[GH-ISSUE #7458] fix: clarify verification report ownership after analysis experiment archival #4455

New issue

Closed

opened 2026-10-31 00:49:56 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-31 00:49:56 +00:00

Owner

Originally created by @gensokyo-dev on GitHub (Oct 27, 2026).
Original GitHub issue: https://github.com/morluto/rea/issues/7458

Area

Documentation / verification ownership after repository archival

Current limitation

The earlier inventory at 59dc48db2 found 251 tracked files under experiments/analysis-efficiency/. By the refreshed baseline e99b6c2c3, commit 365fb03e7 had already archived that directory and moved 35 files needed by deterministic tests to tests/support/analysis-efficiency/. This issue follows up that completed archival; it does not propose restoring the retired benchmark suite.

Source inspection found a remaining ownership and documentation defect: the historical allocation report and receipt JSON still live inside test support, the report's relative mcp-input-catalog.md link no longer resolves, and its documented current regeneration commands still use the removed experiment path. Historical measurement commands also need an explicit archived-checkout context. The existing report producer and conformance test should use the report's proper documentation home.

Desired outcome and success criteria

Keep archived experiment tools retired and keep necessary deterministic helpers under test support.
Move the historical allocation report and its unchanged receipt JSON to docs/verification/analysis-efficiency/.
Update the report producer, existing conformance consumer, current regeneration commands and documentation entry points together; point the catalog-study reference at an exact archived source revision.
Explain which retained files support deterministic tests and distinguish their scope from real-provider verification.
Remove obsolete Turbo input globs for the retired experiment tree while retaining coverage of test-support JavaScript.
Preserve measured numbers, original revisions/hashes, negative results, malformed-input/lifecycle assertions and complete Evidence provenance. No runtime API, default, cache or new execution authority changes.

Expected interaction with targets

The change reorganizes documentation and updates development-only consumers. Validation reads selected public/self-owned artifacts through existing CLI/MCP interfaces. No new automatic target interaction is proposed.

Verification available so far

The stale local link and stale regeneration path are source-observed at e99b6c2c3; this is not an analysis performance defect. Planned: baseline and migrated focused/Python checks, exact historical JSON hash comparison, report regeneration/check, test discovery, static/docs checks and a bounded real-application CLI/MCP smoke task. Report measurements and limitations without claiming a speedup from relocation. Follow current-head CI/review before merging.

Alternatives considered

A wholesale move to a new benchmark directory would revive tools deliberately retired on latest main. Leaving reports inside test support preserves the ownership ambiguity and broken guidance. Deleting the report would remove publicly referenced evidence and its conformance protection.

References

Archive commit: 365fb03e73ffb23f4394b7fb82776f7dd2045301.
Historical efficiency work: #2016.
tests/support/analysis-efficiency/mcp-integer-allocation.md, mcp-integer-report.mjs, and tests/conformance/readmeConsistency.test.ts.

Originally created by @gensokyo-dev on GitHub (Oct 27, 2026). Original GitHub issue: https://github.com/morluto/rea/issues/7458 ### Area Documentation / verification ownership after repository archival ### Current limitation The earlier inventory at `59dc48db2` found 251 tracked files under `experiments/analysis-efficiency/`. By the refreshed baseline `e99b6c2c3`, commit `365fb03e7` had already archived that directory and moved 35 files needed by deterministic tests to `tests/support/analysis-efficiency/`. This issue follows up that completed archival; it does not propose restoring the retired benchmark suite. Source inspection found a remaining ownership and documentation defect: the historical allocation report and receipt JSON still live inside test support, the report's relative `mcp-input-catalog.md` link no longer resolves, and its documented current regeneration commands still use the removed experiment path. Historical measurement commands also need an explicit archived-checkout context. The existing report producer and conformance test should use the report's proper documentation home. ### Desired outcome and success criteria - Keep archived experiment tools retired and keep necessary deterministic helpers under test support. - Move the historical allocation report and its unchanged receipt JSON to `docs/verification/analysis-efficiency/`. - Update the report producer, existing conformance consumer, current regeneration commands and documentation entry points together; point the catalog-study reference at an exact archived source revision. - Explain which retained files support deterministic tests and distinguish their scope from real-provider verification. - Remove obsolete Turbo input globs for the retired experiment tree while retaining coverage of test-support JavaScript. - Preserve measured numbers, original revisions/hashes, negative results, malformed-input/lifecycle assertions and complete Evidence provenance. No runtime API, default, cache or new execution authority changes. ### Expected interaction with targets The change reorganizes documentation and updates development-only consumers. Validation reads selected public/self-owned artifacts through existing CLI/MCP interfaces. No new automatic target interaction is proposed. ### Verification available so far The stale local link and stale regeneration path are source-observed at `e99b6c2c3`; this is not an analysis performance defect. Planned: baseline and migrated focused/Python checks, exact historical JSON hash comparison, report regeneration/check, test discovery, static/docs checks and a bounded real-application CLI/MCP smoke task. Report measurements and limitations without claiming a speedup from relocation. Follow current-head CI/review before merging. ### Alternatives considered A wholesale move to a new benchmark directory would revive tools deliberately retired on latest main. Leaving reports inside test support preserves the ownership ambiguity and broken guidance. Deleting the report would remove publicly referenced evidence and its conformance protection. ### References - Archive commit: `365fb03e73ffb23f4394b7fb82776f7dd2045301`. - Historical efficiency work: #2016. - `tests/support/analysis-efficiency/mcp-integer-allocation.md`, `mcp-integer-report.mjs`, and `tests/conformance/readmeConsistency.test.ts`.

bsakaguchi closed this issue

2026-10-31 00:49:57 +00:00

bsakaguchi referenced this issue

2026-10-31 01:13:05 +00:00

[PR #4463] [MERGED] fix(pe): ignore zero-extent sections in resource interval ownership #6256

bsakaguchi referenced this issue

2026-10-31 01:13:08 +00:00

[PR #4478] [CLOSED] fix(pe): ignore zero-extent sections in resource mapping #6264