Split Player Camera from Debug Camera and Restore G3D Scene Editor Functionality

[bboudaoud-nv]

We have been using activeCamera() for our player camera, which when defaulting to a debug camera removes the ability to use the developer window/F2 to move to a free-moving camera in the scene. Here we add in a playerCamera intended to split this functionality again.

All player-based actions (aside from waypoint editing) now use the player camera, leaving the debug/additional cameras free for future implementation.

Merging this closes #116.

[bboudaoud-nv]

Once complete, this should resolve issue #116. Still some work to do in getting a FirstPersonManipulator to control the debug camera.

[bboudaoud-nv]

Upon discussion w/ @morgan3d I think we may not be using the FirstPersonManipulator as intended. It seems we should consider making some changes here/that G3D may make some changes to make it more explicit the default FirstPersonManipulator provided by GApp should only be used with the debug camera.

This (likely) means we should be constructing our own manipulator to use in order to switch between mouse-as-cursor and mouse-as-camera control mode.

[jspjutNV]

Oh, okay, so we should do a FirstPersonManipulator::create() call somewhere and keep that along side the default one provided by GApp? I assume there should also be changes made to the input handling to change the routing of input depending on which camera/manipulator is active.

[jspjutNV]

It's been a while since we've seen an update on this pull request. Should we include this in the v20.06.01 release milestone, or is there still enough work to do here that we defer this one?

[bboudaoud-nv]

This issue should be closed. I can look into creating a new FirstPersonManipulator to avoid overriding the debug camera one.

On second thought I may need some assistance with understanding how this should be done. I believe "just" creating a new FirstPersonManipulator won't do what we are looking for here.

[jspjutNV]

This latest merge broke the testing. We should make sure there's no regression on testing before we consider this pull request ready.

[bboudaoud-nv]

I got the FPM moving the debug camera (we didn't copy over this line from the GApp::onSimulation() call into our FPSciApp::onSimulation() (which doesn't call GApp::onSimulation(), but instead reproduces a subset of the calls from within it).

This PR can now be reviewed.

[jspjutNV]

The same 4 tests that are passing in master are passing again with the latest commit.

[jspjutNV]

I believe this is ready to merge with the exception of the crash that is now documented in #177 . This bug probably isn't added by this pull request and instead is likely a bug that was always present but wasn't visible because the larger bug this pull request solves was hiding it.