Ævar Arnfjörð Bjarmason [off-list ref] writes:
In any case, your patch clearly undoes whatever canary for gc issues
df4c0d1a792 was trying to put into the test-lib, but didn't say so in
its commit message.
So I figured it was something that was missed at the time, and that I
should send a quick E-Mail to both authors to see if anyone cared, maybe
nobody does.
It's just something I ran into while reviewing test-lib.sh for some
unrelated changes I was making...
Good eyes.
I am not quite sure if we are better off catching a transitory
failure, which could go away when retried in 5 seconds, to remove
the test directory (in which case, the change in question is a clear
regression from what df4c0d1a (test-lib: abort when can't remove
trash directory, 2017-04-20) intended to catch), or we only want to
catch permanent failures (in which case the patch in question is a
clear improvement).
In either way, the decision and the rationale behind it should be in
the log message.
Thanks.