On Sun, Aug 11, 2013 at 12:14 AM, Roy Smith <r...@panix.com> wrote:
> Maybe you've got two different handlers which are both getting the same
> logging events and somehow they both end up in your stderr stream.
> Likely? Maybe not, but if you don't have any logging code in the test
> at all, it becomes impossible. You can't have a bug in a line of code
> that doesn't exist (yeah, I know, that's a bit of a handwave).
Likely? Very much so, to the extent that it is, if not a FAQ,
certainly a Not All That Uncommonly Asked Question. So many times
someone places logging code in something that gets called twice, and
ends up with two handlers. Personally, I much prefer to debug with
straight-up 'print' - much less hassle. I'd turn to the logging module
only if I actually need its functionality (logging to some place other
than the console, or leaving the log statements in and {en|dis}abling
them at run-time).
ChrisA
--
http://mail.python.org/mailman/listinfo/python-list
