Repository navigation
docs(rfc): keep a slash-command name a rename gave up bound to its workflow - #28632
jeeshofone wants to merge 1 commit into
Conversation
6b9a5f9 to
5720ea9
Compare
First Principles Review (Opus 5.5) — ⏭️ skippedRevision |
UX Review (Opus 5.5) — ⏭️ skippedRevision |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS This PR only adds a draft RFC. The design reuses #27528's Notes
[DESIGN-REVIEWED] 8f98010 Description read[DESCRIPTION-READ] 7945e6dda893137238d0f5a7cecd792881571a72ea5ba19562fa13d82c370297 That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
GPT 6.1 Review — ✅ no blocking findingsGPT 6.1 completed its review of This comment is updated in place on each push. Review detailsNo findings. False positive or not applicable? A repository writer can comment: |
Opus 5.5 Review — ✅ no blocking findingsReviewed Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
|
Hold: merge conflict, plus one design point to settle. Reviewed at 5720ea9. GitHub reports this PR as conflicting with main, in Before the next round, please answer the Design Review note on destructive routes. §1.3 lets DELETE resolve through a former slug, so a stale — crew-pr-reviewer (r143) |
5720ea9 to
706df70
Compare
|
Rebased onto main (706df70), both index rows kept. §1.3 now limits PATCH and DELETE to id or current slug, so a stale |
706df70 to
2165b45
Compare
bolichen97
left a comment
There was a problem hiding this comment.
Design brief for rfc-workflow-rename-reservation (read in full at 8f2fa40).
What it does. When a rename changes a saved workflow's slug, the record keeps the old slug in a new former_slugs list. _refs then counts former slugs, so slug allocation never hands one to another workflow, and removal reserves them with the id and slug. A /workflow <former slug> run reaches the renamed workflow through a separate resolve_for_run lookup and says which command it now runs as. Plain get, PATCH and DELETE still match only an id or a current slug.
Pros.
- It fixes the silent rebinding in #23680: a written-down
/workflow deploycan no longer start a different workflow. - Rename becomes as safe as removal, closing the gap the removal RFC's section 4 records.
- The redirect keeps scheduled messages working and tells the person to update them.
- Destructive calls cannot reach a record through a former slug.
- The field is optional on read, so no migration is needed.
Cons.
- A name given up by a rename is never released, so a new workflow asking for it gets a numbered slug such as
deploy-2, or anexact_slugrefusal. That is a small visible change, and releasing a name is left as a non-goal. - It is blocked on #27528 (or needs the fallback in section 3).
- Renames made before it lands are not back-filled.
Tech debt. No. It adds one optional list on the record and one lookup step on the run path.
Maintainer signs the direction. This is a comment, not an approval.
8f2fa40 to
a987769
Compare
|
HOLD (conflict) at a987769. The branch conflicts with main again in The RFC itself needs nothing more from me. Our r143 point is answered: PATCH and DELETE match only an id or current slug (§1.3). The Design note asking whether an older build drops — crew-pr-reviewer (r146) |
a987769 to
15ba375
Compare
|
@iamwhatever thanks for the r146 HOLD. Rebased onto main as 15ba375, still one commit. |
iamwhatever
left a comment
There was a problem hiding this comment.
Approve at 15ba3759ea3537787725debfe2877ba345166fe1.
This is a docs-only draft RFC. Our r146 hold was only for the README conflict, and that conflict is gone; the index keeps main's rows. PATCH and DELETE still match only an id or a current slug (rfc#L66). All checks are green, and @bolichen97 signed the direction.
AI findings judged non-blocking
- Design: §6 does not say whether an older build that rewrites a record drops
former_slugs. Does not block: main's rename builds the record from**current, so the field survives (library.py#L314). A sentence in §6 saying so is optional.
Note: #28664 and #28633 edit the same README sentence, so the PRs that merge after this one will need another rebase.
— crew-pr-reviewer (r148)
15ba375 to
9576075
Compare
iamwhatever
left a comment
There was a problem hiding this comment.
APPROVE at 7920fb2.
This is a docs-only draft RFC plus its index row. The RFC text is identical to 15ba375, which I approved in r148; this head is a rebase, and the README change is the one new row and audit sentence with no conflict. PATCH and DELETE still match only an id or a current slug, and former-slug resolution stays on the run paths (rfc#L66). All checks are green, and @bolichen97 signed the direction.
AI findings judged non-blocking:
- Design (note): §6 does not say whether an older build that rewrites a record drops
former_slugs. Main's rename builds the record from**current, so the field survives (library.py#L314). A sentence in §6 saying so is optional.
— crew-pr-reviewer (r149)
|
A maintainer has already signed off the design direction on this PR. It now waits for a maintainer approval on this head, because the earlier approvals were dismissed by the rebases. Review result on bff19ca: no blocking findings. This is a docs-only draft RFC plus one index row and one audit sentence in the RFC README. The review checked the main-side claims against Optional:
|
|
HOLD (conflict) at bff19ca.
The RFC itself is unchanged since 7920fb2, which we approved in r149, and PATCH and DELETE still match only an id or a current slug (§1.3). All checks are green. The Captain's three optional notes do not block a draft RFC; a sentence for the §1.3 tie-break would help the implementation PR. A new head needs a fresh review. — crew-pr-reviewer (r150) |
Code change approver recommendation: ApproveWhy: the decider estimates a 6% chance that this change introduces a defect, within the 12% limit for approving on its own, and every check passes. This PR only changes docs. It adds a draft RFC and one index row in the README, and it changes no code, so no test is needed. I checked the RFC's claims about main against
The design hangs together:
An automated recommendation, not a review: a maintainer still decides. It approves on its own only when the estimated chance of a defect is at most 12% and the change has at most 400 changed lines. Decider v3c (217fe607) · commit |
|
Rebased onto current main, past the #39262 guard fix that was failing Backend Tests shard 8. The RFC text is unchanged. |
Part of #23680
Problem / Motivation
Goal: Main records a design for keeping a slash-command name that a saved workflow gave up by renaming bound to that workflow, before any code for it is written.
Renaming a saved workflow frees its old slash-command name at once, so a later workflow can take it and a written-down
/workflow <old name>silently runs the later one. The removal reservation in #27528 covers only the names a workflow holds when it is removed. The maintainers parked #23680 and asked for a design through docs/request-for-change.Why it matters
A scheduled message or prompt that names a workflow by its command can start a different workflow after a rename, with nothing to say it changed.
Not a goal
What changed (motivation → approach → change)
Adds
docs/request-for-change/rfc-workflow-rename-reservation.md(statusdraft) and its row in the index indocs/request-for-change/README.md. The document covers the problem, alternatives including doing nothing, the recommended design (aformer_slugslist on each record, counted as taken by slug allocation and reserved on removal through #27528's_refs), how it builds on #27528, a migration plan and a test plan.Backwards compatibility
Compatible: documentation only.
Tests
N/A: documentation only.
scripts/docs_lint.pyruns in Source Gates.Manual verification
N/A: documentation only.
Screenshots / video
N/A: no UI change.