Monster Software Repositories

by Martin Monperrus

What are the biggest software repositories ever made?

We are used to hear that we build software systems with millions of lines of codes, blablabla. Yet, this is a very abstract statement, with no clear meaning since no big software system has clear boundaries. To that extent, the "million lines" is intangible, unreal.

Yet, there are things that we can tangibly observe and measure, such as the number of commits in a software repository. We can identify the biggest software repositories ever made. This is very concrete, one can measure them, one can even browse their list of commits.

linux

This post lists the biggest public software repositories to date. It seems that the biggest repositories produced by humanity so far (Feb 2019) are in the order of "hundreds of thousands" commits (>10 6 ). (and 10 6 is actually the threshold in this table).

Table 1. Some of the biggest code repositories ever in the open-source world (data from Jan 2019)
Software Commits Contributors
https://github.com/torvalds/linux 798710    unknown
https://github.com/chromium/chromium 744581    unknown
https://github.com/mozilla/gecko-dev 631802 3851
https://github.com/LibreOffice/core 433945 853
https://github.com/WebKit/webkit 208041    unknown
https://github.com/Homebrew/homebrew-core 135248 7310
https://github.com/NixOS/nixpkgs 166699 1935
https://github.com/odoo/odoo 122698 873

Of course, one can have arbitrary big monorepos, and some companies (notably Google) maintain gigantic monorepos. Yet, this list gives a sense of where we are today: we are moving fast towards one million commits in monster software repositories.

Do you know other monster repos? Drop a comment below :-)

–Martin Monperrus

Stockholm, Feb 2019

Tagged as: