Small changes to Py5Vector set_mag() and normalize() to make it more consistent with PVector & Processing

[hx2A]

    > In Processing you can set the magnitude of a vector to a negative scalar... this will "flip" the heading...

That's interesting! Right now the Py5Vector will reject a negative value for the set_mag() method. I think we should allow negative values, like Processing does. This is an easy change.

Also Processing "fails silently" to set magnitude to a "null vector". Should we follow that?

I did some experimenting and I see that this is valid Processing code:

PVector v = new PVector();
v.normalize();
println(v);
// [ 0.0, 0.0, 0.0 ]

py5 would reject the similar code, not allowing you to normalize a vector with a magnitude of zero. This does bother me a bit, but being consistent with Processing is probably the better choice. And if this is allowed in py5, using set_mag() on a zero vector will also be allowed.

Originally posted by @hx2A in #242 (comment)

[hx2A]

To summarize:

• Py5Vector should let you set the magnitude to a negative magnitude. This would reverse the direction of the vector
• Normalize should be allowed on zero vectors, but issue a Python warning

[hx2A]

Fixed with #261. thank you, @AsadNizami !