[Python-bugs-list] int() broken? (PR#370)

[M.-A. Lemburg]

wzdd@sesgroup.net wrote:
> 
> Full_Name: Nicholas FitzRoy-Dale
> Version: 1.5.2
> OS: Linux (Debian, Red Hat)
> Submission from: cpe-24-192-0-130.vic.bigpond.net.au (24.192.0.130)
> 
> The following typescript illustrates the problem. I understand int() truncates
> towards zero - but can't see why it would round down by a whole int value!
> 
> The same behaviour is apparent on a Red Hat 6.2 system, where Python was built
> with egcs.
> 
> ~$ python
> Python 1.5.2 (#0, Apr  3 2000, 14:46:48)  [GCC 2.95.2 20000313 (Debian
> GNU/Linux)] on linux2
> Copyright 1991-1995 Stichting Mathematisch Centrum, Amsterdam
> >>> test=591.92
> >>> test*100
> 59192.0
> >>> int (test*100)
> 59191
> >>>
> 
> ... and yet:
> >>> int (test*100+0.1)
> 59192
> >>>

While waiting for Tim Peters to explain the details,
here's the obvious part of the answer:

>>> test = 591.92
>>> '%.17f' % test
'591.91999999999995907'
>>> '%.17f' % (test*100)
'59191.99999999999272404'
>>> '%.17f' % (test*100+0.1)
'59192.09999999999126885'

-- 
Marc-Andre Lemburg
______________________________________________________________________
Business:                                      http://www.lemburg.com/
Python Pages:                           http://www.lemburg.com/python/