Develop features across your whole stack
Clone your entire development system for every feature. Every repository, the database with its data and all services, each clone on its own machine.
Feature branches stop at the repository edge.
A real feature touches the API, the frontend and the schema at once. git worktree gives you a second checkout of one repository, but the database, the other repositories, the installed dependencies and the ports are still shared. Switching features means stashing, migrating back and forth and restarting services.
How it works
- 01
Build the baseline
Put the repositories, toolchains, a database with seed data and the service configuration together in one environment. Take a snapshot when it is ready.
- 02
Clone per feature
Each new feature gets a clone of that snapshot, up and running in under half a minute, with the same files and data and ports of its own.
- 03
Work across the stack
Create a branch in each repository, change API, frontend and schema together, and run the full stack on its usual ports, next to other clones doing the same.
- 04
Keep or discard
Push the branches you want and delete the clone. An approach that went nowhere is gone with it, and the baseline stays untouched.
Like git worktree, but for the whole machine.
Good to know
- A clone copies disks, not running processes. Services start fresh in the clone.
- Cloning does not create or sync git branches. You branch and pull in each clone as usual.
- Clones are independent. Changes in one do not reach the others or the baseline, and clones do not share a network with each other.
- Every clone counts against your account's CPU, memory and storage limits.
