Repository navigation
build!: move to Python 3.15 and update dependencies and tools - #17
Merged
Merged
Conversation
- aqua registry v4.538.0 -> v4.575.0 - uv 0.12.13 -> 0.13.5, betterleaks v1.7.4 -> v1.9.0 (checksums pruned) - CI: aqua v2.61.0 -> v2.65.2 with aqua-installer v4.0.5 -> v4.1.1, which is required to install aqua v2.65.0 or later
- Support Python 3.15.x only (requires-python >=3.15,<3.16, .python-version, ruff target, doctor check); Python 3.14 is dropped. - openai-codex 0.160.0 -> 0.163.0 (bundled runtime included) and mcp 2.2.0 -> 2.3.0; refresh all other locked dependencies. - Review of the SDK from 0.154.0 through 0.163.0 found no API change in what the bridge uses; server-initiated request kinds are unchanged. Pin the new MessagePhase.partial_answer handling, which matches the SDK's own final-response selection, with a unit test. - Align the fake App Server version labels and the documented pins.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
requires-pythonis now>=3.15,<3.16, and.python-version, the ruff target, the classifier and doctor's version check follow. Python 3.14 is dropped, which is a breaking change for 3.14 users. The repo keeps its one-stable-major policy.openai-codex0.160.0 → 0.163.0, with its bundled runtimeopenai-codex-cli-bin.mcp2.2.0 → 2.3.0.uv lock --upgrade. Notable bumps: pydantic 2.14.1 / pydantic-core 2.50.1, which ships cp315 wheels, and ruff 0.17.1.0.154.0.openai-codex review (0.154.0 → 0.163.0)
The code was last validated against 0.154.0, and the bump to 0.160.0 came from Dependabot. I diffed the 0.154.0, 0.160.0 and 0.163.0 wheels and the upstream protocol definitions.
generated/v2_all.py,generated/notification_registry.pyand docstrings inapi.pychanged.MessagePhase.partial_answer. The new notifications go to the general queue, which the bridge drains and drops.declineis still valid for both approval requests. The fail-closed behaviour is therefore unchanged.partial_answerphase is never used as the final text, which matches the SDK's own final-response scan; a unit test now pins this.Follow-ups I suggest but have not done:
item/tool/requestUserInputmid-turn. The bridge rejects it, which is safe but late.Notes
uv.lockmoves to revision 5 (uv 0.13). Older uv releases, and possibly Dependabot's uv updater, may not read it.v0.1.1tag, which requires Python 3.14. The requirements line up only after a new release.Validation (Python 3.15.1, aqua-pinned tools)
aqua exec -- uv run --frozen pytest -m 'not live' -q: 161 passed, 4 deselected. They also pass with-W error::DeprecationWarning.ruff checkandruff format --check(0.17.1): clean.uv build: OK.betterleaks dir .andbetterleaks git . --platform github: no leaks.codex-app-mcp doctor --skip-runtime: python 3.15.1 and SDK 0.163.0 are reported ok.All evidence comes from unit tests, contract tests against the pinned SDK and a fake App Server, and subprocess tests. I did no live-runtime or real-application validation.