* More warning fixes for physics and math

* fix a crash introduced with the warning changed.

Co-authored-by: Jeffery Myers <JefMyers@blizzard.com>

Not exposing some data structures

* explained a bit more the core_window_letterbox example

* fixed a few 'ups' moments that could lead to mild head pain and time loss

Now it's required on Windows if not using a busy wait loop

Generating the projects using CMake, targeting VS2017, results in an error when compiling.
This is due to physac.h trying to make a 'vertices' array of size 'int count', making it const does not work, either.

This changes the static declaration to a malloc/free combo.

Tested using the physics-demo.

Feature macros need to be defined before #including any headers,
preferably through the build system, but this is good enough.

Fixes a compile error on my fork's Travis CI.

Physac wasn't working on OS X.  It looks like the necessary timing code wasn't being compiled in by the preprocessor.

fixes #340

If we compile with c99 without gnu extensions (gnu99) we need this
define, to have CLOCK_MONOTONIC and similar macros available

Working in modules configuration flags...

Signed-off-by: Saggi Mizrahi <saggi@mizrahi.cc>

and remove static from PhysicsThread().

and added PHYSAC_NO_THREADS define. Improved physac example drawing
frames per second in screen.

... and now physac thread creation is done in InitPhysics() and it is
destroyed in ClosePhysics(). User just need to call these functions to
use physac module.

Old physics update system used a static number of steps to calculate
physics (450 for desktop and 64 for android). It was too much and it was
limited by target frame time...

Now physics update runs in a secondary thread using a fixed delta time
value to update steps. Collisions are perfectly detected and resolved
and performance has been improved so much.

To avoid additional dependencies

Finally, physics update is handled in main thread using steps to get
accuracy in collisions detection instead of moving it to a new thread.

Examples are finished as simple and clear as I could. Finally, physac
module is MORE simpler than in the first version, calculation everything
by the same way for both types of physic objects.

I tryed to add rotated physics a couple of times but I didn't get
anything good to get a base to improve it. Maybe for the next version...

No bugs or strange behaviours found during testing.