I tested this series on our repository and found the following speedup in a case when there are 6k heads and 15k csets missing:

However when merge is aborted there is no unresolved merge in progress and hence

It will be nice to add information to commit message about how thing unshelving of a subset works.

The commit message is outdated.

This is fone so as to avoid the return statement

This logic is registered to the statesetection API as abortfunc.

hg abort currently supports --dry-run/-n flag only.
It is used to dry run hg abort

Also, can you add an entry about this to relnotes/next?

Maybe this deserves and entry in relnotes/next as a followup.

In D6582#96537, @taapas1128 wrote:

Okay I will do that. So should I remove --no-backup flag altogether from this series of patches because none of the commands support --no-backup then.

In D6582#96533, @taapas1128 wrote:

@pulkit Yeah I know that but look at D3965. histedit once did support --no-backup flag. It was dropped off for a different reason.

Looking closely, the backup value comes from reading ui.configbool('rewrite', 'backup-bundle') instead of a --no-backup flag. It means we can get rid of the --backup flag to hg abort as there exists no such flag for histedit.

It will be nice to mention about our motivation to move code in the commit message.

Tests have been shown as test-abort.t

Since you did the nice work of moving shelve to core, how about rebasing this patch on tip of default with shelve in core?

Can you add link of mailing list thread in your commit message? Link from markmail or https://www.mercurial-scm.org/pipermail/mercurial-devel/.

Since we both are discussing and working on this, we can understand why this patch does what it does. It will be nice to have more explained commit descriptions which explain why we need to have an function independent of rebase-runtime-state object. This applies to all the patches in the series. In the hg abort patch, explaining how it works will be good.

Didn't look at the actual implementation yet.

This patch needs to be rebased on tip of hg-committed as shelve is in core now.

Review ping for this stack.

Thanks a lot for working on this. This needs to be rebased on tip of default branch. Also, can you add an entry about shelve extension being moved to core in relnotes/next file.