AUTOMATION TO BUILD THE PROJECT ANYWHERE:
As we are always working in the different labs and changing machines or even showing demos at any random pcs, it's always great to have an automated way to set up the project.
The engine currently supports Linux and Windows builds.
Premake:
We used premake scripting for the engine (premake ). Its a bit simpler than CMake and quicker to get things off the ground.
Python Scripts:
The project uses a lot of third party APIs and libs and they are necessary for the project compilation. Also we have a lot of assets which for obvious reasons we don’t keep on git repo.
It is cumbersome to download them and place them in an appropriate folder manually every time you are on a new machine or if you are in a different folder for testing.
So we created a small python utility which would do this for us. It downloads all the relevant files from drive, unzips them in proper directory structure and cleans up after itself.
AZURE PIPELINES:
The team takes pull requests seriously and tries to triple check anything that goes into master. But there are still possibilities of human errors and that's where Azure devops pipelines come in.
Its very easy to setup a pipeline with Azure and it works perfectly for the current requirements of the project.
Currently we are able to fire a build on creation of Pull requests and it’s easier to track the failures and with azure’s intuitive UI and logging mechanism. Following screenshots should give you the idea,
Currently we are only checking for build sanity with pipelines. Team will add unit tests incrementally and that’ll be reflected on the azure pipeline side as well.
Unfortunately, As our project is private, Azure gives us only one pipeline. So team is considering option of Github actions or dockers for the Linux.
Kommentare