Mozilla Version Control Tools latest.
Initially, a one-off server to run the headless repositories will be created.
Clients still push to the repository as before.
Testing Ruby's CGI
A pointer to this external data is stored, allowing the repository to serve up this data to clients that request it.
A headless repository is conceptually a single repository with thousands, but with repository data stored outside the repository itself.
Description: This approach was simple and worked for a long time without significant issues.