Fix: Revert with allow empty

[N-o-Z]

Closes #8801

Waiting for a single approval

[github-actions]

E2E Test Results - DynamoDB Local - Local Block Adapter

[github-actions]

E2E Test Results - Quickstart

[itaigilo]

Nice catch, thanks for fixing.

Looking good, approving:

Though when looking at the code, CherryPick seems to use the same pattern.
Should we handle it as well, or open an Issue for it?

[N-o-Z]

Nice catch, thanks for fixing.

Looking good, approving:

Though when looking at the code, CherryPick seems to use the same pattern. Should we handle it as well, or open an Issue for it?

We should probably open a separate issue for it

[yonipeleg33]

Surely out of scope for this PR, but git-revert doesn't have this flag. Why did we add it instead of aligning with whatever git is doing when reverting an empty commit?

[itaigilo]

Surely out of scope for this PR, but git-revert doesn't have this flag. Why did we add it instead of aligning with whatever git is doing when reverting an empty commit?

AFAICT, it's not about adding this as a flag, but it allows an empty commit for revert -
Assuming that if the original commit was empty, it's ok to allow the reverted commit, which is also empty.

[N-o-Z]

Surely out of scope for this PR, but git-revert doesn't have this flag. Why did we add it instead of aligning with whatever git is doing when reverting an empty commit?

Definitely out of scope of this PR.
I don't remember the requirement consideration then but I'd like to give a more general answer as this question keeps popping up frequently.
Short answer: lakeFS is not Git.
It has its principles based on Git and it draws concepts from it but we're not trying to "comply to a Git spec". Building a versioning engine for data engineers we will and should diverge from Git behavior or principles of it serves the mainstream user better.