apply tests: use "test_must_fail" instead of ad-hoc pattern
Change a fragile test pattern introduced in 6b763c424e4 (git-apply: do not read past the end of buffer, 2007-09-05). Before this we wouldn't distinguish normal "git apply" failures from segfaults or abort(). I'd previously marked this test as passing under SANITIZE=leak in f54f48fc074 (leak tests: mark some apply tests as passing with SANITIZE=leak, 2021-10-31). Let's remove that annotation as this test will no longer pass. Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Ævar Arnfjörð Bjarmason committed
Mar 7, 2022 at 13:48 UTC
1ff543acd01b9e8cfe9af7c262d118ab0bf4d5e7
1 file changed
+2
-16
t/t4123-apply-shrink.sh
+2
-16
@@ -2,8 +2,6 @@
2
3
test_description='apply a patch that is larger than the preimage'
4
5
-
6
-TEST_PASSES_SANITIZE_LEAK=true
5
. ./test-lib.sh
6
7
cat >F <<\EOF
@@ -41,20 +39,8 @@ test_expect_success setup '
39
'
40
41
test_expect_success 'apply should fail gracefully' '
44
-
45
- if git apply --index patch
46
- then
47
- echo Oops, should not have succeeded
48
- false
49
- else
50
- status=$? &&
51
- echo "Status was $status" &&
52
- if test -f .git/index.lock
53
- then
54
- echo Oops, should not have crashed
55
- false
56
- fi
57
- fi
42
+ test_must_fail git apply --index patch &&
43
+ test_path_is_missing .git/index.lock
44
'
45
46
test_done