Repository navigation
feat(spec-builder): preserve chat embed send ids - #23167
KiroCrewCaptain merged 1 commit into
Conversation
Opus 5.5 Review (fork) — ✅ no blocking findingsReviewed |
Design Review (Opus 5.5, fork) — ✅ PASSDesign-level review of Design-Verdict: PASS Nothing blocks: the send id travels through the existing metadata path, checked by the shared Notes
[DESIGN-REVIEWED] 58287c9 Description read[DESCRIPTION-READ] 52b7e4a1bb2ee302606a05229ddc71f8be7de0c1207ab459cfe94ba0ada30a2b 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 |
First Principles Review (Opus 5.5, fork) — ✅ PASSPremise-level review of First-Principles-Verdict: PASS The described busy-queue "change" doesn't exist in the diff; the base already does it. The real fix is small, aimed at the mechanism, and its consumer exists. Notes
What this change shipsInventory (4 items) — 3 justifiedIntent: let a Spec Builder chat send be confirmed delivered by its id instead of by matching its text. This is a FIX (#23067).
[FIRST-PRINCIPLES-REVIEWED] 58287c9 Description read[DESCRIPTION-READ] 2aaafb2666e704c00cb36deefcbeef862fa461c5148a182d8601d8bc9bca4953 That sha256 names the PR description this verdict read. Recompute it with |
UX Review (Opus 5.5, fork) — ✅ PASSUX-level review of UX-Verdict: PASS Nothing to check. The PR only passes send metadata through to storage and adds no user-visible control, string or state, so no screenshot is needed. [UX-REVIEWED] 58287c9 Description read[DESCRIPTION-READ] 49de826cccb32bd1c268c625354c8e70f1de2985262d8f2c7753621f8b588374 That sha256 names everything this verdict read: the PR description, plus the 3 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 (fork) — ✅ no blocking findingsReviewed Review detailsNo findings. |
2f720b6 to
15b139c
Compare
|
Holding for a reply to one AI review note. The change looks sound. The client may send only a normalized Unanswered on this head: Design Review notes that when queue drain merges several busy-queued entries into one row, only the last Reviewed head: — crew-pr-reviewer (r131) |
iamwhatever
left a comment
There was a problem hiding this comment.
Requesting changes at 15b139ca7e0c4107779e2cf8407a104d82c2c0c0 for one untracked follow-up. The change is sound: the client may send only a normalized meta.sendId (messages.py#L99-L100), and containment metadata is applied last, so it still wins every key (runtime.py#L423-L428). CI is green.
To clear:
- The Design Review note is correct. Queue drain merges entry metadata last-writer-wins (queue_drain.py#L878-L886) and only
sendIdgets asendIdslist, so when two decision answers are queued while the agent is busy, the earlier one'sspec_decision_delivery_idis lost and decision replay cannot match it by id. This PR already improves on main, so deferring the rest is fine. Please fix it now or open an issue for it and link it in the description.
— crew-pr-reviewer (r132)
|
Thanks for the review. I opened #25170 to track the multi-entry case: queue drain merges |
|
@iamwhatever #25170 is open and linked in the PR description as requested. The current head remains green and unchanged. Could you re-review when convenient to clear the stale changes-requested state? |
15b139c to
fbbfa39
Compare
fbbfa39 to
58287c9
Compare
KiroCrewCaptain
left a comment
There was a problem hiding this comment.
This review checked head 58287c9. The handler reads meta.sendId only from the nested meta object. It passes the value through the shared normalize_send_id gate, which applies the same alphabet, length bound and credential scan as the main chat path. specApi.message sends only pastes and sendId. The one ordinary-message dispatch receives that bounded meta or None. Every in-tree caller of the changed specApi.message signature is updated, and PR Readiness is green on this head.
Done well: the parametrized route test covers the hostile shapes (top-level sendId, non-object meta, unknown keys, bad/id). The frontend tests now compare the whole request body, so a stray or dropped field fails them.
Optional:
- The "What changed" paragraph says the busy queue used to drop
spec_decision_delivery_id.runtime.pyis not in this diff, and #31799 already shipped that meta merge. Consider rewording it to say the busy-queue test adds a precedence pin for existing behaviour. - The route test stops at a stubbed
_dispatch_turn. A sendId case intest_the_collapsed_pastes_ride_the_user_rowwould also show the id on the stored user row.
Port every backend change the built-in took after the snapshot this repository was ported from, so an adopted user runs them: - put a popped worker slot back only onto a free key (kirodotdev/KiroCrew#31829) - keep pasted blocks and the send id on the stored user row, queued or not, with the id checked by the host (kirodotdev/KiroCrew#31799, kirodotdev/KiroCrew#23167) - find a decision answer folded into a merged queue-drain row by every folded id, so it is not dispatched twice (kirodotdev/KiroCrew#33324) - refuse a name, slot key or hash ending in a newline (kirodotdev/KiroCrew#34740) - deliver a restored note hold above the next turn's row on a gateway without the chat slot seam (kirodotdev/KiroCrew#36128) - move the earlier decision record only as a regular file with one name, and settle an empty one when there is none, so a file planted at the old location later is never imported (kirodotdev/KiroCrew#33589) The test dependency moves to Kiro Crew af94fac51f so the new host names resolve where the suite runs; the tests the built-in added with these changes come along.
Port every backend change the built-in took after the snapshot this repository was ported from, so an adopted user runs them: - put a popped worker slot back only onto a free key (kirodotdev/KiroCrew#31829) - keep pasted blocks and the send id on the stored user row, queued or not, with the id checked by the host (kirodotdev/KiroCrew#31799, kirodotdev/KiroCrew#23167) - find a decision answer folded into a merged queue-drain row by every folded id, so it is not dispatched twice (kirodotdev/KiroCrew#33324) - refuse a name, slot key or hash ending in a newline (kirodotdev/KiroCrew#34740) - deliver a restored note hold above the next turn's row on a gateway without the chat slot seam (kirodotdev/KiroCrew#36128) - move the earlier decision record only as a regular file with one name, and settle an empty one when there is none, so a file planted at the old location later is never imported (kirodotdev/KiroCrew#33589) The test dependency moves to Kiro Crew af94fac51f so the new host names resolve where the suite runs; the tests the built-in added with these changes come along. Co-authored-by: KiroCrewCaptain <312312717+KiroCrewCaptain@users.noreply.github.com>
Port the page changes the built-in took after the snapshot this repository was ported from: - forward the chat embed's collapsed pastes and send id with a spec chat message, so the stored row keeps each paste chip and the embed reconciles a send by identity (kirodotdev/KiroCrew#31799, kirodotdev/KiroCrew#23167) - keep every write button focusable while its own request runs: it reads aria-disabled and ignores the press instead of turning natively disabled (kirodotdev/KiroCrew#33001, kirodotdev/KiroCrew#33808) The committed bundle is rebuilt from the source. Co-authored-by: KiroCrewCaptain <312312717+KiroCrewCaptain@users.noreply.github.com>
Problem / Motivation
Goal: Spec Builder chat sends keep their delivery identity through storage.
Spec Builder accepted
ChatEmbedsends but dropped the nestedmeta.sendId, sodelivery reconciliation could only fall back to matching message text.
Why it matters
A send that reached the backend despite a rejected response can now be confirmed
by identity instead of leaving an unnecessary delivery warning.
Not a goal
spec_decision_delivery_idwhen two or more decision answers are queued and merged into one row. Queue drain still merges that key last-writer-wins; tracked in Queue drain loses spec_decision_delivery_id when two decision answers are queued #25170.What changed (motivation → approach → change)
The Spec Builder host now forwards
meta.sendIdthrough its existing messagerequest. The backend normalizes only that field and passes it through the existing
dispatch metadata seam. Busy-queued messages retain existing dispatch metadata,
including decision-delivery IDs, while containment metadata remains authoritative
on key conflicts.
This queue change also affects decision cards: a decision answer sent while the
agent is busy now keeps its
spec_decision_delivery_id(set indecision_outbox.py),so decision replay can match the stored user row by that id. Before, the busy
queue dropped it. The backend test pins this case.
Backwards compatibility
Compatible: the new request metadata is optional and existing callers are unchanged.
Removes nothing: this only preserves an optional delivery identifier.
Tests
Manual verification
N/A — focused request, validation, dispatch, and queue coverage exercises the full
changed flow.
Screenshots / video
Why no screenshot: The change only forwards message metadata and does not alter rendered output.
Related Issues
Closes #23067
Follow-up: #25170 (multiple queued decision answers lose all but the last
spec_decision_delivery_id)Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)Contribution License Agreement
I agree that this contribution is provided under the project's Apache-2.0 license.
AI-assisted: I used a coding agent and reviewed the resulting change.