The action restores LLVM from local or cloud cache. If a cache miss occurs, it downloads and caches the tools right away from Hylo's LLVM Builds.

• Supported OSs: Windows, Linux, MacOS 14+
• Supported architectures: x86_64, arm64/aarch64
• Supported LLVM build configurations: MinSizeRel, Debug

For the definitive set of options, see action.yml or the table below.

# Option 1: using 'latest' branch, the most recent supported LLVM is installed.
- uses: hylo-lang/get-llvm@latest

# Option 2: Install a specific LLVM version
- uses: hylo-lang/get-llvm@latest
  with:
    llvmVersion: '20.1.6'
    llvmBuildRelease: '20250910-063105'

# Option 3: Install LLVM with Debug configuration
- uses: hylo-lang/get-llvm@latest
  with:
    llvmVersion: '20.1.6'
    llvmBuildRelease: '20250910-063105'
    llvmBuildConfig: 'Debug'

# Option 4: For self-hosted runners, use local cache instead of cloud cache
- uses: hylo-lang/get-llvm@latest
  with:
    useCloudCache: false
    useLocalCache: true

# Option 5: Install LLVM without adding to PATH (for custom integration)
- uses: hylo-lang/get-llvm@latest
  with:
    addToPath: false
    addToPkgConfigPath: false
  id: llvm-setup

# Then use the outputs for custom integration
- name: Use LLVM
  run: |
    echo "LLVM installed at: ${{ steps.llvm-setup.outputs.llvmRootDirectory }}"
    echo "LLVM bin directory: ${{ steps.llvm-setup.outputs.llvmBinDirectory }}"
    echo "LLVM version: ${{ steps.llvm-setup.outputs.llvmVersion }}"
    
# Option 6: Complete example with matrix strategy for multiple LLVM versions
strategy:
  matrix:
    build-config: ['MinSizeRel', 'Debug']
    
steps:
- uses: hylo-lang/get-llvm@latest
  with:
    llvmVersion: 20.1.6
    llvmBuildRelease: '20250910-063105'
    llvmBuildConfig: ${{ matrix.build-config }}

There are two kind of caches:

• The cloud based GitHub cache. Enabled by default, it can be disabled using the input useCloudCache:false.
• The local self-hosted runner cache, stored locally using tool-cache. Disabled by default, it can enabled with the input useLocalCache:true.

Steps of get-llvm:

1. If a cache-hit occurs (either local or cloud cache), LLVM is restored from the cache.
   1. if both local and cloud are enabled, the local cache check goes first.
2. If a cache-miss occurs, i.e. none of the enabled caches hit:

   1. the action downloads and installs the desired version of LLVM.

   2. the action stores LLVM for the enabled caches:

      1. if enabled, on the cloud based GitHub cache. This is beneficial for the next run of the workflow especially on GitHub-hosted runners.
      2. if enabled, on the local GitHub runner cache. This is beneficial for the next run of the workflow on the same self-hosted runner.

      Note: when there is a cache-hit, nothing will be stored in any of the caches.

3. Adds to the PATH environment variable the binary directories for LLVM.

Build with tsc running:

npm run build

Launch lint by:

npm run lint

To build, lint validate and package the extension (and embed the release catalog) for release purpose, run:

npm run pack

To build, pack and run all tests:

npm run test

To run all tests:

npx jest

The software is provided as is, there is no warranty of any kind. All users are encouraged to improve the source code with fixes and new features.

All the content in this repository is licensed under the MIT License.

Copyright (c) 2020-2025 Luca Cappa + Hylo Community