"d" instead of reverting submodule, deletes it, inside a commit

[ilan-schemoul]

Describe the bug

Usually if I click on a commit, I see a modified file and I click "d" it reverts the changes. But in case of submodule, it deletes the entire submodule.

To Reproduce

Steps to reproduce the behavior:

1. Create a commit with a modified sumodule and commit the new SHA1 of the modified submodule in the parent
2. Go in the list of commits [4] and enter on the commit
3. Go down the submodule click delete
   The entire submodule is deleted

Expected behavior

Same as regular files, checkout the HEAD~, not delete it

Screenshots

If applicable, add screenshots to help explain your problem.

Version info:

0.55
git 2.43.0

Additional context

Add any other context about the problem here.

Note

Please try updating to the latest version or manually building the latest master to see if the issue still occurs.

[stefanhaller]

@ilan-schemoul Wow, that's pretty bad, thanks for reporting. It's especially bad because it's a pretty common scenario; it's easy to accidentally include submodule changes in a commit by pressing a when the submodules are not up to date, and then it would be useful to drop these again quickly using d.

Here's a draft PR with a proposed fix: #4937. It's not quite done yet (tests are missing), but reviews and/or tests are welcome already.

[stefanhaller]

The PR should be ready.

Note that it only really fixes the behavior for modifications to a submodule. If you press d on either the addition or the removal of a submodule, dropping the change itself is not enough; the change to .gitmodules should also be dropped, which is likely contained in the same commit. I'm not sure how to deal with this; it feels to me that automatically doing this is too much to ask of lazygit, but maybe it should at least warn the user that they should do it manually. Both is out of scope for this fix, that's also why I only added a test for a modified submodule to the PR, not for added or removed ones.

[stefanhaller]

@ilan-schemoul Can you please confirm that #4937 fixes the issue for you?

[systec-ad]

I noticed the same issue and can confirm that #4937 fixes the behavior.

Thanks! (for the fix and this amazing piece of software ❤️)