Repository navigation
refactor(net)!: drop moq-net exports without a consumer - #6744
Conversation
Quest-Claim: 1b0b0942e3a0d7b3
Delete the dead and test-only moq-net, @moq/net, and moq-ffi items the
API audit found, moving their tests onto what remains: Epoch::time and
moq_epoch_time_ms with every binding wrapper, broadcast::Demand,
goaway::Consumer::{peek, recv}, Version::{names, from_alpn, is_ietf},
Server::with_origin, and the unused stats, frame, path, and error items.
announce::Consumer::is_synced goes test-only and varint::MAX_QUIC
crate-private; JS Group.Demand and the extensions props go @internal.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Warning Review limit reachedYou've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Next included review available in 36 minutes. View limit details
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. Comment |
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
kixelated
left a comment
There was a problem hiding this comment.
Automated review by review (OpenAI)
Reviewed commit: ccfd7f4
No actionable bugs found. The direction is coherent: removing unused surface area rather than preserving redundant wrappers. The inspected Rust polling replacements preserve recovery/closure behavior, current protocol variants remain covered by the migrated enumeration tests, and the epoch API removal is aligned across the FFI and language wrappers.
Verification limits: static GitHub review of the changes and relevant surrounding code only. I did not run builds, tests, binding generation, or interoperability checks.
|
Maintainer decisions (2026-10-25):
(Written by Claude Opus 5.5) |
# Conflicts: # quest/m1/api-audit/README.md # rs/moq-net/src/model/group.rs # rs/moq-net/src/stats.rs # rs/moq-net/src/version.rs # rs/moq-relay/src/websocket.rs # rs/moq-tokio/src/websocket.rs
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Conflicts: # kt/moq/src/jvmAndAndroidTest/kotlin/dev/moq/SmokeTest.kt # quest/m1/js-reload-config.md
|
Merged
(Written by Claude Opus 5.5) |
# Conflicts: # quest/m1/api-audit/ietf-private.md # rs/moq-net/MIGRATE.md
|
Dropped from the queue on a (Written by Claude Opus 5.5) |
Take #6744's export removals (Version::from_alpn, is_ietf, Error::TooManyParameters, SetupOptions remove*). The SETUP codec gates on is_lite() instead. Restore quest/m1/ietf-drop/d14-17.md, since #6707 added an open decision on the request-ID history to it. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Finishes quest/m1/api-audit/net-exports.md, the follow-up to moq-dev/moq#6470: the moq-net,
@moq/net, and moq-ffi items with no consumer outside tests go private, test-only, or away. Tests that used them now poll what remains.Changes
Epoch::time, FFImoq_epoch_time_ms, and the wrappers in every binding are removed (Pythonepoch_time, Swift/Kotlin/DartepochTime, GoEpochTime, C++epoch_time). The Dart bindings are regenerated. Binding tests now check the invalid-epoch refusal throughannounce.moq_mint_epochstays.announce::Consumer::is_syncedis now#[cfg(test)]andannounce::Consumer::syncedis removed. Integration tests and benches callpoll_synced.Error::TooManyParameters,DecodeError::{ExpectedData, Missing, Unsupported},frame::Consumer::poll_read_all(now inlined intoread_all),track::Ordered::poll_finished,Path::next_part,stats::Registry::exclude,stats::Traffic::{is_announced, total_bytes},stats::Tier::is_default,setup::Token::CAT. In JS:Varint.sizeLeadingOnes,SetupOptions.{removeBytes, removeVarint}, thedeliveryTimeoutaccessors, andEpoch.time, which mirrors the Rust removal.broadcast::Demandand bothdemand()getters, plus the loom and integration assertions that only exercised them.goaway::Consumer::{peek, recv}. Callers poll the consumer; the moq-net integration support has a smallDrainingtrait.bandwidth::Producer::{abort, closed}. Its write-onlyabortfield goes too.Version::{names, from_alpn, is_ietf}. Tests list versions withVersion::from_codeover the wire-code ranges.Server::with_origin,stats::Handle::{parent, tier}, and groupProducer::{poll_unused, abort_unused}.varint::MAX_QUICispub(crate). JSGroup.Demand/demand()andConnectProps/AcceptProps.extensionsare@internal.rs/AGENTS.mdnow usestrack::Demandas its example lifecycle watcher, andrs/moq-net/AGENTS.mdno longer mentionsis_ietf(). MIGRATE entries were added for moq-net and@moq/net. Three quests that linked here were updated.Choices made unattended (review these)
transport::poll::Buffered::get_mut: the shared mock harness reads below the buffering, in integration tests and benches.bandwidth::Producer::{used, unused}: the wasm-only moq-ffi dialer (session/dial.rs) awaits them.frame::Buffer::{is_full, filled_mut}:group::Consumer::poll_read_framesandread_framesuse them, so they arepub(crate)rather than deleted.Epoch.timeis also removed, so the two languages stay mirrored. It was unreleased and only tests read it. The sharedepoch.jsonvectors dropunix_ms.Version::from_codeover0xff0dad00..=0xff0dadffand0xff000000..=0xff0000ff. This avoids hardcoding the-wipname. The CLI help tests in moq-cli and moq-tokio use the same approach.Impact
@moq/net, moq-ffi, and every binding wrapper. Migration notes are inrs/moq-net/MIGRATE.mdandjs/net/MIGRATE.md. The epoch-time removals were never released, so they have no entry. Onlymoq_mint_epochremains.Checks
just checkpasses: 8898 Rust tests, Python and JS tests, and kt, swift, go, cpp, dart, and obs checks.just fixincludes the wasm build.cargo check --all-targets --all-featurespasses on the touched crates.just rs loom(two loom models were deleted, the rest untouched), the binding test suites beyondcheck, and the interop lane (no wire change).Follow-ups
RecvBandwidth::pollin moq-tokio still handles a closed output, butoutcan no longer close while the call borrows it. The arm can go if that stays true.(Written by Claude Opus 5.5)
🤖 Generated with Claude Code