[Python-Dev] test_minidom non-failure failure? (take 2)

Jeremy Hylton jeremy@beopen.com
Thu, 12 Oct 2000 16:38:29 -0400 (EDT)

>>>>> "NS" == Neil Schemenauer <nas@arctrix.com> writes:

  NS> In any case, regrtest should probably print a more clueful
  NS> message when objects are found in gc.garbage.  That would go a
  NS> long way in clearing up the confusion.  I'll see about a patch.

I think there is some value to issueing reports when a test creates
trash that can't be cleaned up by the GC, but it doesn't make sense to
report this by default.  If "make test" reports an error, it should
mean that something has gone wrong with the user's build and the
interpreter is broken.  That's not the case here.  The build was fine;
it's just the test that is iffy.

I think we should turn of the -l option by default *and* patch it so
that the output is more useful.  I would suggest printing a count of
the objects, perhaps organized by type/class name.  Something like

    Test created 6 uncollectable objects
    4 foo.Bar instances
    2 foo.Baz instances