diff --git a/flake.lock b/flake.lock index 0f0ac9a871d..d42fd913099 100644 --- a/flake.lock +++ b/flake.lock @@ -35,11 +35,11 @@ }, "nixpkgs": { "locked": { - "lastModified": 1791487292, - "narHash": "sha256-tqM7xDEsaeK1p6VvRygeZ9LkQRpUUkbRT02BHv6hnTM=", + "lastModified": 1792488032, + "narHash": "sha256-RkAXQ3+a5rHMiP97Er7AawR1kvEXX9tK1mhiWedOPTc=", "owner": "NixOS", "repo": "nixpkgs", - "rev": "8edc0c72e3a38faf5434e40d1f19431125fe4b30", + "rev": "e513db7b0a65def6f6e17bb55195dfe510871944", "type": "github" }, "original": { diff --git a/flake.nix b/flake.nix index 7c2406df250..94a6cd3d471 100644 --- a/flake.nix +++ b/flake.nix @@ -158,7 +158,13 @@ cargo-hack cargo-nextest # Source-based test coverage (`just rs coverage`). - cargo-llvm-cov + # Its LLVM test helpers must read the compiler's profile format. + (cargo-llvm-cov.override { + rustPlatform = makeRustPlatform { + cargo = rust-toolchain; + rustc = rust-toolchain; + }; + }) # Browser/WASM bindings (rs/moq-wasm -> @moq/wasm via `just js wasm`). # Cargo.lock selects the matching schema version, independently of # nixpkgs' default wasm-bindgen-cli. diff --git a/quest/m0/README.md b/quest/m0/README.md index 72ee972a4b8..a057d07a2fe 100644 --- a/quest/m0/README.md +++ b/quest/m0/README.md @@ -68,7 +68,7 @@ source, and unwinding so a panic closes only its connection. Nightly on main (triaged 2026-10-13): two build breaks from 2026-10-11 get quests here. The nixpkgs bump broke the Linux zigbuild release builds (its -GCC 16 miscompiles lld 21, fixed upstream and waiting on `nixpkgs-unstable`), +GCC 16 miscompiles lld 21; fixed by bumping nixpkgs past the upstream fix), and fat LTO in the release profile broke the fuzz build. Audio quality has been red since it landed, against budgets recorded locally; re-measuring them on the runner moved here from Watch (maintainer, 2026-10-13). The other red jobs @@ -83,7 +83,6 @@ Each lands its fixes in its own PR and plans only what needs a decision. ## Required - [CI cost and latency](/quest/m0/ci/README.md) - cloud CI runs only in the merge queue as one `just ci` pass, expensive checks run nightly and gate `release`, caches move to R2, and runners are chosen last -- [Release builds link](/quest/m0/release-link.md) - the Linux zigbuild release binaries and moq-gst packages link again after the 2026-10-11 nixpkgs bump, so the next release cut can publish - [qmux 0.5.3](/quest/m0/qmux-0.5.3.md) - crates.io publishes qmux 0.5.3 with the builder session config the release backport of #5609 needs - [Release shutdown backport](/quest/m0/ffi-shutdown-backport.md) - `release` publishes a non-breaking moq-ffi with an async `drain()` and qmux sessions on the QUIC stream limit - [Broadcast epochs](/quest/m0/broadcast-epoch/README.md) - every first-party publisher that can restart mints a fresh route epoch, the newest wins a path, and only routes with the same epoch resume a subscription diff --git a/quest/m0/ci/nightly-ffi-reuse.md b/quest/m0/ci/nightly-ffi-reuse.md index a2c990f8427..574826b84ff 100644 --- a/quest/m0/ci/nightly-ffi-reuse.md +++ b/quest/m0/ci/nightly-ffi-reuse.md @@ -15,9 +15,8 @@ quest. ## Plan `moq-ffi-v0.4.16` (2026-10-16) was cut from a `release` without #5549, so it -ran the per-language workflows and verified nothing here. `Shipping sizes` -skips while the Linux release builds fail; see -[Release builds link again](/quest/m0/release-link.md). +ran the per-language workflows and verified nothing here. The Linux release +toolchain now links; verify `Shipping sizes` on the next scheduled Nightly. The first scheduled Nightly after the change (run 38897397428, 2026-10-16) already verified the dry-run side: the nested job names match the publisher @@ -35,7 +34,3 @@ From the first `Shipping sizes` run: - The Go static libraries keep their `release/go-lib-/` history keys, so the comparison shows no removed or added Go entries. - -## Related - -- [Release builds link again](/quest/m0/release-link.md) - unblocks `Shipping sizes` diff --git a/quest/m0/ffi-shutdown-backport.md b/quest/m0/ffi-shutdown-backport.md index aa36cac9e7b..8d1256fd1d8 100644 --- a/quest/m0/ffi-shutdown-backport.md +++ b/quest/m0/ffi-shutdown-backport.md @@ -51,8 +51,3 @@ none. ## Required - [qmux 0.5.3](/quest/m0/qmux-0.5.3.md) - crates.io publishes qmux 0.5.3 with the builder session config #5609 needs - -## Related - -- [Release builds link](/quest/m0/release-link.md) - the release binaries - must link for the cut that publishes this diff --git a/quest/m0/release-link.md b/quest/m0/release-link.md deleted file mode 100644 index 5778b20923a..00000000000 --- a/quest/m0/release-link.md +++ /dev/null @@ -1,46 +0,0 @@ -# [XS] Release builds link again on the current Nix toolchain - -## Goal - -The Linux release binaries (`moq-cli`, `moq-relay`) and the `moq-gst` -.deb/.rpm packages build with `cargo zigbuild` on `main`'s pinned Nix inputs, -for x86_64 and aarch64, so the next release cut from `main` can publish. - -## Plan - -Seen in the 2026-10-12 nightly (run 38276133321): `moq-cli / release / -Build`, `moq-relay / release / Build` and `moq-gst / Package .deb/.rpm` -fail on both Linux targets with `ld.lld: error: undefined symbol:` (an empty -name), referenced by `bcm.o:(aws_lc_0_45_0_OPENSSL_cpuid_setup)` in -`aws-lc-sys`. The nixpkgs bump in #5316 (d0736ec04) caused it. - -Root cause (found 2026-10-13): nixpkgs' lld 21.1.8 built with GCC 16 is -miscompiled. lld leaves `Symbol` fields uninitialized behind a memset GCC 16 -may drop, so a relocatable link (`ld.lld -r`) writes unnamed `STT_SECTION` -symbols with `SHN_UNDEF`, a different count each run. `zig cc -c` runs -`zig ld.lld -r` on every object it compiles, and zig links lld statically, -so zig's `bcm.o` carries the bad symbols. clang's own `bcm.o` is clean. -`ld.lld -r -o out.o bcm.o` on that clean object reproduces it with the -current nixpkgs lld 21 and not with the pre-bump one, lld 22, or -`staging-next`'s lld 21. Tracked upstream as NixOS/nixpkgs#571232; fixed by -NixOS/nixpkgs#568680. With `staging-next`'s (aaba2f7c9) `zig` and -`cargo-zigbuild`, `cargo zigbuild --locked --release --target -x86_64-unknown-linux-gnu.2.34 -p moq-relay` links. - -NixOS/nixpkgs#568680 is now in `nixpkgs-unstable`. Remaining work: bump -`flake.lock` (or take the weekly Dependabot nix bump), then confirm the same -command for `moq-relay`, `moq-cli` and `moq-gst` on both Linux targets. - -`release` still pins the pre-bump nixpkgs, so shipping releases are fine; -only the next `main` -> `release` merge carries the break. - -Decisions (maintainer, 2026-10-13): - -- ✅ No stopgap pin of `zig` or `cargo-zigbuild` to the pre-bump nixpkgs. - `release` is unaffected and the fix is already in `staging-next`. Revisit - only if a release from `main` must ship before the fix lands. -- ✅ No PR-time zigbuild check on `flake.lock` changes. The nightly caught - this within a day, and a full zigbuild per nix bump works against the CI - cost plan (#5426). - -Public API: none. Wire: none. diff --git a/quest/m1/README.md b/quest/m1/README.md index 7b42ed8e640..ca3befde409 100644 --- a/quest/m1/README.md +++ b/quest/m1/README.md @@ -48,7 +48,6 @@ lands before the cut. ## Required - [Epoch correctness](/quest/m0/broadcast-epoch/README.md) - publisher identity, duplicate suppression, and catalog pinning gate the breaking cut -- [Release builds](/quest/m0/release-link.md) - package artifacts can be built - [Sink restart follow](/quest/m1/sink-restart-follow.md) - moq-cli's server and push sinks follow an epochless `Restart` where their output can mark it, and fail loud otherwise - [Cluster routing](/quest/m1/cluster-routing/README.md) - prove shared routing contracts, implement per-origin routes, and connect outbound local clusters diff --git a/quest/m1/audit-ci.md b/quest/m1/audit-ci.md index 66bdedc257e..2e8bace7cbd 100644 --- a/quest/m1/audit-ci.md +++ b/quest/m1/audit-ci.md @@ -69,9 +69,9 @@ Check first (from the test cost audit's m0 run, 2026-10-19): fixture in the entry itself, or a semantic conflict with a PR merged after it was queued. Check the entry's head branch and what merged around it before calling it a flake. -- Owned elsewhere: release builds by - [release builds link](/quest/m0/release-link.md), and Rust advisories by - [the dependency audit](/quest/m1/audit-deps.md). +- Owned elsewhere: Rust advisories by + [the dependency audit](/quest/m1/audit-deps.md). The release toolchain now + links. - Re-running a single job in place cannot redeploy Pages: the earlier attempt's `github-pages` artifact remains and `deploy-pages` refuses two. Delete the stale artifact first, or wait for the next Nightly.