The testing to ensure a vector of zeros of type floating point isn't...

The testing to ensure a vector of zeros of type floating point isn't misclassified as negative zero can be simplified, as pointed out by Duncan Sands.

llvm-svn: 177386