Skip to content

fix(ci): skip no-commit-to-branch in the full-suite CI run - #47

Merged
nstarman merged 7 commits into
mainfrom
fix-no-commit-to-branch-ci
Sep 15, 2026
Merged

nstarman merged 7 commits into
mainfrom
fix-no-commit-to-branch-ci

Conversation

@nstarman

Copy link
Copy Markdown
Contributor

Urgent: fixes broken CI on main

no-commit-to-branch (added in ##45) is currently failing every push to main — CI checks out a real local branch literally named main for push events, so the hook (which exists to stop a human committing directly to a protected branch) fires every single time, unconditionally. This is a hotfix, not a revert: the hook stays fully active for local git commit/git push (via prek install), which is its actual purpose — it's only skipped in the "run every hook over all files" CI invocation, where it can never do anything but fail.

Test plan

  • SKIP=no-commit-to-branch prek run --all-files (and prek run --skip no-commit-to-branch --all-files) verified locally to skip only this hook and run everything else normally
  • prek run no-commit-to-branch --all-files (without SKIP) still fails on main / versions/* and passes on a feature branch — confirming the hook itself is untouched

🤖 Generated with Claude Code

@github-actions github-actions Bot added the 🔧 Add / update configuration Add or update configuration files. label Sep 15, 2026
no-commit-to-branch was blocking every push to main: CI checks out a
real local branch literally named `main` for push events, so the hook
always fired. It's a client-side guard for a human running `git
commit`/`git push` locally (or via installed git hooks) -- not
something a full "run every hook" CI invocation should re-evaluate
after the fact. Skips it there via SKIP=no-commit-to-branch; the hook
itself is untouched and still fully active locally.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
@nstarman
nstarman force-pushed the fix-no-commit-to-branch-ci branch from 3927bb7 to 99a04b2 Compare September 15, 2026 15:28
…anch

Appends no-commit-to-branch to any SKIP a developer already has set
(e.g. via their shell) rather than overwriting it wholesale, matching
the same fix applied in response to Copilot review feedback on
GalacticDynamics/coordinax#885.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
@codecov

codecov Bot commented Sep 15, 2026 •

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (8e265d1) to head (bb88425).
⚠️ Report is 3 commits behind head on main.

Additional details and impacted files
@@            Coverage Diff            @@
##              main       #47   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            1         1           
  Lines            6         6           
=========================================
  Hits             6         6           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

nstarman and others added 5 commits September 15, 2026 12:30
Addresses Copilot review feedback on GalacticDynamics/galax#847: the
comment said "CI checks out the real main branch," but the skip
applies unconditionally, including local `nox -s lint` runs -- which
is correct (a CI-only skip would leave the same false failure for any
local dev running the full suite while on `main`). Fixes the wording
to match the actual, intended behavior instead of narrowing it.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Addresses Copilot review feedback on GalacticDynamics/dataclassish#94:
the docstring "Run prek." on a session still named `precommit` could
read as though the session itself was renamed. Spells out that it
runs the pre-commit hooks, now via prek.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Addresses Copilot review feedback on GalacticDynamics/galax#847: this
comment still said no-commit-to-branch guards `git push`, but the
earlier stages: [pre-commit] fix means it no longer runs on push at
all. Clarifies that explicitly instead of leaving stale wording.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
… trigger

Addresses Copilot review feedback on GalacticDynamics/galax#847: "it
never fires on push" reads as a claim about this workflow's own
`on: push:` trigger (which is false -- that's why the SKIP exists at
all), when it actually means the git pre-push hook stage. Spells that
out explicitly.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
The archaeology (why pre-commit's nodeenv/pyyaml floors mattered, why
--skip clobbers, the full CI-checkout explanation) belongs in commit
history, not permanently inline. Keeps just enough to orient a future
reader without re-litigating the whole investigation.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
@nstarman nstarman added this to the v1.3.0 milestone Sep 15, 2026
@nstarman
nstarman merged commit bb16191 into main Sep 15, 2026
23 checks passed
@nstarman
nstarman deleted the fix-no-commit-to-branch-ci branch September 15, 2026 18:07
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

🔧 Add / update configuration Add or update configuration files.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant