[BlueboxC]

Problem and value

The full regression shard on #1147 and #1148 fails because the checked-in Checkout Architecture comparison and README motion-proof receipts do not match current dev inputs. Both failures reproduce on unchanged base 847146aeb93b492592adb78dc220f57e4a8b4636, independently of those evidence-link fixes. Example: failed full-regression shard.

Regenerate the Checkout HTML/receipt with the existing comparison command and recapture the README GIF/receipt from the current Gallery artifacts. This is a separate generated-only repair; it does not change runtime, validators, test acceptance, or source fixtures.

Related: #900. This PR refreshes existing outputs; it does not duplicate that issue's proposed freshness-gate work. Checked all 128 open PR file lists and relevant Viewer/builder patches. Other PRs regenerate these paths for their own feature changes; none supplies this four-file refresh from the current unchanged base.

The maintainer agreed to this scope. The agreed integration order is #1150 first, then #1151; shared generated outputs should be rebuilt from the combined source when integration requires it.

Contributor review

• Problem origin: actual CI failures, reproduced on the target base.
• I reviewed the submitted changes and explanation, understand their impact, and can answer questions about them.

Stability impact

Generated-only maintenance. The authoritative source and existing behavior are unchanged. Existing tests already discriminate the stale outputs, so they are reused without adding duplicate assertions or weakening checks. The code-simplification check is inapplicable to generated-only output.

Tests run

Base: 847146aeb93b492592adb78dc220f57e4a8b4636. Candidate: 4c22ebef101f6e60c0d660e659e215d7759c605f.

• Before: node --test --test-name-pattern='checked-in Checkout compare artifact|README motion proof is compact' test/architecture-delta.test.mjs test/readme-showcase.test.mjs fails for both stale artifacts.
• After: node --test test/architecture-delta.test.mjs test/readme-showcase.test.mjs test/gallery.test.mjs: 55 pass, 0 fail, 1 skipped. The skipped Git-for-Windows discovery scenario requires native Windows; no Windows result is claimed here.
• Canonical generation used official Node 22.23.2 (zlib 1.3.1-e00f703), native Chrome, and temporary ffmpeg 7.1 from the hash-verified imageio-ffmpeg 0.7.0 macOS ARM64 wheel.
• The GIF is 960×540, 54 frames, 10 fps, 5.4 seconds, looping, 843870 bytes. Its three source hashes and output hash are verified by the existing test.
• Final-head CI completed successfully on this candidate. Responsibility selection omits the full regression and effective browser execution for this generated-only change; the 55 affected local tests above establish the artifact repair.
• The independently reproduced Sequence fitting failure observed in fix(evidence): use plain GitHub Markdown line links #1147 is repaired separately in #1151. Required final-head CI remains authoritative.

Visual evidence

The before/after GIFs can be compared directly in the diff. Matched frames 9, 27 and 45 were inspected at the same 960×540 dark embedded scene settings; the three scenes, captions and selected relationships remain visible. Contributor personal review, including the recorded visible outcomes, is confirmed. This is fixed-input output freshness, not an Agent generation-quality benchmark.

Generated artifacts

Only examples/checkout-platform-delta.html, its receipt, and docs/assets/archify-live-proof.gif/.json are regenerated with the existing builders. Gallery inputs are unchanged and their existing reproduction tests pass. Packaged archify/ source is unchanged, so archify.zip is unchanged.

BlueboxC

[coderabbitai]

No actionable comments were generated in the recent review.

Recent review info

Run configuration

• Source mode: Auto
• Repository: tt-a1i/archify
• Branch: dev
• File: .coderabbit.yaml
• Commit: 847146a
• Layers: Repository YAML
• Review profile: CHILL
• Plan: Advanced
• Run ID: 39117fed-9b76-48d7-8ecd-8b95ad88e7d5

Commits

Reviewing files that changed from the base of the PR and between 847146a and 4c22ebe.

Files ignored due to path filters: 1

• docs/assets/archify-live-proof.gif is excluded by !**/*.gif

Files selected for processing: 3

• docs/assets/archify-live-proof.json
• examples/checkout-platform-delta.html
• examples/checkout-platform-delta.receipt.json

Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 9 remain after this review.

📝 Summary

Refreshes stale generated Checkout comparison and README motion-proof artifacts, including their receipts. The change does not alter runtime behavior or validation rules, so compatibility impact is limited to updated generated files. Reviewed base: 847146a; head: 4c22ebe. The author reports that targeted artifact checks failed before regeneration and that three static test files passed 55 tests with 1 skipped afterward. The author also reports successful CI at the head; no CI result was independently observed here. The author cites reused hash and visual-inspection evidence, but its original revision is unspecified. Full regression and browser acceptance are not established.

The proof manifest updates the GIF byte count and digest, plus the digests for the signal-flow, blueprint, and classic scenes. The Checkout receipt updates its artifact SHA-256 and byte count.

Priority: Normal

Estimated code review effort: 1 (Trivial) | ~3 minutes

Possibly related PRs

• tt-a1i/archify#1087: Updates the same proof manifest and motion-proof GIF.
• tt-a1i/archify#1143: Records the same Checkout artifact digest and size, and updates the shared proof manifest.
• tt-a1i/archify#1100: Regenerates the motion-proof GIF and receipt after source-drift failures.

Merge Risk: Minimal · up to 4c22e

The generated proof artifacts are ready to merge with no identified correctness or operational risk.

Pre-merge checks: 2

Check name	Status	Explanation
Contribution Scope		The PR states the current-base problem, the intended artifact refresh, preserved runtime and compatibility behavior, test results, and the applicable generated-artifact scope. The description records …
Validation Evidence		PASS. Evaluated head 4c22ebef101f6e60c0d660e659e215d7759c605f against base 847146aeb93b492592adb78dc220f57e4a8b4636. The authoritative diff changes only four generated outputs. No fixtures, builde…

Full details: Contribution Scope

Explanation

The PR states the current-base problem, the intended artifact refresh, preserved runtime and compatibility behavior, test results, and the applicable generated-artifact scope. The description records related issue context and maintainer scope agreement. The reviewed diff changes only the four named generated outputs, with no source, validator, default, or public contract changes. The refreshed GIF manifest and Checkout receipt match the actual artifact bytes and SHA-256 values. Failure/rollback behavior is not applicable to this generated-only refresh.

Full details: Validation Evidence

Explanation

PASS. Evaluated head 4c22ebef101f6e60c0d660e659e215d7759c605f against base 847146aeb93b492592adb78dc220f57e4a8b4636. The authoritative diff changes only four generated outputs. No fixtures, builders, validators, tests, or runtime sources changed. I observed that the GIF and Checkout receipt byte counts and SHA-256 values match their files, and all three README scene hashes match their gallery artifacts. The existing tests exact-compare the Checkout artifact and receipt, verify GIF dimensions, looping, digest, and scene freshness, and verify green gallery receipts. The author reports 55 passing tests, 0 failures, 1 native-Windows skip, and successful final-head CI. The PR separates author-reported CI and perceptual frame inspection from automated evidence and does not claim full-regression or browser-gate coverage. The available evidence is proportionate to this generated-only freshness repair.

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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.}

[tt-a1i]

Thank you for the carefully scoped fixes and for distinguishing constructed reproductions, CI results, and personal acceptance. We have reviewed the value and maintenance cost of #1146–#1151 and agree with these bounded directions:

• fix(ci): refresh stale Checkout and README proof artifacts #1150 first: refresh the stale Checkout and README proof artifacts using the existing generators, without changing runtime behavior or weakening checks. This addresses a shared integration blocker; we independently observed the Checkout reproducibility failure in community: add archify-index listing #884's CI.
• fix(reader): recover compact Sequence fit after header reflow #1151 next: allow the optional bottom gap to yield for compact automatic Sequence diagrams, while preserving source-text size, actual viewport/dock clearance, and scrolling for genuinely long diagrams. Reuse the recorded browser evidence; this is not a request to force every diagram onto one screen.
• fix(class): preserve authored parameter delimiters #1146: preserve authored parameter delimiters instead of inserting spaces into literal values.
• fix(evidence): use plain GitHub Markdown line links #1147: use plain-view GitHub Markdown line links, preserving other providers and path-only links. Coordinate the shared helper with feat: GitHub Enterprise source links for repository evidence #1063 when integrating.
• fix(evidence): reject non-commit revision objects #1148: enforce the existing commit-object identity contract using the existing Git query, without adding fetching, silent identity normalization, or a broader validation framework.
• fix(timeline): wrap event titles at grapheme boundaries #1149: the small grapheme-safe Timeline wrapping correction is acceptable at lower priority. Keep it separate from a general CJK/Unicode line-breaking system and coordinate overlapping work with fix(timeline): keep the year on multi-year labels and the header in frame #1032/[Bug]: Wrapped CJK text can start a line with closing punctuation (Tree, Timeline, Sequence notes) #1067.

These are scope agreements, not final approval or a claim that we reran your full suites. Please provide your own review/accountability confirmation for the submitted changes and explanations, including the recorded visible outcomes where applicable, and mark the ready items ready for review. One concise reply explicitly covering these six PRs is sufficient; no six separate acknowledgements are needed. Keep anything you have not personally reviewed as Draft and identify the remaining gap.

Please sequence the generated-output overlap between #1150 and #1151, rebuilding from the combined source when integration requires it rather than independently chasing the same conflicts. Reuse valid existing evidence at its recorded revision and rerun only what a new diff/base invalidates, together with required final-head CI. No new five-project Agent benchmark is requested for these bounded fixes, and skipped CI coverage should continue to be identified accurately.

Thanks again. The intended integration target remains dev, with #1150 and #1151 taking priority.

[BlueboxC]

Scope acknowledged. I have personally reviewed the changes, explanations and recorded visible outcomes of #1150 and #1151, understand their impact, and can answer questions about them. Both are now ready for review, with successful CI on unchanged heads 4c22ebef101f6e60c0d660e659e215d7759c605f and e1ff91f8c752700250a8c6a77be25806e5cf0cc9.

The agreed integration order is #1150 first, then #1151. Generated-output overlap should be resolved by rebuilding from the combined source when integration requires it. The existing evidence and explicit skip limits remain pinned to their tested revisions; any new diff or base must be accounted for together with required final-head CI.

This personal confirmation covers #1150 and #1151 only; #1146–#1149 are not included.

BlueboxC