The locale.getdefaultlocale() function only relies on environment variables. At Python startup, Python calls setlocale() is set the LC_CTYPE locale to the user preferred encoding.

Since Python 3.7, if the LC_CTYPE locale is "C" or "POSIX", PEP-538 sets the LC_CTYPE locale to a UTF-8 variant if available, and PEP-540 ignores the locale and forces the usage of the UTF-8 encoding. The *effective* encoding used by Python is inconsistent with environment variables.

Moreover, if setlocale() is called to set the LC_CTYPE locale to a locale different than the user locale, again, environment variables are inconsistent with the effective locale.

In these cases, locale.getdefaultlocale() result is not the expected locale and it can lead to mojibake and other issues.

For these reasons, I propose to deprecate locale.getdefaultlocale(): setlocale(), getpreferredencoding() and getlocale() should be used instead.

For the background on these issues, see recent issue:

• bpo-43552
• bpo-43557

cal_locale.py: Test calendar.LocaleTextCalendar() default locale, manual test for #75349.

New changeset 04dd60e by Victor Stinner in branch 'main':
bpo-46659: Update the test on the mbcs codec alias (GH-31168)
04dd60e

New changeset 06b8f16 by Victor Stinner in branch 'main':
bpo-46659: test.support avoids locale.getdefaultlocale() (GH-31167)
06b8f16

For these reasons, I propose to deprecate locale.getdefaultlocale(): setlocale(), getpreferredencoding() and getlocale() should be used instead.

Please see the discussion on https://bugs.python.org/issue43552: locale.getpreferredencoding() needs to be deprecated as well. Instead we should have a single locale.getencoding() as outlined there... perhaps in a separate ticket ?! Thanks.

Please see the discussion on https://bugs.python.org/issue43552: locale.getpreferredencoding() needs to be deprecated as well. Instead we should have a single locale.getencoding() as outlined there... perhaps in a separate ticket ?! Thanks.

Yeah, I read this issue. But these things are too complicated :-) I prefer to move step by step.

Once locale.getencoding() (or a similar function) is added, we can update the deprecation message.

I hope to be able to deprecate getdefaultlocale() and to add such new function in Python 3.11.

New changeset 04dd60e by Victor Stinner in branch 'main':
bpo-46659: Update the test on the mbcs codec alias (GH-31168)

This change is not correct, I created bpo-46668 to fix it.

New changeset 7a0486e by Victor Stinner in branch 'main':
bpo-46659: calendar uses locale.getlocale() (GH-31166)
7a0486e

getdefaultlocale() falls back to LANG and LANGUAGE. It allows also to specify a list of looked up environment variables. How could this use case be covered with getlocale()?

getdefaultlocale() falls back to LANG and LANGUAGE.

_Py_SetLocaleFromEnv(LC_CTYPE) (e.g. setlocale(LC_CTYPE, "")) gets called at startup, except for the isolated configuration [1].

I think calendar.Locale*Calendar should try the LC_CTYPE locale if LC_TIME is "C", i.e. (None, None). Otherwise, it's introducing new default behavior. For example, with LC_ALL set to "ru_RU.utf8":

3.8:

    >>> locale.getlocale(locale.LC_TIME)
    (None, None)
    >>> locale.getlocale(locale.LC_CTYPE)
    ('ru_RU', 'UTF-8')
    >>> cal = calendar.LocaleTextCalendar()
    >>> cal.formatweekday(0, 15)
    '  Понедельник  '

3.11.0a5+:

    >>> locale.getlocale(locale.LC_TIME)
    (None, None)
    >>> locale.getlocale(locale.LC_CTYPE)
    ('ru_RU', 'UTF-8')
    >>> cal = calendar.LocaleTextCalendar()
    >>> cal.formatweekday(0, 15)
    '     Monday    '
    >>> locale.setlocale(locale.LC_TIME, '')
    'ru_RU.utf8'
    >>> cal = calendar.LocaleTextCalendar()
    >>> cal.formatweekday(0, 15)
    '  Понедельник  '

[1] https://docs.python.org/3/c-api/init_config.html?#isolated-configuration

Serhiy: "getdefaultlocale() falls back to LANG and LANGUAGE. It allows also to specify a list of looked up environment variables. How could this use case be covered with getlocale()?"

What's your use case to use env vars rather than the current LC_CTYPE locale?

My concern is that when setlocale() is called, the current LC_CTYPE locale is inconsistent and you can get mojibake and others issues.

See for example:
https://bugs.python.org/issue43552#msg389069

Marc-Andre Lemburg wants to deprecate it:
https://bugs.python.org/issue43552#msg389076

I think calendar.Locale*Calendar should try the LC_CTYPE locale if LC_TIME is "C", i.e. (None, None). Otherwise, it's introducing new default behavior. For example, with LC_ALL set to "ru_RU.utf8": (...)

Oh. Serhiy asked me to use LC_TIME rather than LC_CTYPE.

See also my example in the PR:
#31166 (comment)

Oh. Serhiy asked me to use LC_TIME rather than LC_CTYPE.

Since Locale*Calendar is documented as not being thread safe, __init__() could get the real default via setlocale(LC_TIME, "") when locale=None and the current LC_TIME is "C". Restore it back to "C" after getting the default. That should usually match the behavior from previous versions that called getdefaultlocale(). In cases where it differs, it's fixing a bug because the default LC_TIME is the correct default.

Eryk: I created #75397 which uses the user preferred locale if the current LC_TIME locale is "C" or "POSIX".

Moreover, it no longer gets the current locale when the class is created. If locale=locale is passed, just use the current LC_TIME (or the user preferred is the locale is "C" or "POSIX").

New changeset ccbe804 by Victor Stinner in branch 'main':
bpo-46659: Fix the MBCS codec alias on Windows (GH-31218)
ccbe804

New changeset b899126 by Victor Stinner in branch 'main':
bpo-46659: Deprecate locale.getdefaultlocale() (GH-31206)
b899126

New changeset 4fccf91 by Victor Stinner in branch 'main':
bpo-46659: Enhance LocaleTextCalendar for C locale (GH-31214)
4fccf91

locale.getdefaultlocale() is now deprecated.

calendar now uses locale.setlocale() instead of locale.getdefaultlocale().

The ANSI code page alias to MBCS now has better tests and better comments.

Thanks Eryk Sun for your very useful feedback!

Thanks, Victor.

locale.getlocale does not return a correct RFC 1766 language code on Windows, which makes it incompatible with locale.getdefaultlocale. (#82986)

Our project relies on locale.getdefaultlocale to get the default system language code, which won't work using locale.getlocale due to the above bug. This breaks downward compatibility.

May I request NOT REMOVING locale.getdefaultlocale until there is a good solution for the bug of locale.getlocale?

Return of locale.getlocale() is not compatibile with typical locale string (Python 3.11.7 on Windows):

locale.getdefaultlocale()
('it_IT', 'cp1252')

locale.getlocale()
('Italian_Italy', '1252')

'it_IT' permits to fastly use gettext ./locale/<it_IT>/... translations mo files.

Return of locale.getlocale() is not compatibile with typical locale string

That's why the deprecation suggests using locale.setlocale(). Hum, the API is surprising, but locale.setlocale() should be used to get the current locale: locale.setlocale(locale.LC_CTYPE, None) gets the current LC_CTYPE locale for example.

Maybe the deprecation message should be clarified. I dislike locale.getlocale(), it can return invalid locale which is not accepted by setlocale(). I wanted to deprecate it.

The point of having getlocale() is to be able to normalize the returned values to make them compatible to setlocale().

The C lib API of setlocale() to actually fetch the current locale values is completely counterintuitive, which is why I had added getlocale() instead. On top of this, setlocale() does not always return values which you can feed back into it.

@shinewook: Could you please check what locale._setlocale(locale.LC_TYPE) returns on your Windows system ? It's possible that we'll have to add more aliases to the table to fix this. Thanks.

BTW: Perhaps better to open a separate ticket for that problem...