release: 0.9.2.3 — run it before you ship it - #180
Merged
Conversation
Rolls Unreleased into a named block and bumps both version sites. The batch is two user-reported bugs and the checklist change they forced. chimera code took 5.1s to reach a prompt because create_provider imported all ten provider modules — and the anthropic + openai SDKs — to build one; lazy lookup takes that to 566ms on the box that hit it. And quitting dumped a KeyboardInterrupt traceback after "Bye!", so a clean exit looked like a crash. The name is the lesson. 0.9.2.2 was verified after publish: clean venv outside the repo, correct version inside and out, fixes confirmed present in the wheel. Every assertion true, and it exercised nothing a user does first. Both defects were one command away. This is the first release to go through the stricter gate, and it passed on its own terms: time to prompt 0.79s SIGINT no traceback all-extras 11,032 passed / 79 skipped CI posture 10,515 passed / 133 skipped cold mypy 736 files, 0 errors scrubs 7/7 claims+hygiene 140 passed Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Cuts 0.9.2.3. Two user-reported bugs and the checklist change they forced.
The bugs
chimera codetook 5.1 s to reach its prompt.create_providerimported all ten built-in provider modules — and theanthropic(445 ms) andopenai(247 ms) SDKs with them — to construct one. Lazy lookup: 1942 ms → 566 ms on the box that hit it, 3.4×.Quitting dumped a traceback after "Bye!" —
asyncio.runre-raises when a Ctrl+C lands during shutdown, so a clean exit rendered as a crash.Why the name
0.9.2.2 was verified after publish: clean venv outside the repo, correct version inside and out, fixes confirmed present in the wheel. Every assertion true — and it exercised nothing a user does first. Both defects were one command away.
Playbook 14 and CLAUDE.md now require launching the entry point, timing it, and SIGINT-ing it, on a machine that is not the dev box.
The first version of that gate was itself half-blind:
chimera code < /dev/nullraisesEOFError, notSIGINT, so it exits 0 and silent on the exact build that crashes on Ctrl+C. Corrected to require both lines.First release through the stricter gate — and it passed on its own terms
mypySingle local failure is
test_validation_split, documented as env-sensitive locally and green in CI.Release notes:
docs/releases/0.9.2.3.md.After merge
Tagging
v0.9.2.3publishes to PyPI. Held for explicit confirmation as usual.🤖 Generated with Claude Code