Hello, so I am trying to setup an automated test suite that will run on the standalone client on a build machine.
For tests that are triggered from CPP I got those working decently how ever following this guide
I am running into a few issues…
As you can see adding the tests to the map adds it to the session frontend (which should mean I can run these from CMDLine, Sweet!) But when I actually run it, the actual tests does not start, it does however open the map…
in fact the only way to trigger the tests is to call the Run All Functional Tests node in blueprint. Hooking this node up to something like BeginPlay seems like a terrible idea.
I expected that the session frontend would pick up on these tests or if its not meant to be used this way and correctly launch even the functional tests (blueprint) tests correctly or not display it in the view…
So, this leaves me with some questions, What is the correct workflow for this? Because this is driving me nuts for a few reasons
- I guess I can write a separate test from CPP code and load the map manually and then call the Run All Functional Tests from CPP but this seems like a bit of a workaround perhaps this is the correct flow?
If this is a broken flow, the documentation shouldn’t advertise it as “working” but that is a separate issue.
-
I am not convinced placing tests in the level blueprint will strip these in shipping based on the PP defines in FunctionTest.h is this to be expected?
-
Has anyone managed to fully integrate this into a production pipeline because right now the tools as I understand them makes very little sense?