Message 189005 - Python tracker

➜

This issue tracker has been migrated to GitHub , and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

In-reply-to
Author	mark.dickinson
Recipients	Alex.Rhatushnyak, mark.dickinson
Date	2013年05月12日.10:33:03
SpamBayes Score	-1.0
Marked as misclassified	Yes
Message-id	<1368354784.01.0.346172609972.issue17958@psf.upfronthosting.co.za>

Content
Sorry: this is not a bug, but a difficult-to-avoid consequence of floating-point imprecision: math.log(n, 2) is computed as log(n) / log(2), and each of the log computations and the division can introduce small errors. For what it's worth, Python 3.3 has a `log2` function, which has accuracy that's a little bit better than math.log(n, 2), and gives the 'correct' answer exact powers of 2. Closing as invalid.

Content

Sorry: this is not a bug, but a difficult-to-avoid consequence of floating-point imprecision: math.log(n, 2) is computed as log(n) / log(2), and each of the log computations and the division can introduce small errors.
For what it's worth, Python 3.3 has a `log2` function, which has accuracy that's a little bit better than math.log(n, 2), and gives the 'correct' answer exact powers of 2.
Closing as invalid.

History
Date	User	Action	Args
2013年05月12日 10:33:04	mark.dickinson	set	recipients: + mark.dickinson, Alex.Rhatushnyak
2013年05月12日 10:33:04	mark.dickinson	set	messageid: <1368354784.01.0.346172609972.issue17958@psf.upfronthosting.co.za>
2013年05月12日 10:33:03	mark.dickinson	link	issue17958 messages
2013年05月12日 10:33:03	mark.dickinson	create

homepage