Author zach.ware
Recipients brett.cannon, eric.araujo, ezio.melotti, python-dev, r.david.murray, serhiy.storchaka, terry.reedy, zach.ware
Date 2013-01-07.22:07:43
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <>
I've come up with a semi-comprehensive list of the tests that cause ugly failures with test discovery.  Tests were run on 3.4 debug, on Windows 7 32bit without several of the 'external' projects built, using the command ``PCbuild\python_d.exe -m unittest discover Lib/test/ "test_x*.py"`` where 'x' was whatever it took to get a specific test or group of tests to run.

There are two basic causes of failure, and a couple of oddballs.

First up, tests that fail with discovery, but run fine with ``PCbuild\python_d.exe -m test test_name``:

test_array test_asyncore test_bisect test_bufio test_bytes test_codecs test_concurrent_futures test_configparser test_ctypes test_dbm test_decimal test_file test_format test_ftplib test_future3 test_hash test_imaplib test_import test_index test_io test_iterlen test_locale test_multiprocessing test_module test_osx_env test_pickle test_random test_robotparser test_runpy test_set test_shelve test_socket test_sqlite test_sys test_tarfile test_time test_univnewlines  test_warnings test_xml_etree

These 39 should be relatively straightforward fixes like heapq and genericpath have been.

The next group are tests that are properly skipped by the test package, but fail noisily with unittest discovery:

test_codecmaps_cn test_codecmaps_hk test_codecmaps_jp test_codecmaps_kr test_codecmaps_tw test_crypt test_curses test_epoll test_fcntl test_gdb test_grp test_hashlib test_ioctl test_kqueue test_largefile test_nis test_openpty test_ossaudiodev test_pipes test_poll test_posix test_pty test_pwd test_readline test_resource test_smtpnet test_socketserver test_ssl test_syslog test_tcl test_threadsignals test_tk test_ttk_guionly test_ttk_textonly test_urllib2net test_wait3 test_wait4 test_winsound test_zipfile64

Some of these are skipped due to certain resources not being allowed, some don't have the proper module available.  In any case, discovery does not respond well to the skip methods used for each of them.

The 'oddballs' I mentioned before are as follows:

json_tests.test_fail leakers.test_gestalt test_glob test_json test_shutil test_urllib2_localnet

Each of these fail on my machine whichever way I run them.  I'm not sure if any of them actually have any issues with discovery, I'm merely listing them here for completeness' sake.  I will try to test them again myself on Linux later, and report back what I find.

If anyone else feels led to tackle any of these, please add me to the nosy list of any new issues filed.  I plan to eventually work through all of these myself, but don't want to duplicate effort :)


Zach Ware
Date User Action Args
2013-01-07 22:07:44zach.waresetrecipients: + zach.ware, brett.cannon, terry.reedy, ezio.melotti, eric.araujo, r.david.murray, python-dev, serhiy.storchaka
2013-01-07 22:07:44zach.waresetmessageid: <>
2013-01-07 22:07:44zach.warelinkissue16748 messages
2013-01-07 22:07:43zach.warecreate