diff --git a/quest/m1/README.md b/quest/m1/README.md index fa88b1bc5f6..af15501687b 100644 --- a/quest/m1/README.md +++ b/quest/m1/README.md @@ -181,6 +181,7 @@ per-frame CPU growth since #5926. - [HLS export lists a same-codec reconfigure's first group under the new init](/quest/m1/hls-reinit-same-codec.md) - the hand-over pick, whichever of the catalog and the records arrives first, after the catalog `hls` removal is planned - [moq-quic deletes its remaining 0-RTT code](/quest/m1/quic-zero-rtt-leftovers.md) - only the code that drops a 0-RTT packet remains, with a test - [Gate the no-em-dash rule](/quest/m1/em-dash-gate.md) - `just check` fails on an em dash in any tracked text file +- [The resume model passes its long run again](/quest/m1/resume-model-7542.md) - fix #7542's regression that drops groups a long-budget reader can still get - [A resume more than one group behind never hangs](/quest/m1/ietf-resume-behind.md) - a resumed read gets every group or a drop on every wire version - [moq-quic's idle connections keep no burst-sized state](/quest/m1/quic-readable-dedupe.md) - readable events are deduplicated and a quiet connection keeps no per-peak-stream bytes - [Stats name the route instance's epoch with a type](/quest/m1/stats-epoch-type.md) - `GroupFrame::epoch` gets a dedicated type and tests stop mislabelling the instance diff --git a/quest/m1/ietf-resume-behind.md b/quest/m1/ietf-resume-behind.md index e97e1693143..c2f417d81d3 100644 --- a/quest/m1/ietf-resume-behind.md +++ b/quest/m1/ietf-resume-behind.md @@ -24,3 +24,10 @@ found it and kept its two direct bounded-resume tests on draft 22 only. [moq-dev/moq#7625](https://github.com/moq-dev/moq/pull/7625)); a refused run must come back as dropped, not hang. - Enable the draft-18 cases of #7631's bounded-resume tests as the regression. +- A relay loses a group on a history SUBSCRIBE (folded in 2026-11-02, + maintainer, from [moq-dev/moq#7625](https://github.com/moq-dev/moq/pull/7625)): + with the `fetched_copy` fixture in `fetch_only` on draft-22, where the relay + already fetched and cached group 2, a fresh session's `1..3` read gets only + group 2. The relay asks upstream for groups 1 to 2 and upstream sends only + group 2, so group 1 is lost silently (`main` hangs instead). Group 1 must + arrive or come back as dropped. The `1..2` case #7631 tests is fine. diff --git a/quest/m1/resume-model-7542.md b/quest/m1/resume-model-7542.md new file mode 100644 index 00000000000..3eae4d31d94 --- /dev/null +++ b/quest/m1/resume-model-7542.md @@ -0,0 +1,23 @@ +# [S] The resume model passes its long run again + +## Goal + +`moq-net`'s route-switching resume model +(`model::resume_model::readers_match_the_logical_track`) passes at 3000 cases +on every seed again, fixed at its cause, so a reader with a long budget never +gets groups back as dropped that its route can still deliver. + +## Plan + +Planned 2026-11-02 (maintainer, follow-up planning after the 2026-11-02 /quest-complete run), from [moq-dev/moq#7625](https://github.com/moq-dev/moq/pull/7625)'s +checks. + +- It fails on `main` on seeds 12345, 777 and 4242 with the same minimal cases, + and bisects to [moq-dev/moq#7542](https://github.com/moq-dev/moq/pull/7542) + (lite-07 frame bounds, timed datagram admission, sparse reverse search, and + shared time-anchor walks). The 64-case run inside `just check` still passes. +- Symptom: a reader with an hour budget gets groups 0 to 10 back as dropped + when the route only wrote group 11. +- Repro: `MOQ_RESUME_SEED=12345 MOQ_RESUME_CASES=3000 cargo test -p moq-net + --all-features --release --lib model::resume_model::readers_match_the_logical_track`. +- Keep #7542's flat per-range costs; add the minimal case as a fixed regression.