✅ Assert the fast emitter's output always re-parses in differential fuzz

[frenck]

Breaking change

None. This change is fuzz-harness only; it touches no library code and no public API.

Proposed change

The differential fuzz oracle (check_emit_roundtrip, shared by the differential and differential_options targets) used to silently return when the re-emitted YAML failed to re-decode, split into several documents, or was invalid UTF-8. That hid the worst dumps bug class: output a reader rejects or reshapes, with no crash to flag it.

This tightens the oracle so each of those is a hard finding. The fast emitter takes a freshly decoded Value tree (no user edits), so it must always produce valid YAML that loads reads back as the same single document; anything else is a dumps bug. The lenient roundtrip target (the round-trip emitter, which re-emits modified documents and can legitimately produce composer-rejected output for exotic inputs) is left untouched.

This is the closing step of the fuzz-and-fix pass: the three emitter bugs the strict oracle surfaced are already fixed and merged (#144 quoting a ... document-end marker, #145 keeping a folded scalar off a marker line, #146 indenting a tagged sequence item under its dash). With those in, both differential targets now run clean under the strict oracle (702k and 1.5M executions locally, zero findings).

Type of change

• Dependency or tooling upgrade
• Bugfix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• Deprecation (replaces or removes a feature, with a migration path)
• Breaking change (a fix or feature that changes existing behavior)
• Code quality, refactor, or test-only change
• Documentation only

Additional information

• This PR fixes or closes issue: None
• This PR is related to: 🐛 Quote a string that re-reads as a document-end marker #144, 🐛 Keep a folded scalar off the start of a marker line #145, 🐛 Indent a tagged sequence item under its dash when indentless #146 (the emitter bugs this strict oracle surfaced)
• Link to a separate documentation pull request: None

Checklist

• I have read the AI Policy, and this pull request was not created by an autonomous agent.
• I fully understand the code in this pull request and can explain every line, including any AI-assisted changes.
• The change is covered by tests, and uv run pytest passes locally. A pull request cannot be merged unless CI is green.
• uv run ruff check . and uv run ruff format --check . pass.
• cargo fmt --check and cargo clippy --all-targets -- -D warnings pass.
• Round-trip fidelity is preserved: an unmodified document still re-emits byte-for-byte.
• No commented-out or dead code is left in the pull request.

If the change is user-facing:

• Documentation under docs/ is added or updated, and docs/verify_examples.py still passes.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: 9d713f60-99b4-455a-83b3-3a87b4d9af99

📥 Commits

Reviewing files that changed from the base of the PR and between fc31ce5 and 9852a20.

📒 Files selected for processing (1)

• src/lib.rs

---

📝 Walkthrough

Walkthrough

The differential fuzz harness now treats emitter roundtrip failures as hard assertions. It panics on invalid UTF-8, YAML re-decode failures, and reparsing into anything other than exactly one document, then continues with the existing values_equiv comparison.

Possibly related PRs

• frenck/YAMLRocks#98 — This PR changes the same check_emit_roundtrip logic in src/lib.rs, so it is directly related to the stricter roundtrip validation here.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly reflects the main change: making the differential fuzz emitter round-trip checks stricter.
Description check	✅ Passed	The description is detailed and directly matches the changes to the differential fuzz oracle.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

---

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.}

[codspeed-hq]

Merging this PR will not alter performance

✅ 16 untouched benchmarks

---

_{Comparing frenck/fuzz-assert-emit-reparse (9852a20) with main (fc31ce5)}

[Copilot]

Pull request overview

This PR tightens the differential fuzz oracle (check_emit_roundtrip, shared by the differential and differential_options targets) so that the fast emitter's output is held to a strict re-parse invariant. Previously, invalid UTF-8, re-decode failures, and multi-document output were silently returned (skipped); now each is a hard panic. Since the fast emitter operates on a freshly decoded Value tree with no user edits, any such outcome is a genuine dumps bug. The change is fuzz-harness only and touches no library code or public API. It is the closing step of an emitter hardening pass whose three surfaced bugs were already fixed in #144, #145, and #146.

Changes:

• Convert the three lenient return skips (UTF-8 decode, re-decode, document-count) into hard assertions with descriptive panic messages that name the triggering options, input, and emitted text.
• Update the differential and check_emit_roundtrip doc comments to reflect the strict-assertion contract and document the three bugs this oracle surfaced.

---

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