fix(mothership): keep any queued send the server may already hold from being edited

[waleedlatif1]

Summary

Follow-up to #8717.

• No queued send the server may already hold can be edited. fix(mothership): keep a withdrawn first message held at the queue head as written #8717 blocked edits only for a first message the unmount queued itself. Other entries carrying an earlier attempt's id stayed editable, and an edit dropped that id, so the edited text could go out as a second message:
  • a send another surface withdrew, handed over through the send event or the stored handoff, that lands in a busy chat's queue;
  • a queued follow-up re-queued after its dispatch got no answer.
• One place enforces it. The queue store marks every message written with a resumeUserMessageId as admissionUnknown, unless the writer says the server is known not to have it. Queues restored from the session go through the same guard, so entries saved before this change are covered too. The existing edit guards (store, hook, Edit button) then apply.
• Only the server refusing that id clears it. A busy 409 or a superseded 409 means the server doesn't have the id. It checks the send's id before it takes the chat lock, so it answers a retry of an admitted id as a duplicate, never as busy. A message held this way becomes editable again once the server refuses it.
• A Send-now whose Stop never settles keeps the earlier uncertainty. Nothing was sent, which proves nothing about the earlier attempt the message resumes, so the message stays uneditable.
• Moves a test doc comment back above its own test.

Type of Change

• Bug fix

Testing

• New DOM tests, each failing before its fix:
  • does not let a withdrawn send handed to a busy chat be edited
  • does not let a queued follow-up be edited after its send got no answer
  • lets a held message be edited again once the server refuses it as busy
  • keeps a resumed message uneditable when its Send-now Stop does not settle
• Store tests: treats any message resuming an earlier attempt as possibly sent, unless told otherwise, and a rehydration test (store.dom.test.ts) for queues saved before the guard.
• The existing held-Stop edit test now carries admissionUnknown: false: that entry was never sent and resumed no earlier attempt, so the server can't have it.
• The full gate and integration tests ran on the CI runner.

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated
docs	Skipped		Oct 7, 2026 5:05am UTC

[waleedlatif1]

@greptile

[waleedlatif1]

@cubic-dev-ai review this PR

[cubic-dev-ai]

@cubic-dev-ai review this PR

@waleedlatif1 I have started the AI code review. It will take a few minutes to complete.

[cubic-dev-ai]

All reported issues were addressed across 5 files

<sub>Reply with feedback, questions, or to request a fix.

Fix all with cubic | Turn on auto-fix | Re-trigger cubic</sub>

[greptile-apps]

Confidence Score: 5/5

[Medium risk] Prevents editing of queued messages that may already be sent.

The PR appears safe to merge; both previously reported edit-guard gaps are fixed.

What we checked:

• Saved sends keep their guard: Saved entries pass through withAdmissionGuard. An earlier attempt ID gains admissionUnknown: true unless a writer already supplied the flag.
• Failed Stop keeps earlier uncertainty: Stop failure returns neverSent. Queue restoration keeps the entry's previous admissionUnknown value rather than treating the earlier send as refused.

Summary

This PR prevents edits to queued messages the server may already hold.

• Queue writes and saved queues apply the same edit guard.
• A failed Stop preserves the earlier send's uncertainty instead of treating it as refused.
• Both previous, unnumbered findings are fixed. New regression tests cover these paths.
• No new actionable issues were found. Tests were inspected, not run.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart TD
  A[Queue write or restore] --> B{Earlier attempt ID?}
  B -->|No| C[Editable]
  B -->|Yes| D{Known refused?}
  D -->|Yes| C
  D -->|No| E[Cannot edit]
  E --> F{Retry outcome}
  F -->|Server refuses ID| C
  F -->|Stop fails before send| E

Loading

_{Reviews (2) · Last reviewed commit: "fix(mothership): keep earlier-attempt un..." · Reviewed by Greptile}

[waleedlatif1]

@greptile

[waleedlatif1]

@cubic-dev-ai review this PR

[cubic-dev-ai]

@cubic-dev-ai review this PR

@waleedlatif1 I have started the AI code review. It will take a few minutes to complete.

[cubic-dev-ai]

No issues found across 6 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Turn on auto-fix | Re-trigger cubic}