beads/release-gates/be-al8j-dolt-sql-argv-e2big-gate.md
Jim Wordelman 4274815c6c
Fix E2BIG when runDoltSQL passes large schema SQL as a single argv (#5846)
* test(dolt): red — Fix: runDoltSQL passes 134KB of schema SQL in one argv, past Linux MAX_ARG_STRLEN (E2BIG) - 10 dolt tests red (refs be-go6d)

* feat(dolt): green — Fix: runDoltSQL passes 134KB of schema SQL in one argv, past Linux MAX_ARG_STRLEN (E2BIG) - 10 dolt tests red (refs be-go6d)

* test(dolt): wire TestRunDoltSQLHandlesLargeScript into the default lane (refs be-wq49)

Reviewer follow-up on be-go6d (runDoltSQL argv E2BIG fix): the regression
test proving it lived in git_remote_test.go under //go:build integration,
gated by skipIfNoDolt — which checks for the shared containerized Dolt SQL
server (testServerPort). That test never touches the server, only the
local dolt CLI, so the gate was wrong and the test only ever ran on the
non-blocking nightly cron, never in the PR-gating/merge-gating lane.

Extract TestRunDoltSQLHandlesLargeScript + its runDoltSQL helper into a
new file, dolt_sql_large_script_test.go, with no tag and gated on
testutil.RequireDoltBinary(t) instead. Remove the originals from
git_remote_test.go (kept as a same-package call site — setupGitRemote
still calls runDoltSQL) to avoid duplicate symbols when both files
compile together under -tags=integration.

No production code changes: runDoltSQL's fix was already proven correct
in be-go6d's own tdd_red/tdd_green (cc407e186/fceb1c69a). This closes the
CI-wiring gap only.

Verified: test absent from the default lane before (mechanically-verified
red state via a targeted -run + -v check), present and passing after;
default and integration-tagged compiles both stay duplicate-symbol free;
full ./internal/storage/dolt/... default-lane suite still green with no
regression.

* test(feat): red — Fix: runDoltSQL passes 134KB of schema SQL in one argv, past Linux MAX_ARG_STRLEN (E2BIG) - 10 dolt tests red (refs be-wq49)

TestRunDoltSQLHandlesLargeScript now calls testutil.RequireDoltCLIOnly,
which does not exist yet. This fails the build (RED), proving the test
is not currently gated correctly.

Root cause: RequireDoltBinary conflates "needs containerized
dolt-sql-server" with "needs local dolt CLI only" by honoring
BEADS_TEST_SKIP=dolt before checking for the CLI binary. That env var
is unconditionally set by scripts/ci/lib/test-env.sh's
beads_test_env_enter(), so this test silently SKIPs instead of
running PASS/FAIL in the required pr-core-wrapper CI job — even
though it has no containerized-server dependency.

* feat: green — Fix: runDoltSQL passes 134KB of schema SQL in one argv, past Linux MAX_ARG_STRLEN (E2BIG) - 10 dolt tests red (refs be-wq49)

Add RequireDoltCLIOnly, which checks for the dolt CLI binary but does
NOT honor BEADS_TEST_SKIP=dolt. Extract the shared binary-presence
check into requireDoltBinaryPresent so RequireDoltBinary's existing
behavior (skip on BEADS_TEST_SKIP=dolt, fail under GITHUB_ACTIONS if
missing) is unchanged for its other 20 call sites.

TestRunDoltSQLHandlesLargeScript now gates on RequireDoltCLIOnly
instead of RequireDoltBinary, so it actually runs PASS/FAIL in the
required pr-core-wrapper CI job instead of silently SKIPping — it has
no dependency on the shared containerized dolt-sql-server that
BEADS_TEST_SKIP=dolt exists to exclude.

Verified under bash (matching scripts/ci/pr-core.sh's real shebang
and trap semantics — sourcing test-env.sh into an interactive zsh
session instead triggers zsh's function-local EXIT-trap behavior,
which tears the hermetic env down early; not a bug in the script):

  source .buildflags && source scripts/ci/lib/test-env.sh && \
    beads_test_env_enter && \
    go test -run ^TestRunDoltSQLHandlesLargeScript$ -v ./internal/storage/dolt/...

  --- PASS: TestRunDoltSQLHandlesLargeScript (1.03s)

* release-gate: PASS record for runDoltSQL E2BIG fix (deploy/be-al8j-gate)
2026-08-19 08:51:07 -07:00

8.1 KiB

Release gate — Fix: runDoltSQL passes 134KB of schema SQL in one argv, past Linux MAX_ARG_STRLEN (E2BIG)

  • Build beads (CLOSED): be-go6d (round 1 fix), be-wq49 (follow-up: wire TestRunDoltSQLHandlesLargeScript into the default lane)
  • Deploy bead: be-al8j
  • Review bead: be-d3zx — round 3, verdict PASS, recorded on commit 1cde0f0931025ea3783ffd757a444f3ff88084a0
  • Commits: 5 commits over origin/main — 9cc80debf (red, be-go6d), 876b58cf8 (green, be-go6d), ed302a5f3 (wire into default lane, be-wq49), e6fd799c6 (red, be-wq49), 1cde0f093 (green, be-wq49 — the reviewed/deploy SHA)
  • Branch: builder/be-go6d (provenance only, not a push target) — deploy cut to deploy/be-al8j-gate, pushed to fork (quad341/beads-sec003-contrib)
  • Evaluated: 2026-08-19 by beads/deployer

Scope

runDoltSQL (test helper that shells out to dolt sql -q <query>) was passing large generated schema SQL as a single CLI argument, exceeding Linux's MAX_ARG_STRLEN and failing with E2BIG — breaking 10 dolt-backed tests. The fix, plus the follow-up making the regression test itself run in the default (untagged) lane instead of only under -tags=integration.

Diff scope, confirmed directly via git diff --name-only origin/main...HEAD (3 files) and independently reviewed file-by-file, not just diffstat'd:

  • internal/testutil/testdoltcommon.go — extracts requireDoltBinaryPresent as a shared helper; adds RequireDoltCLIOnly (skips the BEADS_TEST_SKIP=dolt check for tests that only need the local dolt binary, not the containerized SQL server); RequireDoltBinary's external behavior unchanged.
  • internal/storage/dolt/dolt_sql_large_script_test.go (new) — the regression test, TestRunDoltSQLHandlesLargeScript, plus a locally-scoped runDoltSQL fixed to avoid the argv-length fault.
  • internal/storage/dolt/git_remote_test.go — not called out in the reviewer's style_findings list, independently caught and verified here. Removes the old runDoltSQL (identical fix content) from this integration-tagged file, replaced with a comment pointing to its new home. Pure relocation, zero net logic change — necessary so the new default-lane test can call it without requiring the integration tag. Same single theme as the other two files.

assert_deploy_ancestry_scope origin/main 1cde0f0931025ea3783ffd757a444f3ff88084a0 be-wq49 be-go6d run as a mechanical second check on top of the manual diff read: rc=0 — zero .claude/** paths introduced, every one of the 5 commits cites be-wq49 or be-go6d.

Gate criteria

# Criterion Verdict Evidence
0 Pre-flight: not already merged PASS gh api repos/gastownhall/beads/commits/1cde0f0931025ea3783ffd757a444f3ff88084a0/pulls → [], no existing PR contains this commit.
6 Clean divergence from origin/main PASS git merge-base --is-ancestor origin/main <deploy-sha> confirms origin/main is a strict ancestor; git rev-list --left-right --count origin/main...<deploy-sha> → 0 5 (0 behind, 5 ahead). Trivial fast-forward, no self-rebase needed.
1 Review PASS present PASS be-d3zx closed, verdict: pass, round 3, on this exact commit.
2 Acceptance criteria met PASS Reviewer's round-3 exit_contract (5 criteria) independently re-verified below, plus full-suite re-run beyond what the reviewer scoped to.
3 Tests pass PASS See "Tests run" below — diff-owned test independently re-run, plus a full-repo make test sweep: 94/94 packages ok, 0 FAIL.
3a Non-diff-owned failures correctly attributed N/A No test failures occurred at all (diff-owned or otherwise) — this clause has nothing to attribute. Applies to the lint/policy lane instead; see criterion 3b.
3b Policy/lint lane PASS (non-diff-owned failures present, correctly attributed) ci-pr-lint and ci-pr-policy both exit nonzero, but both failures trace to files with zero diff against origin/main — see below. Diff-owned lint/policy surface is clean.
4 No unresolved HIGH findings PASS Reviewer: zero HIGH/MEDIUM findings, diff is pure test-infrastructure with no new dependencies or injection surface. The only findings anywhere in scope for this gate (3 gosec G602) are non-diff-owned, see below.
5 Clean working tree PASS git status at the deploy SHA shows no staged/unstaged changes on tracked files; only the pre-existing, unrelated untracked scratch files already documented in this worktree (scripts/rebase-resolve-lib.sh, prior gate leftovers), never staged.
7 Single feature theme PASS All 3 files serve exactly one purpose — see Scope above. Confirmed two independent ways: manual file-by-file diff read, and the mechanical assert_deploy_ancestry_scope guard (rc=0).

Tests run on release branch (independent re-verification)

Diff-owned test, run standalone under the real harness (matching the reviewer's own documented methodology, bash not zsh):

source .buildflags && source scripts/ci/lib/test-env.sh && beads_test_env_enter && \
  go test -run '^TestRunDoltSQLHandlesLargeScript$' -v ./internal/storage/dolt/...
Test Result
TestRunDoltSQLHandlesLargeScript PASS (matches reviewer's independent PASS)

Full-repo canonical CI-equivalent (make test = TEST_COVER=1 ./scripts/test.sh, the Makefile test: target), run against the checked-out deploy SHA with DOCKER_HOST=unix:///run/user/1000/podman/podman.sock and TESTCONTAINERS_RYUK_DISABLED=true (cached dolthub/dolt-sql-server:2.2.0 image confirmed present beforehand, no pull needed):

  • 94/94 packages ok, 0 FAIL. Skipping: printed empty — no BEADS_TEST_SKIP override applied, this is the real default lane.
  • internal/storage/dolt (the diff-owned package): ok 10.519s.
  • internal/testutil (the other diff-owned package): ok 0.117s.
  • Total coverage 38.5%, no regressions observed anywhere in the sweep.

This exceeds the reviewer's own scope (package-level regression only); a full-repo sweep found nothing the reviewer's narrower run wouldn't have caught, and adds confidence beyond it.

Non-diff-owned lint/policy failures (criterion 3b)

make ci-pr-lint → golangci-lint reports 3 gosec G602 (slice index out of range) findings, all in backend/conformance/{importer_contract,relations_contract}.go. make ci-pr-policy → version-consistency check reports .githooks/commit-msg missing BEGIN/END BEADS INTEGRATION markers.

Both attributed, not waved through:

  • git diff --stat origin/main..HEAD -- backend/conformance/ and -- .githooks/commit-msg are both empty — neither path appears anywhere in this diff at all.
  • git blame on the flagged importer_contract.go lines: last touched 2026-08-09 by Julian Knutsen (d38cd9d435). relations_contract.go: last touched 2026-08-04, same author (d801ec43dc). Both predate this diff's commit chain (2026-08-18/19) entirely.
  • Neither file shares any import or symbol with the two diff-owned packages (internal/storage/dolt, internal/testutil) — a pure test-infra argv fix in one package cannot plausibly have introduced a slice-bounds issue or a githook marker gap in unrelated files.

Both are real, pre-existing repo-wide issues, worth a follow-up bead, but neither is owned by or attributable to this diff. Not filed by the deployer directly (outside gate scope) — flagging to the operator/mayor in the close-out instead.

Findings from reviews (no action required)

From be-d3zx (round 3): no HIGH or MEDIUM findings. Security: pure test-infrastructure diff, zero production code, zero new dependencies; requireDoltBinaryPresent/RequireDoltCLIOnly only call exec.LookPath and os.Getenv — no injection surface.

Verdict

PASS — all 7 applicable criteria clear. Proceeding to cut deploy/be-al8j-gate from 1cde0f0931025ea3783ffd757a444f3ff88084a0, push to fork, and open a contributor PR against gastownhall/beads. Per the merge-authority carve-out for repos this deployer role does not maintain, the job ends at the opened PR — no merge-request, no deploy-clearance status.