[PATCH] t5551: "GIT_TEST_LONG=Yes make test" is broken

[PATCH] t5551: "GIT_TEST_LONG=Yes make test" is broken

[Junio C Hamano]

The "test_expect_success 'tag following always works over v0 http'"
test in t5551 fails when it tries to run "git init tags", but this
happens only when EXPENSIVE test is allowed to run.  

This is because the step tries to create a repository with "git init
tags" but the EXPENSIVE test that runs way before it creates and
leaves around a temporary file "tags".  Have the EXPENSIVE test
clean it up after itself.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
---

 * As we should give individual contributors a stable base, I have
   identified tests that fail with EXPENSIVE enabled in "master".
   With this fixed, we should add GIT_TEST_LONG=YesPlease to the CI
   jobs run upon PRs are created by contributors.

 t/t5551-http-fetch-smart.sh | 1 +
 1 file changed, 1 insertion(+)

diff --git c/t/t5551-http-fetch-smart.sh w/t/t5551-http-fetch-smart.sh
index a26b6c2844..e236e526f0 100755
--- c/t/t5551-http-fetch-smart.sh
+++ w/t/t5551-http-fetch-smart.sh
@@ -481,6 +481,7 @@ test_expect_success 'test allowanysha1inwant with unreachable' '
 '
 
 test_expect_success EXPENSIVE 'http can handle enormous ref negotiation' '
+	test_when_finished "rm -f tags" &&
 	(
 		cd "$HTTPD_DOCUMENT_ROOT_PATH/repo.git" &&
 		create_tags 2001 50000

Re: [PATCH] t5551: "GIT_TEST_LONG=Yes make test" is broken

[Jeff King]

On Fri, May 08, 2026 at 02:31:03PM +0900, Junio C Hamano wrote:

> The "test_expect_success 'tag following always works over v0 http'"
> test in t5551 fails when it tries to run "git init tags", but this
> happens only when EXPENSIVE test is allowed to run.  
> 
> This is because the step tries to create a repository with "git init
> tags" but the EXPENSIVE test that runs way before it creates and
> leaves around a temporary file "tags".  Have the EXPENSIVE test
> clean it up after itself.

I wondered how and when this broke, but it looks like it has been since
2024 when the later v0 test was added. And it looks like I am the
culprit for both the EXPENSIVE test adding the "tags" file and the v0
test assuming the path is available. :)

Anyway, this seems like the obviously correct solution.

-Peff