[kixelated]

Problem

Reconfiguring a published rendition with jitter or delay unset could resolve below its advertised floor and abort import with JitterDecreased or DelayDecreased.

Approach

• Floor only unsupplied jitter and delay in Rendition::set at the current catalog values, including measurement and modify raises. Explicit lower values still fail without changing the catalog.
• Keep those fields owned by detection, so later measurements can raise them. Other estimate fields retain their existing resolution rules.
• Add deterministic coverage for mixed supplied/unsupplied fields, lower subsequent measurements, H.264/H.265 resolution changes through container::Producer::reconfigure, and a new track identity accepting its own lower values. Update the existing text-jitter and video-delay reset expectations.
• Complete and remove the quest and its links.

Impact

• Public API: no signature changes. Resetting an omitted jitter or delay retains its published floor; explicitly lowering either remains an error.
• Wire: no format changes.

Overlap inspected

• moq-dev/moq#7436 changes HLS playlist retention and predecessor routing. Its remaining hand-over race is separate. No catalog source changes overlap.
• moq-dev/moq#7387 adds the nightly conformance fixture and raises jitter via modify before an inline resolution switch. This fixes that import failure without copying its fixture or claiming its remaining HLS/DASH conformance failures are resolved. Its quest dependency is removed here.

Alternatives

Flooring explicit fields would silently accept forbidden decreases. Keeping the previous supplied hints would pin fields the new config deliberately leaves to measurement. This uses the same published-floor rule as measurement, only for omitted fields.

Choices made without asking

Test the shared importer reconfiguration boundary with both H.264 and H.265 configs and real timestamp/reorder observations, without adding ffmpeg or HLS export dependencies to this unit regression. No new public helper or dependency.

Validation

• Reproduction before the fix: both new tests fail with JitterDecreased on the unset-field reconfiguration.
• Standard quest gate just rs test -p moq-mux: 1,963 passed, 3 skipped at 38c47dac423b08232e38098f1aa7995f15c02a1c.
• Standard quest gate just ci origin/main: exit 1 in this harness environment. mbx cannot launch Cargo (Argument list too long), so Rust docs and the CI lint-warning regression fail before their intended validation. Existing repo-side work: moq-dev/moq#7594. No upstream posts or fixes were made.
• Final-head diagnostic: quest gate env -u BASH_ENV just ci origin/main: exit 0, all selected checks pass, 4,400 tests passed and 66 skipped at 654b1bfa9b. This retains mbx and the Nix toolchain but disables the harness startup script in child shells. It is not an unmodified standard-gate pass.
• Earlier diagnostic, before the review cleanup: the same gate with BASH_ENV unset and pinned Cargo ahead of the mbx shim also passes, 4,400 tests. The final diagnostic above uses mbx.
• quest check --base origin/main: pass. No full HLS conformance/ffprobe run: its fixture and unrelated export failures remain owned by the existing PRs.

Review

Claude Code, Claude Opus 5.5 (claude-opus-5-5), medium effort, through quest review --base origin/main --harness claude.

• Latest: round 2, reviewed head 654b1bfa9bfe114a52726b5a907e8f5a50357c30. No findings.
• Round 1: one low-severity maintainability finding, fixed. set now reads the published estimate once and shares it between flooring and decrease validation; stage uses the same comparison helper.
• No open decisions or unaddressed findings.

Follow-ups

None created. Existing HLS PRs retain their owners and remaining work.

(Written by GPT-6.1 Sol)

[kixelated]

Automated review by review (OpenAI)

Reviewed commit: 654b1bf

No actionable bugs found. Flooring only omitted jitter/delay in the shared rendition setter addresses the reconfiguration failure at its source without silently accepting explicit decreases or turning measured fields into supplied hints. Keeping the existing identity-change path gives a new track its own floor. The mixed-field and importer-boundary regressions cover the important distinctions without adding production machinery.

Verification: static GitHub review of the changes and surrounding catalog/reconfiguration paths. I did not run tests, builds, or HLS/ffprobe conformance checks; the PR's reported validation was not independently reproduced.

(Written by OpenAI)

[coderabbitai]

No actionable comments were generated in the recent review.

Recent review info

Run configuration

• Source mode: Auto
• Repository: Unavailable
• Branch: Unavailable
• File: No configuration file from this repository applied
• Commit: Unavailable
• Layers: Organization UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: 6d19af93-25f2-44db-9911-1354569bbfaf

Commits

Reviewing files that changed from the base of the PR and between f8c608a and 654b1bf.

Files selected for processing: 5

• quest/m1/README.md
• quest/m1/catalog-jitter-floor.md
• quest/m1/hls-conformance.md
• rs/moq-mux/src/catalog/tracks.rs
• rs/moq-mux/src/container/producer.rs

Files with no reviewable changes: 3

• quest/m1/hls-conformance.md
• quest/m1/README.md
• quest/m1/catalog-jitter-floor.md

Included review availability: This review used your included allowance. Your plan provides up to 4 included reviews per hour; 1 remain after this review.

Walkthrough

Rendition reconfiguration now retains published jitter and delay values when the replacement config omits those fields. Explicitly lower values still return decrease errors. Catalog and producer tests cover omitted fields, explicit decreases, and changes to codec identity. The related quest entries and planning document were removed.

Severity of issue fixed: Low

Merge Risk: Minimal · up to 654b1

No merge-blocking code risk was established. The reported standard gate did not complete, although its diagnostic rerun passed.

Pre-merge checks: 4 · 1

Failed checks: 1 warning

Check name	Status	Explanation	Resolution
Docstring Coverage		Docstring coverage is 57.14% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff with cyclomatic complexity of at least 2. Documenta…	Write docstrings for the functions missing them to satisfy the coverage threshold.

Passed checks: 4 passed

Check name	Status	Explanation
Title check		The title clearly and concisely describes the main change: retaining published jitter and delay floors during mux reconfiguration.
Description check		The description directly explains the reconfiguration problem, implementation, behavior, tests, validation, and impact. It is fully related to the changeset.
Linked Issues check		Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check		Check skipped because no linked issues were found for this pull request.

Full details: Docstring Coverage

Explanation

Docstring coverage is 57.14% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff with cyclomatic complexity of at least 2. Documentation-only edits and deletions within surviving functions are included; unchanged diff context and fully deleted functions are excluded. Analyzed 7 functions across 2 files.

Resolution

Write docstrings for the functions missing them to satisfy the coverage threshold.

Finishing Touches

Simplify code

• Commit to this branch
• Create a new PR

• 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

[pullfrog]

✅ No new issues found.

Reviewed changes

• Rendition::set floors omitted fields — rs/moq-mux/src/catalog/tracks.rs: on an already-published rendition, unsupplied jitter/delay are floored at the catalog's current value before validation, so a same-codec reconfigure no longer resolves below the advertised floor and aborts with JitterDecreased/DelayDecreased. Explicitly supplied lower values still fail, and self.supplied keeps the un-floored fields so detection still owns them.
• check_decrease takes the published estimate explicitly — refactored from a method to check_decrease(previous, next); the !self.present guard moved to callers (set guards with if self.present, stage already rejects NotPublished). One catalog read per call.
• Public set doc — rs/moq-mux/src/container/producer.rs documents the floor and the explicit-lower refusal.
• Tests — new reconfigure_floors_only_unsupplied_fields (2x2 supplied/omitted matrix) and reconfigure_keeps_floors_until_the_track_identity_changes (H.264/H.265 via reconfigure, modify() raise, explicit-lower refusals, new identity taking its own floor); updated published_delay_never_decreases and published_text_jitter_never_decreases so set(None) retains the floor while replace(None) still refuses.
• Quest cleanup — removed quest/m1/catalog-jitter-floor.md, its m1/README.md row, and the hls-conformance.md link; no dangling references remain.

Verified locally: the three reconfigure tests pass, cargo clippy --locked -p moq-mux --all-targets -- -D warnings is clean, and the full moq-mux --lib suite passes except 20 pre-existing ffmpeg/ffprobe-dependent failures (those tools are absent from this shell, unrelated to the diff).

  ^{｜ View workflow run ｜ Using deepseek-v4.1-flash (free via Pullfrog for OSS) ｜ 𝕏}