PPR: marker-in-content regression test and version-skew error message (#4890)

[AbanoubGhadban]

What

Closes #4890 — the remaining acceptance criteria after the core PostponedState-on-metadata work was absorbed by the pipeline re-land (#4888).

1. Marker-in-content regression test (JS)

Adds a test proving the metadata-based protocol is immune to the classic in-band signaling flaw: a component whose rendered HTML contains the literal old PPR_POSTPONED_STATE delimiter text (injected via dangerouslySetInnerHTML so the exact bytes land in the output) round-trips correctly through prerender and resume.

This is the test that would fail under the old #4659 prototype's delimiter-scanning approach — any rendered content containing the marker text would cause the split to land in the wrong place.

2. Version-skew error message (Ruby)

Improves the error message in ppr_check_prerender_protocol! to name both sides' expectations:

• The expected chunk metadata key (pprPrerenderComplete)
• The Rails gem version (react_on_rails_pro vX.Y.Z)

This makes version-skew scenarios (old renderer ↔ new Rails) diagnosable.

Not included (by design)

Parser stripping of PPR keys — discussed and decided to skip. The current design correctly keeps the parser generic (parseLengthPrefixedStream.ts and length_prefixed_parser.rb strip only payloadType). PPR keys pass through as opaque metadata for the application layer, consistent with hasErrors, consoleReplayScript, etc. Stripping them would couple the generic parser to PPR-specific constants with no architectural benefit.

Verification

• JS PPR tests: 12/12 pass (pnpm --filter react-on-rails-pro exec jest tests/pprServerRenderedReactComponent.test.jsx)
• Ruby version-skew test: 1/1 pass (bundle exec rspec -e 'lacks the PPR protocol metadata')

🤖 Generated with Claude Code

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 44bc3b40-4b9a-409a-adb9-897cb52b8de2

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

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.}

[greptile-apps]

Greptile Summary

The PR strengthens PPR protocol regression coverage and makes version-skew failures easier to diagnose.

• Adds a marker-in-content test covering prerender metadata extraction and resume.
• Includes the expected PPR metadata key and Rails gem version in protocol-mismatch errors.
• Updates the Ruby helper spec for the expanded diagnostic.

Confidence Score: 4/5

The PR appears safe to merge, with only a non-blocking gap in the specificity of the version-message assertion.

The production diagnostic uses available, correctly loaded constants and the marker regression exercises the real metadata protocol; only the Ruby test can pass if the actual version value is omitted.

Files Needing Attention: react_on_rails_pro/spec/dummy/spec/helpers/react_on_rails_pro_helper_spec.rb

Important Files Changed

Filename	Overview
packages/react-on-rails-pro/tests/pprServerRenderedReactComponent.test.jsx	Adds a focused regression fixture proving legacy delimiter text in rendered content does not corrupt metadata-based postponed state.
react_on_rails_pro/app/helpers/react_on_rails_pro_helper.rb	Expands the existing protocol-mismatch diagnostic with the expected metadata key and Rails gem version.
react_on_rails_pro/spec/dummy/spec/helpers/react_on_rails_pro_helper_spec.rb	Covers the expanded diagnostic, but does not verify that the actual version value is present.

_{Reviews (1): Last reviewed commit: "PPR: marker-in-content regression test a..." | Re-trigger Greptile}

[github-actions]

size-limit report 📦

Path	Size
react-on-rails/client bundled (gzip)	63.61 KB (0%)
react-on-rails/client bundled (gzip) (time)	63.61 KB (0%)
react-on-rails/client bundled (brotli)	54.61 KB (0%)
react-on-rails/client bundled (brotli) (time)	54.61 KB (0%)
react-on-rails-pro/client bundled (gzip)	64.96 KB (0%)
react-on-rails-pro/client bundled (gzip) (time)	64.96 KB (0%)
react-on-rails-pro/client bundled (brotli)	55.77 KB (0%)
react-on-rails-pro/client bundled (brotli) (time)	55.77 KB (0%)
registerServerComponent/client bundled (gzip)	135.45 KB (0%)
registerServerComponent/client bundled (gzip) (time)	135.45 KB (0%)
registerServerComponent/client bundled (brotli)	81.77 KB (0%)
registerServerComponent/client bundled (brotli) (time)	81.77 KB (0%)
wrapServerComponentRenderer/client bundled (gzip)	127.9 KB (0%)
wrapServerComponentRenderer/client bundled (gzip) (time)	127.9 KB (0%)
wrapServerComponentRenderer/client bundled (brotli)	75 KB (0%)
wrapServerComponentRenderer/client bundled (brotli) (time)	75 KB (0%)

[claude]

Review Summary

Reviewed the diff (JS regression test + Ruby error-message improvement + spec update). This is a small, well-scoped PR and I didn't find any bugs, security issues, or performance concerns.

What I verified:

• PprMarkerInContent correctly uses dangerouslySetInnerHTML to inject the literal <!--PPR_POSTPONED_STATE--> bytes into rendered HTML, proving the metadata-based protocol (chunk-level pprPrerenderComplete/pprPostponedState/pprRenderErrored keys) is immune to in-band marker collisions — a real regression test for the Experimental: PPR prerender + resume rendering pipeline #4659-style flaw, not just a smoke test.
• The new test reuses existing helpers/constants (runPrerender, runResume, collectStreamResult, PPR_*_CHUNK_KEY) consistently with the rest of the suite, and exercises both the prerender and resume paths end-to-end.
• ReactOnRailsPro::Ppr::PRERENDER_COMPLETE_CHUNK_KEY and ReactOnRailsPro::VERSION referenced in the improved error message both exist and resolve as expected (react_on_rails_pro/lib/react_on_rails_pro/ppr.rb, version.rb).
• The Ruby spec's Regexp.escape(ReactOnRailsPro::VERSION) correctly handles the dots in the version string (e.g. 17.0.0.rc.6) that would otherwise be interpreted as regex wildcards — good catch already incorporated from the prior review round.
• The raised error message concatenates cleanly (no missing spaces/words) and the assertion order in the spec (did not report completion metadata → pprPrerenderComplete → react_on_rails_pro v<version>) matches the actual message order.

Minor, non-blocking nitpick:

• The phrase "Rails (react_on_rails_pro vX.Y.Z) requires the renderer to emit a trailing protocol chunk..." reads a little awkwardly (parenthetical mid-sentence). Not a functional issue — just a possible wording tweak if you want to polish it further, e.g. "The installed react_on_rails_pro gem (vX.Y.Z) requires...".

No blocking issues found. LGTM.