We’ve talked about having both a test-a-thon and a doc-a-thon to get tests and documentation in shape for the upcoming tutorials at KHQ in July.
One thing that would be interesting would be to have the tests generate images for the documentation (or at least using the testing framework to generate images or even movies for the documentation). Ken Martin has done some work that exposes portions of the Qt UI to ParaView inside virtual reality; we could probably use the same technique to grab images or movies of Qt XML “tests” for the documentation.
By making portions of the documentation contribute to test results, we could catch at least some places where the documentation becomes out-of-sync with the implementation. It would also reduce the manual effort required to keep figures up-to-date as the modelbuilder UI changes.
Would that require us to be on preview’s master branch?
Ken mentioned that currently it includes displaying parts of the PV GUI directly in VR. Wondering how much needs to be done to show CMB specific panels
I assume by “preview” you mean “ParaView”? If so, then yes. This is definitely for after that change (although I think that will be coming shortly after the tutorial in July, yes?
Assuming OpenGL widgets are handled properly, very little work… if you look at the classes in Ken’s MR, you just pass in QWidget and it generates a vtkImage texture. So we could try passing in QMainWindow. You are correct that if we want to show dialogs (file dialogs, popups, etc.) that would be more work. But it may even be possible to grab those without much work since the QtTesting framework knows when events are being passed to widgets not in the main-window hierarchy.