Skip to content

test: characterize replay application behavior and coordinator ownership - #2137

Open
thymikee wants to merge 1 commit into
mainfrom
codex/issue-2130
Open

test: characterize replay application behavior and coordinator ownership#2137
thymikee wants to merge 1 commit into
mainfrom
codex/issue-2130

Conversation

@thymikee

Copy link
Copy Markdown
Member

Summary

Adds one representative application-boundary contract through handleSessionReplayCommands for native and Maestro replay/test flows. It covers admission-before-mutation, routing, session lifetime, failure projection and artifacts, format-neutral suite hosting, progress, and cancellation. The existing AST/import coordinator-ownership proof remains unchanged; no production files or exports changed.

Closes #2130

Validation

  • Focused replay and ownership run: 8 files, 84 tests passed.
  • pnpm check:affected --run: all runnable checks passed (format, lint, typecheck, fallow, related Vitest).
  • Temporary planted-red behavior probes, all restored: session lifetime, exact failure message, failed-step artifact ledger, target admission before dispatch, Maestro routing, progress emission, and active cancellation each failed its named application test when violated.
  • Temporary planted-red structural probes, all restored: a second createReplayCoordinator import failed both ownership tests; a SessionStore import in the resume chain failed the named no-binding test.
  • Independent adversarial review found no valid findings; no changes were needed.

Scope / size

  • Touched files: 1.
  • Production files and exports changed: 0.
  • Gross diff: +296 / -0 lines.
  • Net test LOC: +296. Existing routing/runtime tests either bypass this handler seam or cover distinct contracts, so no identical coverage was removed.

Residual risks

  • The focused contract uses real session/source fixtures and mocks only the device snapshot boundary; native/device/provider/live-network behavior and exact-head GitHub lanes remain CI responsibilities.
  • Coordinator identity remains intentionally covered by the separate AST/import proof because behavior tests cannot distinguish two coordinators backed by the same store and session.

@github-actions

Copy link
Copy Markdown

Size Report

Metric Base Current Diff
JS raw 2.51 MB 2.51 MB 0 B
JS gzip 841.4 kB 841.4 kB 0 B
npm tarball 968.2 kB 968.2 kB 0 B
npm unpacked 3.35 MB 3.35 MB 0 B

npm unpacked components

Component Base Current Diff
JS / dist source 2.67 MB 2.67 MB 0 B
Apple runner source/project 581.1 kB 581.1 kB 0 B
macOS helper source 54.8 kB 54.8 kB 0 B
Android helper artifacts 0 B 0 B 0 B
Other package files 45.6 kB 45.6 kB 0 B

Startup median (7 runs, lower is better):

Scenario Base Current Diff
CLI --version 27.2 ms 26.6 ms -0.6 ms
CLI --help 73.2 ms 73.9 ms +0.6 ms

Top changed chunks: no changes in the largest emitted chunks.

Top changed packed files

No changed packed files.

@thymikee

Copy link
Copy Markdown
Member Author

Review found no code or characterization-design issues. The new suite asserts application behavior at handleSessionReplayCommands without pinning engine decomposition, while the existing AST/import gate remains the correct separate proof of single coordinator ownership. The 296 test LOC cover the issue's distinct admission, routing, failure/artifact, lifetime, format-neutral suite, progress, and cancellation rows; no production surface changed. Remaining status is CI only: Android Smoke and Coverage were still running at review time.

@thymikee thymikee added the ready-for-human Valid work that needs human implementation, judgment, or maintainer merge label Aug 28, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ready-for-human Valid work that needs human implementation, judgment, or maintainer merge

Projects

None yet

Development

Successfully merging this pull request may close these issues.

test: characterize replay application behavior and coordinator ownership

1 participant