use t_float instead of float

[umlaeute]

Pd is (slowly) moving towards double-precision.

as such, you should never use float or double of any numbers that communicate with Pd, as this will guarantee to break once Pd changes it's internal number representation to double precision.

instead, always use t_float for numbers and t_sample for signals.

[josephlarralde]

Got it, I'll update this.

BTW, thanks for your last PR, it seems to solve most parts of the problem.
I still need to figure out a way to deal with the .w_float needed to access the actual array values, though. I'll test your patch but I bet it will need a few more tweaks to get the object back working (and maybe some modifications to jl.cpp.lib, too).