code-mode: extend test coverage to lock in cell lifecycle

[cconger]

This PR establishes the intended behavior as an executable contract before a refactor of the cell runtime begins. It also fixes cases where a second observer or termination request could replace an existing response channel and leave the original caller unresolved.

Behavior codified

• A cell can yield output and subsequently resume to completion.
• A caller can run a cell until it has no immediately runnable work, receive its accumulated output and outstanding tool-call IDs, and then resume the same cell when the awaited work is available.
• Each cell admits one active observer:
  • a second observer receives an explicit busy error
  • the existing observer remains registered and is not displaced
• A natural result (conclusion of the js module) that has already reached the cell controller wins over a later termination request.
• Otherwise, termination preempts execution and resolves both:
  • the active observer, if present
  • the caller requesting termination
• Repeated termination requests are rejected while termination is already in progress.
• Terminal responses are sent only after outstanding callback work has been handled:
  • natural completion drains notifications and cancels outstanding tool calls
  • termination cancels and drains both notification and tool callbacks.
• Cell removal and cell_closed notification happen after callback cleanup

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 03eec933ea

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[anp-oai]

nit: if we're not actually protecting any data with synchronization points between these atomic ops, might as well use relaxed everywhere

[anp-oai]

nit: this seems like not many iterations, could this flake if timing gets weird in a cloud vm that's running tests?

[cconger]

Bumped this to a much larger cycle, proved failure without and readded, agree that this is kinda "magic", there is a deterministic test to verify the behavior in the code-mode tests, but this tries to capture it at the suite level too.