Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions quest/m1/api-audit/resolve/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -52,6 +52,7 @@ Wire: none.

- [Rust request and follow](/quest/m1/api-audit/resolve/rust.md) - Rust names the two resolutions `request` and `follow`, deletes `routed` and `routed_broadcast`, and makes the route-level `Follow` private
- [Bindings request and follow](/quest/m1/api-audit/resolve/bindings.md) - moq-ffi, the C ABI, and every wrapper mirror `request` and `follow` with an optional epoch
- [The watch player's follow prop](/quest/m1/api-audit/resolve/watch.md) - `@moq/watch`'s `Broadcast` names its resolution choice `follow` instead of `announced`
- [Callers resolve once](/quest/m1/api-audit/resolve/callers.md) - no caller hand-rolls synced resolution or decides table absence by wall clock

## Related
Expand Down
20 changes: 20 additions & 0 deletions quest/m1/api-audit/resolve/watch.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,20 @@
# [S] The watch player's follow prop

## Goal

`@moq/watch`'s `Broadcast` names its resolution choice `follow` (default
true), mirroring `@moq/net`'s `follow` and `request`, instead of
`announced`. Passing `announced` throws a rename error, as `reload` already
does.

## Plan

Decided 2026-10-29 (maintainer), a follow-up of
[moq-dev/moq#7173](https://github.com/moq-dev/moq/pull/7173), which renamed
the `@moq/net` side and left the player surface out of scope. One vocabulary
across `@moq/net` and the player, so the prop reads as the method it picks.

Update the watch docs and README, and any in-tree caller that passes the prop.

Public API: breaking in `@moq/watch`, with an Impact migration note. Wire:
none.
1 change: 1 addition & 0 deletions quest/m1/ts-reliability/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -102,6 +102,7 @@ catalog linger (#5305).
- [A timebase step restarts the TS import](/quest/m1/ts-reliability/step-restart.md) - an unflagged step or a rewind republishes under a new epoch instead of ending the import and every subscriber's export
- [A restart keeps its catalog](/quest/m1/ts-reliability/restart-catalog.md) - a restarted import publishes its catalog with its first media
- [PAT changes](/quest/m1/ts-reliability/pat-lifecycle.md) - a running export follows a moved PMT PID
- [A small step back keeps the TS export's video](/quest/m1/ts-reliability/step-back.md) - the export re-acquires on a sub-group step back from any source instead of dropping its video
- [A demoted subscriber priority reaches the publisher](/quest/m1/ts-reliability/stitch-catch-up.md) - a mid-stream priority demotion reaches the publisher on lite-07 and moq-transport, including an old instance after an epoch switch

## Related
Expand Down
38 changes: 38 additions & 0 deletions quest/m1/ts-reliability/step-back.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,38 @@
# [S] A small step back keeps the TS export's video

## Goal

When a source's decode time steps back by more than the export's delay but
within one group, the TS export re-acquires and keeps rendering video,
instead of dropping nearly every video frame until the next group.

## Plan

Decided 2026-10-29 (maintainer), a follow-up of
[moq-dev/moq#7165](https://github.com/moq-dev/moq/pull/7165), which made the
jitter buffer re-acquire on an unmarked step in either direction.

Found while landing #7165: with 600 ms groups and a 500 ms delay, a −560 ms
step drops 105 of 120 video frames over 8 groups, on main's engine as well.
The export's video decode-time reconstruction clamps DTS to stay
non-decreasing, so only the first frame after the step regresses, and it is
still in time. The rest arrive late without a regression, so the jitter
buffer's backward rule never sees the step. Audio is unaffected.

The fix belongs where video decode time is reconstructed, not in the jitter
buffer: the step has to reach the jitter buffer as a regression it can act
on.

[A timebase step restarts the TS import](/quest/m1/ts-reliability/step-restart.md)
turns any rewind from a TS or SRT import into a new epoch, so this remains
for every other source (fMP4, MKV, FLV, RTMP, hang publishers). Kept a
separate quest for that reason.

Tests: an end-to-end TS export of an H.264 + AAC source with the −560 ms
step, which fails on main and keeps video rendering with the fix.

Public API: none. Wire: none.

## Related

- [A timebase step restarts the TS import](/quest/m1/ts-reliability/step-restart.md) - handles a TS or SRT source's rewind at import
Loading