Skip to main content
Version: 2.6 (deprecated)

Run programs

How to add a new implementation to the run goal.


The run goal runs a single interactive process in the foreground, such as running a script or a program.

Example repository

This guide walks through adding a simple run implementation for Bash that runs the equivalent /bin/bash ./script.sh. See here for the final implementation.

1. Set up a binary target type

Usually, you will want to add a "binary" target type for your language, such as bash_binary or python_binary. Typically, both the run and package goals operate on binary target types.

When creating a binary target, you should usually subclass the Sources field and set the class property expected_num_files = 1.

See Creating new targets for a guide on how to define new target types.

from pants.engine.target import COMMON_TARGET_FIELDS, Dependencies, Sources, Target

class BashSources(Sources):
expected_file_extensions = (".sh",)


class BashBinarySources(BashSources):
required = True
expected_num_files = 1


class BashBinary(Target):
"""A Bash file that may be directly run."""

alias = "bash_binary"
core_fields = (*COMMON_TARGET_FIELDS, Dependencies, BashBinarySources)

2. Set up a subclass of RunFieldSet

As described in Rules and the Target API, a FieldSet is a way to tell Pants which Fields you care about targets having for your plugin to work.

Usually, you will require the binary target's Sources subclass from Step 1, such as BashBinarySources or PythonBinarySources. Add this Sources subclass to the class property required_fields of your new FieldSet. This means that your binary implementation will run on any target with that sources field or a subclass of it.

Create a new dataclass that subclasses RunFieldSet:

from dataclasses import dataclass

from pants.core.goals.binary import BinaryFieldSet

@dataclass(frozen=True)
class BashRunFieldSet(BinaryFieldSet):
required_fields = (BashBinarySources,)

sources: BashBinarySources

Then, register your new RunFieldSet with a UnionRule so that Pants knows your binary implementation exists:

from pants.engine.rules import collect_rules
from pants.engine.unions import UnionRule

...

def rules():
return [
*collect_rules(),
UnionRule(RunFieldSet, BashRunFieldSet),
]

3. Create a rule for your logic

Your rule should take as a parameter the RunFieldSet from Step 1. It should return RunRequest, which has the fields digest: Digest, args: Iterable[str], and extra_env: Optional[Mapping[str, str]].

The RunRequest will get converted into an InteractiveProcess that will run in the foreground.

The process will run in a temporary directory in the build root, which means that the script/program can access files that would normally need to be declared by adding a files or resources target to the dependencies field.

The process's environment will not be hermetic, meaning that it will inherit the environment used by the ./pants process. Any values you set in extra_env will add or update the specified environment variables.

from dataclasses import dataclass

from pants.core.goals.run import RunFieldSet, RunRequest
from pants.core.target_types import FilesSources, ResourcesSources
from pants.core.util_rules.source_files import SourceFiles, SourceFilesRequest
from pants.engine.addresses import Addresses
from pants.engine.rules import Get, MultiGet, rule
from pants.engine.target import Sources, TransitiveTargets
from pants.util.logging import LogLevel

from examples.bash.target_types import BashBinarySources, BashSources

...

@rule(level=LogLevel.DEBUG)
async def run_bash_binary(field_set: BashRunFieldSet) -> RunRequest:
# First, we find the `bash` program.
bash_program_paths = await Get(
BinaryPaths, BinaryPathRequest(binary_name="bash", search_path=("/bin", "/usr/bin")),
)
if not bash_program_paths.first_path:
raise EnvironmentError("Could not find the `bash` program on /bin or /usr/bin.")
bash_program = bash_program_paths.first_path

# We need to include all relevant transitive dependencies in the environment. We also get the
# binary's sources so that we know the script name.
transitive_targets = await Get(TransitiveTargets, Addresses([field_set.address]))
binary_sources_request = Get(SourceFiles, SourceFilesRequest([field_set.sources]))
all_sources_request = Get(
SourceFiles,
SourceFilesRequest(
(tgt.get(Sources) for tgt in transitive_targets.closure),
for_sources_types=(BashSources, FilesSources, ResourcesSources),
),
)
binary_sources, all_sources = await MultiGet(
binary_sources_request, all_sources_request
)

# We join the relative path to our program with the template string "{chroot}", which will get
# substituted with the path to the temporary directory where our program runs. This ensures
# that we run the correct file.
# Note that `BashBinarySources` will have already validated that there is exactly one file in
# the sources field.
script_name = os.path.join("{chroot}", binary_sources.files[0])

return RunRequest(
digest=all_sources.snapshot.digest,
args=[bash_program.exe, script_name],
)

In this example, we run the equivalent of /bin/bash ./my_script.sh. Typically, your args will include the program you're running, like /bin/bash, and the relative path to the binary file. For some languages, you may use values other than the file name; for example, Pants's python_binary target has an entry_point field, and the run implementation sets args to the equivalent of python -m entry_point.

When using relative paths in args or extra_env, you should join the values with the template string "{chroot}", e.g. os.path.join("{chroot}", binary_sources.files[0]). This ensures that you run on the correct file in the temporary directory created by Pants.

Finally, update your plugin's register.py to activate this file's rules.

pants-plugins/bash/register.py
from bash import run_binary


def rules():
return [*run_binary.rules()]

Now, when you run ./pants run path/to/binary.sh, Pants should run the program.

4. Add tests (optional)

Refer to Testing rules. TODO