beads/release-gates/be-3b4e-gate.md
Jim Wordelman ac8086429e
fix(cmd/bd): label created_by as 'Created by:' instead of 'Owner:' (#5800)
* test(bd): red — bd show text view labels created_by as 'Owner:' while --json 'owner' is a different field with a different value (refs be-ss66)

* fix(cmd/bd): label created_by as 'Created by:' instead of 'Owner:'

The text view rendered issue.CreatedBy under an "Owner:" label, but a
distinct Owner field (a git author email) also exists on types.Issue
with a different value. Relabel to avoid conflating the two.

be-ss66

* docs(release-gates): add be-3b4e gate record — PASS

bd show text view labels created_by as "Owner:" while --json exposes a
distinct owner field. Review be-r0za passed; independently re-verified
diff-owned test, full-suite + isolated + origin/main-baseline test runs,
and the ci-pr-policy lane. Found and losslessly corrected an unrelated
automated commit that had landed on this deploy branch.

Agent-Signature: claude-code-claude-sonnet-5-unknown-reasoning on behalf of Jim Wordelman

* docs(changelog): record the created_by label change under [Unreleased]

Review ask on PR #5800: this changes a label every `bd show` reader sees,
and the repo logs smaller output changes than this.

refs be-ss66, be-xm0ly

* docs(release-gates): note the be-3b4e gate describes 092947695, not the head

The gate's branch-integrity section claimed the branch held "exactly the two
reviewed commits over origin/main". The review round for PR #5800 added a
CHANGELOG entry and a main merge, so that line no longer describes the head.

Corrected by annotation rather than by editing the verdict: the PASS is still
accurate about the SHA it examined, and no post-gate commit touches
production code.

refs be-xm0ly

---------

Co-authored-by: Jim Wordelman <james@wordelman.name>
2026-08-20 05:31:58 -07:00

12 KiB
Raw Permalink Blame History

Release Gate: be-3b4e — bd show text view labels created_by as "Owner:"

  • Deploy bead: be-3b4e
  • Review bead: be-r0za (closed, verdict: pass)
  • Builder branch (provenance only): builder/be-ss66
  • Repo: gastownhall/beads
  • Deploy commit: 0929476954e3bec6516d12d1e8e792bd9375209a
  • Deploy branch: deploy/be-3b4e-gate (cut from the commit above)
  • Base: origin/main @ 7505e173f2659ba6e1f955b86d81a4f9e21810ca
  • Evaluated: 2026-08-15 by beads/deployer

Scope

bd show's text view labeled the created_by field as "Owner:", which is misleading — --json exposes a distinct owner field with a different value. The fix relabels the text-view line to "Created by:" so it no longer implies the two fields are the same thing.

Diff, confirmed via git diff --name-only origin/main...HEAD (2 files):

  • cmd/bd/show_format.go — relabels Owner: %s → Created by: %s in formatIssueMetadata; updates the adjacent comment.
  • cmd/bd/show_format_metadata_test.go — adds TestFormatIssueMetadata_CreatedByLabel, asserting "Created by: alice" is present, "Assignee: carol" is present, and "Owner: " is absent, given a fixture with distinct CreatedBy/Owner/Assignee values — directly proving the two fields are not conflated.

Repo-wide grep for other "Owner:" call sites found exactly one other hit, cmd/bd/github.go:313, which formats an unrelated GitHub-repo-config field — correctly out of scope for this fix.

Gate criteria

# Criterion Verdict Evidence
6 Clean divergence from origin/main (checked first) PASS git log --oneline origin/main..HEAD shows exactly the 2 expected commits, 0 behind. assert_deploy_ancestry_scope origin/main HEAD be-3b4e be-ss66 → rc=0 (widened to include be-ss66, the builder bead both commits cite — matches be-3b4e's own documented provenance, branch: builder/be-ss66).
1 Review PASS present PASS be-r0za status=closed, verdict: pass; deploy_bead: be-3b4e / deploy_commit: 0929476954e3bec6516d12d1e8e792bd9375209a match exactly.
2 Acceptance criteria met PASS Diff matches the bead's title/description exactly: a single mislabeled string, fixed, with a regression test directly guarding the described bug. uncovered_criteria: none (independently re-confirmed via the repo-wide "Owner:" grep above).
3 Tests pass PASS See "Tests run" below — diff-owned test passes; all FAILs independently confirmed pre-existing/environmental, none touch the diff's 2 files.
3b Policy/lint lane (make ci-pr-policy) PASS See "Policy lane" below.
4 No unresolved HIGH findings PASS be-r0za's OWASP-style walkthrough (9 categories) is clean; style findings (gofmt/go vet/golangci-lint) clean. Diff is a 2-line string-literal relabel in a formatting function — trivially low risk, consistent with that assessment.
5 Clean branch PASS git status --porcelain=v1 on the deploy branch is empty aside from 3 pre-existing untracked files belonging to other beads' sessions (release-gates/be-hi97-no-workspace-tests-gate.md, release-gates/be-uoat-dolt-diff-export-gate.md, scripts/rebase-resolve-lib.sh) — not staged, not part of this branch's history, deliberately left untouched (see "Branch-integrity incident" below).
7 Single feature theme PASS 2 files, 1 production fix + 1 direct regression test, one bug, one label.

Branch-integrity incident: unrelated commit found and removed

Before writing this record, deploy/be-3b4e-gate's tip (28c259392) did not match the expected deploy commit. Investigation (git reflog show deploy/be-3b4e-gate) showed the branch had originally been cut correctly at 0929476954e3bec6516d12d1e8e792bd9375209a, then had a second commit appended on top:

28c259392 witness: salvage uncommitted work (be-3b4e)
092947695 fix(cmd/bd): label created_by as 'Created by:' instead of 'Owner:'

This "witness" commit was not created by any action taken while evaluating this gate. Its content (git show --stat) was exactly the three untracked files noted above — a git add -A-style sweep, almost certainly an automated session/worktree safety net reacting to a session boundary in this shared worktree, with no knowledge of which bead each file belonged to. Left in place, it would have carried two other sessions' internal gate records (including one with a FAIL verdict and detailed internal investigation notes) into a public upstream PR.

Confirmed no other ref pointed at 28c259392 (git log --all --oneline), then corrected with a plain git reset (mixed — index only, no --hard, no working-tree content touched) back to the reviewed SHA:

git reset 0929476954e3bec6516d12d1e8e792bd9375209a

This is lossless: all three files remain intact on disk as untracked worktree state, exactly as they were before the witness commit fired, for their owning sessions to commit under their own beads. deploy/be-3b4e-gate now has exactly the two reviewed commits over origin/main, re-confirmed above under criterion 6.

Tests run

Independently re-verified rather than trusted from be-r0za's report, with the diff-owned test resolved by name, not inferred from counts.

Diff-owned test — TestFormatIssueMetadata_CreatedByLabel: PASS, confirmed via targeted verbose run and via absence from every FAIL list below. be-r0za independently confirms the same result via its own targeted verbose run.

Full suite (scripts/ci/pr-core.sh-equivalent, -race -short -skip '^TestEmbedded' -p 4 -parallel 4, default 10m per-package timeout): 88 ok / 5 FAIL packages — cmd/bd (timed out: panic: test timed out after 10m0s, 600.560s), internal/beads, internal/config, internal/formula, internal/metrics. 0 SKIP. This is one package worse than be-r0za's own baseline (4 FAIL: cmd/bd, internal/beads, internal/config, internal/formula) — both discrepancies (the cmd/bd timeout and the extra internal/metrics FAIL) were investigated to a specific, confirmed root cause rather than waived:

  • System load at time of run: uptime showed load average 42–52 on this shared multi-agent host (multiple concurrent dolt sql-server processes, concurrent go test/go build for unrelated packages, other active agent sessions) — the immediate suspect for both anomalies.
  • cmd/bd timeout: re-ran in isolation (./cmd/bd/... alone, -timeout 30m, no other change). Completed cleanly, no timeout, no panic — directly confirms transient resource contention, not a diff-caused hang. This isolated run surfaced 72 individual --- FAIL: tests; every one cross-checked by name against be-r0za's documented pre-existing failure categories (workspace/worktree, config, migration, backup, hooks-install, doctor-health, init/bootstrap, recipe-loading) — full match, zero overlap with show_format.go, including the same representative test (TestWhereNoWorkspace) be-r0za used for its own baseline proof.
  • internal/metrics: the full-suite log's apparent failing test names for this package (e.g. TestMigrateToServer_*, TestReset_Worktree*) turned out to be my own extraction artifact — go test -p 4's concurrent, interleaved output made naive line-based attribution unreliable. Re-ran internal/metrics alone in isolation: clean ok, 0 failures. Not a real failure, diff-unrelated regardless.
  • cmd/bd/doctor flip-flop: passed in the full-suite run above (ok 5.489s) but failed in the isolated cmd/bd retest (FAIL 5.452s), the inverse pattern from internal/metrics. Resolved by direct origin/main baseline reproduction (see below) rather than assumption — the pass/fail flip between two runs of identical code is itself evidence of environment-sensitivity, not a deterministic diff-caused regression.

Origin/main baseline reproduction (own reproduction, mirroring be-r0za's methodology — git worktree add --detach <scratch> origin/main, detached HEAD at 7505e173f, zero diff applied, same TMPDIR/GOTMPDIR workaround):

  • TestWhereNoWorkspace (representative test, matching be-r0za's own baseline-reproduction proof): FAILS identically on origin/main with zero diff.
  • cmd/bd/doctor full package: FAILS identically on origin/main with zero diff — FAIL 4.829s, with the same 6 test names as the isolated retest above (TestRunDoltHealthChecks_DoltBackendNoServer, TestCheckFederationRemotesAPI_ServerNotRunning, TestCheckFreshClone_ServerModeUnreachable, TestCheckRepoFingerprint_UsesTargetRepoOutsideCWD, TestCheckBeadsRole_NotConfigured, TestCheckBeadsRole_NotGitRepo). Test names (NoServer / ServerModeUnreachable / NotConfigured) indicate sensitivity to ambient Dolt-server/config state, consistent with the contention diagnosis above. This directly proves the flip-flop is pre-existing environmental flakiness, not caused by this diff.

Scratch worktree removed after use (git worktree remove).

Conclusion: diff-owned test passes; every FAIL package/test is either independently reproduced against a zero-diff origin/main baseline, or directly attributable to shared-host contention/output-interleaving with a confirmed root cause, and none touches either of the diff's 2 files.

Policy lane (make ci-pr-policy)

Initial run failed on "check version consistency": .githooks/commit-msg: no 'BEGIN/END BEADS INTEGRATION' marker found. Root-caused, not waived: .githooks/commit-msg is an untracked gc-rig session shim ("rewritten on every session start"; confirmed absent from origin/main via git ls-tree / git cat-file -e), unrelated to the beads project. scripts/check-versions.sh globs .githooks/* on the raw filesystem rather than git ls-tree, so it picks up the shim.

Remediated reversibly: moved the shim to scratch, re-ran the check (clean, rc=0), restored the shim immediately after, confirmed restoration (executable, correct size). git status --porcelain=v1 showed no stray changes before or after. Full make ci-pr-policy (check-build-tags, check-go-install-guidance, check-versions, doc-flags/doc-freshness, check-testing-short, workapi/frontend boundary check, no-beads-jsonl-changes, make api-check): clean.

Findings from review (no action required)

From be-r0za: no HIGH or MEDIUM findings across security (9 OWASP-style categories, all clear) or style (gofmt/go vet/golangci-lint clean).

Result: PASS

All 7 criteria plus the 3b policy/lint lane pass, independently re-verified rather than trusted from be-r0za's report. Two genuine discrepancies between my results and the reviewer's baseline (internal/metrics, cmd/bd/doctor) were investigated to specific, confirmed root causes rather than asserted away. One branch-integrity incident (an unrelated automated commit) was found and losslessly corrected before push.

Merge authority: gastownhall/beads is a contributor-only repo (we do not maintain it — origin remote push is disabled, DISABLED-upstream-is-fetch-only-push-to-fork-and-PR). Deployer's job ends at the opened PR; no release-gate/deploy-clearance status posted, no merge-request routed to mayor. Upstream maintainers own the merge.

Post-gate history

The criteria walk above examined 0929476954e3bec6516d12d1e8e792bd9375209a. Commits landed on deploy/be-3b4e-gate after that SHA, in order:

Commit What
f0e8b8f7d This gate document.
5f5767c71 docs(changelog): record the created_by label change under [Unreleased] — the one blocking ask from the PR #5800 review.
fd6286ec0 Merge origin/main. The new CHANGELOG entry collided with two entries main had prepended to the same [Unreleased] / ### Changed list; resolved by keeping all three. Without it the PR was no longer cleanly mergeable.
this commit This annotation.

None of them touches production code, so the criteria walk still describes the shipped diff (cmd/bd/show_format.go plus cmd/bd/show_format_metadata_test.go) exactly, and the PASS stands for the SHA it examined. This section exists so the "exactly the two reviewed commits over origin/main" line in the branch-integrity section above is not read as a claim about the current head.