jj-parallelize

restructures a series of sequential commits to be parallel siblings

TLDR

Make a range of revisions siblings of each other

$ jj parallelize [rev1]::[rev2]

Parallelize the current revision and its parent

$ jj parallelize @-::@

Parallelize specific revisions by change ID

$ jj parallelize [xyz] [abc]

jj parallelize declares that a set of revisions are independent of each other: they stop being ancestors or descendants of one another and become siblings. Running jj parallelize 1::2 on a linear history 0-1-2-3 makes 1 and 2 both children of 0, and 3 becomes a merge of 1 and 2.Revisions outside the set keep their relationships: former ancestors of a revision in the set stay ancestors, and former descendants stay descendants. Because of this, parallelizing non-adjacent revisions such as '1 | 3' with 2 in between is a no-op.Useful for splitting a stack of independent changes so they can be reviewed or merged separately.