Re: [RFC PATCH v4 0/6] CI: js/ci-github-workflow-markup rebased on "use $GITHUB_ENV"
- From
Ævar Arnfjörð Bjarmason <avarab@gmail.com>
- Date
- Apr 14, 2022, 12:47 UTC
- Message-ID
- <220414.86tuavde26.gmgdl@evledraar.gmail.com>
- In-Reply-To
- <xmqq35ig2c6l.fsf@gitster.g>
On Thu, Apr 14 2022, Junio C Hamano wrote:
Show 46 quoted lines
> Ævar Arnfjörð Bjarmason <avarab@gmail.com> writes:
>
>> 6: 1d2b94436fc ! 6: b291f64821c ci: call `finalize_test_case_output` a little later
>> @@ t/test-lib.sh: trap '{ code=$?; set +x; } 2>/dev/null; exit $code' INT TERM HUP
>> test_failure=$(($test_failure + 1))
>> say_color error "not ok $test_count - $1"
>> shift
>> - printf '%s\n' "$*" | sed -e 's/^/# /'
>> - test "$immediate" = "" || _error_exit
>> +@@ t/test-lib.sh: test_failure_ () {
>> + say_color error "1..$test_count"
>> + _error_exit
>> + fi
>> + finalize_test_case_output failure "$failure_label" "$@"
>> }
>
> This part, specifically the following from [6/6], looks iffy.
>
> @@ -782,13 +782,13 @@ trap '{ code=$?; set +x; } 2>/dev/null; exit $code' INT TERM HUP
> # the test_expect_* functions instead.
>
> test_ok_ () {
> - finalize_test_case_output ok "$@"
> test_success=$(($test_success + 1))
> say_color "" "ok $test_count - $@"
> + finalize_test_case_output ok "$@"
> }
>
> test_failure_ () {
> - finalize_test_case_output failure "$@"
> + failure_label=$1
> test_failure=$(($test_failure + 1))
> say_color error "not ok $test_count - $1"
> shift
> @@ -798,18 +798,19 @@ test_failure_ () {
> say_color error "1..$test_count"
> _error_exit
> fi
> + finalize_test_case_output failure "$failure_label" "$@"
> }
>
>
>
> With the other 29-patch series applied on the same base as before,
> test_failure_ does not have such "fi" inside. Misapplication of
> rebase or something?This re-submission was rebased on "master", so that "fi" in the context is from the now-landed ab/test-tap-fix-for-immediate.
I saw you'd fixed that conflict already, but figured rebasing before submission (as usual) would be helpful anyway, sorry about the confusion.
> In any case, I've wiggled both series in and rebuilt 'seen'. > Looking good as before.
Thanks, the end-state of the resolution looks good, and matches what I have locally.