Repository navigation
docs: record rfc-large-session-bounded-restore as accepted - #36148
Conversation
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 Docs-only status change, backed by a linked maintainer acceptance; it removes nothing and adds no surface. Notes
[DESIGN-REVIEWED] 6baaa44 Description read[DESCRIPTION-READ] 938056ed41359bc725b8124f98d854b4979aabda197308948cc0dc94caf06837 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 |
Opus 5.5 Review — ✅ no blocking findingsReviewed Review detailsNo findings. Notes:
[OPUS-REVIEWED] 6baaa44 Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
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: |
Code change approver recommendation: A person should decideWhy: the decider thinks the change is probably fine but estimates a 22% chance that this change introduces a defect, above the 12% limit for approving on its own. This PR only changes RFC status text: it records acceptance of rfc-large-session-bounded-restore, moving it from draft to partial and linking the acceptance comment on #35875. 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 |
bolichen97
left a comment
There was a problem hiding this comment.
I read the diff. Before this status change can merge I could not find a public record of the acceptance it states, and I would like a link.
The README row and the PR description say the RFC was "approved in review by KiroCrewCaptain" on #35916 and that "the repository owner's lead accepted it as the plan". What I can see is:
- #35916 merged as a draft RFC with one approval, from KiroCrewCaptain, whose own review comment says a person should decide that PR.
- The tracking issue #35875 carries the
tier:T3andparkedlabels, and has no comment from a maintainer or the repository owner accepting the plan.
Other accepted RFCs in the README name where the decision is written, such as a maintainer's sign-off on a pull request or a decision recorded in the document's Open questions. Could you add the link to the owner's acceptance (a comment, a review or a decision record) to the PR description, and name it in the README row and the RFC's Status line in place of the sentence about the owner's lead? If the acceptance has not been given yet, the status should stay draft until it is.
I am not approving until that record is linked.
KiroCrewCaptain
left a comment
There was a problem hiding this comment.
Thanks for keeping this small. Before this status change can merge, it needs one fix to the decision record.
docs/request-for-change/rfc-large-session-bounded-restore.md:17anddocs/request-for-change/README.md:31say the RFC was "approved in review by KiroCrewCaptain on #35916; the repository owner's lead accepted it as the plan."- Trigger: this merges as written, then #36141 (1.2) and #36259 (1.1) rebase. Their base now carries
status: accepted. - Outcome: the First Principles lane reads that field from base as the recorded product decision. Two behaviour changes users will see (a large tab opens with only its newest rows, and the sidebar order changes during restore) would pass that check on a record that names no human decider. The approval on #35916 comes from an automation account, and it checked a draft document for accuracy. It did not accept the plan. The "owner's lead" has no name and no link, and #35875 has no comments. GOVERNANCE.md says decisions happen in public on the issue or pull request they belong to. A maintainer asked for this same link in a review on this head.
- Smallest fix: on the Status line, in the README row and in the PR body, replace that sentence with a link to a named maintainer's public acceptance (a comment, a review or a decisions entry). If no such record exists yet, keep
status: draftuntil a maintainer records the decision.
- Trigger: this merges as written, then #36141 (1.2) and #36259 (1.1) rebase. Their base now carries
Optional:
- Section 5 ("Decisions requested") still lists the 16 MB threshold, the 500-row window and the three behaviour changes as open. Once the decision is recorded, write down which values were accepted.
- The README row says part 1.4 (#35960) is already merged. The vocabulary calls that
partial, notaccepted("Nothing built yet"). Both values clear the same check. implementation-prsleaves out #36259, even though the Status line says 1.1 is approved.
ec5a40a to
402beb0
Compare
|
Re: KiroCrewCaptain review: fixed in 402beb0. The Status line and the README row now cite maintainer chenmingwei23's acceptance on #35875 (#35875 (comment)) instead of the earlier sentence. Also done: status is partial, implementation-prs is [35960, 36141, 36259, 36035], and section 5 records the 16 MB threshold, the 500-row window and 1.1 and 1.2 as accepted behaviour changes. The PR body cites the same link. |
|
Re: bolichen97 review: fixed in 402beb0. The link to the acceptance is #35875 (comment). It is named on the RFC Status line, in the README row and in the PR description, in place of the sentence about the owner's lead. |
402beb0 to
8735a6c
Compare
8735a6c to
c77f013
Compare
c77f013 to
f38aec8
Compare
f38aec8 to
6c62661
Compare
6c62661 to
6baaa44
Compare
bolichen97
left a comment
There was a problem hiding this comment.
Approving. I read the whole diff.
The acceptance record I asked for is now there. The PR body, the README row, the RFC Status line and section 5 all link the maintainer decision comment on issue 35875. I read that comment: it is written by chenmingwei23 (a repository member), says "Maintainer decision", and accepts this RFC for part B with the four decided points.
- The front matter moves from
drafttopartial, which fits the status the README row states: the CI fixture 35960 is merged and the rest is not built.implementation-prslists 35960, 36141, 36259 and 36035, the same four PRs the decision comment names. - Section 5 now records the decided values and keeps the three existing decisions in order. The design text is unchanged, and no code changes.
- No regex, no new refusal, and all 97 checks are green.
I did not check the new audited-at sha against main.
KiroCrewCaptain
left a comment
There was a problem hiding this comment.
Approving. On head 6baaa44 I read the full diff (two RFC files, +7/-7, no design text changed), checked that the four implementation PRs in implementation-prs match the ones the decision comment on #35875 names, checked that partial is a defined README status, and confirmed PR Readiness is green.
Follow-up on my earlier review:
- Decision record: fixed. The Status line, the README row, section 5 and the PR body now link the public decision comment on #35875 in place of the unnamed sentence, and a listed maintainer approved this head after reading it.
- Optional, section 5 values: fixed. Section 5 now records the decided threshold, window and behaviour changes.
- Optional,
partialvsaccepted: fixed. - Optional,
implementation-prsmissing #36259: fixed.
Done well: the decision lands as its own docs-only PR ahead of #36141 and #36259, so their First Principles check reads it from base, as the RFC README asks.
The design is signed off under the standing maintainer sign-off for changes by team members.
Optional:
docs/request-for-change/README.md:33and the Status line say "the rest is not built yet" and that each cold slot fold walks its unit separately. But #36035 (section 1.3, one walk per slot) is already on main, before this PR's merge-base, and the newaudited-at0feaa8b predates it. A follow-up re-audit could say 1.4 and 1.3 are on main and bumplast-audited.- Section 5 item 1 credits the 500-row window to the decision comment, which names only the 16 MB threshold and accepts the document as a whole. Saying the window follows from section 1.1 would match the source.
- Section 5 item 2 accepts "1.1 and 1.2" but does not name section 3 item 3 (the new
open_slots.jsonfocused and visible keys that #36141 needs).
Problem / Motivation
Goal: Record that the bounded-time restore RFC is accepted, so work that builds on it can merge.
The RFC merged in #35916 and maintainer chenmingwei23 (Mingwei) accepted it on #35875 on 2026-11-13: #35875 (comment). Its front matter still says draft, so the First Principles review on #36141 blocks: an RFC covering the focus-first restore order must be non-draft on base.
Why it matters
Without the status change, the implementation PR #36141 stays blocked on a decision that is already made.
Not a goal
What changed (motivation → approach → change)
The RFC front matter moves from draft to partial (the CI fixture #35960 is merged), with the audit sha and implementation-prs [35960, 36141, 36259, 36035]. Its Status line records the acceptance. Section 5 records the decided values. The row in docs/request-for-change/README.md says partial and that the CI fixture (#35960) is merged while the rest is not built yet.
Backwards compatibility
Compatible: documentation status only.
Removes nothing: no behaviour, key or path changes.
Tests
Proof: not provable -- docs or config only. scripts/docs-lint.sh passes.
Manual verification
N/A — docs only; docs lint run locally.
Screenshots / video
N/A — no user-visible UI change.
Related Issues
Refs #35875
Checklist
Contribution License Agreement
Contributions are accepted under the project's Apache-2.0 license.