Matthew DeVore <[email protected]> writes:
> t/t4202-log.sh | 4 ++++
> t/t8002-blame.sh | 4 ++++
> 7 files changed, 14 insertions(+), 1 deletion(-)
> ...
> diff --git a/t/t4202-log.sh b/t/t4202-log.sh
> index 153a506151..819c24d10e 100755
> --- a/t/t4202-log.sh
> +++ b/t/t4202-log.sh
> @@ -1703,4 +1703,8 @@ test_expect_success 'log --source paints symmetric
> ranges' '
> test_cmp expect actual
> '
>
> +test_expect_success '--exclude-promisor-objects does not BUG-crash' '
> + test_must_fail git log --exclude-promisor-objects source-a
> +'
> +
> test_done
> diff --git a/t/t8002-blame.sh b/t/t8002-blame.sh
> index 380e1c1054..eea048e52c 100755
> --- a/t/t8002-blame.sh
> +++ b/t/t8002-blame.sh
> @@ -118,4 +118,8 @@ test_expect_success '--no-abbrev works like --abbrev=40' '
> check_abbrev 40 --no-abbrev
> '
>
> +test_expect_success '--exclude-promisor-objects does not BUG-crash' '
> + test_must_fail git blame --exclude-promisor-objects one
> +'
> +
> test_done
OK. We used to be hitting BUG() which is an abort() in disguise, so
must-fail would have caught it without the fix in this patch. Now
we would see a more controlled failure.
... goes and makes sure that is the case ...
Not really. We were already doing a controlled failure via die(),
so these two tests would not have caught the problem in the code
before the fix in this patch.
But nevertheless this is a good change; I do not think it is worth
grepping for "unrecognized option" to differentiate the two cases.
Thanks.
