reset --hard" that the test
does.

-Peff