fleet/fleetd
3
0
Fork 0
Code Issues 73 Pull Requests Actions Packages Projects Releases 1 Wiki Activity

CB-577: correlate async questions to their turns #49

Closed
agent wants to merge 1 commits from worker/cb577-f36fdc-18 into main
pull from: worker/cb577-f36fdc-18
merge into: fleet:main
Conversation 1 Commits 1 Files Changed 2
+27 -2
agent commented 2026-08-15 06:25:51 +02:00
Member
Copy Link
Copy Source

Bind each async question to the Task registered for its exact forward rendezvous waiter. On an unanswered question, keep the ticket PENDING but remove stale target tracking so a later send can proceed. Tests run: 683, Failures: 0, Errors: 0, Skipped: 0; BUILD SUCCESS.

Bind each async question to the Task registered for its exact forward rendezvous waiter. On an unanswered question, keep the ticket PENDING but remove stale target tracking so a later send can proceed. Tests run: 683, Failures: 0, Errors: 0, Skipped: 0; BUILD SUCCESS.
agent added 4 commits 2026-08-15 06:25:52 +02:00
CB-577: correlate async questions to turns e186c7945a
CB-577: handle questions without async waiters 5275922d1d
CB-577: test async question waiter ownership 927e0151d4
CB-577: model async question ownership race
CI / build (pull_request) Successful in 1m29s
Details
CI / contract (pull_request) Failing after 0s
Details
74b0087ebb
ltms added 1 commit 2026-08-15 06:28:56 +02:00
CB-577: remove obsolete async target tracking
CI / contract (pull_request) Successful in 43s
Details
CI / build (pull_request) Successful in 56s
Details
c884802b13
ltms added 2 commits 2026-08-15 06:32:05 +02:00
M4: correct unit 2 criterion 1 — TurnToken cannot carry the session turn 0edc6615fc 
The criterion required the token to bind the session turn number. Three
independent refusals from the implementer showed why that is not
implementable at this layer: MessageService owns acceptance but never
learns of delivery, and CompletionResolver.onDelivered runs before
SessionManager.onDelivered, so the turn number does not exist yet at the
only point the token could capture it.

Records both rejected alternatives and why, so the next reader does not
re-derive them: a target-keyed registry restores the ambiguity the token
exists to remove, and injecting a turn counter couples layers to fill a
field nothing reads yet.
M4: fail tickets on terminal health states
CI / contract (pull_request) Successful in 44s
Details
CI / build (pull_request) Successful in 57s
Details
3b2f395d3d
ltms commented 2026-08-15 07:12:52 +02:00
Owner
Copy Link
Copy Source

Closing — this is the one PR whose code is genuinely not in main, and that is deliberate.

The CB-577 work itself landed: main has 74b0087, 927e015, merge 5f5573a, plus the follow-up c884802 / 2f48e08. What is left on worker/cb577-f36fdc-18 is exactly one commit, 3b2f395 ("M4: fail tickets on terminal health states"), which was rejected on review. Three defects, none fixed:

  1. The silent-default trap, a sixth time. It adds a 6-argument FleetHealthMonitor overload defaulting failTarget to (_, _) -> { }, so every existing caller keeps the no-op and the feature ships turned off. The commit adds zero tests — the tell that the overload exists only so old tests still compile.
  2. It fires every tick, not on transition. failTarget.accept sits outside reportTransition, so a member that stays GONE has abandon called every interval. The brief asked for a bounded retry.
  3. A duplicate AtomicReference import, and an AtomicReference<MessageService> null window — a turn failing before messagesRef.set(messages) is silently skipped.

Worth recording: CI passed on this commit (run #1162, 3b2f395, conclusion success). A green build cannot see any of the three defects — defect 1 is precisely "no tests were added", which a build does not measure.

Re-briefed as a fresh unit in #52 (CB-580). Do not start from this commit; it will carry the overload back in. Verified 2026-08-15.

Closing — **this is the one PR whose code is genuinely not in `main`, and that is deliberate.** The CB-577 work itself landed: `main` has `74b0087`, `927e015`, merge `5f5573a`, plus the follow-up `c884802` / `2f48e08`. What is left on `worker/cb577-f36fdc-18` is exactly one commit, `3b2f395` ("M4: fail tickets on terminal health states"), which was **rejected on review**. Three defects, none fixed: 1. **The silent-default trap, a sixth time.** It adds a 6-argument `FleetHealthMonitor` overload defaulting `failTarget` to `(_, _) -> { }`, so every existing caller keeps the no-op and the feature ships turned off. The commit adds **zero tests** — the tell that the overload exists only so old tests still compile. 2. **It fires every tick, not on transition.** `failTarget.accept` sits *outside* `reportTransition`, so a member that stays `GONE` has `abandon` called every interval. The brief asked for a bounded retry. 3. A duplicate `AtomicReference` import, and an `AtomicReference<MessageService>` null window — a turn failing before `messagesRef.set(messages)` is silently skipped. Worth recording: **CI passed on this commit** (run #1162, `3b2f395`, conclusion `success`). A green build cannot see any of the three defects — defect 1 is precisely "no tests were added", which a build does not measure. Re-briefed as a fresh unit in **#52 (CB-580)**. Do not start from this commit; it will carry the overload back in. Verified 2026-08-15.
ltms closed this pull request 2026-08-15 07:12:54 +02:00
Some checks are pending
CI / contract (pull_request) Successful in 44s
Details
CI / build (pull_request) Successful in 57s
Details

Pull request closed

Please reopen this pull request to perform a merge.
Sign in to join this conversation.
Reviewers
No Reviewers
Labels
Clear labels
blocked
Cannot start until something else lands. The body says what.
 needs-live-proof
Merged and green, but never shown working on the running daemon. Not the same as done.
 ready-to-delegate
Scope, files and acceptance criteria are written. A worker can be briefed from the body alone.
 silent-default
A feature that compiles, passes tests, and ships turned off. Nine recurrences and counting.
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
agent kevin ltms
No Assignees
2 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: fleet/fleetd#49
Delete Branch "worker/cb577-f36fdc-18"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?