Skip to content

chore(deps): consolidate Dependabot updates for openenv-core - #1605

Closed
cursor[bot] wants to merge 0 commit into
mainfrom
dependabot/openenv-core-consolidated
Closed

cursor[bot] wants to merge 0 commit into
mainfrom
dependabot/openenv-core-consolidated

Conversation

@cursor

@cursor cursor Bot commented Nov 12, 2026 •

Copy link
Copy Markdown
Contributor

Summary

Consolidates all open Dependabot PRs that touch the core library / repo-level tooling (outside envs/) into a single branch, rebased on latest main. Each original dependency bump is preserved as its own commit (merged via individual db/<PR#> branches) so history stays traceable.

Included (merged cleanly, no conflicts):

Deliberately excluded:

  • chore(deps): update fastmcp requirement from <4.0.0,>=3.0.0 to >=3.0.0,<5.0.0 #1331 — fastmcp requirement <4.0.0 → <5.0.0. Merging this alone breaks envs/posttrain_env/server/posttrain_environment.py (ModuleNotFoundError: No module named 'fastmcp.tools.tool' — fastmcp 4.x moved/removed ToolResult from that import path). Verified by syncing with the new constraint (resolves fastmcp 4.1.8) and running the test suite; with the old constraint (fastmcp 3.4.11) tests pass. Since this touches a core library dependency, I left it out rather than merge a breaking change — #1331 is closed with an explanation; it will need a follow-up code change before it can be safely merged.

Type of Change

  • Bug fix (dependency maintenance)
  • New feature
  • Breaking change
  • Documentation
  • New environment
  • Refactoring

Alignment Checklist

  • I have read .claude/docs/PRINCIPLES.md and this PR aligns with our principles
  • I have checked .claude/docs/INVARIANTS.md and no invariants are violated
  • I have run tests and lint and addressed all issues

RFC Status

  • Not required (dependency bumps only)

Test Plan

  • uv sync --all-extras resolves cleanly with all included dependency changes.
  • PYTHONPATH=src:envs uv run pytest tests/ -q → 5463 passed, 235 skipped, 5 failed. The 5 failures (test_auto_env.py HF Space integration tests requiring network/interactive confirmation, test_websockets.py step_count assertions) are pre-existing and reproduce identically on a clean origin/main checkout — confirmed via a side-by-side baseline run. No new failures introduced.
  • uv run ruff check src/ tests/ → all checks passed.
  • uv run usort check src/ tests/ and uv run ruff format src/ tests/ --check → only the two pre-existing issues documented in AGENTS.md (plus one pre-existing README.md formatting note), none introduced by this change.
  • uv lock --check passes for tests/validation_runtime.

Claude Code Review

N/A — triage automation run; individual Dependabot PRs were reviewed for mergeability and consolidated manually, with the one breaking change (#1331) excluded and left open for a manual fix.

Open in Web View Automation 

Note

Low Risk
Routine CI action and test-runtime dependency pins with no auth, data, or application logic changes; the one risky fastmcp bump was explicitly excluded.

Overview
Rolls multiple pending Dependabot bumps for repo-level tooling into one branch while keeping each original bump as its own commit.

Dependency / lockfile: bumps Pillow in tests/validation_runtime (uv.lock).

GitHub Actions: pins newer commits for the reusable huggingface/doc-builder workflows used for main, PR, and PR-upload documentation builds; bumps astral-sh/setup-uv to v10.3.3 in discovery-catalog.yml; and updates docker/login-action in docker-build.yml.

Out of scope: the fastmcp <5.0.0 constraint change is intentionally not included because it breaks posttrain_env imports against fastmcp 4.x and needs a separate code fix.

Reviewed by Cursor Bugbot for commit 5b2da64. Bugbot is set up for automated code reviews on this repo. Configure here.

@burtenshaw burtenshaw added enhancement New feature or request size: small Small pull request labels Nov 12, 2026 — with Cursor
@cursor cursor Bot closed this Nov 12, 2026
@cursor
cursor Bot force-pushed the dependabot/openenv-core-consolidated branch from 845ed6a to 5b2da64 Compare November 12, 2026 07:22

@cursor cursor Bot left a comment

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

Alignment Review Report

Scope reviewed: git diff 5b2da641...845ed6a9 — this consolidated/aggregate Dependabot PR touches only .github/workflows/*.yml (5 reusable-workflow/action SHA bumps) and tests/validation_runtime/uv.lock (no pyproject.toml change in this repo). No src/, envs/, or docs changes.

Automated Checks

  • Lint: N/A for this diff's actual content (0 .py/.md files changed; ruff/usort can't touch .yml/.lock). Ran bash .claude/hooks/lint.sh anyway after installing uv — it exit-1's on ~2 pre-existing usort files (tests/envs/test_grid_world.py, test_julia_env.py, already called out as known baseline noise in AGENTS.md) and ~47 pre-existing README/docstring reformat candidates. None of those paths are in this PR's diff.
  • Debug code: CLEAN — bash .claude/hooks/check-debug.sh only reports pre-existing prints/TODOs in src/openenv/core/containers/test_local_docker_provider.py, harness/rollout.py, and runtime/uv_provider.py, none of which this PR touches.
  • Stale-branch/merge-conflict check: clean. git fetch origin main confirms origin/main == this PR's base commit exactly; git merge-tree origin/main <head> returns a single clean tree (no conflict).
  • GitHub Actions SHA pins: verified authentic. docker/login-action -> 8aa5831f... and astral-sh/setup-uv -> 68559e99... both match their claimed tags (v4.8.0, v10.3.3) exactly via gh api repos/<owner>/<repo>/commits/<tag>. All 5 bumps stay on full 40-char SHAs (good security hygiene, consistent with PRINCIPLES.md's reproducibility stance).

Open RFCs Context

Scanned all 11 RFCs (000–012). None cover CI/GitHub Actions, dependency management, or the tests/validation_runtime toolchain's own composition — but tests/validation_runtime/ itself is the fixture for RFC 008 (Environment Auto-Validation, In Review, @zkwentz), so a correctness regression inside its lock is read against RFC 008's reproducibility goals (see below).

Tier 1: Fixes Required

  • tests/validation_runtime/uv.lock:418 — gradio is pinned to the pre-release 7.0.0.dev0 instead of the current latest stable release. Verified live against pypi.org/pypi/gradio/json: stable releases 6.28.0 through 6.31.0 exist and were skipped entirely in favor of a .dev0 build.
  • tests/validation_runtime/uv.lock:458 — same issue, gradio-client pinned to 2.8.0.dev0 instead of the latest stable 2.7.2.

Root cause, reproduced locally (not a guess): I copied tests/validation_runtime/ to a scratch dir, pointed its path-dependency at an absolute path, deleted uv.lock, and ran a clean-room uv lock with no flags → result: gradio==6.31.0, gradio-client==2.7.2, pillow==12.3.0 (unchanged), safehttpx==0.1.7 (unchanged), no psutil. Then I ran uv lock --upgrade-package gradio --prerelease allow on the same tree, and it reproduced every line of this PR's actual lock diff exactly: gradio→7.0.0.dev0, gradio-client→2.8.0.dev0, psutil added, safehttpx→0.2.0, and the httpx→httpx2 dependency-edge swap on gradio/gradio-client/safehttpx. So the pillow 12.3.0→13.0.0 bump at uv.lock:1088 and the new psutil entry at uv.lock:1136 are not independent legitimate bumps — they're side effects pulled in only because gradio's pre-release build requires them; the latest stable gradio does not. Whatever step regenerated this lock during the Dependabot consolidation appears to have run with --prerelease allow (or equivalent) instead of a default/stable-only resolve.

Note: httpx2 (pulled in via the gradio pre-release) is a real, legitimate package — it's the httpx-stewardship successor maintained by Pydantic/the original httpx author, not a typosquat — so that part is fine in isolation; it's only in the lock because of the pre-release jump.

Suggested fix: re-run the lock regeneration for tests/validation_runtime/uv.lock without pre-release resolution so it lands on gradio==6.31.0 / gradio-client==2.7.2 (pillow/safehttpx/psutil will then revert to their pre-PR state automatically — no separate action needed on those). uv lock --check --project tests/validation_runtime currently passes on the committed lock since it's internally consistent, it's just resolved against the wrong policy, so a plain consistency check won't catch this class of issue.

Tier 2: Alignment Discussion

Principle Conflicts

None identified. No API/reward/client-server/MCP/Pydantic surface touched.

RFC Conflicts

ALIGNMENT FLAG: Pinning a pre-release dependency inside the RFC-008 reproducible-validation fixture

  • Principle/RFC at stake: RFC 008 (Environment Auto-Validation, In Review) — this fixture exists specifically to make validation runs reproducible.
  • The concern: Even once fixed per Tier 1, it's worth noting why this matters here specifically: tests/validation_runtime/ is the dedicated reproducible-validation harness for RFC 008. PyPI pre-release (.devN) builds are far more likely to be withdrawn, superseded, or have their artifacts change upstream than stable pins, which is in tension with a fixture whose whole purpose is dependable, repeatable builds. This is a secondary, non-blocking note riding on the same root cause as the Tier 1 fix above, not a separate design question.
  • Suggested reviewer: @burtenshaw (fixture owner, authored the original tests/validation_runtime/ docker validation lab and the dependabot consolidation merges) and @zkwentz (RFC 008 author)

Summary

  • 2 mechanical issues to fix (the gradio/gradio-client pre-release pins; pillow/psutil/safehttpx changes resolve themselves once those two are fixed)
  • 1 alignment point for human review (RFC-008 reproducibility context, non-blocking)
  • 0 RFC design conflicts

The GitHub Actions portion of this PR (5 SHA-pin bumps) is clean and verified authentic — no concerns there.

Open in Web View Automation 

Sent by Cursor Automation: Pre-review

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

enhancement New feature or request size: small Small pull request

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant