fix: retarget PR to master when parent PR has been landed

[edmondop]

Summary

• After landing the bottom of a stack and rebasing, jj spr diff now retargets the new bottom PR from its synthetic base branch to master
• Extract determine_base_branch() function for testability
• Deletes the stale synthetic base branch after retargeting
• 6 unit tests covering all base branch determination scenarios

Previously, after landing PR #1 in a stack and rebasing PR #2 onto master, running jj spr diff would keep PR #2 targeting the old synthetic base branch. GitHub would then show the full diff (including already-landed changes) instead of just PR #2's changes.

Test plan

• cargo test — all tests pass
• cargo clippy --all-features --all-targets -- -D warnings — clean
• Land bottom of a stack, rebase, run jj spr diff, verify PR is retargeted to master

[jennings]

Nice! I noticed this the other day and could have sworn it was supposed to already do this.

[jennings]

I tried this and the pull request closed before it was able to retarget.

It looks like we're deleting the old branch before retargeting the pull request to main. Am I reading it right?

If so, I think we need to update the PR first, then delete the previous target branch.

This is the output I got:

❯ jj spr diff -r u
57100fd <snip>
  #️⃣   Pull Request #418: https://github.com/<snip>/pull/418
Message (leave empty to abort): rebase
  ⚾  Commit was rebased - updating Pull Request #418
  🎯  Retargeting Pull Request #418 to main
  🛑  GitHub: Validation Failed
      Documentation URL: https://docs.github.com/rest/pulls/pulls#update-a-pull-request
      Errors:
      - {"code":"invalid","field":"base","message":"Cannot change the base branch of a closed pull request.","resource":"PullRequest"}

[edmondop]

Thank you, this is actually something that I had fixed on my fork, and I forgot to update the PR! Let me update the PR

[edmondop]

Should be fixed with 4068da5

[jennings]

I've played with the new version and it worked at least once, but today this happened:

$ jj spr diff

# oops, that should have been cherry picked.

$ jj spr diff --cherry-pick

This is what the PR says:

The "deleted the branch main" is alarming. It didn't actually get deleted because it's protected.

If you don't fix it first, I hope to find time to fix it myself because I really like this feature.

[edmondop]

I have never faced this issue can you try to find out how you got there? Happy to fix it once I have a repel

…

On Mon, Mar 30, 2026 at 11:28 AM Stephen Jennings ***@***.***> wrote: *jennings* left a comment (LucioFranco/jj-spr#110) <#110?email_source=notifications&email_token=AAGGOKMD2IL5ITI6N7SUMD34TKHAVA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMJVGU4TENZSGA4KM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2LK4DSL5RW63LNMVXHIX3POBSW4X3DNRUWG2Y#issuecomment-4155927208> I've played with the new version and it worked at least once, but today this happened: $ jj spr diff # oops, that should have been cherry picked. $ jj spr diff --cherry-pick This is what the PR says: image.png (view on web) <https://github.com/user-attachments/assets/bbdaccf0-8f9e-4ba3-b504-2bdba94d4094> The "deleted the branch main" is alarming. It didn't actually get deleted because it's protected. If you don't fix it first, I hope to find time to fix it myself because I really like this feature. — Reply to this email directly, view it on GitHub <#110?email_source=notifications&email_token=AAGGOKMD2IL5ITI6N7SUMD34TKHAVA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMJVGU4TENZSGA4KM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2LK4DSL5RW63LNMVXHIX3POBSW4X3DNRUWG2Y#issuecomment-4155927208>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAGGOKOSJELEDPNAEGB27AD4TKHAVAVCNFSM6AAAAACWF34CQKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DCNJVHEZDOMRQHA> . You are receiving this because you authored the thread.Message ID: ***@***.***>