git-machete-intellij-plugin

git-machete-intellij-plugin copied to clipboard

In case when slide in failure (e.g. submodule misconfiguration) new branch entry is added to machete file anyway

@micpiotrowski please add the specific steps to reproduce this one

1. create git repository and subrepository in it
2. corrupt submodule settings (e.g. edit patch to submodule in main-repo/.git/modules/<submodule-name>/config in worktree key
3. try to to slide in new branch in that submodule repo

As effect branch will not been created (creation failure) but branch entry will be added to machete file

This is one of examples why branch creation may failed (in this case it is missconfiguration)

corrupt submodule settings (e.g. edit path to submodule ...)

Hmm to be fair very little can be guaranteed (also by git itself) in case someone corrupts vital settings like this one...

I think this one is at most minor (if it's an issue at all) unless it's also likely to affect users in more frequent corner cases

Hmm to be fair very little can be guaranteed (also by git itself) in case someone corrupts vital settings like this one...

Yes, I agree but the same can happen if by any other reason branch will not been created. But ok, it's not so important as now.

@micpiotrowski but do we know any other reason that might lead to this issue?

After a quick look at the present approach, I can say that there is no way to handle branch creation failures. Ofc we could get deeper and use GitBranchWorker (where the handling seems to be possible) directly. We lose the reused dialog then.

@micpiotrowski but do we know any other reason that might lead to this issue?

Actually, no.

After a quick look at the present approach, I can say that there is no way to handle branch creation failures. Ofc we could get deeper and use GitBranchWorker (where the handling seems to be possible) directly. We lose the reused dialog then.

So leave as is. It seems to be really not so important that I think at the beginning.

Whoops, actually this seems to also happen when the checkout is declined by the user (due to possible file conflicts)... branch is still getting slid in. Which I think merits NOT considering this issue minor anymore :/

I have a little bit of problem in reproducing the issue.

Quite likely the issue no longer exists, might've been fixed over the last two years without us noticing ;) Closing then

new issue: the entry does not get deleted from the machete file, upon the deletion of the branch.

upon the deletion of the branch

You mean when branch is deleted e.g. via git branch -d ..., outside of git-machete right? that's expected in general!