Building a package locally

In this tutorial, we will clone an existing feedstock and build a package for it locally. Building locally is often a convenient way to quickly test your changes.

The tutorial assumes that you are running a Linux or macOS system, and have Docker, Git and Python installed.

Clone the feedstock

Once a new recipe is accepted for conda-forge, a new feedstock repository is created for it. All subsequent work on the recipe happens in that repository. Let's start by cloning the xmltodict-feedstock repository at a commit corresponding to version 1.0.4, build 0.

Enter a directory of your choice, clone the repository and then enter it:

git clone --revision=728537f91186b70e7ac27cc7d5c036c82486d942 https://github.com/conda-forge/xmltodict-feedstock/
cd xmltodict-feedstock/

Build the recipe

Feedstocks provide a convenient build-locally.py script that can be used to build recipes locally. Start it:

python build-locally.py

Normally, the script would ask you to choose a configuration to build at this point. However, pure Python packages such as xmltodict have only one configuration, so the script proceeds immediately. It will pull the conda-forge Docker image, use it to provision all the build tools and then will build the package, while outputting the build process to the console.

Alternatively, for feedstocks having conda_install_tool: pixi in conda-forge.yml, you may use pixi run build instead.

Near the end of the process, output validation will be performed and the packages produced will be printed as part of that, for example:

validation results:
{
  "noarch/xmltodict-1.0.4-pyhcf101f3_0.conda": true
}

The built package will be found at the appropriate subdirectory of build_artifacts:

$ ls -lh build_artifacts/noarch/
total 28K
-rw-r--r-- 1 user user 461 02-23 16:33 repodata.json
-rw------- 1 user user 21K 02-23 16:33 xmltodict-1.0.4-pyhcf101f3_0.conda

Clean up

Once you are done working with a feedstock, you want to remove the build artifacts, as they can take a lot of space:

rm -r build_artifacts/

You may also want to clean up leftover Docker containers and images. However, note that this may also remove other stopped containers and unused resources (which may matter if you are using Docker for other purposes):

docker system prune

This will leave the downloaded Linux image used for the builds, though. To remove all images, use:

docker system prune -a