The Mill Evaluation Model

Evaluating a Mill task typically goes through the following phases:

  1. Compilation: Mill compiles the build.mill to classfiles, following the The Mill Bootstrapping Process to eventually produce a RootModule object

  2. Resolution: Mill resolves the list of Tasks given from the command line, e.g. resolve or foo.compile or {bar,qux}._.test, to a list of concrete Task objects nested on Modules within the RootModule along with their transitive dependencies

    • In the process, the relevant Mill Modules are lazily instantiated

  3. Evaluation: Mill evaluates the gathered Tasks in dependency-order, either serially or in parallel

Limitations of the Mill Evaluation Model

This three-phase evaluation model has consequences for how you structure your build. For example:

  1. You can have arbitrary code outside of Tasks that helps set up your task graph and module hierarchy, e.g. computing what keys exist in a Cross module, or specifying your def moduleDeps

  2. You can have arbitrary code inside of Tasks, to perform your build actions

  3. But your code inside of Tasks cannot influence the shape of the task graph or module hierarchy, as all Resolving and Planning happens first before any Tasks are evaluated.

This should not be a problem for most builds, but it is something to be aware of. In general, we have found that having "two places" to put code - outside of Tasks to run during Planning or inside of Tasks to run during Evaluation - is generally enough flexibility for most use cases.

The hard boundary between these two phases is what lets users easily query and visualize their module hierarchy and task graph without running them: using inspect, plan, visualize, etc.. This helps keep your Mill build discoverable even as the build.mill codebase grows.

Caching at Each Layer of the Evaluation Model

Apart from fine-grained caching of Tasks during Evaluation, Mill also performs incremental evaluation of the other phases. This helps ensure the overall workflow remains fast even for large projects:

  1. Compilation:

    • Done on-demand and incrementally using the Scala incremental compiler Zinc.

    • If some of the files build.mill imported changed but not others, only the changed files are re-compiled before the RootModule is re-instantiated

    • In the common case where build.mill was not changed at all, this step is skipped entirely and the RootModule object simply re-used from the last run.

  2. Planning:

    • If the RootModule was re-used, then all previously-instantiated modules are simply-re-used

  3. Evaluation:

    • Tasks are evaluated in dependency order

    • Targets only re-evaluate if their input Tasks change.

    • Task.Persistents preserve the T.dest folder on disk between runs, allowing for finer-grained caching than Mill’s default task-by-task caching and invalidation

    • Task.Workers are kept in-memory between runs where possible, and only invalidated if their input Tasks change as well.

    • Tasks in general are invalidated if the code they depend on changes, at a method-level granularity via callgraph reachability analysis. See #2417 for more details

This approach to caching does assume a certain programming style inside your Mill build: we may-or-may-not re-instantiate the modules in your build.mill and we may-or-may-not re-execute any particular task depending on caching, but your code needs to work either way. Furthermore, task defs and module `object`s in your build are instantiated lazily on-demand, and your code needs to work regardless of which order they are executed in. For code written in a typical Scala style, which tends to avoid side effects, this is not a problem at all.

One thing to note is for code that runs during Resolution: any reading of external mutable state needs to be wrapped in an interp.watchValue{…​} wrapper. This ensures that Mill knows where these external reads are, so that it can check if their value changed and if so re-instantiate RootModule with the new value.

The Mill Bootstrapping Process

Mill’s bootstrapping proceeds roughly in the following phases:

  1. If using the bootstrap script, it first checks if the right version of Mill is already present, and if not it downloads it to ~/.mill/download

  2. It instantiates an in-memory MillBuildRootModule.BootstrapModule, which is a hard-coded build.mill used for bootstrapping Mill

  3. If there is a meta-build present mill-build/build.mill, it processes that first and uses the MillBuildRootModule returned for the next steps. Otherwise it uses the MillBuildRootModule.BootstrapModule directly

  4. Mill evaluates the MillBuildRootModule to parse the build.mill, generate a list of ivyDeps as well as appropriately wrapped Scala code that we can compile, and compiles it to classfiles

  5. Mill loads the compiled classfiles of the build.mill into a java.lang.ClassLoader to access it’s RootModule

Everything earlier in the doc applies to each level of meta-builds in the Mill bootstrapping process as well.

In general, .sc files, import $file, and import $ivy can be thought of as a short-hand for configuring the meta-build living in mill-build/build.mill:

  1. .sc and import $file are a shorthand for specifying the .scala files living in mill-build/src/

  2. import $ivy is a short-hand for configurin the def ivyDeps in mill-build/build.mill

Most builds would not need the flexibility of a meta-build’s mill-build/build.mill, but it is there if necessary.

Mill supports multiple levels of meta-builds for bootstrapping:

  • Just build.mill

  • One level of meta-builds: mill-build/build.mill and build.mill

  • Two level of meta-builds: mill-build/mill-build/build.mill, mill-build/build.mill and build.mill

The Mill Meta Build works through a simple use case and example for meta-builds.