We are so close to having the build process finished. It has presented a fair share of issues and has forced me to learn a lot of new things that I didn't think I would have to when I first began this journey. While we do have the vast majority of the build and unit test process finished for the GSource branch, we're still missing what is arguably the most important part.
We got the whole build and process set up so that we now have to project being created, built, deployed, installed, and launched.
The meat of what's going on is happening here. The creation of the project in the build phase is pretty standard. compared to the platforms, but the main differences are here in the testing portion. Firstly, we are navigating into the build folder so that we have access to the solution file. We then use Push-Location to save the directory that we are currently in. We do this because the next line of code opens up the Visual Studio Developer Powershell, and doing so changes our working directory to some other random directory. But we are then able to use Pop-Location to get us back into the directory we had saved. We then use the devenv command to deploy the project (it also rebuilds the project), which will generate the AppX folder. After this point, we thought we would then have to take the AppX folder and compress it into a .appx package and then use the package to install the application onto the machine. But through some testing, I discovered that the deploy command actually goes ahead and installs the application onto your machine. And so then we were able to skip those two other steps, which I'm very happy about since that install step was giving me a lot of trouble. Finally, the bottom command then goes ahead and launches the application.
The problem we are having is that after the application is launched, the yml just ends there. On the other platforms, the Unit Test is a console application, so when it launches the executable, it opens in the command line, and it can't continue until the unit test completes. It is also then able to see whether or not all the unit tests have passed. But UWP isn't able to run as a console application, so the yml doesn't wait for it to finish. So right now, the unit tests only pass and fail based on whether or not the application is able to be launched, not on the individual unit tests.
So now we have to find a way to not only get the yml to wait for the application to go through all the unit tests but also be able to read the unit tests and determine when and which unit test fails. The Start-Process has several commands that are actually perfect for what we are looking for it to do, but unfortunately, they all seem to be incompatible with UWP applications. We are so close to the end; I can taste it. But this is looking to be one of the more major issues and could potentially take some time to complete.