This directory contains everything needed to build the local git2cpp source code as an
Emscripten-forge package and create a local JupyterLite deployment
that can run it in a terminal.
It works on Linux and macOS but not Windows.
It uses a separate micromamba environment defined in deploy-environment.yml. To set this up use:
micromamba create -f deploy-environment.yml
micromamba activate git2cpp-deployThen to build git2cpp and create the JupyterLite deployment use:
makeThis performs the following steps:
- Clones the
emscripten-forge/recipesrepository. - Modifies the
git2cpprecipe (usingmodify_recipe.py) to build from the localgit2cppsource code in../src/. - Builds the package from the recipe using
pixi. - Builds the JupyterLite deployment in the
dist/directory using the locally-built package.
The built package will be in the em-forge-recipes/output/emscripten-wasm32 directory with a name
something like git2cpp-0.0.3-h2072262_3.tar.bz2. If you compare this with the latest
Emscripten-forge package on https://prefix.dev/channels/emscripten-forge-dev/packages/git2cpp,
the local package should have the same version number and the build number should be one higher.
To serve the JupyterLite deployment use:
make serveand open a web browser at the URL http://localhost:8080/. Start a terminal and run the
cockle-config command (typing coc, then the tab key then enter should suffice). Amongst the
displayed information should be the git2cpp package showing that it is from a local directory
such as file:///something-or-other/git2cpp/lite-deploy/em-forge-recipes/output rather than from
prefix.dev such as https://repo.prefix.dev/emscripten-forge-dev.
After making changes to git2cpp source code, to rebuild the package and deployment use:
make rebuildand then re-serve using:
make serve