Workflows for rebasing all changes after a fetch?

[bcspragu]

TL;DR - What's the equivalent of git-branchless's git sync in jj?

Longer explanation:

I frequently have five or six anonymous branches that I'm working on in a given repo, e.g.

@  ooozwmnq brandon@sprague.mx 2024-12-30 20:47:29 c0505549
│  Change #5
│ ○  vywnqtkx brandon@sprague.mx 2024-12-30 20:47:26 df4a26df
├─╯  Change #4
│ ○  vslmknsz brandon@sprague.mx 2024-12-30 20:47:15 bc0eeff7
├─╯  Change #3
│ ○  lqwuuksv brandon@sprague.mx 2024-12-30 20:47:13 98fe18c2
├─╯  Change #2
│ ○  lyoszovo brandon@sprague.mx 2024-12-30 20:46:58 8a60a6c1
├─╯  Change #1
○  mkqwwrry brandon@sprague.mx 2024-12-30 20:46:43 main b1750983
│  Create README.md
◆  zzzzzzzz root() 00000000

Now I merge in Change #1 (or I just fetch other people's changes), and main moves, e.g.

@  ssopxytu brandon@sprague.mx 2024-12-30 20:53:24 d82fb96d
│  (empty) (no description set)
◆  lyoszovo brandon@sprague.mx 2024-12-30 20:46:58 main 8a60a6c1
│  Change #1
│ ○  ooozwmnq brandon@sprague.mx 2024-12-30 20:47:29 c0505549
├─╯  Change #5
│ ○  vywnqtkx brandon@sprague.mx 2024-12-30 20:47:26 df4a26df
├─╯  Change #4
│ ○  vslmknsz brandon@sprague.mx 2024-12-30 20:47:15 bc0eeff7
├─╯  Change #3
│ ○  lqwuuksv brandon@sprague.mx 2024-12-30 20:47:13 98fe18c2
├─╯  Change #2
◆  mkqwwrry brandon@sprague.mx 2024-12-30 20:46:43 b1750983
│  Create README.md
~

What workflow are people using to rebase all those changes back onto main? For example:

1. Do people just manually jj rebase -s <first revision> -d main for each branch?
2. Something like jj rebase -b 'all:heads(all() ~ main)' -d main gets the job done (no idea if that's the best way to do it, I'm bad at revsets)
3. On one project, I have some alias like:

jj rebase -s 'all:roots(main..bookmarks(glob:"brandon/*"))' -d main --ignore-immutable

And then there's the approach Chris Krycho uses, where he has a wip revision with all his actual branches as parents, and then he does jj rebase -b all:wip -d main

All of these work, but I'm wondering if there's some canonical way that makes sense for the common case and people recommend for one reason or another.

[PhilipMetzger]

There's #1039 for that command and @essiene is working on it. It also is a long desired feature for a longtime user like me. It does mean that you'll need to use the workarounds you posted above until exists.

[bcspragu]

Ah excellent, thanks for the info!