Repository navigation
fix(net): align whole-group IETF FETCH end locations - #6295
Conversation
Quest-Claim: 153fd45bc4bedc3d
Co-Authored-By: GPT-6 <noreply@openai.com>
|
Checkpointed the complete three-sided FETCH candidate: canonical whole-group request/answer handling, dual-accepting reader, and overflow-free maximum-group range refusal. Required cross-language interop passes; all selected JS tests, drafts and lint checks pass. Default Rust tests pass 7,981/7,982, with only the known base CLI readiness failure handled in #6290; 511 capture/playback and 219 CLI playback feature tests pass. Public Rust/JS API is unchanged. Drafts 14-19 whole-group FETCH_OK changes from Keep draft for Decisions needed: staged reader-first rollout (recommended pending explicit compatibility approval) versus the quest's all-at-once exception. The body records a concrete pinned Moxygen trace: request End bytes (Written by GPT-6) |
…nd-location # Conflicts: # doc/concept/standard.md # rs/moq-net/MIGRATE.md
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The moq-ietf draft describes the moq-lite-07 relay extension, whose wire this change does not touch, and the JS FETCH decoder is unchanged. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
@codex review (Written by Claude Opus 5.5) |
|
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 31 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. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 19d4a36a5b
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Problem
On moq-transport drafts 14-19 the publisher widens a whole-group FETCH end
{g, 0}to{g+1, 0}, while the subscriber rejects a canonical{g, 0}answer. The widening also overflows at the maximum Group ID, including draft 20+ Absolute filters, returning INVALID_RANGE for a valid but unsupported range.Closes #6199.
Approach
Keep the standalone request's whole-group form, compare groups without widening or subtraction, and echo
{g, 0}for a covered whole group. The subscriber accepts both the canonical{g, 0}and the legacy{g+1, 0}answer, keeping the final sequence and object bounds. Draft 20+ stays inclusive.Tests reproduce all three failures before the fix and cover drafts 14-22: canonical and legacy answers with and without End Of Track, nonzero frame starts, bounded ranges, cancellation, and the wire's maximum Group ID. The moq-net MIGRATE note and
doc/concept/standard.mddescribe the drafts 14-19 meaning.Retires
quest/m1/ietf-fetch-end-location.md, which unblocksquest/m1/js-fetch.md.Decision
The maintainer settled this on behavior (recorded in the merged quest): all three sides land together, with no staged rollout.
{g+1, 0}only because itstoExclusiveEndhelper widens the request before comparing; its parser and fetch handle keep{g, 0}, and no data path fails. Behavior wins over that validator.Impact
{g, 0}rather than{g+1, 0}; receivers still accept the old form. End Of Track sets the final sequence to the group after the fetched one. Draft 20+ encoding is unchanged.drafts/describe moq-lite and its extensions;draft-lcurley-moq-ietfcovers the moq-lite-07 relay hop, whose wire this does not touch, so the earlier note added there was dropped. The JS FETCH decoder (JS refuses every FETCH) is unchanged too.Validation
{g+1, 0}, and a maximum-group multi-range returns INVALID_RANGE.just checkpasses JS, Rust clippy/lint, quest and drill; its flake module stops at the Nix runner-image build because this host deniesunsharewriting/proc/self/uid_map, unrelated to this change.quest check --base origin/mainpasses.just test interop --all: 127 PASS, 0 FAIL across Rust, Python, Go, C++, JS (browser, node, bun), C and GStreamer publishers and subscribers, auth refusals, CLI FETCH, finite tails, compression and drop cases.🤖 Generated with Claude Code
(Written by Claude Opus 5.5)