README.md

GitHub Action: aidasoft/run-lcg-view

This GitHub Action executes user payload code inside a LCG view environment, specified by the user.

Instructions

Prerequisites

This action depends on the user to call the companion action uses: cvmfs-contrib/github-action-cvmfs@v2 before using uses: aidasoft/run-lcg-view@v4, which will install CVMFS on the node. GitHub Actions currently do not support calling the action github-action-cvmfs from within run-lcg-view, this needs to be done explicitly by the user.

Example

You can use this GitHub Action in a workflow in your own repository with uses: aidasoft/.

A minimal job example for GitHub-hosted runners of type ubuntu-latest:

jobs:
  run-lcg:
    runs-on: ubuntu-latest
    steps:
    - uses: actions/checkout@v2
    - uses: cvmfs-contrib/github-action-cvmfs@v2
    - uses: aidasoft/run-lcg-view@v4
      with:
        release-platform: "LCG_99/x86_64-centos7-gcc10-opt"
        run: |
          gcc --version
          which gcc

In this case the action will automatically resolve the correct container image (in this case centos7) and spawn an instance with Docker from GitHub Container Registry or with Singularity from /cvmfs/unpacked.cern.ch/. The Dockerfile for the supported images can be found in the AIDASoft/management repository.

The action mounts the checkout directory into the mentioned container and wraps the variable run in the script:

#!/usr/bin/env bash
export LC_ALL=C
set -e

source ${VIEW_PATH}/setup.sh

${RUN} # the multi-line variable specified in the action under run: |

which is executed in the container and thus giving the user an easy and direct access to run arbitrary code on top of LCG views.

macOS

It is currently not possible to mount CVMFS on the github CI machines with macOS.

Parameters

The following parameters are supported:

• container: Which container to use as base to setup a view. By default the container is inferred from view-path (default: auto)
• platform: LCG view platform you are targeting (e.g. x86_64-centos8-gcc10-opt)
• release: LCG view release you are targeting (e.g. LCG_99)
• release-platform:LCG view release platform string you are targeting (e.g. LCG_99/x86_64-centos8-gcc10-opt)
• unpacked: Use image from /cvmfs/unpacked.cern.ch with Singularity, or if false use GitHub Container Registry with Docker (default: false)
• run: They payload code you want to execute on top of the view
• setup-script: Initialization/Setup script for a view that sets the environment (e.g. setup.sh)
• view-path: Path where the setup script for the custom view is location. By specifying this variable the auto-resolving of the view based on release and platform is disabled. Furthermore the full path has to contain the architecture of the build in the form /dir1/dir2/x86_64-{arch}-gcc../dir4/dir5. The system will try to resolve the docker container equal to the string {arch} (the string after x86_64-).

Please be aware that you must use the combination of parameters release and platform together or use just the variable release-platform alone. These two options are given to enable more flexibility for the user to form their workflow with matrix expressions.

Minimal Example

There are minimal examples, which are also workflows in this repository in the subfolder .github/workflows/.

Limitations

The action will always resolve the correct image to execute your code on top the requested view, therefore you must always set the top level GitHub Action variable runs-on: ubuntu-latest.

Coverity Scan extension

Prerequisites

It is also possible to automatize Coverity Scan with this action. There are several steps that you need to do before being able to use this specific feature of this action:

• Register you Open Source Project with Coverity Scan
• Create a private Github Container Registry image with the Coverity Scan binaries (see Dockerfile and workflow) embedded within. This will remove the need to always download the binary for each scan
• Create a Personal Access Token to be able to access the private image (add this variable as a secret)
• Add to the project secrets the Coverity Scan token from your project

Example

You can use this feature from this GitHub Action in a workflow in your own repository with uses: aidasoft/run-lcg-view@v4.

jobs:
  run-coverity:
    runs-on: ubuntu-latest
    steps:
    - uses: actions/checkout@v2
    - uses: cvmfs-contrib/github-action-cvmfs@v2
    - uses: aidasoft/run-lcg-view@v4
      with:
        coverity-container: 'ghcr.io/aidasoft/coverity:latest'
        coverity-cmake-command: 'cmake -DCMAKE_CXX_STANDARD=17 ..'
        coverity-project: 'AIDASoft%2Fpodio'
        coverity-project-token: ${{ secrets.PODIO_COVERITY_TOKEN }}
        github-pat: ${{ secrets.READ_COVERITY_IMAGE }}
        release-platform: "LCG_99/x86_64-centos7-gcc10-opt"

The user needs to take care that the release-platform is compatible with the coverity-container.

The action mounts the checkout directory into the selected container and wraps the variable coverity-cmake-command in the script:

#!/usr/bin/env bash

set -e

source ${VIEW_PATH}/${SETUP_SCRIPT}
${RUN}
mkdir build
cd build
${COVERITY_CMAKE_COMMAND}
cov-build --dir cov-int make -j4
tar czvf /myproject.tgz cov-int

It is expected that the cov-build command is already in the PATH of the selected image.

The action finishes by uploading the tar ball to the server as per instruction given by Coverity Scan.

Parameters for Coverity extension

The following parameters are supported:

• coverity-container: Location of container that has Coverity Scan installed (default: ghcr.io/aidasoft/coverity:latest but not public)
• coverity-cmake-command: CMake command for building your project, assuming in source build.
• coverity-project: Coverity project name in URL encoding ( / -> %2F e.g. AIDASoft%2FDD4hep)'. Name under which your project is registered with the Coverity Scan server.
• coverity-project-token: Coverity project token to interact with the server
• github-pat: GitHub Personal Access Token for reading your custom image given under coverity-container