Skip to content

fix: complete V3 runtime isolation and acceptance evidence - #327

Open
EvanProgramming wants to merge 164 commits into
mainfrom
Evan/v3-finalization-completion
Open

EvanProgramming wants to merge 164 commits into
mainfrom
Evan/v3-finalization-completion

Conversation

@EvanProgramming

Copy link
Copy Markdown
Owner

V3 could still lose cancellation/persistence evidence, trust worker-provided execution authority, or let edited benchmark source influence the final oracle. This candidate preserves task/result identities and lifecycle state, keeps actual permissions and one-use execution tickets in a privileged parent, and executes frozen fixture assertions outside edited source processes. It also retains legacy full-output retrieval and keeps public installers on published v2.0.8 while V3 is unreleased.

The signed head is 0109494a87448ce345dc69751b314b0aeab88ef0. This consolidates the existing V3 candidate stack and subsequent hardening for review against main; merging remains blocked until the formal release evidence is complete. The intended integration is a non-forced fast-forward to the same verified SHA.

Validation on this exact head:

  • Six-platform CI and original 30-minute soaks passed on Linux/macOS/Windows × Python3.12/3.13. Original artifact digests, CRCs, all549 tracked-source identities and300 process-audit checkpoints were independently checked. Windows retains its documented supported-feature/POSIX exceptions.
  • Four privileged Linux/macOS isolation combinations passed174 methods each, with explicit platform skips. Linux host AppArmor/user-namespace settings remained unchanged. Replay, forged authority, protected-file bypass, malicious fixture/verifier and detached-process cleanup regressions are included.
  • Local full suite:1,611 tests,1,583 passed and28 explicit platform/privilege skips; check/lint/docs/architecture/Go checks passed. Docker persistence, wheel/sdist installation outside the checkout, CLI, authenticated Web, TUI backend and retained package/TUI artifacts passed their respective checks. Cross-builds are distinguished from target execution.
  • R258/fix: avoid cumulative tool-result replay in legacy adapter #319 additionally has independently reviewed actual production runtime/SDK loopback-HTTP evidence for legacy archival, complete paging, typed failures, scope denial and application/SQLite reopening. Controlled upstream responses are not external model-quality evidence.

Not ready to merge or publish: R260 still needs the complete bounded native SQLite/VFS causal evidence review; the fixed180-cell matched DeepSeek evaluation,54-cell routing experiment and independent outcome/workflow review have not run. Historical settled plus held DeepSeek exposure was re-audited as CNY0.582914960007; no new paid calls were made. The cumulative cap remains CNY15, including smoke/retry/unknown usage, with CNY2.032768 reserved before each request. No statistical threshold or original acceptance is waived.

Evidence, signed artifact checksums, the266-row requirement index and unpublished release-note preparation are retained outside the checkout. No passing17-gate manifest, main advance, tag, draft/public Release, package upload or preview-publisher execution is claimed or performed.

@ghfind-review ghfind-review Bot added the review: high ghfind author score; see https://ghfind.com label Nov 7, 2026
@coderabbitai

coderabbitai Bot commented Nov 7, 2026 •

Copy link
Copy Markdown

Important

  • Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

Run configuration

  • Source mode: Auto
  • Repository: Unavailable
  • Branch: Unavailable
  • File: No configuration file from this repository applied
  • Commit: Unavailable
  • Layers: defaults
  • Review profile: CHILL
  • Plan: Advanced
  • Run ID: 19627591-d3f7-4f93-8d19-b9f841c08fa3
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

Share

@EvanProgramming
EvanProgramming marked this pull request as ready for review November 7, 2026 11:24
@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Nov 7, 2026 •

Copy link
Copy Markdown
Codex Review Summary: ✅ Completed

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

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-11-07T11:27:28.703517Z 0109494 Draft marked ready
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.

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

review: high ghfind author score; see https://ghfind.com

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant