Repository navigation
fix(mcp): define the SSE port out-parameter in the transport-less build - #2643
Merged
Merged
Conversation
12 of 18 tasks
lusoris
added a commit
that referenced
this pull request
Oct 8, 2026
…ld (#2643) * fix(mcp): define the SSE port out-parameter in the transport-less build The stub of vmafx_mcp_start_sse never wrote its port out-parameter, so the cuda tidy lane reported readability-non-const-parameter at mcp_server.c:204. The parameter is a documented output in mcp.h, so it stays non-const and the stub now stores 0 (no bound port). * docs(state): record the cuda lane mcp_server finding as closed * test(mcp): pin the transport-less vmafx_mcp_start_sse refusal and its port output Signed-off-by: Lusoris <lusoris@proton.me>
lusoris
force-pushed
the
fix/tidy-mcp-server-const
branch
2 times, most recently
from
October 8, 2026 22:48
6104d99 to
8645fed
Compare
…026-6617 and the stdlib advisories (#2651) * fix(deps): move golang.org/x/net to v0.60.0 for GO-2026-6617 The Go vulnerability database published GO-2026-6617 (HTTP/2 server crash from an HPACK encoder race, fixed in v0.60.0). govulncheck found it reachable from the controller, the tune executor, pkg/libvmaf and tools/obssmoke, so the pre-push security hook refused every push. go vet ./... and govulncheck ./... pass after the bump. * fix(deps): move the Go toolchain to 1.27.2 for the standard-library advisories The vulnerability database release that published GO-2026-6617 also published twelve standard-library advisories (GO-2026-6599/6600/6603/6604/6605/6607/6608/ 6609/6610/6611/6612/6613), and the standard library carries GO-2026-6617 in its own HTTP/2 copy. With x/net v0.60.0 alone the CI go vet + go test job still failed at the govulncheck gate on Go 1.27.1, and so did the local pre-push security hook. Go 1.27.2 fixes all thirteen. go.mod declares go 1.27.2. RELEASE_GO_BASE and DEV_GO_BASE move to the golang:1.27-trixie digest that carries Go 1.27.2, and the Dockerfile and compose copies are rewritten by check-base-image-single-source.sh --write. The state row no longer claims govulncheck passed on 1.27.1. * build(go): pin gosec to a commit built with x/tools v0.50.0 so it reads Go 1.27.2 export data Go 1.27.2 writes export data version 5. gosec v2.29.0, the newest release, builds with x/tools v0.49.0, which reads at most version 4, so under Go 1.27.2 it reported "could not import" for every package and checked nothing. Commit 9e8e5d91f2f3 of gosec's main branch builds with x/tools v0.50.0; under Go 1.27.2 it type-checks the whole tree and reports the same five findings v2.29.0 reports on master under Go 1.27.1. Signed-off-by: Lusoris <lusoris@proton.me>
…2641) * port(upstream): MSVC: install static library as vmaf.lib (3b4dd350e) A static MSVC, clang-cl or icx-cl build names its installed libraries vmaf.lib and vmafx.lib, the files the MSVC linker opens for -lvmaf and -lvmafx, instead of Meson's libvmaf.a and libvmafx.a. The names are set with explicit name_prefix / name_suffix keywords because Meson's namingscheme option needs Meson 1.10 and this project's floor is 1.4.0 (ADR-2752). Every other build keeps its names. scripts/ci/check_msvc_library_names.py checks the installed files and pkg-config files; the MSVC CUDA, SYCL and ARM64 legs run it after installing. Signed-off-by: Lusoris <lusoris@proton.me>
…ld (#2643) * fix(mcp): define the SSE port out-parameter in the transport-less build The stub of vmafx_mcp_start_sse never wrote its port out-parameter, so the cuda tidy lane reported readability-non-const-parameter at mcp_server.c:204. The parameter is a documented output in mcp.h, so it stays non-const and the stub now stores 0 (no bound port). * docs(state): record the cuda lane mcp_server finding as closed * test(mcp): pin the transport-less vmafx_mcp_start_sse refusal and its port output Signed-off-by: Lusoris <lusoris@proton.me>
lusoris
force-pushed
the
fix/tidy-mcp-server-const
branch
from
October 9, 2026 02:11
8645fed to
13fe86c
Compare
This was referenced Oct 8, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
The cuda clang-tidy lane reported
core/src/vmafx/mcp_server.c:204:43 readability-non-const-parameteron a clean master (8493fcd9d): the transport-less stub ofvmafx_mcp_start_ssenever wrote itsportparameter.portis a documented output of the public function (core/include/vmafx/mcp.h, read bycore/src/compat/libvmaf/mcp.con success only), so it stays non-const; the stub now stores 0 (no bound port) whenportis non-NULL, as the real implementation does when it binds. No baseline edit. HostedTidy Ratchetruns the cpu lane and was green on master; only the cuda lane (run locally) is affected.Type
fix— bug fixChecklist
git commit -s; fix a branch withgit rebase --signoff origin/master). See DCO sign-off.make format && make lintis green locally.test_vmafx_mcp_absent(suitefast): positive (sentinel*portbecomes 0, NOTSUP, error namesmcp), boundary (NULLport), NULL error pointer; skips with a reason in an MCP build. Against the unfixed stub it fails (port output defined (0), not left as the sentinel); with the fix 3 of 3 pass.Bug-status hygiene (ADR-0165)
docs/state.mdupdated in this PR with a row in Recently closed.Netflix golden-data gate (ADR-0024)
assertAlmostEqual(...)score in the Netflix golden Python tests.Deep-dive deliverables (ADR-0108)
AGENTS.mdinvariant note — no rebase-sensitive invariants.port).Reproducer
On master: exit 2/4 with
mcp_server.c: warnings 0 -> 1(and, until #2636 lands,core/src/hip/dispatch_strategy.c: not measured). On this head merged with #2636:598 TUs, 0 warnings (baseline 0), ... baseline matches measurement, exit 0.Known follow-ups
None.