puzzled about floats

[Marc 'BlackJack' Rintsch]

On Mon, 24 Sep 2007 14:13:18 -0400, Neal Becker wrote:

> from math import modf
> 
> class nco (object):
>     def __init__ (self, delta):
>         self.delta = delta
>         self.phase = 0.0
>     def __call__ (self):
>         self.phase += self.delta
>         f,i = modf (self.phase)
>         print  modf (self.phase)
>         if (self.phase > 1.0):
>             self.phase -= 1.0
>         return self.phase
> 
> n = nco (0.1)
> for x in xrange (100):
>     print '%.12f' % n()
> 
> prints out
> [...]
> (0.99999999999999978, 0.0)   <<< from modf
> 1.000000000000   << from n()
> 
> I'm baffled as to why 'print modf (self.phase)' prints out the first value,
> but the result printed in the 2nd case is different.  Without any precision
> spec on the first print, an approximate float value was printed, but even
> with %.12f, the second gives exactly 1.000....

Tuples are printed with calling `repr()` on the objects:

In [144]: str(0.1)
Out[144]: '0.1'

In [145]: repr(0.1)
Out[145]: '0.10000000000000001'

In [146]: '%.12f' % 0.1
Out[146]: '0.100000000000'

In [147]: '%.50f' % 0.1
Out[147]: '0.10000000000000000555111512312578270211815834045410'

Ciao,
	Marc 'BlackJack' Rintsch