I think we should improve the developer experience and testing feedback loop by explicitly adding fail-fast: false to our multi-OS testing matrix in .github/workflows/rust.yml.

The Problem

Currently, our tests job in rust.yml runs a parallel execution matrix across four different platforms (macos-14, windows-latest, ubuntu-24.04-arm, and ubuntu-latest).

Because we haven't explicitly configured the fail-fast value inside the strategy block, GitHub Actions defaults to fail-fast: true. This means that if a developer pushes a bug that breaks only on Windows, but the ubuntu-latest runner happens to finish compiling and failing first, GitHub will instantly force-cancel the other three platforms mid-run.

This creates massive blind spots: developers lose their test results for the other platforms and have no idea if their bug is platform-specific or completely universal until they push a second fix.

The Fix

We should simply append fail-fast: false underneath the strategy block for the tests job.

    strategy:
      fail-fast: false
      matrix:
        include:
          - os: macos-14
          - os: windows-latest
          - os: ubuntu-24.04-arm
          - os: ubuntu-latest