[PR #5409] [MERGED] ci(windows): run environment registration regressions #6921

New issue

Closed

opened 2026-11-01 21:44:04 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-11-01 21:44:04 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/morluto/rea/pull/5409
Author: @gensokyo-dev
Created: 10/20/2026
Status: ✅ Merged
Merged: 10/20/2026
Merged by: @gensokyo-dev

Base: main ← Head: audit/ci-tests-20261020-round4

📝 Commits (2)

fdabb50 ci(windows): run environment registration regressions
282e5e5 ci(windows): select registration implementation owners

📊 Changes

4 files changed (+67 additions, -0 deletions)

View changed files

📝 .github/workflows/ci.yml (+3 -0)
📝 scripts/ci/scopes.mjs (+6 -0)
📝 tests/boundary/filesystem/ciChangeScope.test.ts (+48 -0)
📝 tests/process-global/curatedWindowsLane.test.ts (+10 -0)

📄 Description

Summary

The environment-registration suites contain 13 Windows-only regressions that the portable Linux suite skips. The Windows curated invocation omitted their three owning files, and changes to either their tests or key implementations did not select that lane. Include those existing suites and assign their test paths plus the ClientConfiguration*, ClientRegistration* and DshCordis* production owners to the established package/native-host scope.

Add one required-membership guard and twelve actual Git-diff classification cases covering the three test files and all nine current production files. Existing assertions, deadlines, retries and platform conditions remain intact. This is CI/test maintenance within existing scope; no production contract or provider authority changes. Found during a latest-main audit; no separate issue.

Executed validation

Baseline main: e84468c0.

Before the workflow/scope fix, the new controls produced 4 failed / 242 passed: missing Windows membership and package=false/runtime=false for each test-only path.
Production-owner controls reproduced the review finding on initial PR head fdabb5073: 9 failed, with package=false.
Final CI guard run: 255 passed across both changed test files, including all nine production-owner controls and the original three test-only controls.
The unchanged runtime suites and discovery controls were included in the preceding six-file Linux run: 301 passed / 13 Windows-only skipped. The follow-up modifies only scope data and CI classification tests; runtime code and these suites are identical. This is not Windows runtime qualification.
npm run check: all six static tasks plus module/architecture boundaries passed.
npm run verify:test-discovery: 1,782 files exactly once.
git diff --check: passed.

Final head 282e5e521: CI 39404203264 completed successfully with 38/38 jobs passed, including all five portable shards, Portable tests, CI required, Windows and every selected native/provider lane. CI executed merge preview 9f8fb1a8c with main e84468c0.

The hosted Windows job passed all three added files: 34 + 10 + 16 = 60 passed, no skips, including all 13 Windows-only regressions. The complete curated step passed 43 files / 598 tests, with 31 existing skips in other suites; subsequent Node 22/native/package verification also passed.

Post-merge integration: normal squash merge aa314ab682 also includes upstream #5390 and #5415, which landed while PR CI was running; its complete tree therefore differs from the PR merge preview. Main CI 39406988130 checked out this exact merge commit and passed 38/38 jobs. Its Windows job again passed all 60 target tests with zero skips, including the 13 Windows-only regressions. This qualifies the integrated merge commit in addition to the PR head.

Codex review completed with no findings on final head; Greptile scored 5/5. The production-owner finding was fixed with the three owner families and nine fail-before/pass-after controls. The existing native artifact, capability prerequisites, complete portable suite and required report gates are retained. No real external DSH/MCP client or Hopper/Ghidra engine claim is made.

Measured resource cost

Pinned Node 24.18.0/npm 11.16.0, Linux, two CPUs; heavy commands serialized. The focused run reused a warm runtime build and performed fresh tests. Locked installed dependencies were reused only after an identical lock digest and the dependency-install validator passed.

Final completed workload	Wall s	Family CPU s	Sampled family RSS MiB
Six-file runtime/CI discovery run	110.39	161.29	1,061
Final two CI guard files	36.37	41.63	654
Final static/architecture checks	91.19	169.89	4,855
Final exact-once discovery	0.92	1.04	287

RSS sums observed descendants every 100 ms, may double-count shared pages or miss short/reparented children, and is not an aggregate hard cap. No observed OOM. Baseline controls included a cold 70.85 s compilation and failed, so their total is not an equivalent performance comparison. Added Windows coverage intentionally adds work; no speedup claim. Baseline Windows curated step took 74 s; the qualified final head took 126 s with the added 60 cases. These single observations run different coverage and do not isolate host/cache variance or establish a speedup. The initial head recorded 91 s, demonstrating that host/run variation also matters. Host variance, agent token usage and CI-wide cost are not measured. LeanToken retrieval failed internally; inspection used the pinned local checkout.

Author readiness

Human owner: @gensokyo-dev

The owner explicitly delegated these declarations and authorized opening ready, following CI/review and normal merge in this session. Codex reviewed the final diff and evidence; independent human review was not observed.

I reviewed the final diff and relevant existing code, understand the change and its edge cases, and take responsibility for the complete PR, including AI-assisted work.
I checked the evidence above, confirmed it supports the current change, and recorded meaningful validation gaps.
I declare this PR ready for initial maintainer review.

Before merge

This PR addresses a concrete problem or an agreed enhancement.
Material scope and authority questions are resolved, and any required maintainer scope decision covers the final diff.
The PR has one focused outcome and the title follows type(scope): outcome.
Related issue is linked, or the reason for not linking one is stated above.
Tests cover changed observable behavior and meaningful failure paths.
Owning docs, contracts, and generated metadata are updated where needed.
User-visible CLI/MCP changes include representative output.
I checked the final diff for secrets, unrelated cleanup, and unsupported claims.

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/morluto/rea/pull/5409 **Author:** [@gensokyo-dev](https://github.com/gensokyo-dev) **Created:** 10/20/2026 **Status:** ✅ Merged **Merged:** 10/20/2026 **Merged by:** [@gensokyo-dev](https://github.com/gensokyo-dev) **Base:** `main` ← **Head:** `audit/ci-tests-20261020-round4` --- ### 📝 Commits (2) - [`fdabb50`](https://github.com/morluto/rea/commit/fdabb50730505c8968603e3bd1542a0bc2c2e40a) ci(windows): run environment registration regressions - [`282e5e5`](https://github.com/morluto/rea/commit/282e5e521a0abaa4ce134b37d2edfdf158127838) ci(windows): select registration implementation owners ### 📊 Changes **4 files changed** (+67 additions, -0 deletions) <details> <summary>View changed files</summary> 📝 `.github/workflows/ci.yml` (+3 -0) 📝 `scripts/ci/scopes.mjs` (+6 -0) 📝 `tests/boundary/filesystem/ciChangeScope.test.ts` (+48 -0) 📝 `tests/process-global/curatedWindowsLane.test.ts` (+10 -0) </details> ### 📄 Description ## Summary The environment-registration suites contain **13 Windows-only regressions** that the portable Linux suite skips. The Windows curated invocation omitted their three owning files, and changes to either their tests or key implementations did not select that lane. Include those existing suites and assign their test paths plus the `ClientConfiguration*`, `ClientRegistration*` and `DshCordis*` production owners to the established package/native-host scope. Add one required-membership guard and twelve actual Git-diff classification cases covering the three test files and all nine current production files. Existing assertions, deadlines, retries and platform conditions remain intact. This is CI/test maintenance within existing scope; no production contract or provider authority changes. Found during a latest-main audit; no separate issue. ## Executed validation Baseline main: `e84468c0`. - Before the workflow/scope fix, the new controls produced **4 failed / 242 passed**: missing Windows membership and `package=false`/`runtime=false` for each test-only path. - Production-owner controls reproduced the review finding on initial PR head `fdabb5073`: **9 failed**, with `package=false`. - Final CI guard run: **255 passed** across both changed test files, including all nine production-owner controls and the original three test-only controls. - The unchanged runtime suites and discovery controls were included in the preceding six-file Linux run: **301 passed / 13 Windows-only skipped**. The follow-up modifies only scope data and CI classification tests; runtime code and these suites are identical. This is not Windows runtime qualification. - `npm run check`: all six static tasks plus module/architecture boundaries passed. - `npm run verify:test-discovery`: **1,782 files exactly once**. - `git diff --check`: passed. Final head `282e5e521`: [CI 39404203264](https://github.com/morluto/rea/actions/runs/39404203264) completed successfully with **38/38 jobs passed**, including all five portable shards, `Portable tests`, `CI required`, Windows and every selected native/provider lane. CI executed merge preview `9f8fb1a8c` with main `e84468c0`. The [hosted Windows job](https://github.com/morluto/rea/actions/runs/39404203264/job/118569637178) passed all three added files: **34 + 10 + 16 = 60 passed, no skips**, including all **13 Windows-only regressions**. The complete curated step passed 43 files / 598 tests, with 31 existing skips in other suites; subsequent Node 22/native/package verification also passed. Post-merge integration: normal squash merge `aa314ab682` also includes upstream #5390 and #5415, which landed while PR CI was running; its complete tree therefore differs from the PR merge preview. [Main CI 39406988130](https://github.com/morluto/rea/actions/runs/39406988130) checked out this exact merge commit and passed **38/38 jobs**. Its [Windows job](https://github.com/morluto/rea/actions/runs/39406988130/job/118579260782) again passed all **60 target tests with zero skips**, including the 13 Windows-only regressions. This qualifies the integrated merge commit in addition to the PR head. [Codex review](https://github.com/morluto/rea/pull/5409#issuecomment-6288140145) completed with no findings on final head; [Greptile](https://github.com/morluto/rea/pull/5409#issuecomment-6288196656) scored **5/5**. The production-owner finding was fixed with the three owner families and nine fail-before/pass-after controls. The existing native artifact, capability prerequisites, complete portable suite and required report gates are retained. No real external DSH/MCP client or Hopper/Ghidra engine claim is made. <details> <summary>Measured resource cost</summary> Pinned Node 24.18.0/npm 11.16.0, Linux, two CPUs; heavy commands serialized. The focused run reused a warm runtime build and performed fresh tests. Locked installed dependencies were reused only after an identical lock digest and the dependency-install validator passed. | Final completed workload | Wall s | Family CPU s | Sampled family RSS MiB | | --- | ---: | ---: | ---: | | Six-file runtime/CI discovery run | 110.39 | 161.29 | 1,061 | | Final two CI guard files | 36.37 | 41.63 | 654 | | Final static/architecture checks | 91.19 | 169.89 | 4,855 | | Final exact-once discovery | 0.92 | 1.04 | 287 | RSS sums observed descendants every 100 ms, may double-count shared pages or miss short/reparented children, and is not an aggregate hard cap. No observed OOM. Baseline controls included a cold 70.85 s compilation and failed, so their total is not an equivalent performance comparison. Added Windows coverage intentionally adds work; no speedup claim. Baseline Windows curated step took 74 s; the qualified final head took 126 s with the added 60 cases. These single observations run different coverage and do not isolate host/cache variance or establish a speedup. The initial head recorded 91 s, demonstrating that host/run variation also matters. Host variance, agent token usage and CI-wide cost are not measured. LeanToken retrieval failed internally; inspection used the pinned local checkout. </details> ## Author readiness Human owner: @gensokyo-dev The owner explicitly delegated these declarations and authorized opening ready, following CI/review and normal merge in this session. Codex reviewed the final diff and evidence; independent human review was not observed. - [x] I reviewed the final diff and relevant existing code, understand the change and its edge cases, and take responsibility for the complete PR, including AI-assisted work. - [x] I checked the evidence above, confirmed it supports the current change, and recorded meaningful validation gaps. - [x] I declare this PR ready for initial maintainer review. ## Before merge - [x] This PR addresses a concrete problem or an agreed enhancement. - [x] Material scope and authority questions are resolved, and any required maintainer scope decision covers the final diff. - [x] The PR has one focused outcome and the title follows `type(scope): outcome`. - [x] Related issue is linked, or the reason for not linking one is stated above. - [x] Tests cover changed observable behavior and meaningful failure paths. - [x] Owning docs, contracts, and generated metadata are updated where needed. - [x] User-visible CLI/MCP changes include representative output. - [x] I checked the final diff for secrets, unrelated cleanup, and unsupported claims. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi

2026-11-01 21:44:04 +00:00