bpo-44031: fix test_tabnanny failure in non-ascii CWD

[iritkatriel]

I was able to reproduce the failure on a Mac, and this change fixes it.

https://bugs.python.org/issue44031

[vstinner]

I suggest to use os.fsdecode() instead. subprocess uses the locale encoding, it can be different than UTF-8. Moreover, fsdecode() uses the surrogateescape error handler for undecodable bytes.

[vstinner]

I confirm that this PR fix https://bugs.python.org/issue44031

Without this PR, the test fails if the Python directory contains Загрузки:

$ pwd /home/vstinner/python/Загрузки $ ./python -m test -v test_tabnanny (...) ERROR: test_command_usage (test.test_tabnanny.TestCommandLine) ---------------------------------------------------------------------- Traceback (most recent call last): File "/home/vstinner/python/Загрузки/Lib/test/test_tabnanny.py", line 325, in test_command_usage self.validate_cmd(stderr=stderr) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/vstinner/python/Загрузки/Lib/test/test_tabnanny.py", line 297, in validate_cmd err = err.decode('ascii') ^^^^^^^^^^^^^^^^^^^ UnicodeDecodeError: 'ascii' codec can't decode byte 0xd0 in position 29: ordinal not in range(128) (...) 

With this PR, it works as expected:

$ ./python -m test -v test_tabnanny (...) Tests result: SUCCESS 

[vstinner]

LGTM.

[iritkatriel]

@raghavthind2005 FYI - there is a discussion on python-dev about spam activity on GitHub (which wastes our time) and your recent activity on the cpython repo was mentioned there.

See this link: https://mail.python.org/archives/list/python-dev@python.org/message/UO6ZSNWLLXWU7AZ7NGQDTOQ2WVX2ZAZN/

[miss-islington]

Thanks @iritkatriel for the PR 🌮🎉.. I'm working now to backport this PR to: 3.9, 3.10.
🐍🍒⛏🤖 I'm not a witch! I'm not a witch!

[bedevere-bot]

GH-31047 is a backport of this pull request to the 3.10 branch.

[bedevere-bot]

GH-31048 is a backport of this pull request to the 3.9 branch.