Skip to content

Conversation

@basile-henry
Copy link

This small PR adds the option to start from <base> commit rather than from the beginning of the repo's history.

This has the advantage of being much faster when we only want to split a branch over the more recent commits. And it also makes it possible to keep the history before that <base> commit intact, which is useful when splitting a big PR into small PRs that all start from the same point in the repo's history.

I understand the current interface is not ideal, maybe basecommit..headref would be preferable to the current -b basecommit headref.
I'll see if I can implement this. Any tips to do that kind of parsing in bash? 😄

@djpohly
Copy link
Owner

djpohly commented May 21, 2018

I wonder if the parsing might be done by Git itself, which would be more ideal than shell parsing. For instance, if "base" is the merge-base of master and massive-pr, "master..massive-pr" would be the same as "base..massive-pr" when passed to most Git commands (including rev-list, which drives the main loop in split-branch). The original filter-branch does something like this and might be worth looking into, and our case is simpler, as we only accept one source branch.

FYI, done in shell, it'd be something like:

case $rev in
  *..*)
    start=${rev%%..*}
    end=${rev#*..}
    ;;
  *)
    start=$rev
    end=$rev
    ;;
esac

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants