Skip to content

fix(ci): skip no-commit-to-branch in the full-suite CI run - #73

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 ##71) 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

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 c2ed01e to 43a888d 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>
@github-code-quality

github-code-quality Bot commented Sep 15, 2026 •

Copy link
Copy Markdown

Code Coverage Overview

Languages: Python

Python / code-coverage-agent

The overall line coverage in commit 4959328 in the fix-no-commit-to-bra... branch remains at 100%, unchanged from commit 7878258 in the main branch.


Updated September 15, 2026 17:05 UTC

@nstarman nstarman added this to the v0.3.0 milestone Sep 15, 2026
nstarman and others added 5 commits September 15, 2026 12:29
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 merged commit bf149c2 into main Sep 15, 2026
16 checks passed
@nstarman
nstarman deleted the fix-no-commit-to-branch-ci branch September 15, 2026 18:06
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.

1 participant