Packaging and Testing with Crossbow¶
The content of
arrow/dev/tasks directory aims for automating the process of
Arrow packaging and integration testing.
- Integration tests:
Various docker tests
Individual jobs are executed on public CI services, currently:
Linux: GitHub Actions, Travis CI, Azure Pipelines
macOS: GitHub Actions, Azure Pipelines
Windows: GitHub Actions, Azure Pipelines
Because of the nature of how the CI services work, the scheduling of
jobs happens through an additional git repository, which acts like a job
queue for the tasks. Anyone can host a
queue repository (usually
A job is a git commit on a particular git branch, containing the required
configuration files to run the requested builds (like
crossbow.yml for GitHub Actions ).
Crossbow handles version generation, task rendering and
submission. The tasks are defined in
The following guide depends on GitHub, but theoretically any git server can be used.
If you are not using the ursacomputing/crossbow repository, you will need to complete the first two steps, otherwise procede to step 3:
Clone either ursacomputing/crossbow if you are using that, or the newly created repository next to the arrow repository:
By default the scripts looks for a
crossbowclone next to the
arrowdirectory, but this can configured through command line arguments.
git clone https://github.com/<user>/crossbow crossbow
Important note: Crossbow only supports GitHub token based authentication. Although it overwrites the repository urls provided with ssh protocol, it’s advisable to use the HTTPS repository URLs.
Create a Personal Access Token with
workflowpermissions (other permissions are not needed)
Locally export the token as an environment variable:
or pass as an argument to the CLI script
Add the previously created GitHub token to Travis CI:
CROSSBOW_GITHUB_TOKENencrypted environment variable. You can set it at the following URL, where
ghuseris the GitHub username and
ghrepois the GitHub repository name (typically
Confirm the auto cancellation feature is turned off for branch builds. This should be the default setting.
Install Python (minimum supported version is 3.8):Miniconda is preferred, see installation instructions:
Install the archery toolset containing crossbow itself:
$ pip install -e "arrow/dev/archery[crossbow]"
Try running it:
$ archery crossbow --help
The script does the following:
Detects the current repository, thus supports forks. The following snippet will build kszucs’s fork instead of the upstream apache/arrow repository.
$ git clone https://github.com/kszucs/arrow $ git clone https://github.com/kszucs/crossbow $ cd arrow/dev/tasks $ archery crossbow submit --help # show the available options $ archery crossbow submit conda-win conda-linux conda-osx
Gets the HEAD commit of the currently checked out branch and generates the version number based on setuptools_scm. So to build a particular branch check out before running the script:
$ git checkout ARROW-<ticket number> $ archery crossbow submit --dry-run conda-linux conda-osx
Note that the arrow branch must be pushed beforehand, because the script will clone the selected branch.
Reads and renders the required build configurations with the parameters substituted.
Create a branch per task, prefixed with the job id. For example, to build conda recipes on linux, it will create a new branch:
Pushes the modified branches to GitHub which triggers the builds. For authentication it uses GitHub OAuth tokens described in the install section.
Query the build status¶
Build id (which has a corresponding branch in the queue repository) is returned
$ archery crossbow status <build id / branch name>
Download the build artifacts¶
$ archery crossbow artifacts <build id / branch name>
Submit command accepts a list of task names and/or a list of task-group names to select which tasks to build.
Run multiple builds:
$ archery crossbow submit debian-stretch conda-linux-gcc-py37-r40 Repository: https://github.com/kszucs/arrow@tasks Commit SHA: 810a718836bb3a8cefc053055600bdcc440e6702 Version: 0.9.1.dev48+g810a7188.d20180414 Pushed branches: - debian-stretch - conda-linux-gcc-py37-r40
Just render without applying or committing the changes:
$ archery crossbow submit --dry-run task_name
conda package builds and a Linux one:
$ archery crossbow submit --group conda centos-7
$ archery crossbow submit --group wheel
There are multiple task groups in the
tasks.yml like docker, integration
and cpp-python for running docker based tests.
archery crossbow submit supports multiple options and arguments, for more
see its help page:
$ archery crossbow submit --help