Bug in polyfit

Hi, I'm using matplotlib 0.71 and I think I found a bug in

    > polyfit.

    > This simple linear regression on two data points gives the
    > correct answer:
    >>>> polyfit([731.924,731.988],[915,742],1)
                       ^^^^
                     floats
    > However, if I multiply my x values by 1000 the result is
    > wrong:
    >>>> polyfit([731924,731988],[915,742],1)
                       ^^^^
                     integers

Both of these should work

print polyfit([731.924,731.988],[915.,742.],1)
print polyfit([731924.,731988.],[915.,742.],1)

I fixed the polyfit code to explicitly convert the input arrays to
floats arrays, which fixes this bug.

Thanks for the report.

JDH