[Python-Dev] time.clock() on windows (original) (raw)

Scott Dial scott+python-dev at scottdial.com
Wed Oct 21 23:52:26 CEST 2009

Curt Hagenlocher wrote:

But it makes more sense to understand why someone chose to implement time.clock() on Windows the way they did -- this seems very broken to me, and I think it should be changed.

Some SVN detective work takes this to all the way back to r7713 (1997-04-02). The original implementation checked by Guido and attributed to Mark Hammond. So, we should ask Mark why he did that.

Can anyone honestly use it, as it is, without already having normalize it across platforms themselves?

I don't know how much of an impact it is, but the current implementation of clock() does not require argument parsing, so the proposal to add a "absolute" boolean-flag argument is perhaps bad. This is generally a function used for performance timing and that proposal adds some amount of latency to the query. The proposal to add a clockbase() function is perhaps better because of this, you need only call it once, and you can cache the result for the life of your process.

-- Scott Dial scott at scottdial.com scodial at cs.indiana.edu

More information about the Python-Dev mailing list