Conversation
The rail per 20.06 §PR-1, built by an opus author and verified by a distinct
opus reviewer (author != verifier), with the author's reasoning handed to the
reviewer to attack.
WHAT SHIPS
- bin/qtest: root layouts/ diffs no longer exit 0 silently - layouts/next/ and
content/next/ map to the next-pilot key; any other root layouts/ path
escalates to --all. This closes a TODAY hole: SEO partial edits shipped
ungated before this.
- content/next/: cascade {noindex, private}. Verified by breaking it - strip
the cascade and BOTH guards fail (robots flips to index,follow; sitemap
gains the pages). All the safety is the cascade; the build: block the spec
suggested was proven inert by ablation (byte-identical output) and deleted.
- layouts/next/baseof.html: functional partials verbatim (meta, schemas,
analytics, favicons, service worker), zero fl-* (grep of built page: 0
hits), clean chrome reused. Analytics additionally gated on (not
.Params.noindex) so /next/ testing never pollutes the ADR-0004 engagement
read.
- layouts/next/single.html + pages/next-pilot.css: the smoke template and a
44-line tokens+reset bundle. Unique bundleName verified.
- enhanced-meta-tags.html: canonical resolves source_page when set - proven
byte-inert on legacy by full-tree diff of all 1768 built files (only the
/next/ page differs). The relative-canonical scare was refuted: legacy pages
emit the identical relative form in test builds; production is absolute via
canonifyURLs.
- content/next/services/fractional-cto.md stub, test_next_pilot (desktop,
mobile, noindex + sitemap unit guards), ADR-0006, ownership-map row,
new-page.md fork section.
REVIEW FINDINGS APPLIED (9)
- The pilot renders 47 of the real page's 536 words - service copy lives in
frontmatter the smoke template does not read. Relabelled everywhere: PR-1
ships a RAIL SMOKE PAGE; porting overview/faqs is PR-2's first task, and the
flip instruction now carries the checklist (flipping type:next early would
gut the live page to the smoke subset).
- The canonical guard could not fail - %r{/services/fractional-cto/\z} also
matches the /next/ fallback. Now paired with refute_match /next/.
- content/next/ mapped to nil in qtest - the same false-green class one
directory over. Fixed.
- navigation.css:20 leaks a legacy body rule (18px/1.6/#121212) into the rail -
named in ADR-0006 Consequences so PR-2 owns the override instead of
rediscovering it.
- macOS next/pilot baselines are committed as FIRST baselines with the
machine-local caveat (nothing prior to corrupt; CI-Linux is the fleet
authority). The course/chapter macOS re-record stays blocked - different
case, prior good baselines exist.
Known red until the CI dispatch lands: Linux next/pilot baselines do not exist
and fail_if_new hard-errors in CI - recorded on the branch next.
Gates: bin/hugo-build clean; rake test:unit 283 runs 0 failures (reviewer run);
next_rail_test 3 runs 17 assertions green after the guard fix; leak check of
the full built tree: zero /next/ references outside /next/.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Contributor
|
Important
This repository does not receive automatic reviews because it has fewer than 10 stars. ⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro Plus Run ID: 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. Comment |
Contributor
Screenshot diffs detected
|
The dispatch recorded 71 baselines for a PR that adds two new page keys. Kept exactly the 2 next/pilot FIRST baselines (Bin 0 -> N: no prior baseline existed, so nothing can be corrupted) and dropped the other 69 whole-site rewrites, which would fold unrelated drift into this PR. Bot commit replaced via force-with-lease; its ref survives in run 32483366996. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Paul 2026-08-21: highly opinionated questions get multiple subagents with different personas and an organized vote. Recorded in CLAUDE.md's autonomy grant (widened from big/critical calls to any question where reasonable people disagree) and applied to the imminent design judging: four distinct personas - the burned founder, the CFO skeptic, the mobile commuter, the brand strategist - each voting with its strongest objection and evidence, plus the fable /impeccable critic. Independent agreement valid; manufactured dissent is not; each panellist names the strongest finding AGAINST its own preference. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…e pilot Codex final gate on #555 returned 2 findings, both about the NEXT page rather than this one, both accepted. 1. The v2 baseof dropped the mermaid/font feature block - my own PR-1 brief said keep it, the author omitted it, and the internal reviewer missed it too (the pilot's source has no diagrams, so nothing was visibly broken). Restored verbatim from the theme baseof with one necessary difference: a /next/ stub renders the SOURCE page's content, so the render hook sets the Store flags on the source page - the block now resolves source-or-self ($mp) before reading the flags. Verified the conditional still holds: the pilot ships zero mermaid bytes. 2. bin/qtest mapped ALL layouts/next/ and content/next/ changes to the hardcoded next-pilot key - a future v2 page's own test would never run for shared-template changes, the same silent-narrowing class this PR exists to close. Now derives every ^next- key from the PAGE_TESTS registry, so adding a page's key IS its registration. Gates: bin/hugo-build clean; next_rail_test green; ruby -c on qtest. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
PR-1 of the clean-slate plan (20.06): the v2 rail — root
layouts/next/,/next/noindexed staging on production, and the qtest false-green fix.Built by an opus author, verified by a distinct opus reviewer handed the author's reasoning to attack. 9 findings, all applied before this PR opened.
What ships
bin/qtestroot-layouts/fixlayouts/next/foo.html → ["next-pilot"], SEO partial →:all; on HEAD both werenil→ silent exit 0. Closes a today hole — SEO partial edits shipped ungated./next/noindex + sitemap exclusionindex,followAND sitemap gains the pages. The spec'sbuild:block was proven inert by ablation (byte-identical output) and deleted — all the safety is the cascade.baseof(74 lines)fl-*in built output; conversion events intact (data-lead-form,data-cta-locationpresent); analytics additionally gatednot .Params.noindexso staging never pollutes the ADR-0004 engagement readenhanced-meta-tagscanonical/next/page. The relative-canonical scare was refuted: legacy emits the identical form in test builds; production is absolute viacanonifyURLs./next/references outside/next/across the built tree — no RSS, menus, taxonomies, sitemap, sw precacheThe honest label
The pilot is a rail smoke page, not a service-page mirror — measured: it renders 47 of the real page's 536 words, because service copy lives in frontmatter (
overview,faqs) the smoke template doesn't read. ADR-0006 andnew-page.mdnow say so, and the flip checklist forbidstype: nexton a real page until the v2 template reads every frontmatter section it uses. Porting those sections is PR-2's first task, before styling.Review disposition
9 findings from the verifier, including: a canonical guard that could not fail (the regex also matched the fallback — now paired with a
refute_match),content/next/mapping to nil in qtest (the same false-green class one directory over), and the legacybodyrulenavigation.css:20leaks into the rail (named in ADR-0006 so PR-2 owns the override). macOS pilot baselines committed as first baselines with the machine-local caveat; the reviewer's verdict distinguishes this from the still-blocked course/chapter re-record.Gates
bin/hugo-buildclean ·rake test:unit283 runs, 0 failures ·next_rail_test3/17 green after the guard fix · Linuxnext/pilotbaselines: CI record dispatched on this branch, will be screened to the pilot keys only before merge.🤖 Generated with Claude Code