docs(rfc): RFC 0033 green→red — run #20: .6 corpus arm undischarged, §9.14 record

[jensholdgaard]

What

Comparative dispatch run #20 (the first on main after the §7 freeze and the RFC 0033 merges) reported, and it carries an honest correction:

The good: all frozen gates pass on main (the dispatch works as the regression gate), equivalence held, and the registry component is byte-identical to run #8's 513,862 B — the cache regressed nothing.

The finding: every pair ran cold — no artifact published. The write-through abstained on the very corpus the cache was built for, consistent with §3.2's own size guard: the artifact is uncompressed JSON of every (template_id, version) template string, while the fold it must beat is zstd-compressed Parquet of the same strings. Publishing would have made warm acquisition cost more bytes than the fold — the guard was right to refuse.

The correction: RFC0033.6's corpus arm ("a warm published artifact … warm/cold ≤ 1/10 on that corpus") is therefore undischarged, and I flipped the RFC green prematurely (#514, before the run reported). This PR reverts the status to red, adds a dated run-#20 note under §5.6 (scenario text unchanged — the criterion is right, the artifact encoding isn't there yet), and records the full finding as docs/benchmarks.md §9.14.

Also recorded: M_L2 stays frozen-deferred (§7's unfreeze condition unmet), and the lever is a format_version: 2 compressed artifact body — cheap by the RFC's own unknown-version-is-absent rule, expected (to be measured, not assumed) to land well under the fold size.

Checks run

mdbook build clean. Docs only.

🤖 Generated with Claude Code

https://claude.ai/code/session_01WQY9wfrfRggqSpMLH8Xj3Y

[coderabbitai]

Warning

Review limit reached

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

Next review available in: 45 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: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: a9e88a33-1710-405f-8a16-52124ada8fc8

📥 Commits

Reviewing files that changed from the base of the PR and between e1ef0e2 and aa63988.

📒 Files selected for processing (2)

• docs/benchmarks.md
• docs/rfcs/0033-cached-template-map.md

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch rfc0033-run20-finding

---

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.}

[Copilot]

Pull request overview

Reverts RFC 0033’s status from green back to red after comparative dispatch run #20 showed the cache artifact was never published on the headline corpus (size-guard abstention), leaving the RFC0033.6 corpus arm undischarged. Adds permanent run #20 documentation so the current state is recorded and traceable from the RFC scenario to the benchmarks log.

Changes:

• Flip RFC 0033 frontmatter status green → red and add a run #20 note under RFC0033.6 explaining why the corpus arm is undischarged.
• Record the full comparative run #20 finding in docs/benchmarks.md as a new §9.14 entry.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
docs/rfcs/0033-cached-template-map.md	Reverts RFC status to red and documents run #20’s “no artifact published” outcome under the RFC0033.6 scenario.
docs/benchmarks.md	Adds §9.14 capturing run #20 results and the implications for RFC 0033 gating and the freeze/unfreeze conditions.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.