fix(ci)(interim): re-add continue-on-error to platform-build (mc#664 fix-forward in flight)
Some checks failed
Block internal-flavored paths / Block forbidden paths (pull_request) Successful in 8s
Lint curl status-code capture / Scan workflows for curl status-capture pollution (pull_request) Successful in 11s
Secret scan / Scan diff for credential-shaped strings (pull_request) Successful in 11s
CI / Detect changes (pull_request) Successful in 18s
E2E API Smoke Test / detect-changes (pull_request) Successful in 21s
E2E Staging Canvas (Playwright) / detect-changes (pull_request) Successful in 22s
Handlers Postgres Integration / detect-changes (pull_request) Successful in 22s
Runtime PR-Built Compatibility / detect-changes (pull_request) Successful in 20s
gate-check-v3 / gate-check (pull_request) Successful in 18s
qa-review / approved (pull_request) Failing after 13s
security-review / approved (pull_request) Failing after 13s
sop-tier-check / tier-check (pull_request) Successful in 17s
E2E Staging Canvas (Playwright) / Canvas tabs E2E (pull_request) Successful in 9s
E2E API Smoke Test / E2E API Smoke Test (pull_request) Successful in 10s
Handlers Postgres Integration / Handlers Postgres Integration (pull_request) Successful in 7s
Runtime PR-Built Compatibility / PR-built wheel + import smoke (pull_request) Successful in 4s
CI / Shellcheck (E2E scripts) (pull_request) Successful in 17s
audit-force-merge / audit (pull_request) Successful in 21s
CI / Python Lint & Test (pull_request) Successful in 7m20s
CI / Platform (Go) (pull_request) Failing after 8m35s
CI / Canvas (Next.js) (pull_request) Successful in 10m33s
CI / Canvas Deploy Reminder (pull_request) Has been skipped
CI / all-required (pull_request) Failing after 5s

Phase-3-masked test failures in workspace-server/internal/handlers/ surfaced
when #656 (RFC internal#219 Phase 4) flipped platform-build continue-on-error
from true to false on 0e5152c3. The pre-#656 main was masking these:

  4x delegation_test.go (lines 1110/1176/1228/1271):
      TestExecuteDelegation_DeliveryConfirmedProxyError_TreatsAsSuccess
      TestExecuteDelegation_ProxyErrorNon2xx_RemainsFailed
      TestExecuteDelegation_ProxyErrorEmptyBody_RemainsFailed
      TestExecuteDelegation_CleanProxyResponse_Unchanged
    Root cause: expectExecuteDelegationBase/Success/Failed helpers do not
    mock the DB queries production has issued since ~2026-04-21:
      - UPDATE workspaces SET last_outbound_at (commit 2f36bb9a, 2026-04-18,
        async goroutine fired from logA2ASuccess in a2a_proxy_helpers.go)
      - SELECT delivery_mode / SELECT runtime FROM workspaces (lookup* in
        a2a_proxy_helpers.go since file split in 64ccf8e1, 2026-04-21)
      - INSERT INTO activity_logs (a2a_receive) via LogActivity in
        logA2ASuccess/logA2AError (preexisting, not mocked)
      - recordLedgerStatus writes (RFC #2829 #318)
    Symptoms: sqlmock unexpected query → production short-circuits → trailing
    ExpectExec for completed/failed never fires → mock.ExpectationsWereMet()
    reports unmet remaining expectations. 8.11s uniform wall time is the
    delegationRetryDelay × 2 attempts after the first unexpected-query causes
    a transient retry path. Halt cond #3 applies (>7 days masked → broader
    sweep needed; many subsequent commits stacked on top).

  1x mcp_test.go:433 (TestMCPHandler_CommitMemory_GlobalScope_Blocked):
    Commit 7d1a189f (2026-05-10) hardened mcp.go:427 to scrub err.Error()
    from JSON-RPC error.Message (OFFSEC-001 / #259) — returning the constant
    string "tool call failed" instead. The test asserts the message contains
    "GLOBAL". Production-vs-test contract collision; needs a design call
    (revert OFFSEC scrub for this code class, or update the test to assert a
    different oracle e.g. captured logs / specific error code). Halt cond #2
    applies (alternate-class finding, not sqlmock-mismatch).

Time-boxed Option A (90 min sqlmock update) does not fit either failure class
within scope. Choosing Option B per brief: interim re-mask of platform-build
only — the other 4 #656 flips (changes, canvas-build, shellcheck, python-lint)
retain continue-on-error: false. This is a sequenced revert→fix→reflip per
feedback_strict_root_only_after_class_a emergency clause, NOT a permanent
re-mask. mc#664 stays open as the fix-then-reflip tracker.

Process note for charter SOP-N (companion to vendor-truth-review-discipline):
before flipping a job continue-on-error: true → false, do not trust the
combined-status "success" signal alone — pull the actual run log and grep
for --- FAIL / FAIL <package> to confirm the tests really pass. The masked
green on 0e5152c3 came from continue-on-error suppressing the per-job status
to neutral, which the combined-status aggregator counted as not-failure.

Cross-links:
- mc#664 (hongming-pc2 04:35Z Phase-3-masked defect filing)
- mc#656 (the flip that surfaced this; 0e5152c3 first commit to actually run
  the Go tests against internal/handlers/* since the silent stack-up began)
- feedback_strict_root_only_after_class_a (revert→fix→reflip discipline)
- feedback_return_contract_change_audit_caller_tests (mcp case applies)
- feedback_no_such_thing_as_flakes (these are real bugs, not flakes)

Evidence (run 17810 / job 33895 / task 34532 on 0e5152c3):
- 5x --- FAIL lines confirmed in actions_log/molecule-ai/molecule-core/e4/34532.log
- delegation_test.go:1110/1176/1228/1271: "unmet sqlmock expectations"
- mcp_test.go:433: "error message should mention GLOBAL, got: tool call failed"

Gitea 1.22.6 quirk #10 confirmation: per the run, job-level continue-on-error
DID still allow the combined commit-status to show neutral/success when the
job logically failed — so the #656 PR check showed green even with these
underlying failures masked. Reproduced.

Co-Authored-By: Hongming Wang <hongmingwang.rabbit@users.noreply.github.com>
This commit is contained in:
Molecule AI · core-lead 2026-05-12 04:40:32 +00:00
parent 0e5152c342
commit 9aa2b13934

View File

@ -126,8 +126,29 @@ jobs:
name: Platform (Go)
needs: changes
runs-on: ubuntu-latest
# Phase 4 (RFC #219 §1): confirmed green on main 2026-05-12.
continue-on-error: false
# mc#664 (interim): re-mask platform-build pending fix-forward. Phase 4
# (#656) flipped this to continue-on-error: false based on a Phase-3-masked
# "green on main 2026-05-12" — the prior continue-on-error: true had
# been hiding failing tests in workspace-server/internal/handlers/.
# Two distinct failure classes surfaced on 0e5152c3:
# (1) 4x delegation_test.go (lines 1110/1176/1228/1271): helpers
# expectExecuteDelegationBase/Success/Failed are missing sqlmock
# expectations for queries production has issued since ~2026-04-21
# (last_outbound_at UPDATE, lookupDeliveryMode/Runtime SELECTs,
# a2a_receive INSERT activity_logs, recordLedgerStatus writes).
# Halt cond #3 applies (regression > 7 days → broader sweep).
# (2) 1x mcp_test.go:433 (TestMCPHandler_CommitMemory_GlobalScope_Blocked):
# commit 7d1a189f (2026-05-10) hardened mcp.go to scrub err.Error()
# from JSON-RPC responses (OFFSEC-001), but the test asserts the
# error message contains "GLOBAL". Production-vs-test contract
# collision — needs design call, not mock update.
# Time-boxed Option A (90 min) did not fit the cross-cutting scope.
# This is a sequenced revert→fix→reflip per
# feedback_strict_root_only_after_class_a emergency clause — NOT
# a permanent re-mask. Re-flip blocked on mc#664 fix-forward landing.
# Other 4 #656 flips (changes, canvas-build, shellcheck, python-lint)
# retain continue-on-error: false; only platform-build regresses.
continue-on-error: true # mc#664 fix-forward in flight; re-flip when tests pass
defaults:
run:
working-directory: workspace-server