Hi Kristoffer
On 31/10/2024 17:25, Kristoffer Haugsbakk wrote:
On Thu, Oct 31, 2024, at 17:30, Phillip Wood wrote:
The diff below shows a fix and a new test that fails without the
sequencer changes. The fix is based on master, so it might need
updating to go on top of Junio's series. The test could probably
be improved to use the existing setup.
Thank you! That those lines apparently worked had kind of been
bothering me. It’s nice to get some clarity on the issue.
But shouldn’t there be a signoff somewhere if I am to incorporate that
diff into the series? Or is the signoff implied?
Here's a nicer test case, the key is that we need a conflicting
"fixup -C" after a "fixup" otherwise we just end up using the
fixup message file which only ever contains a single commit message.
Here's my signoff for this and the sequencer changes in my previous mail
Signed-off-by: Phillip.Wood <phillip.wood@xxxxxxxxxxxxx>
Best Wishes
Phillip
---- >8 ----
diff --git a/t/t3437-rebase-fixup-options.sh b/t/t3437-rebase-fixup-options.sh
index 7929e2e2e3a..a4b90e881e3 100755
--- a/t/t3437-rebase-fixup-options.sh
+++ b/t/t3437-rebase-fixup-options.sh
@@ -127,6 +127,21 @@ test_expect_success 'fixup -C with conflicts gives correct message' '
test_cmp expected-author actual-author
'
+test_expect_success 'conflicting fixup -C after fixup with custom comment string' '
+ test_config core.commentString COMMENT &&
+ test_when_finished "test_might_fail git rebase --abort" &&
+ git checkout --detach A3 &&
+ test_must_fail env FAKE_LINES="1 fixup 2 fixup_-C 4" git rebase -i A &&
+ echo resolved >A &&
+ git add A &&
+ FAKE_COMMIT_AMEND=edited git rebase --continue &&
+ test_commit_message HEAD <<-\EOF
+ A3
+
+ edited
+ EOF
+'
+
test_expect_success 'skipping fixup -C after fixup gives correct message' '
test_when_finished "test_might_fail git rebase --abort" &&
git checkout --detach A3 &&
