Test test_input_string_transform_factory_removes_thousands_separator_and_is_float_aware_with_LOCALE_and_FLOAT_example is failing on last release 5.3.3

[agustinhenze]

Hi there, I'm the debian maintainer of this module :). I was packaging last version and I can repeat in a controlled environment the test mentioned on the subject is failing for both Python versions. Python2.7 and Python3.6.

=================================== FAILURES ===================================
 test_input_string_transform_factory_removes_thousands_separator_and_is_float_aware_with_LOCALE_and_FLOAT_example 
test_natsort/test_input_string_transform_factory.py:126: in test_input_string_transform_factory_removes_thousands_separator_and_is_float_aware_with_LOCALE_and_FLOAT_example
    assert _input_string_transform_factory(ns.LOCALE | ns.FLOAT)(x) == '12543642642.534,534980'
E   AssertionError: assert '12,543,642,642.534,534,980' == '12543642642.534,534980'
E     - 12,543,642,642.534,534,980
E     ?   -   -   -           -
E     + 12543642642.534,534980

You can see the detailed log here

https://salsa.debian.org/debian/natsort/-/jobs/40303

[SethMMorton]

What is your LOCALE set to? While cleaning up the test suite recently, I noticed that the tests that use locale were making the assumption that the locale was "C" or "en_US". If that is not the case for you it would explain the failure.

[agustinhenze]

Hi @SethMMorton locale is unset 🤷‍♂️.

In [1]: import locale

In [2]: locale.getdefaultlocale()
Out[2]: (None, None)

So we found the failure, thanks! :)

[SethMMorton]

I will update the tests to not make any assumptions about the environment.

[SethMMorton]

Is it possible for you to try running the tests as of c749c79 (branch "clarify-code-and-simplify-tests")? All tests using LOCALE should be explicitly setting a locale.

[SethMMorton]

Does this still give you problems on 5.4.0?

[agustinhenze]

Hi @SethMMorton, sorry for the delay. I've packaged last version and it works fine :). Here you can see the output for the tests. So, I'm closing this issue. Thank you very much!