The float arrays on 64 bit machines do not work correctly.
The problem comes from the size of the t_word union,
which is exactly the size of a float on 32 bit
machines, but its double the
size of a float on 64 bit machines. This means that
routines
which treat the array as float* (the ones in d_array)
treat them
differently as those in g_array (which think of them as
t_word elements).
This could be fixed in several (all suboptimal) ways.
I'm leaving this up to Miller.
Anonymous
Logged In: YES
user_id=564396
Originator: NO
seems to be fixed in 0.41