On 3/14/2013 2:58 AM, Chris Angelico wrote:
While putting together some timing stats for the latest Python 3.3
string representation thread, I ran into an oddity in how IDLE handles
timeit. The normal way to profile Python code, according to stuff I've
found on the internet, is timeit.timeit:
import timeit
timeit.timeit("s=s[:-1]+u'\u1234'","s=u'asdf'*10000",number=10000)
0.57752172412974268
Now I knew that the module I wanted was timeit, but I didn't remember
the full incantation, so I did the obvious thing:
import timeit
timeit.<ctrl-space>
Only one thing came up: Timer. And help(timeit) doesn't mention
timeit.timeit either. Whereas the documentation:
http://docs.python.org/2/library/timeit.html
http://docs.python.org/3/library/timeit.html
clearly states that timeit.timeit() is the way to do things.
Snooping the source shows that timeit.timeit() is just a tiny
convenience function (alongside timeit.repeat()), but it'd still be
nice to have that come up in the Ctrl-Space list, since it's the
most-normal way to run timing tests.
Would there be a problem with adding timeit (and possibly repeat) to __all__?
-__all__ = ["Timer"]
+__all__ = ["Timer", "timeit"]
I believe everything in the doc should be in __all__. Open an issue on
the tracker and add terry.reedy as nosy.
Note on the issue that timeit.default_timer is not callable, contrary to
the doc (which should also be fixed).
--
Terry Jan Reedy
--
http://mail.python.org/mailman/listinfo/python-list
