In this example, we will walk through a simple monorepo setup that can be used with Earthly. The entire code of this exercise is available in the examples/monorepo directory on GitHub.

In this example, let's assume we've organized our monorepo such that each root-level directory is a sub-project.

├── proj1
└── proj2

As such, each sub-project might have its own self-contained Earthfile, specific to the programming language and setup of the sub-project. A root-level Earthfile can tie everything together or surface important targets of the sub-projects.

├── Earthfile
├── proj1
│ └── Earthfile
└── proj2
└── Earthfile

Here is an example of a possible root-level Earthfile file, where an all target simply calls the +docker target of each sub-project.

# Earthfile
FROM alpine:3.11
BUILD ./proj1+docker
BUILD ./proj2+docker

Note that the directory hierarchy may be as vast and deeply-nested as appropriate for your setup. In addition, build targets within projects may depend on targets from other projects. As an example, consider the case where one project builds a library and another takes the built library as an artifact and imports it in order to use it in its own build.

Further, through the use of caching, the build setup is able to infer automatically which sub-projects to rebuild because of local changes, and which ones to reuse cache for.

To review this example with its complete code, check out the examples/monorepo directory on GitHub.