Skip to content

ci: run the python 3.9 client job on ubuntu-24.04 - #39351

Merged
Pearcekieser merged 1 commit into
mainfrom
pearcekieser/client-py39-runner
Nov 19, 2026
Merged

Pearcekieser merged 1 commit into
mainfrom
pearcekieser/client-py39-runner

Conversation

@Pearcekieser

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: The Python client's 3.9 test job runs again, so a client change can pass CI and a merge group.

The kirocrew-client (py3.9) job fails at actions/setup-python with "The version '3.9' with architecture 'x64' was not found for Ubuntu 26.04." The kirocrew-linux-4-core label boots ubuntu-26.04, and setup-python has no 3.9 build for it. Seen on #39347, run 44437435061.

Why it matters

Every PR that touches packages/kirocrew-client-py gets this red, and merge-queue-readiness.yml polls client-py.yml, so a merge group with such a change cannot land. The package declares Python 3.9 as its floor, so dropping the job would stop testing a supported version.

Not a goal

  • Changing the package's supported Python versions.
  • Moving other jobs off the ubuntu-26.04 image.

What changed (motivation → approach → change)

The job fails because the image has no 3.9. The 3.9 matrix entry now sets runner: ubuntu-24.04, the newest hosted image that setup-python has 3.9 for, and runs-on reads matrix.runner before the larger-runner label. The 3.12 entry sets no runner and keeps the label. The ci-and-reviews.md paragraph about these label sites describes the client-py.yml exception.

Backwards compatibility

Compatible: the job names, the triggers and the 3.12 job do not change. Only the runner of the 3.9 job changes.

Removes nothing: both matrix entries still run.

Tests

Proof: not provable -- docs or config only

The workflow tests that read client-py.yml or ci-and-reviews.md pass, 530 tests, among them test/test_ci_fleet_routing_expression_parity.py and test/test_ci_runner_pool_hygiene.py. CI on this PR runs both matrix entries, and the 3.9 job is the check that this works.

Manual verification

N/A -- the change is to a workflow, and this PR's own CI run is the verification.

Real-machine verification

N/A -- the diff changes a CI workflow and a doc, and runs nothing on a machine.

Related Issues

Sibling sweep

Swept: git grep -nE "['"]3.(9|10)['"]" -- .github/workflows
Siblings: none -- the only hits are the client-py.yml matrix lines this PR changes

Pattern harvest

Not generalizable: one matrix entry pins the oldest Python the package supports, and only that entry outlived the runner image's toolcache.

Checklist

  • A Conventional Commits title (feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...); PR Hygiene does not count commits
  • Existing tests pass and new tests added for new functionality
  • Self-review completed; code follows project style guidelines
  • Documentation updated (if applicable)
  • No secrets, credentials, or internal references in the diff

Fixes #39350

The kirocrew-client (py3.9) job fails in actions/setup-python: the
runner boots ubuntu-26.04, and setup-python has no Python 3.9 build for
that image. The package declares requires-python >= 3.9, so the floor
job must keep running.

The 3.9 matrix entry sets runner: ubuntu-24.04, and runs-on reads
matrix.runner before the larger-runner label. The 3.12 entry keeps the
label.
@Pearcekieser
Pearcekieser requested review from a team and KiroCrewCaptain as code owners November 19, 2026 01:45
@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Nov 19, 2026
@github-actions

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ⏭️ skipped

Revision f6250505ae391d39c7ade3f244aa7de4810b423e touches no user-facing surface (no changes under website/ or committed screenshots), so the UX review was skipped. Does not block merge.

@github-actions

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of f6250505ae391d39c7ade3f244aa7de4810b423e — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

Design-Verdict: PASS

Nothing blocks. Pinning the 3.9 floor job to ubuntu-24.04 is a small, reversible CI change aimed at the actual cause, and the 3.12 entry keeps the larger-runner routing.

Notes

  • ubuntu-24.04 will also be retired eventually. When that happens, this same red returns with no warning. Adding a dated comment, or linking a tracking issue that names when to revisit the 3.9 floor, would make that a planned decision instead of a surprise.

[DESIGN-REVIEWED] f625050

Description read

[DESCRIPTION-READ] e0848c2b38548400f9ca019f7ac0c0338ca23624156e48a3e53d511fde98761f

That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ✅ PASS

Premise-level review of f6250505ae391d39c7ade3f244aa7de4810b423e — why this exists and whether the shipped surface is the smallest honest version. Updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

First-Principles-Verdict: PASS

Nothing blocks: this fixes the one matrix entry that fails, and no other workflow pins Python 3.9.

Notes

  • Item 3 — one consumer, generalized: matrix.runner || is a general per-entry runner slot, and only the "3.9" entry sets it (1 consumer; grep python-version:.*3\.(9|10)|"3.9" over .github/ finds only the 2 lines in client-py.yml). Subtraction: none smaller. A matrix.python == '3.9' check would be the same size and would hard-code the version in a second place.

What this change ships

Inventory (3 items) — 2 justified

Intent: get the Python client's 3.9 CI job running again so client PRs and merge groups can land. This is a FIX.

  1. The kirocrew-client (py3.9) job runs on ubuntu-24.04 instead of the 4-core label's ubuntu-26.04 image — justified
  2. The ci-and-reviews.md paragraph on the label sites now describes the client-py.yml exception — justified
  3. A per-entry runner key on the client-py matrix — one consumer, generalized; nothing smaller exists

[FIRST-PRINCIPLES-REVIEWED] f625050

Description read

[DESCRIPTION-READ] de92026de25b9b3364bec5d3cf48cc2e2fd491c5cbddff6b017b0405add61980

That sha256 names the PR description this verdict read. Recompute it with .github/scripts/pr-description-capture.sh; if it differs, the description changed after this verdict was formed and any finding drawn from the description is unproven.

@github-actions

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed f6250505ae391d39c7ade3f244aa7de4810b423e — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] f625050

Verdict parsed from the review's SHA-scoped output markers for commit f6250505ae391d39c7ade3f244aa7de4810b423e.

False positive or not applicable? A repository writer can comment:
/ai-review override fable f6250505ae391d39c7ade3f244aa7de4810b423e false-positive: <reason, 20 to 500 characters>

@github-actions

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of f6250505ae391d39c7ade3f244aa7de4810b423e and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] f625050

False positive or not applicable? A repository writer can comment:
/ai-review override gpt f6250505ae391d39c7ade3f244aa7de4810b423e false-positive: <reason, 20 to 500 characters>

@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed readiness: checking Automated validation is still running labels Nov 19, 2026
@Pearcekieser
Pearcekieser enabled auto-merge (squash) November 19, 2026 02:51

@KiroCrewCaptain KiroCrewCaptain left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Thanks for the fix. The review checked these things on this head:

  • The new include entry changes only the existing 3.9 matrix leg and adds no third job. The 3.12 leg has no matrix.runner, so it falls back to the unchanged label expression.
  • On this head's CI, kirocrew-client (py3.9) ran on ubuntu-24.04 and passed. kirocrew-client (py3.12) stayed on kirocrew-linux-4-core and also passed. On the base, the 3.9 job failed at setup-python.
  • Job names, triggers, permissions and steps are unchanged. No workflow test pins this runs-on.

Done well: you override only the one leg that needs it, and the doc paragraph that lists this label site changed in the same commit, so it stays accurate.

Optional:

  • ubuntu-24.04 will be retired some day, so this pin will need another look then.

@Pearcekieser
Pearcekieser merged commit 4bf9581 into main Nov 19, 2026
100 checks passed
@Pearcekieser
Pearcekieser deleted the pearcekieser/client-py39-runner branch November 19, 2026 04:00
@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Nov 19, 2026
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.

ci: the kirocrew-client (py3.9) job fails on the ubuntu-26.04 runner image

2 participants