MBTI-PDF-SNAPSHOT-PRINT-LAYOUT-POLISH-10: polish MBTI snapshot print layout

[fermatmind]

Summary

• Implements PR-G print-layout polish for the real MBTI result-page snapshot tree.
• Compresses decorative narrative illustration slots into print-only accent strips, reduces hero/section spacing, and relaxes large influence/unlock panels so they can paginate instead of forcing opener pages to leave large blank areas.
• Keeps small cards/list items together and keeps all real detail content in the result-page component tree.
• Registers the PR-train item and adds focused contract coverage for no summary shell, real tree presence, decorative media compression, and break behavior.

Why

PR-F restored the real result-page component tree and current PDFs are HeadlessChrome/Skia, not mPDF or the old summary shell. The remaining issue is print layout: Page 3/Page 7 section openers leave large blank areas because decorative media and large avoid-break cards consume/force page space.

Companion backend PR

• fap-api companion: MBTI-PDF-SNAPSHOT-PRINT-LAYOUT-POLISH-10: add snapshot PDF render version fap-api#2577
• Cache isolation uses X-Pdf-Render-Version / artifact hash input instead of clearing production cache or bumping surface v5.

Validation

• ./node_modules/.bin/vitest run tests/contracts/mbti-pdf-snapshot-print-layout-polish.contract.test.ts tests/contracts/mbti-result-page-pdf-visual-pagination.contract.test.ts tests/contracts/result-gotenberg-print-route.contract.test.ts tests/contracts/result-client-view-state.contract.test.tsx tests/contracts/mbti-result-page-pdf-smoke-quality-gate.contract.test.ts tests/contracts/result-private-print-chrome.contract.test.ts
• ./node_modules/.bin/tsc --noEmit
• ruby -e "require 'yaml'; YAML.load_file('docs/codex/pr-train.yaml')" && python3 -m json.tool docs/codex/pr-train-state.json >/dev/null
• git diff --check

Note: pnpm typecheck was blocked before running by local approve-builds state for esbuild/sharp/unrs-resolver. package.json typecheck is tsc --noEmit, and that exact payload passed directly. I did not take ownership of the unrelated untracked pnpm-workspace.yaml.

Deferred

• No deploy, env, DB/CMS/import, cache clearing, Gotenberg timeout/route/endpoint changes, content injection changes, mPDF fallback, summary PDF, or deletion of real result detail content.
• Optional follow-up: rendered PDF visual smoke after both PRs are merged/deployed asynchronously.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: b36b6b8d26

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]

Record the actual reviewed commit SHA

This ledger entry records d336ef... as the fap-web commit, but the commit being reviewed is f13e0e3a994726b5f3adafab3116898a3cb8ba18 and git cat-file -e d336ef... does not resolve in this repository. Any PR-train closeout/reconciliation that trusts this field will look for the wrong revision, so update the ledger to the actual commit SHA before relying on it for merge or cleanup state.

Useful? React with 👍 / 👎.