There are still a bunch of tests failing in the py3k (sic) branch: test_class fails because hash() of a new-style class that defines __eq__ or __cmp__ doesn't raise an exception. There's a design issue here: should the default hash be changed to raise an exception in this case (perhaps triggering other test failures) or should the tests be deleted? (This is from a test module that used to test classic classes.)
test_exception fails with a SystemError for popping a non-exception off the stack test_importhooks fails test_pickletools fails test_pyclbr fails test_scriptpackages fails a bunch of tests for XML modules all fail (perhaps some of these are related): minidom, sax, xdrlib, xml_etree I recently fixed repr.py and test_repr.py, but I wonder if we shouldn't delete that module... If anyone wants to help out debugging these, please do! If you solve one, just mail me a patch (or check it in if you've got checkin perms and feel confident -- just still drop me a line). I'd really like the p3yk branch to be error-free so that if further changes cause new errors, the failures are easily attributable. -- --Guido van Rossum (home page: http://www.python.org/~guido/) _______________________________________________ Python-3000 mailing list [email protected] http://mail.python.org/mailman/listinfo/python-3000 Unsubscribe: http://mail.python.org/mailman/options/python-3000/archive%40mail-archive.com
