PDM Multirun

A PDM plugin to run a command on multiple Python versions.

:warning: Looking for maintainers. I do not use PDM nor this plugin nowadays, and don't have the time to fix CI failures or upgrade the code base to support new versions of PDM. If you wish to be added as a collaborator on this project in order to help maintaining it, or if you wish the repository to be transfered to you, please open an issue or discussion, with enough information about you for me to trust you :wink:

Installation

With pipx:

pipx install pdm
pipx inject pdm pdm-multirun

With PDM:

pdm self add pdm-multirun

As a local-only plugin:

# pyproject.toml
[tool.pdm]
plugins = [
    "pdm-multirun",
]

pdm install --plugins

Usage

This plugin adds a multirun command to PDM. The command accepts the same parameters as the run command, with an additional -i, --interpreters, --versions parameter that allows to specify the interpreters to use.

pdm multirun pytest tests/

To specify interpreters, pass a comma-separated string of Python versions:

pdm multirun -i 3.10,3.11 pytest tests/

If you use virtual environments instead, pass their names to the --interpreters option and add the -e, --venvs flag:

pdm multirun -ei 3.10,3.11 pytest tests/

pdm multirun -ei tests38,tests39 pytest tests/

You can set PDM Multirun to use virtual environments by default by setting the PDM_MULTIRUN_USE_VENVS environment variable to 1.

By default, PDM Multirun reads Python versions (or venv names) from the PDM_MULTIRUN_VERSIONS environment variable. It is a string of {major}.{minor} versions (or venv names), separated by spaces, that can be found and called by PDM.

export PDM_MULTIRUN_VERSIONS="3.8 3.9 3.10 3.11 3.12"
pdm multirun pytest tests/

export PDM_MULTIRUN_VERSIONS="tests38 tests39 tests310"
pdm multirun pytest tests/

PDM Multirun sets a number of environment variables that can be used by code run in each version.

• PDM_MULTIRUN set to 1 whenever PDM Multirun is being used.
• PDM_MULTIRUN_CURRENT is set to the name of the current interpreter or virtual environment (such as that passed using -i or PDM_MULTIRUN_VERSIONS).

You can use these variables, for example, to output metadata about the current Python version or interpreter, like in the example below, which if invoked by PDM Multirun, would start by printing the name of the virtual environment or interpeter, and the version of Python being used.

script.py

import os
import sys

MULTIRUN = os.getenv("PDM_MULTIRUN", "0") == "1"

if MULTIRUN:
    int_name = os.getenv('PDM_MULTIRUN_CURRENT', '')
    py = f"{int_name}: {sys.version_info[0]}.{sys.version_info[1]}"

    print(f"{py} - Hello from python! 👋")

# continue script as required...

In a scenario where you had two virtual environments, tests38 and tests39, and saved this script in a file named example.py, running the command:

pdm multirun -e tests38,tests39 python example.py

Would output the text below.

tests38: 3.8 - Hello from python! 👋
tests39: 3.9 - Hello from python! 👋

PDM Multirun successively runs the pdm use then pdm run internal actions. By default, if PDM cannot "use" an interpreter/venv, it continues with the next.

# will continue with 3.8 even if 3.7 is not available
pdm multirun -i 3.7,3.8 pytest tests/

You can tell it to fail instead with the -f, --fail-fast flag:

# will stop at 3.7 if it's not available
pdm multirun -fi 3.7,3.8 pytest tests/

If the command you run fails on a Python version, PDM Multirun stops there. In any case, PDM Multirun will restore the Python interpreter saved in .pdm-python (through the pdm use command) before exiting.