Ævar Arnfjörð Bjarmason <avarab@xxxxxxxxx> writes:
> Change the "t5516-fetch-push.sh" test code to make use of
> "test_when_finished" to remove data instead of having tests clean up
> leftover data from earlier tests, which may or may not be
> there (e.g. depending on the --run=* option).
>
> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@xxxxxxxxx>
> ---
> t/t5516-fetch-push.sh | 50 +++++++++++++++++++++++++------------------
> 1 file changed, 29 insertions(+), 21 deletions(-)
>
> diff --git a/t/t5516-fetch-push.sh b/t/t5516-fetch-push.sh
> index e4bb7581568..fbe0a72b0b2 100755
> --- a/t/t5516-fetch-push.sh
> +++ b/t/t5516-fetch-push.sh
> @@ -23,7 +23,8 @@ D=$(pwd)
>
> mk_empty () {
> repo_name="$1"
> - rm -fr "$repo_name" &&
> + test_when_finished "rm -rf \"$repo_name\"" &&
Any justification to swap between fr and rf?
> + test_path_is_missing "$repo_name" &&
So, the idea is that the philosophy so far was that each test clears
whatever the mess the previous ones created and empties what it cares
to be empty with "rm -fr", but now in the new world order, each test
relies on test_when_finished to clear any and all effects it leaves
to the environment?
OK. Since the clearing the "effects" is to remove the whole thing,
there is little room for such a plan to go wrong.
> @@ -191,32 +192,32 @@ grep_wrote () {
> grep 'write_pack_file/wrote.*"value":"'$1'"' $2
> }
>
> -test_expect_success 'push with negotiation' '
> - # Without negotiation
;-)
> +test_expect_success 'push without negotiation' '
> mk_empty testrepo &&
> git push testrepo $the_first_commit:refs/remotes/origin/first_commit &&
> test_commit -C testrepo unrelated_commit &&
> git -C testrepo config receive.hideRefs refs/remotes/origin/first_commit &&
> - echo now pushing without negotiation &&
> + test_when_finished "rm event" &&
> GIT_TRACE2_EVENT="$(pwd)/event" git -c protocol.version=2 push testrepo refs/heads/main:refs/remotes/origin/main &&
> - grep_wrote 5 event && # 2 commits, 2 trees, 1 blob
> + grep_wrote 5 event # 2 commits, 2 trees, 1 blob
> +'
>
> - # Same commands, but with negotiation
> - rm event &&
Good to split them into two. Presumably the later part wasn't
relying on the leftover side effects created by the early part?
Thanks. Looking good.
