fix(#407): commensurable term-fidelity byte meta + retain per-attempt artifacts

[khaliqgant]

Closes #407. Two harness debts from the term-fidelity program.

(1) meta bytes/offset incommensurable — the exact-2.0 artifact

quiet.activity.bytes and brokerOffset read exactly 2.0 apart in 5/5 codex bundles. That was proven NOT to be double delivery (headless probe: client bytes / snapshot offset = 1.0000; BrokerManager does one IPC send per worker_stream event). The 2.0 was a derivation artifact — the two meta fields measure different things on different baselines:

field	measures	baseline	unit
quiet.activity.bytes	client-received broker:pty-chunk string bytes	activity-probe install (pre-spawn ⇒ agent first byte)	UTF-8 bytes of decoded IPC string
brokerOffset	broker snapshot cumulative offset	worker start (offset 0)	raw PTY bytes

Fix: new dependency-free tests/term-fidelity/byte-accounting.ts derives a self-documenting byteAccounting block that replaces the bare pair. Each figure declares its baseline + unit, and clientToBrokerByteRatio is computed on a shared agent-start baseline (the probe is installed before spawn) so ~1.0 is the meaningful "one-to-one delivery" value. It carries an embedded note:

A ratio near an integer such as 2.0 is NOT proof of double PTY delivery … Never infer a delivery mechanism from this number; confirm by counting BrokerManager IPC sends per worker_stream event.

So the exact-2.0 can no longer be misread. Validated with the client-bytes-vs-snapshot-offset probe pattern (ratio derivation) and unit-tested.

(2) retry-then-pass erased first-attempt artifacts

Playwright retry-then-pass cleaned error-context/trace, and the harness wrote divergence bundles to a fixed test-results/term-fidelity/<cli>/<workload>/ path, so a retry overwrote the diverging first attempt (which twice was a REAL divergence event) — unexaminable afterward.

Fix:

• Bundles + telemetry now write under attempt-<retry>/ (retry index threaded from test.info().retry → launchFidelityHarness → harness.attempt → oracle.ts), so a retry never overwrites a prior attempt.
• playwright.term-fidelity.config.ts: preserveOutput: 'always' + documented trace: 'retain-on-failure' so Playwright's own error-context/trace for a failed attempt survive a later passing attempt.
• README updated to document attempt segregation + byteAccounting.

Verification

• npx vitest run tests/term-fidelity/byte-accounting.test.ts → 5 passed (1.0 ratio, 4dp rounding, the 2.0 forensic case + note, null-on-no-offset, divide-by-zero guard).
• npx tsc --noEmit -p tests/term-fidelity/tsconfig.json → clean.
• New pure unit is wired into the vitest node project via a tests/term-fidelity/**/*.test.ts include; Playwright testMatch is unchanged so it is not run as a browser test.

Note: the full Electron matrix (npm run test:fidelity / real agents, token burn) was not run here — these are harness-instrumentation + config changes with no product-runtime surface; the pure derivation and typechecks cover them. Flag if you want me to burn a real matrix run for end-to-end confirmation.

🤖 Generated with Claude Code

[gemini-code-assist]

Warning

You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again!

[coderabbitai]

Warning

Review limit reached

@khaliqgant, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 50 minutes

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 5b0a4e3f-4a1b-4305-bc2f-2ed8602743b2

📥 Commits

Reviewing files that changed from the base of the PR and between 188c3d5 and 38f92d5.

📒 Files selected for processing (9)

• .gitignore
• playwright.term-fidelity.config.ts
• tests/term-fidelity/README.md
• tests/term-fidelity/byte-accounting.test.ts
• tests/term-fidelity/byte-accounting.ts
• tests/term-fidelity/harness.ts
• tests/term-fidelity/oracle.ts
• tests/term-fidelity/term-fidelity.spec.ts
• vitest.config.mjs

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/term-fidelity-harness-debts

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[gemini-code-assist]

Caution

The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased.

[cursor]

Bugbot is not enabled for your account, so this pull request was not reviewed.

Enable Bugbot in the Cursor dashboard to get automatic reviews on future PRs.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 38f92d56ee

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".

[chatgpt-codex-connector]

Normalize the counters before calling them commensurable

In the exact-2.0 scenario motivating this change, this still divides the unchanged quiet.activity.bytes by the unchanged full broker.offset; the parent harness already installed the activity probe before spawnRealAgent, so wrapping those values introduces no new shared-baseline adjustment. Moreover, re-encoded decoded-string bytes are not necessarily equal to raw PTY bytes. Consequently an existing 2.0 result remains 2.0 while the metadata marks it commensurable and describes it as an artifact, potentially obscuring an actual duplicate or missing-delivery signal. Record comparable offset/delta metadata from the received events or explicitly mark this ratio non-commensurable.

Useful? React with 👍 / 👎.