is a repository of dummy projects, that use supported testing frameworks and build tools to bring it all together, and demonstrate how TLB can be used on a typical project. Its meant to highlight the hookup points and allow users to play with environment variables even before they hook-up TLB with their project's build.
Usually there is one top-level shell-script in these codebases. This shell-script sets necessary environment variables to instruct TLB process to partition tests into N
-partitions(most if not all projects use 2 partitions, but you can tweak it to see how it behaves with different numbers), and runs those partitions one after another(serially).
Obviously this is not what would happen on a real build, it is only meant to demonstrate how tests are partitioned and reordered. On a real build/CI setup, all partitions would start at the same time, run on(well, ideally) different machines, and finish almost at the same time.
Its usually a good idea to run the top level script a few times just to see how the balancing during the very first build is suboptimal(as it is forced to use count-balancing, because of lack of historical data) and how it starts balancing really well next time onwards(once it has test times populated as feedback from the first build, it used time balancing).
A very useful repository to get a first hand feel of how TLB works, and to borrow environment variable configuration from.