On Thu, Jan 31, 2019 at 05:15:39PM +0900, Masashi Honma wrote: > This patch set modifies hwsim test to support python3 because python2 is > deprecated from 2020. After applying this patch set, user can run hwsim > test both python2 and python3. We can run with python3 by modifying > run-tests.py. Thanks, applied with number of fixes and cleanup. I ended up splitting the patches differently to make it easier to review the changes and to be able to pick the parts that looked correct to me. I did replace some areas by not doing unnecessary hexlify/unhexlify conversions (i.e., using binary comparisons where appropriate and hex strings were more convenient). In addition, I did more changes with wpaspy support. > [failing tests] > Some test cases fails. This is not because of python3, it fails with python2 > also. This is the list of the test cases. In addition to this, it looks like number of test cases were being skipped in your setup. Those test cases missed changes needed for python3. I implemented those and the end result can now pass every single test case with both python2 and python3. While the coverage looks more or less identical, there are some differences in particular due to the way dict iteration works. python3 ended up processing the dict entries in different order and that ended up actually finding a real bug in the D-Bus handling.. I did not yet check whether there could be other differences in coverage, so I left the python2 as the default for now. Anyway, the goal from now on would be to maintain working python3 support and likely remove that python2 default eventually. > [known issue] > Writing un-printable code with logger.debug() causes exception. > This occurs only at vm environment and this does not interrupt the test. > The test end with PASS even thought this exception. > > Traceback (most recent call last): > File "/usr/lib/python3.6/logging/__init__.py", line 996, in emit > stream.write(msg) > UnicodeEncodeError: 'ascii' codec can't encode character '\x80' in position 68: ordinal not in range(128) It looks like the logging.FileHandler() is not able to handle these and at least with Python 3.5.2 it did not even handle utf-8 which was a bit surprising. I tried to get that working, but did not find any easy way of doing so. Should try to figure out what exactly is happening here and why utf-8 part did not work.. For non-printable-even-with-utf-8-working cases, those should probably be filtered out by the code that generates such output. It does not really belong in a log file that is supposed to be text. -- Jouni Malinen PGP id EFC895FA _______________________________________________ Hostap mailing list Hostap@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/hostap
