NaN for vertex data in debug_draw starting at 3rd frame

[nrichards]

NaN for vertex data in debug_draw starting at 3rd frame

[viblo]

The reason why it happens is that the bricks doesnt have any mass or moment set. Objects without mass (or with infinite mass) behave very strangely, just as you could imagine a 0 mass object would in real life :)

To fix it I suggest you set a mass on the brick_shape shape. For example brick_shape.mass = 1. http://www.pymunk.org/en/latest/pymunk.html#pymunk.Body.__init__

This is also the reason why I havent tried so hard to fix the NaN drawing issue, since its a very good indication of something bad/strange happening, and just hiding the error would likely lead to an even more difficult to understand error a little later.
Since you are not the first one to experience this I wonder if I should add a better error text or even outright preventing 0 or inf mass of objects added to a space.

[nrichards]

That fixes this. Thank you @viblo for the explanation and fix: Aww yiss, massless physics. !Noob alert! I'll be happily using debug_draw shortly.

Re future: perhaps "Does Shape have a reasonable mass?" + the prior ValueError exception message? Adding guidance text to the current ValueError could be useful for a noob like me. This usage-error occurred when I naively, trivially set a KINEMATIC to be DYNAMIC.

Q: Where would you prefer it to be in the call-stack? ext_cpSpaceDebugDrawPolygonImpl ? According to comments on StackOverflow, an untriggered try: is low-cost, performance wise.