Skip to content

The GitHub Action to setup vcpkg for your C++ based projects. Stores built ports using Binary Caching backed onto GH Cache.

License

Notifications You must be signed in to change notification settings

shinnytech/run-vcpkg

 
 

Repository files navigation

Action Status

Coverage Status

run-vcpkg@v11 requires vcpkg more recent than 2023-03-29 (e.g. commit id 5b1214315250939257ef5d62ecdcbca18cf4fb1c)

Quickstart with a C++ project template

Take a look at this C++ project template that applies all the following instructions, but also shows how to create a pure workflow without using special GitHub action that you cannot run locally on your development machine, but directly using the tools (CMake, Ninja, vcpkg, C++ compilers) you already use daily.

The run-vcpkg action setups (and optionally runs) vcpkg to install the packages specified in the vcpkg.json manifest file. It leverages the vcpkg's Binary Caching backed to GitHub Action cache, delegating cache and key management to vpckg.

Special features which provide added value over a pure workflow are:

  • automatic caching leveraging vcpkg's ability to store its Binary Caching artifacts onto the GitHub Action cache so that packages are built only once and reused in subsequent workflow runs. The user can customize the behavior by setting the environment variable VCPKG_BINARY_SOURCES before vcpkg runs.
  • set the workflow variables ACTIONS_CACHE_URL and ACTIONS_RUNTIME_TOKEN to let users easily running vcpkg in a run step such as - run: vcpkg install or - run: vcpkg integrate install without forcing the users to set the variables manually.
  • automatic dump of log files created by CMake (e.g., CMakeOutput.log) and vcpkg. The content of those files flow into the workflow output log. Customizable by the user by setting the input logCollectionRegExps.
  • automatic parsing of CMake, vcpkg and gcc, clang, msvc errors, reporting them contextually in the workflow summary by means of annotations.
  • although disabled by default (see input doNotCache), run-vcpkg can cache vcpkg's executable and data files to speed up subsequent workflow runs. Since bootstrapping vcpkg already downloads a prebuilt binary saving the time to build vcpkg, this form of caching is useful only when the prebuilt executable is not served as it happens for the ARM platform. Note this cache does not contain the libraries built by vcpkg, which is instead managed by vcpkg itself.

The provided samples use GitHub hosted runners.

Good companions are the run-cmake action and the get-cmake actions.


Quickstart with instructions

It is highly recommended to use:

jobs:
  build:
    steps:
      #-uses: actions/cache@v1   <===== YOU DO NOT NEED THIS!

      # Install latest CMake.
      - uses: lukka/get-cmake@latest
      # Or pin to a specific CMake version:
      # lukka/[email protected]

      # Setup vcpkg: ensures vcpkg is downloaded and built.
      # Since vcpkg.json is being used later on to install the packages
      # when `run-cmake` runs, no packages are installed at this time 
      # (and vcpkg does not run).
      - name: Setup vcpkg (it does not install any package yet)
        uses: lukka/run-vcpkg@v11
        #with:
          # This is the default location of the directory containing vcpkg sources.
          # Change it to the right location if needed.
          # vcpkgDirectory: '${{ github.workspace }}/vcpkg'

          # If not using a submodule for vcpkg sources, this specifies which commit
          # id must be checkout from a Git repo. 
          # Note: it must not be set if using a Git submodule for vcpkg.
          # vcpkgGitCommitId: '${{ matrix.vcpkgCommitId }}'

          # This is the glob expression used to locate the vcpkg.json. 
          # Change it to match a single manifest file you want to use.
          # Note: do not use `${{ github.context }}` to compose the value as it
          # contains backslashes that would be misinterpreted. Instead
          # compose a value relative to the root of the repository using
          # `**/path/from/root/of/repo/to/vcpkg.json` to match the desired `vcpkg.json`.
          # vcpkgJsonGlob: '**/vcpkg.json'

          # This is only needed if the command `vcpkg install` must run at this step.
          # Instead it is highly suggested to let `run-cmake` to run vcpkg later on
          # using the vcpkg.cmake toolchain. The default is `false`.
          # runVcpkgInstall: true

      - name: Run CMake consuming CMakePreset.json and run vcpkg to build dependencies
        uses: lukka/run-cmake@v10
        with:
          # This is the default path to the CMakeLists.txt along side the
          # CMakePresets.json. Change if you need have CMakeLists.txt and CMakePresets.json
          # located elsewhere.
          # cmakeListsTxtPath: '${{ github.workspace }}/CMakeLists.txt'

          # This is the name of the CMakePresets.json's configuration to use to generate
          # the project files. This configuration leverages the vcpkg.cmake toolchain file to
          # run vcpkg and install all dependencies specified in vcpkg.json.
          configurePreset: 'ninja-multi-vcpkg'

          # This is the name of the CMakePresets.json's configuration to build the project.
          buildPreset: 'ninja-multi-vcpkg'

    #env:
    #  By default the action disables vcpkg's telemetry by defining VCPKG_DISABLE_METRICS.
    #  This behavior can be disabled by defining `VCPKG_ENABLE_METRICS` as follows.
    #  VCPKG_ENABLE_METRICS: 1
    #
    #  [OPTIONAL] Define the vcpkg's triplet you want to enforce, otherwise the default one
    #  for the hosting system will be automatically choosen (x64 is the default on all
    #  platforms, e.g. `x64-osx`).
    #  VCPKG_DEFAULT_TRIPLET: ${{ matrix.triplet }}

Action reference: all input/output parameters

Description of all input parameters: action.yml

Flowchart

Flowchart with related input in action.yml which let customize the flow.

┌──────────────────────────┐
|  If running in GH Runner,|   Environment variables:
|  set the env vars:       |   - If any env var is
|  - VCPKG_BINARY_SOURCES  |     already defined
|  - ACTIONS_CACHE_URL     |     it will not be overridden.
|  - ACTIONS_RUNTIME_TOKEN |
└─────────────┬────────────┘
              ▼
┌──────────────────────────┐
|  Skipped by default.     |
│  Compute cache key from: │   Inputs:
│  - vcpkg Git commit      │   - `vcpkgGitCommitId`
│  - platform and OS       │   - `doNotCache`: set to false
└─────────────┬────────────┘     to run this block.
              ▼
 ┌─────────────────────────┐   Inputs:
 │ Locate vcpkg.json.      │   - `vcpkgJsonGlob`
 └────────────┬────────────┘   - `vcpkgJsonIgnores`
              ▼
 ┌─────────────────────────┐   Inputs:
 | Skipped by default.     |   - `vcpkgDirectory`
 │ Restore vcpkg           │   - `doNotCache`: set to false
 │ from the GH cache.      │     to run this block.
 └────────────┬────────────┘
              ▼
 ┌─────────────────────────┐   Inputs:
 │ If vcpkg is not a       │   - `vcpkgDirectory`
 │ submodule, fetch it     │   - `vcpkgGitCommitId`
 │                         │   - `vcpkgGitURL`
 └────────────┬────────────┘   - `doNotUpdateVcpkg`
              ▼
 ┌─────────────────────────┐
 │ Rebuild vcpkg executable│   Inputs:
 │ if not in sync with     │   - `vcpkgGitCommitId`
 │ sources.                │   - `vcpkgGitURL`
 └────────────┬────────────┘
              ▼
  <Is `runVcpkgInstall:true`>┐    Inputs:
          ────┬────        No│   - `runVcpkgInstall`
              │ Yes          │
              ▼              │
 ┌─────────────────────────┐ │
 │ Launch `vcpkg install`  │ │   Inputs:
 │ where vcpkg.json has    │ │   - `runVcpkgFormatString`
 │ been located.           │ │   Environment variables:
 └────────────┬────────────┘ │   - `VCPKG_DEFAULT_TRIPLET` is used. If not yet
              │              │     set, it is set to the current platform.
              │              │   - `VCPKG_INSTALLED_DIR` is used as value for
              │              │     `--x-install-root` when running `vcpkg install`.
              │              │     Check out the `runVcpkgFormatString` input.
              ▼              │   - `VCPKG_BINARY_SOURCES` is used. If not yet 
 ┌─────────────────────────┐ │      set, it is set to leverage the GitHub Action
 │ Set `VCPKG_ROOT` and    │ │      cache storage for Binary Caching artifacts.
 │ `VCPKG_DEFAULT_TRIPLET` │ │
 │ workflow-wide env vars. │ │
 └────────────┬────────────┘ │
              ├───────────── ┘
              ▼
 ┌─────────────────────────┐
 | Skipped by default.     |
 │ If no cache-hit,        │  Inputs:
 │ store vcpkg onto        │  - `doNotCache`: set to false to
 │ GH cache.               │    run this block.
 └────────────┬────────────┘
              |
              ▼
              ⬬

Best practices

Use vcpkg as a submodule of your repository

It is highly suggested to pin the specific version of vcpkg you want to use to keep a consistent development experience between local and remote build environments. This is accomplished by using vcpkg as submodule of your Git repository; this way the version of vcpkg used is implied by the commit id specified by the submodule for vcpkg.

Use vcpkg's vcpkg.json file to specify the dependencies

The vcpkg.json is a manifest file that declaratively specifies the dependencies to be installed. The file is being used automatically by running CMake (e.g. by using run-cmake) when:

  • starting CMake with the vcpkg.cmake toolchain file.
  • the root CMake source directory contains a vcpkg.json file.

Or it can also be used by invoking vcpkg install in a directory where vcpkg.json is located (e.g., input runVcpkgInstall : true).

When conditions are satisfied, the toolchain execution starts vcpkg to install the packages declared in the manifest file.

Putting this manifest-like file under source control is highly recommended as this helps to run vcpkg the same exact way locally and remotely on the build servers. The dependencies specified in the vcpkg.json file are installed when CMake runs (i.e. at run-cmake execution time).

Samples

Checkmarks indicates whether the samples "uses" or specifies the thing in the header or whether it is true.

workflow link vcpkg as submodule explicit triplet vcpkg toolchain CMake's Presets Ninja run-vcpkg runs vcpkg CMake runs vcpkg
link
link
link
link

Who is using run-vcpkg

This graph shows the list of public repositories with more than 25 stars using run-vcpkg.


License

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

Copyright © 2019-2020-2021-2022-2023 Luca Cappa

Disclaimer

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.

Contributing

Read CONTRIBUTING.md

About

The GitHub Action to setup vcpkg for your C++ based projects. Stores built ports using Binary Caching backed onto GH Cache.

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • TypeScript 91.6%
  • JavaScript 8.4%