In all_tests.py, the list of files is shuffled randomly to ensure the tests
are run independently of one another. However, for tracking down ordinary
test failures, it'd be useful to have all_tests.py run the tests in the
same order between executions.
The attached patch adds a --det (meaning "deterministic") command-line
option that skips the shuffling when it's passed. (I'm using
optparse.OptionParser since it looks like from the todo comments that
future work on all_tests.py will eventually require it anyway.)
OptionParser was added to the standard library in Python 2.3, if that's
relevant.
Current output
==============
> python all_tests.py
0 failures in 26 tests in file: test_account_manager.rst
0 failures in 13 tests in file: test_execution.rst
0 failures in 59 tests in file: test_dhtml.rst
[snip]
> python all_tests.py
0 failures in 4 tests in file: test_errors.rst
0 failures in 56 tests in file: test_io_widget.rst
0 failures in 29 tests in file: test_menu.rst
[snip]
Expected output after patch
===========================
> python all_tests.py --det
0 failures in 26 tests in file: test_account_manager.rst
0 failures in 13 tests in file: test_comet.rst
0 failures in 19 tests in file: test_configuration.rst
[snip]
> python all_tests.py --det
0 failures in 26 tests in file: test_account_manager.rst
0 failures in 13 tests in file: test_comet.rst
0 failures in 19 tests in file: test_configuration.rst
[snip]
This wasn't a Python 3-specific issue so I figured I'd file a patch against
the trunk.
Original issue reported on code.google.com by shadytr...@gmail.com on 29 May 2009 at 2:50
Original issue reported on code.google.com by
shadytr...@gmail.com
on 29 May 2009 at 2:50Attachments: