[PATCH] Reassociate x + -0.1234 * y into x - 0.1234 * y

[Erik Verbruggen]

This should have gone to the list (and the reviewers)...

> This does not require -ffast-math, and it gives CSE/GVN more options to
> eliminate duplicate expressions in, e.g.:
> 
>  return ((x + 0.1234 * y) * (x - 0.1234 * y));
> 
> http://reviews.llvm.org/D4904