[SDK Sentinel] Synchronize alternate ordering test with history

[sdk-sentinel-bot]

Caution

This PR contains untrusted AI-generated code. Do not approve or run CI until a maintainer has reviewed the diff. SDK Sentinel verified that GitHub Actions remained approval-gated with zero executable jobs when this PR was opened.

Summary

• CI evidence: failed CI job · workflow run · last known pass · first known failure

Prevent the alternate async-loop ordering test from intermittently ending with an activity ScheduleToStart timeout. The failure occurred in the [first failing main Windows job](https://github.com/temporalio/sdk-python/actions/runs/38939898149/job/117104613026) and was corroborated by Linux and Windows PR jobs. The [preceding same-matrix normal suite](https://github.com/temporalio/sdk-python/actions/runs/38772365734/job/116551198300) passed.

Root cause

The test used fixed sleeps as proof that its workflow task and separate-queue activity had progressed. Worker startup and polling are asynchronous, so loaded runners could advance the test before those milestones.

Fix

Wait for activity-scheduled history before stopping the first workflow worker, then wait for activity-completed history before sending signals and restarting workflow processing. This preserves the intended history ordering without elapsed-time assumptions.

Validation

The unchanged test passed 20/20 locally, so the flake did not naturally reproduce. The candidate passed 20/20, passed once with time-skipping V2, and passed the repository validator. Windows Python 3.14 CI remains required because that environment produced the primary failure.

Validation status: ci-validation-required

• Flake confidence: high
• Fix confidence: high
• Value: high
• Patch scope: standard — 1 changed file; 1.4 KiB
• Local reproduction: not-reproduced — The unchanged focused test passed 20/20 on Linux Python 3.14.
• Regression coverage: pass-after-only — The existing test now uses deterministic history milestones and passed 20 stress iterations plus time-skipping V2.
• Unchanged baseline (Poe lint, bridge lint, and editable build) — passed
• Independent candidate (Poe lint, bridge lint, and editable build) — passed

Investigator-run checks

• poe test -s -k test_alternate_async_loop_ordering --flake-finder --flake-runs=20 — passed; 1 attempt(s). Unchanged checkout passed 20/20; natural reproduction did not occur.
• poe test -s -k test_alternate_async_loop_ordering --flake-finder --flake-runs=20 — passed; 1 attempt(s). Candidate passed 20/20.
• poe test -s --workflow-environment time-skipping-v2 -k test_alternate_async_loop_ordering — passed; 1 attempt(s). Candidate passed in time-skipping V2 mode.
• .ci-flake-runtime/input/validate.sh — passed; 1 attempt(s). Trusted unchanged baseline validation passed.
• .ci-flake-runtime/input/validate.sh — passed; 1 attempt(s). Candidate lint, bridge lint, and editable bridge build passed.

Required target CI

• build-lint-test (3.14, windows-latest), normal workflow environment

Residual risks

• The exact Windows runner failure was not reproduced locally.
• The candidate has not yet been exercised under Windows CI load.

Automation provenance

• Investigator workflow
• Base SHA: da794611dbfc746ca27c65385d00d9089de5c57b

If this finding should not be fixed, apply the sdk-sentinel:false-positive label and close the PR. Sentinel will suppress the finding until its affected source changes.

SDK Sentinel recurrence history

• Unique failed CI run attempts: 3
• First occurrence: 2026-10-16 17:44 UTC
• Latest occurrence: 2026-10-16 23:19 UTC
• Recent occurrences:
  • 2026-10-16 23:19 UTC — failed CI job
  • 2026-10-16 22:47 UTC — failed CI job
  • 2026-10-16 17:44 UTC — failed CI job