How to set up your development environment#

With `devcontainer` in VS Code#

The development workflow is streamlined for Linux thanks to the devcontainer configuration bundled in this repository. You’ll need Docker and VS Code with the Dev Containers extension:

Clone conda/conda and conda/conda-libmamba-solver to your preferred locations (e.g. ~/devel/conda and ~/devel/conda-libmamba-solver, respectively). The location does not matter as long as both repositories have the same parent directory.
Open your conda-libmamba-solver clone with VS Code.
Connect to the DevContainer image via the bottom-left menu (❱❰) and click on “Reopen in Container”. Pick one of the suggested configurations: conda-forge or defaults. The only difference is the base installation (Miniforge and Miniconda, respectively).
The image will be built and after a couple minutes, you’ll be dropped into a Bash shell. Enjoy! Since the local repositories are mounted, you can make modifications to the source live, and they will be reflected in the Docker instance automatically. Run pytest or conda as needed, no need to reload Docker!
If the development environment breaks, click again on ❱❰ and, this time, choose “Rebuild container”. You might need to Retry a couple times.

Note

Developing libmamba The devcontainer configuration also supports libmamba 1.x development. You just need to have the mamba-org/mamba repository (branch 1.x) cloned next to conda and conda-libmamba-solver. Once the container has started, run develop-mamba to set it up. If you are modifying C++ sources, re-run develop-mamba to rebuild the libraries.

With regular Docker#

You can reuse the devcontainer scripts with regular Docker too.

Clone conda/conda and conda/conda-libmamba-solver to your preferred locations (e.g. ~/devel/conda and ~/devel/conda-libmamba-solver, respectively). You should also clone mamba-org/mamba if you need to develop libmamba. The location does not matter as long as all repositories have the same parent directory.

Start a new Docker instance with this command. Adjust the local mounts as necessary.

# For defaults-based images, use:
$ docker run -it --rm \
   -v ~/devel/conda:/workspaces/conda \
   -v ~/devel/mamba:/workspaces/mamba \
   -v ~/devel/conda-libmamba-solver:/workspaces/conda-libmamba-solver \
   continuumio/miniconda3:latest \
   bash
# For conda-forge-based images, use the following instead:
$ docker run -it --rm \
   -v ~/devel/conda:/workspaces/conda \
   -v ~/devel/mamba:/workspaces/mamba \
   -v ~/devel/conda-libmamba-solver:/workspaces/conda-libmamba-solver \
   condaforge/miniforge3:latest \
   bash

Run the post_create and post_start scripts:

$ bash /workspaces/conda-libmamba-solver/.devcontainer/post_create.sh
$ bash /workspaces/conda-libmamba-solver/.devcontainer/post_start.sh

If you want to develop with mamba in editable mode, run:
```
$ source ~/.bashrc
$ develop-mamba
```

General workflow#

We strongly suggest you start with the Docker-based workflow above. It is a better development experience with a fully disposable environment. However, sometimes you might need to debug issues for non-Linux installations. In that case, you can follow these general instructions, but be careful with overwriting your existing conda installations, especially when it comes to shell initialization!

Get yourself familiar with the “Development environment” guide for conda itself.
Fork and clone the conda-libmamba-solver repository to your preferred location:

$ git clone "git@github.com:$YOUR_USERNAME/conda-libmamba-solver" "$REPO_LOCATION"
$ cd "$REPO_LOCATION"

Install the required dependencies for conda-libmamba-solver:

$ conda install \
    --file "$REPO_LOCATION"/dev/requirements.txt \
    --file "$REPO_LOCATION"/tests/requirements.txt

Install conda-libmamba-solver with pip:

$ cd $REPO_LOCATION
$ python -m pip install --no-deps -e .

Debugging `conda` and `conda-libmamba-solver`#

Once you have followed the steps described in the general workflow above you may need to investigate the state in a particular point. Insert a breakpoint() within the code and run a test or conda directly to hit the breakpoint.

Debugging Mamba#

While debugging the conda workflows only requires modifying python code and running conda. Debugging the mamba code requires recompilation and is not as easy to jump into a debugger to investigate state.

Get familiar with the “Local development” guide for mamba itself.
Fork and clone the mamba repository to your preferred location:

$ git clone "git@github.com:$YOUR_USERNAME/mamba" "$REPO_LOCATION"
$ cd $REPO_LOCATION

Use the Docker image for development suggested above and re-run develop-mamba whenever you make change to mamba in $REPO_LOCATION. This should take less than a minute.

We recommend debugging via either breakpoints and using gdb or print statements via std::cout << ... << std::endl. The following shows an example of inserting print statements into the libmamba source in order to debug the libsolv state.