Overview

meta is a tool for managing multi-project systems and libraries. It answers the conundrum of choosing between a mono repo or many repos by saying "both", with a meta repo.

File at the root of the repo call .meta:

{
  "projects": {
    "path/to/module": "<https://github.com/insight-infrastructure/repo>",
....
    "path/to/module2": "<https://github.com/insight-infrastructure/repo2>",
	}
}

A meta repo is a repo consisting of many other nested sub repos. You can have meta repos inside other meta repos and to help with that, we built a little tooling to allow cross cutting changes to each repo when you make git pushes.

How it works under the hood is it clones the repo into your repo and then puts an entry into your gitignore to ignore the folder path. Pretty straight forward but what is nice is that you can then iterate over those directories and perform git actions for cross cutting changes.

Why we use it

• meta is a great tool to break up a repo into smaller pieces
• We can use pieces of code in a variety of contexts in multiple places
  • It is basically like a submodule but easier to work with
• We are in a situation where we need to present fellows with production level systems but don't want to inundate them with a ton of production code.
• With meta, we can break up our infrastructure into smaller pieces and allow users more freedom to stitch together the repo representation that they want
• With all of our code in smaller repos, we can do more targeted testing by testing when changes happen to specific parts instead of having to do a diff on a large repo with a whole bunch of logic to trigger the appropriate test
  • Large monorepos need specific tooling to make things like this happen properly

How we use it