Fix incorrect rotation value multiplication order

[jngbsn]

Without this, the rotation is applied before any translation, causing it act as if (0, 0, 0) is the rotation origin, instead of rotating around its own local origin.

[cart]

While I think this change is useful, it creates an inconsistency with the "apply_scale" method, which "scales" the current matrix relative to the origin. Rotate() was intended to be the corollary to that method. But rotating the "rotation" property (like you do in this pr) makes a lot more sense to me.

However I'm starting to think that the "apply operation directly to current matrix" convenience methods like apply_scale and rotate cause more trouble/confusion than they're worth.

Can we make the following changes in this pr:

1. remove apply_scale and apply_uniform_scale from Transform/GlobalTransform
2. add your "rotate fix" to GlobalTransform

[cart]

Hmm this latest build error isn't your fault. Clippy nightly apparently just got updated and added new rules.

[cart]

I'm honestly getting pretty tired of nightly clippy breakage. I think we should switch back to stable.

[cart]

I'm fixing clippy here: #577

[cart]

We will likely want to close this now that #596 is the direction we're taking the transform in. I'll leave it open for now, but if we merge #596, this will become unnecessary.