fix(ci): run workflows on PRs to release-v1

[MilesCranmerBot]

Re-enables GitHub Actions workflows for pull requests targeting the release-v1 branch by adding release-v1 to the on.pull_request.branches filters (and updating CodeQL's push/pull_request branch filters accordingly).

[MilesCranmer]

@MilesCranmerBot can you please disable the 3.8 Python jobs on release-v1? We should test for 3.9 instead.

[MilesCranmerBot]

Done — I updated the Linux workflow matrix to drop Python 3.8 and test Python 3.9 instead (release-v1 PR runs).

Commit: f805182 (touches .github/workflows/CI.yml)

[MilesCranmer]

@MilesCranmerBot please also address the CI failures with a push

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
⚠️ Please upload report for BASE (release-v1@e9a0406). Learn more about missing BASE report.

Additional details and impacted files

@@              Coverage Diff              @@
##             release-v1    #1154   +/-   ##
=============================================
  Coverage              ?   93.25%           
=============================================
  Files                 ?       21           
  Lines                 ?     1557           
  Branches              ?        0           
=============================================
  Hits                  ?     1452           
  Misses                ?      105           
  Partials              ?        0           

☔ View full report in Codecov by Sentry.
📢 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.

[MilesCranmer]

@MilesCranmerBot any idea why this is failing CI. Did something change?

[MilesCranmerBot]

CI failure is coming from the Windows job only. It’s installing Julia via setup-julia with version: '1' (currently Julia 1.12.5 on windows-latest) and then pulling SymbolicRegression v1.11.3 (release-v1 backend), which is crashing its Distributed workers on Windows (ProcessExitedException during the pipeline test).

I pushed a fix to the PR to pin the Windows workflow to Julia 1.11 for release-v1: 12ae132 (fix(ci): pin Julia to 1.11 on Windows). That should stop the worker-exit crash; rerun should go green if nothing else is lurking.

[MilesCranmer]

@MilesCranmerBot why is Julia 1.12.5 broken on Windows?

[MilesCranmerBot]

@MilesCranmer Not sure it’s fundamentally broken, but in our CI context it’s been flaky on Windows: julia-version: "1" resolves to the latest stable (currently 1.12.5), and with the SymbolicRegression v1.x backend we’ve seen intermittent hard worker crashes / access violations on windows-latest under Julia 1.12.x.

To keep release-v1 PR CI reliable, I pinned the Windows workflow to Julia 1.11 for now. Once we have a minimal reproducer (or an upstream Julia issue to point to), we can unpin back to "1" / 1.12.x.