[Python-Dev] Worse news

Guido van Rossum guido@digicool.com
Mon, 22 Jan 2001 16:18:26 -0500


[Tim]
> So does mine now.  The remaining failures require *unusual* ways of running
> the test suite (with -r to get test_cpickle to fail, confirmed now by Jeremy
> under Linux;
[and later]
> BTW, delighted(?) to hear that test_cpickle fails for you too!

This (test_cpickle) is a red herring -- it's a shallow failure in the
test suite.  test_cpickle imports test_pickle, but test_pickle first
outputs the test output from testing pickle -- unless test_pickle has
been run before!  This succeeds:

  ./python Lib/test/regrtest.py test_cpickle test_pickle

and this fails:

  ./python Lib/test/regrtest.py test_pickle test_cpickle

Use regrtest.py -v to fidn out why. :-)

I'm not sure how to restucture this, but it's not of the same quality
as test_extcall or test_sax failing.  Neither of those has failed for
me on Linux during hours of testing.  However on Windows I get an
occasional appfail dialog box when using rt.bat.

--Guido van Rossum (home page: http://www.python.org/~guido/)