From 425fd7728060e72bb29299561fc5e3347b480bcb Mon Sep 17 00:00:00 2001 From: Ian Jackson Date: Sat, 4 Feb 2023 13:47:30 +0000 Subject: tests: Bail on commits tagged "[dgit ... INTERNAL ...]" We are about to generate some of these. I have tested ad-hoc that this does in fact trigger, as follows: * Start with the tip of my wip branch with the absurd git apply series filtering work completed. * Ad-hoc disable the git rebase in the absurd git apply cleanup, so that the series filtering commit ends up in the output: - if ($use_absurd) { + if (0 and $use_absurd) { * Revert the change to the test case absurd example; this leaves a test case which forces use of absurd git apply, but where the series filtering doesn't in fact alter the series file, so dgit is happy that the trees match. In this situation, the new test *did* complain about the INTERNAL commit. Signed-off-by: Ian Jackson --- tests/lib | 2 ++ 1 file changed, 2 insertions(+) diff --git a/tests/lib b/tests/lib index dc6a5ba..748a335 100644 --- a/tests/lib +++ b/tests/lib @@ -454,6 +454,8 @@ t-git-fsck () { 0) fail "unexpected messages from git-fsck" ;; *) fail "grep of git-fsck failed" ;; esac + + t-output "" git log --all --grep '^\[dgit .*INTERNAL.*]' } t-check-only-bpd () { -- cgit v1.2.3