GitHub commands

What the connection gives you and what each button does.

Three buttons move your project between three places: this server, your repository on GitHub, and the live application your visitors see. Here is what each one does and in which direction.

Push — from this server to GitHub

Sends what exists here to the repository. Used once at the start, to fill an empty repository with the starting point, and afterwards only when work was done in the panel itself.

Pull — from GitHub to this server

Brings in what you pushed from your own machine. Content and pages apply at once; code still needs a deploy to become the running application.

Deploy — from files to the running application

Compiles the application and restarts it. It does not touch the repository at all: it turns the files that are already here into what your visitors see.

Files travel through git — data never does.

Rows, uploaded files and vectors live on the server and are shared with your local copy while you develop. They are not in the repository, so a clone does not carry them and a push does not send them. The trap is that settings changed in this panel are FILES: your local copy will not know about them until you pull.

The rule that prevents conflicts: use one place at a time. Start every session with Pull, finish every session with Push. Working in the panel and on your machine at the same time is the one common way to create a conflict here.

If a conflict happens: you will be told in plain terms. Take the message to your machine, resolve it there with your agent's help, then push and pull as usual.