D1082: split: new extension to split changesets

[quark (Jun Wu)]

quark added inline comments.

INLINE COMMENTS

> martinvonz wrote in split.py:90
> s/ctx/rev/ seems a little clearer (I realize that ctx has a __int__ method, but I didn't know that until I read this)

Good advice. This is probably copy-pasted from old code.

> martinvonz wrote in split.py:95
> Do we want to exclude all troubled commits? Maybe phase- and content-divergent ones are better to not rebase too? (I.e. '%ld - obsolete() - troubled()')

I think it's better to rebase content-divergent changesets here. Since that enables people to check the final result of the divergence and be able to make a final judgement about which one to pick or whether to merge the divergence.

I also think in the future it might make sense to rebase some unstable ones too as long as there is a fast way to test if doing that won't cause merge conflicts.

> martinvonz wrote in split.py:98
> maybe s/will/would/ ?

Will change.

> martinvonz wrote in split.py:109
> Do we really want to allow splitting when there's an unfinished merge? Or is that handled elsewhere?

Good catch. This is unintentional. Will remove.

REPOSITORY
  rHG Mercurial

REVISION DETAIL
  https://phab.mercurial-scm.org/D1082

To: quark, #hg-reviewers
Cc: lothiraldan, martinvonz, dlax, mercurial-devel