Repository navigation
fix(deps): bump grpc to 1.83.2 and thrift to 0.24.0 - #7495
Conversation
Resolve Dependabot alerts #146 (apache/thrift < 0.24.0) and #151 (google.golang.org/grpc <= 1.83.1) on the root module with the minimal bump: no other module moves, and both need only go 1.25, so the Nix flake stays on its current nixpkgs. Recompute default.nix vendorHash for the new go.sum. This lands the security fix independently of the Dolt 2.4.1 upgrade in #7441, which carries the same two bumps. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
bee-ghosttrack
left a comment
There was a problem hiding this comment.
Approving on d84e982.
What I checked locally (GOTOOLCHAIN=go1.26.7):
go mod tidyreproduces the committedgo.mod/go.sumexactly (no diff);go mod verify= all modules verified;go list -mod=readonly -deps ./cmd/bdresolves.- Diff vs base 43b1fb2 is 3 files, +7/-7: the two version lines, their four go.sum lines, and
default.nixvendorHash. No other module moved in go.mod. - grpc v1.83.2 declares
go 1.25.0, thrift v0.24.0go 1.25; the root module isgo 1.26.5, so no toolchain pull. - Bazel:
go_deps.from_file(go_work = ...)reads versions from go.mod;MODULE.bazel.lockcarries no Go module versions (0 grpc mentions), and #7441's equivalent bump touched only go.mod/go.sum, so no lock regen is owed. - vendorHash: no nix here; relying on the
Test Nixjob, which is green on this head.
CI at review time: 30 pass, 5 skipping (cache seed/verify, Bazel Farm, vendorHash updater, authorize), 0 failed/pending.
|
This PR has been added to the review queue. Track progress: https://factory.gascity.com/city/maintainer-city/runs/gcg--9223372036845917344 |
|
Triage: |
Maintainer Adoption ReviewThanks for the contribution, @julianknutsen! This PR lifts This PR was reviewed and accepted as-is. Original PR ReviewDecision: approve. Specific gaps fixed:
Non-blocking follow-ups (shared for visibility — NOT completed in this PR):
Maintainer ChangesNo maintainer changes -- PR accepted as-is. Final Review StatusReady for the merge queue: final head CI: https://github.com/gastownhall/beads/actions/runs/38111111508 Review Iterations1 review pass performed: three independent review lanes plus a synthesis pass, over this PR rebased onto main Adopted via claude-code-claude-opus-5-5-max on behalf of Julian Knutsen |
Brings the branch forward so the PR can merge and GitHub will create pull_request runs for it. At fa94565 the PR was CONFLICTING on four paths, and GitHub does not run the PR suite in that state -- that head got only the three pull_request_target runs and no PR or PR Risk run at all, which is why this is a prerequisite for review rather than tidying. Each conflict resolved as @bee-ghosttrack specified in his 2026-10-11 review: - .github/workflows/proxied-local-smoke.yml -- took main's deletion. c292964 retired that lane and moved it under Bazel. Confirmed nothing is orphaned: TestProxiedLocalSmokeMatchesPinnedDoltVersion, which read this file and which an earlier commit on this branch updated, no longer exists on main either. - tools/bazel/dolt.bzl -- kept both sides. Main made the fetch rule take a `version` attr (merged cleanly) and added a "2.1.8" digest block for the migration corpus's external runtime; this branch added the "2.3.5" block. Both are present, ordered ascending alongside the existing "2.2.0" to match where main placed 2.1.8. DOLT_VERSION stays "2.3.5". - scripts/pull_dolt_image_test.go -- took main's side, which deletes the local const along with TestDoltImagePullWorkflowsUseRetryHelper (4f924fc). NOTE, because resolving this literally would have silently reverted the container-image pin: main did not delete `doltSQLServerImage`, it MOVED it to scripts/repo_helpers_test.go, still reading 2.2.0. Taking main's side alone therefore drops this branch's pin move. The one-line move to 2.3.5 is applied at the const's new home instead, which is what keeps TestPinnedDoltCLIMatchesContainerImage passing -- it compares the CLI pin against the container image, so the literal resolution would have failed it. - default.nix -- placeholder, deliberately, per the review. Main moved its own hash twice while this branch sat (creack/pty, then the grpc/thrift bump in gastownhall#7495), so the merged go.sum differs from both parents and the hash is new. Nothing here can generate it: neither nix nor docker is installed on the authoring machine. The agreed path is one extra push -- placeholder, read the reported value from the Nix Flake job log, push the real one -- and nothing else in the PR changes between those two pushes. `go mod tidy -diff` is empty on the merge result. The named dependencies land as main has them: creack/pty v1.1.24 present, thrift v0.24.0 and grpc v1.83.2 (main's gastownhall#7495 bump supersedes the v1.83.1 that matched dolt v2.3.5's own go.mod, and being the higher version it is what the graph selects). This branch's dolt/go v2.3.5 pin, vitess and go-mysql-server are preserved. One correction to the review's note on dependencies, since following it literally would make the tree untidy: goccy/go-json is in main's go.mod now and was absent from this branch before the merge, so it is this bump that drops it, not main. `go mod tidy -diff` is clean without it and re-adding it would not be tidy. Gates on the merge result: go build -tags gms_pure_go clean, gofmt clean, `go test ./scripts/` green including TestWorkflowsInstallPinnedDolt, TestPinnedDoltCLIMatchesContainerImage, TestBazelDoltJobMirrorsContainerJobs, TestCheckDoltServerRulesPerTarget and TestStorageEmbeddedShardManifestGeneratorNotStale; schema and testutil suites green. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
… the built binary Rebuilt as ONE commit on current main (9c4b3d8), not rebased: upstream's gastownhall#7495 took grpc to v1.83.2 and thrift to v0.24.0 on its own, so the grpc half of this change is now a no-op and the old go.mod hunks only conflicted. Re-measured on main with govulncheck (symbol scan, -tags gms_pure_go): GO-2026-6355 x/crypto v0.55.0 -> fixed in v0.56.0 (reachable) GO-2026-6354 x/crypto v0.55.0 -> fixed in v0.56.0 (reachable) GO-2026-5932 x/crypto v0.55.0 -> NO FIX at any version (module scan) After this change neither GO-2026-6355 nor GO-2026-6354 is reported. GO-2026-5932 is carried forward and named, not counted as resolved. Both example modules are mirrored so they cannot pin a lower floor than the module they demonstrate. All upgrades, no downgrades; upstream's go.mod is otherwise untouched. default.nix vendorHash regenerated with scripts/update-nix-vendorhash.sh (Docker nixos/nix fallback); its nix build of the new hash succeeded. Verified: go build ./... clean, go mod verify reports all modules verified, both example modules build. Agent-Signature: claude-code-claude-opus-5-5-unknown-reasoning on behalf of voxist.executor
… the built binary Rebuilt as ONE commit on current main (9c4b3d8), not rebased: upstream's gastownhall#7495 took grpc to v1.83.2 and thrift to v0.24.0 on its own, so the grpc half of this change is now a no-op and the old go.mod hunks only conflicted. Re-measured on main with govulncheck (symbol scan, -tags gms_pure_go): GO-2026-6355 x/crypto v0.55.0 -> fixed in v0.56.0 (reachable) GO-2026-6354 x/crypto v0.55.0 -> fixed in v0.56.0 (reachable) GO-2026-5932 x/crypto v0.55.0 -> NO FIX at any version (module scan) After this change neither GO-2026-6355 nor GO-2026-6354 is reported. GO-2026-5932 is carried forward and named, not counted as resolved. Both example modules are mirrored so they cannot pin a lower floor than the module they demonstrate. All upgrades, no downgrades; upstream's go.mod is otherwise untouched. default.nix vendorHash regenerated with scripts/update-nix-vendorhash.sh (Docker nixos/nix fallback); its nix build of the new hash succeeded. Verified: go build ./... clean, go mod verify reports all modules verified, both example modules build. Agent-Signature: claude-code-claude-opus-5-5-unknown-reasoning on behalf of voxist.executor
…reaches the built binary (gastownhall#6355) * fix(deps): raise x/crypto past the fixable advisories it carries into the built binary Rebuilt as ONE commit on current main (9c4b3d8), not rebased: upstream's gastownhall#7495 took grpc to v1.83.2 and thrift to v0.24.0 on its own, so the grpc half of this change is now a no-op and the old go.mod hunks only conflicted. Re-measured on main with govulncheck (symbol scan, -tags gms_pure_go): GO-2026-6355 x/crypto v0.55.0 -> fixed in v0.56.0 (reachable) GO-2026-6354 x/crypto v0.55.0 -> fixed in v0.56.0 (reachable) GO-2026-5932 x/crypto v0.55.0 -> NO FIX at any version (module scan) After this change neither GO-2026-6355 nor GO-2026-6354 is reported. GO-2026-5932 is carried forward and named, not counted as resolved. Both example modules are mirrored so they cannot pin a lower floor than the module they demonstrate. All upgrades, no downgrades; upstream's go.mod is otherwise untouched. default.nix vendorHash regenerated with scripts/update-nix-vendorhash.sh (Docker nixos/nix fallback); its nix build of the new hash succeeded. Verified: go build ./... clean, go mod verify reports all modules verified, both example modules build. Agent-Signature: claude-code-claude-opus-5-5-unknown-reasoning on behalf of voxist.executor * fix(deps): raise x/net to v0.60.0 past five reachable advisories govulncheck on main reports GO-2026-6603, 6610, 6611, 6612 and 6617 in golang.org/x/net v0.58.0 as reachable from bd's own code; v0.60.0 fixes all five. It requires x/crypto v0.57.0, x/text v0.42.0 and x/mod v0.41.0, so those rise with it in the root module and both example modules. govulncheck on this branch: 0 vulnerabilities. vendorHash recomputed by scripts/update-nix-vendorhash.sh (Docker path, completed nix build). Agent-Signature: claude-code-claude-opus-5-5-unknown-reasoning on behalf of karel.bourgois@voxist.com * fix(deps): align tools/nogo with the root x/mod and x/text floors Raising x/net to v0.60.0 also moved the root go.mod to x/mod v0.41.0 and x/text v0.42.0, but tools/nogo/go.mod stayed on v0.40.0 and v0.41.0. gazelle.go.work joins the two modules into one go_deps graph, so TestModuleRequirementsAgreeWithRoot (//tools/nogo:nogo_test) failed the Bazel test and integration lanes on this PR. Verified: //tools/nogo:nogo_test passes under Bazel; the same test run on the previous commit's files fails with the CI message; go mod verify is clean. Agent-Signature: claude-code-claude-opus-5-5-unknown-reasoning on behalf of voxist.executor
Resolve Dependabot alerts gastownhall#146 (apache/thrift < 0.24.0) and gastownhall#151 (google.golang.org/grpc <= 1.83.1) on the root module with the minimal bump: no other module moves, and both need only go 1.25, so the Nix flake stays on its current nixpkgs. Recompute default.nix vendorHash for the new go.sum. This lands the security fix independently of the Dolt 2.4.1 upgrade in gastownhall#7441, which carries the same two bumps. Co-authored-by: Julian Knutsen <julianknutsen@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com> (cherry picked from commit 7477ec1) Backport note: hotfix/1.3.1 was on grpc v1.83.0 (main was on v1.83.1), so go.mod/go.sum were re-resolved with `go get google.golang.org/grpc@v1.83.2 github.com/apache/thrift@v0.24.0 && go mod tidy`; no other module moved. default.nix vendorHash recomputed for this go.sum with a nixos/nix build.
…reaches the built binary (gastownhall#6355) * fix(deps): raise x/crypto past the fixable advisories it carries into the built binary Rebuilt as ONE commit on current main (9c4b3d8), not rebased: upstream's gastownhall#7495 took grpc to v1.83.2 and thrift to v0.24.0 on its own, so the grpc half of this change is now a no-op and the old go.mod hunks only conflicted. Re-measured on main with govulncheck (symbol scan, -tags gms_pure_go): GO-2026-6355 x/crypto v0.55.0 -> fixed in v0.56.0 (reachable) GO-2026-6354 x/crypto v0.55.0 -> fixed in v0.56.0 (reachable) GO-2026-5932 x/crypto v0.55.0 -> NO FIX at any version (module scan) After this change neither GO-2026-6355 nor GO-2026-6354 is reported. GO-2026-5932 is carried forward and named, not counted as resolved. Both example modules are mirrored so they cannot pin a lower floor than the module they demonstrate. All upgrades, no downgrades; upstream's go.mod is otherwise untouched. default.nix vendorHash regenerated with scripts/update-nix-vendorhash.sh (Docker nixos/nix fallback); its nix build of the new hash succeeded. Verified: go build ./... clean, go mod verify reports all modules verified, both example modules build. Agent-Signature: claude-code-claude-opus-5-5-unknown-reasoning on behalf of voxist.executor * fix(deps): raise x/net to v0.60.0 past five reachable advisories govulncheck on main reports GO-2026-6603, 6610, 6611, 6612 and 6617 in golang.org/x/net v0.58.0 as reachable from bd's own code; v0.60.0 fixes all five. It requires x/crypto v0.57.0, x/text v0.42.0 and x/mod v0.41.0, so those rise with it in the root module and both example modules. govulncheck on this branch: 0 vulnerabilities. vendorHash recomputed by scripts/update-nix-vendorhash.sh (Docker path, completed nix build). Agent-Signature: claude-code-claude-opus-5-5-unknown-reasoning on behalf of karel.bourgois@voxist.com * fix(deps): align tools/nogo with the root x/mod and x/text floors Raising x/net to v0.60.0 also moved the root go.mod to x/mod v0.41.0 and x/text v0.42.0, but tools/nogo/go.mod stayed on v0.40.0 and v0.41.0. gazelle.go.work joins the two modules into one go_deps graph, so TestModuleRequirementsAgreeWithRoot (//tools/nogo:nogo_test) failed the Bazel test and integration lanes on this PR. Verified: //tools/nogo:nogo_test passes under Bazel; the same test run on the previous commit's files fails with the CI message; go mod verify is clean. Agent-Signature: claude-code-claude-opus-5-5-unknown-reasoning on behalf of voxist.executor (cherry picked from commit 3b49066) Backport notes: - tools/nogo/go.mod and go.sum are dropped (no tools/nogo module on hotfix/1.3.1); the CHANGELOG hunk is dropped (the v1.3.2-rc.3 prep carries the entry). - go.mod/go.sum were re-resolved on this branch with `go get golang.org/x/net@v0.60.0 golang.org/x/crypto@v0.57.0 && go mod tidy`. x/sync, x/sys and x/term also rise here (to v0.23.0, v0.48.0, v0.46.0), the versions main already had; every x/* module now matches main. - The two example modules were re-resolved the same way. Their indirect dolthub and grpc/thrift lines move to the versions the root module already pins (they replace ../..), so their effective build is unchanged. - default.nix vendorHash recomputed for this go.sum with a nixos/nix build.
Summary
Minimal security bump of the root module, closing the two open Dependabot alerts:
github.com/apache/thrift:authority/Hostgoogle.golang.org/grpcBoth are indirect deps (via the Dolt family).
go get google.golang.org/grpc@v1.83.2 github.com/apache/thrift@v0.24.0 && go mod tidychanges 2 lines ingo.modand 4+/4- ingo.sum; no other module moves.default.nixvendorHashis recomputed for the newgo.sum. The flake's nixpkgs is unchanged: thrift 0.24.0 and grpc 1.83.2 need only go 1.25.This is the fallback path that lands the security fix independently of #7441 (Dolt driver 2.4.1 upgrade, which carries the same two bumps but is still blocked on
dolthub/file-locksfreebsd support). After this merges,git merge origin/mainon #7441's branch conflicts ingo.mod(the genprotogoogleapis/api/googleapis/rpcpins beside the grpc line both PRs bump) anddefault.nix(vendorHash);go.summerges cleanly to #7441's copy. #7441 already carries both bumps, so resolve both conflicts to #7441's side ("ours" on that branch): this PR'svendorHashdoes not match #7441'sgo.sumand would break the Nix build. (Corrected during maintainer review: this paragraph previously said #7441 picks this up with a plaingit merge origin/main.)Verification (on main
43b1fb2e4f)go build -tags gms_pure_go ./...andgo vet -tags gms_pure_go ./cmd/bd/ ./internal/...: clean.nixos/nixin Docker): fake-hash build reportsgot: sha256-IL80Qs9c7ZG17R/hRF6D+F6WVmGB0I+k60+nFuUpIJA=; with that hashnix flake check -Lpasses andnix run .#default -- --helpruns.bazel test //tools/nogo:nogo_test //scripts/...://tools/nogo:nogo_test(root/tools module alignment) PASSED, and all other//scripts/...targets PASSED except//scripts:doc_freshness_required_test, which is taggedintegration-onlyand fails without--config=integration(expected; unrelated to this change, green on main's integration lane).tools/nogo/go.modneeds no change: it shares no requirement with grpc/thrift.🤖 Generated with Claude Code