Message106831
| Author |
mark.dickinson |
| Recipients |
belopolsky, mark.dickinson, mcherm, rhettinger, stutzbach, tim.peters, vstinner |
| Date |
2010年06月01日.08:31:10 |
| SpamBayes Score |
0.0013394057 |
| Marked as misclassified |
No |
| Message-id |
<1275381072.68.0.694530842599.issue8860@psf.upfronthosting.co.za> |
| In-reply-to |
| Content |
An example where getting correct rounding is awkward: the following rounds up, with or without the patch:
>>> datetime.timedelta(seconds=0.6112295)
datetime.timedelta(0, 0, 611230)
But in this case, the result should actually be rounded down (with either rounding mode), since the exact value of float('0.6112295') is a little less than 0.6112295:
>>> print(Decimal(0.6112295))
0.61122949999999998116351207499974407255649566650390625
The problem is that leftover_us ends up being exactly 0.5 in this case. Again, this could be worked around if we really care, but I'm not convinced it's worth it. |
|