Skip to content

Latest commit

 

History

History
94 lines (77 loc) · 4.32 KB

docker.md

File metadata and controls

94 lines (77 loc) · 4.32 KB
Raw

Docker

Docker? A TL;DR

Docker is a system that lets you configure a small virtual computer (called a container) through a file (Dockerfile) that can be used to run programs in a deterministic environment on any machine.

This helps solve the problem of "How do I write something that works an MacOS, Windows, and all the flavors of Linux?!"

Setup

  1. Download & Install the Docker Desktop platform for your machine: https://www.docker.com/
  2. Once installed build the container for TGO in a shell from within the TGO repo checkout: 
    ~/Projects/TGO (envy-format-ci) $ docker build -t tgo:check -f support/Dockerfile .
    The initial build will take quite a while because it needs to:
    1. download a full linux install
    2. update the system
    3. install git, python, pip, and the gdscript toolkits

Usage

Once the initial setup is done you should be able to run the GDscript tooling through docker. Run the following commands from the root of your TGO checkout...

gdlint

$ docker run -v .:/TGO/repo -ti tgo:check ./gdlint.sh
Success: no problems found

gdformat

$ docker run -v .:/TGO/repo -ti tgo:check ./gdformat.sh
# the above works in windows powershell, for mac os / linux you may need:
# docker run -v "$(pwd)":/TGO/repo -ti tgo:check ./gdformat.sh
would reformat ./repo/Scripts/Components/Character.gd
would reformat ./repo/Scripts/Components/Detectable.gd
would reformat ./repo/Scripts/Components/Lamp.gd
would reformat ./repo/Scripts/Devin.gd
would reformat ./repo/Scripts/Enums.gd
would reformat ./repo/Scripts/Utils.gd
would reformat ./repo/ZZ_Scratch/Movement/CameraZoom2D.gd
would reformat ./repo/ZZ_Scratch/Movement/GrayBox.gd
would reformat ./repo/ZZ_Scratch/Movement/SpecialTree.gd
would reformat ./repo/ZZ_Scratch/Movement/TestMovement.gd
10 files would be reformatted, 0 files would be left unchanged.

If you'd like to apply the changes you can run it with --apply.

⚠️ Note that you should always commit your work before reformatting in the event the changes break something.

$ docker run -v .:/TGO/repo -ti tgo:check ./gdformat.sh --apply
# the above works in windows powershell, for mac os / linux you may need:
# docker run -v "$(pwd)":/TGO/repo -ti tgo:check ./gdlint.sh
reformatted ./repo/Scripts/Components/Character.gd
reformatted ./repo/Scripts/Components/Detectable.gd
reformatted ./repo/Scripts/Components/Lamp.gd
reformatted ./repo/Scripts/Devin.gd
reformatted ./repo/Scripts/Enums.gd
reformatted ./repo/Scripts/Utils.gd
reformatted ./repo/ZZ_Scratch/Movement/CameraZoom2D.gd
reformatted ./repo/ZZ_Scratch/Movement/GrayBox.gd
reformatted ./repo/ZZ_Scratch/Movement/SpecialTree.gd
reformatted ./repo/ZZ_Scratch/Movement/TestMovement.gd
10 files reformatted, 0 files left unchanged.

Command explanation

If you want we can take a second to understand what's being run:

  • docker — this tells docker that we're about to request it to take some action with a container (virtual computer)
  • run — the command that we want docker to do is "run" something with the container
  • -v .:/TGO/repo — a special directory in the container has been declared as a VOLUME. This means we can configure, at execution, a directory from our computer to be reflected within the container.
    The -v flag is saying "map the first directory to the second in the container." In this case . is "the current directory" and /TGO/repo is telling docker that we should make the contents of the current directory available at /TGO/repo
  • -ti — not strictly necessary but if you wanted to run bash or anything that required interactivity you would need this so I'll often include it out of habit
  • tgo:check — This is the name of the container to use. It should match the name given in docker build
  • ./gdlint.sh, ./gdformat.sh — this is the command to run in the container. In this case these commands are specific to our TGO project and you can find what they do in gdlint.sh and gdformat.sh. You can see where they get copied into the container in the Dockerfile.