Ævar Arnfjörð Bjarmason <ava...@gmail.com> writes:
> Mark the tests where "git fsck" fails at the end with extra test code
> to check the fsck output. There fsck.{err,out} has been created for
> us.
>
> A later change will add the support for GIT_TEST_FSCK_TESTS. They're
> being added first to ensure the test suite will never fail with
> GIT_TEST_FSCK=true during bisect.
I am sympathetic to what step 3/3 (eh, rather, an earlier "let's not
leave the repository in corrupt state, as that would make it
inconvenient for us to later append new tests") wants to do, but not
this one---these markings at the end makes it inconvenient for us to
later add new tests to these script before them.
