Is it possible to squash randomly selected commits?

[pentago]

Is your feature request related to a problem? Please describe.

I, and probably lots of other people, occasionally need to squash/fixup commits that are not directly one below another and would appreciate it if we could select random commits to squash/fixup.

Describe the solution you'd like:

Instead of allowing to squash/fixup only commits that are directly one below another in history, I'd like to be able to squash/fixup random commits.

For example, consider this scenario:

Let's say I only want to squash/fixup only commits YYY and WWW to Init.

I'm not sure how would I do it, perhaps there's an existing lazygit workflow that just isn't that obvious, or at least my understanding of git features and terminology makes me unsure which I could use to achieve the goal.

If my request is not sufficient for a new feature, at least I'd appreciate instructions on how to achieve the same with the existing feature set.

I just thought that it would be great if we could:

1. hit v to select commits to squash
2. hit c-f (to define target commit in a popup)
3. hit enter to perform squash or fixup

Describe alternatives you've considered
N/A

Additional context
N/A

[stefanhaller]

Here's a long discussion of how the multi-selection feature was designed, and why. It explains the reasons why we decided to only support contiguous range selections; I don't think we need to discuss this again here.

Which means that you will have to reorder your commits (using ctrl-j and/or ctrl-k) so that they are contiguous, and then you can squash the whole range.

If you know ahead of time that you are going to want to squash "YYY" into "Init", maybe you should get into the habit of committing it as "fixup! Init", so that you can later auto-squash it easily.

[ChrisMcD1]

/close

To throw another option into here, you can enter an interactive rebase on the bottom commit with e, and then use <c-j> and <c-k> to move the target commit down before choosing s to mark it as squash, and then remove it. This is often faster once you have to move a commit more than a few places, since you are not constantly entering and existing a rebase.