Skip to content

refactor(tools): neutral runner contract with cloud regression evidence - #165

Merged
LIghtJUNction merged 2 commits into
mainfrom
fix/architecture-boundaries-20260929
Sep 28, 2026
Merged

LIghtJUNction merged 2 commits into
mainfrom
fix/architecture-boundaries-20260929

Conversation

@LIghtJUNction

@LIghtJUNction LIghtJUNction commented Sep 28, 2026 •

Copy link
Copy Markdown
Contributor

Scope

Fixes #160. Refs #82 #83 #85; no umbrella issue or production acceptance is complete.

core.tool_execution owns the actual asynchronous callable ToolRunner and local-artifact ToolResult. EffectWorker and BubblewrapRunner use the same contract. Sandbox no longer imports the worker implementation. effects.ToolResult stays the identical class; the unused ToolExecutor.invoke-to-ResourceRef declaration is explicitly retired with an import alias to the actual runner port.

Existing local output validation, publication transaction, current authorization, mandatory isolation, quarantine and attempt/state/deadline fences are unchanged. Returning an existing ResourceRef from an injected runner is tested to produce no published output. No production action.

Verified RED → GREEN

RED b553c59b484a4e5a383df3500ad9e4cab9d51272, tree 59fadd97fb38f4dc60a21cb4dfb0535f913ef808: run36480982432 7 failed / 31 passed. All failures were new boundary tests; old tool and completion-fence regressions passed.

Final head f14fd15, tree 8a62150d8273a2c75775872694749cb529835966, tested as merge 0af4b19 onto main 476e64f. Refreshed PR metadata still reports that head/base/merge and mergeable=true.

  • Full run36481370643: 1,860 passed, shard counts 457 / 484 / 473 / 446, 0 failures, errors or skips. Downloaded all four artifacts; independently verified SHA-256, JUnit actual node IDs, unique/disjoint partition membership, and the exact union against the same collected node set. 8 conformance tests passed and installable wheel/sdist build gate passed.
  • Focused run36481370662: 38 passed / 0 failures, errors or skips; independently verified artifact10996748629 hash and exact source/tree JSON.
  • Recovery safety36481370672: completed success.

Full artifact IDs/hash: 10997336115=5e5d2739a19d3ce1282d3baa00f1361f77763cb5c4e82e40d29bf504e010ec19; 10997431290=034e4b8423ec78681c6d8c371a9e2b3a77433ba48de1a16f79821f9eeb1f3ee4; 10996419745=b60449dc9578633afc64aaa1f114924e1a50c45cc06790b3ac45bd2e8e44b7af; 10997176225=36a5eb7760b4c2c0c19fd7f4976e107622fae5f136d731c9e757ca5816f11da0.

Reviewed the full final diff: expected neutral types/annotations/compatibility export, unchanged worker logic (one incidental whitespace-only market-mail hunk), new tests/docs and focused workflow. Codex security review completed for f14fd15 without posted findings; Copilot could not review because of quota and is not counted as approval. No existing assertions or full CI gates were weakened.

Persistent notes: docs/ARCHITECTURE_PROGRESS_20260929.md (initial checkpoint; updated continuation in #169), docs/TOOL_RUNNER_BOUNDARY.md. All project execution/tests ran in cloud; local work was source inspection/editing and artifact verification. #155/#166 branches were not changed.

Refs #160 #82 #83 #85. Record source-bound CI evidence and an independent progress note before changing implementation.
…-neutral

Keep publication, current authorization and lease fences in EffectWorker; preserve the effects.ToolResult export and explicitly retire the unused ToolExecutor.invoke declaration.

Refs #160 #82 #85.
Copilot AI lite review requested due to automatic review settings September 28, 2026 20:44

Copilot AI left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Sep 28, 2026 •

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
🔒 Security Review ✅ Completed 2026-09-28T20:49:48.570859Z f14fd15 PR opened
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@LIghtJUNction
LIghtJUNction merged commit 08c5892 into main Sep 28, 2026
13 of 23 checks passed
LIghtJUNction added a commit that referenced this pull request Sep 28, 2026
…ontracts

Merge main 08c5892 into #167. Preserve all #165 source, tests, workflows and documentation; resolve core/contracts.py by keeping both QuerySession/domain declarations and the ToolExecutor=ToolRunner compatibility alias. Refs #156 #157 #83.
LIghtJUNction added a commit that referenced this pull request Sep 28, 2026
… implementation

Combine #171 client5f56bbe4 with #167 e768cbc without changing either other work stream's ref. Retain #167 executor, packet/schema, QuerySession/RelationalSession and all regression tests; retain merged #165 tool contract. Core codec owns the one result_wire function; packet re-exports that same object and executor keeps its packet import. Preserve all client helper, installation and signature tests. Apply the complete frozen-JSON comparison correction from #169 3db973a, whose 34 focused regressions passed in cloud.

Refs #158 #156 #157 #83 #85. Final combined tree requires new full/focused cloud gates; no historical GREEN or production acceptance is claimed.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[P2][架构] 工具执行端口与实际 runner 分层不清,结果类型反向依赖 worker 编排模块

2 participants