It seems like my recent work on Python initialization broke Python on FreeBSD. Many test_c_locale_coerce tests now fail with:

"Fatal Python error: get_locale_encoding: failed to get the locale encoding: nl_langinfo(CODESET) failed"

Example of failed buildbot build:

https://buildbot.python.org/all/#/builders/87/builds/1340

It seems like this buildbot started to fail at this build:

https://buildbot.python.org/all/#/builders/87/builds/1332

This build has been triggered by this change:

commit 177d921c8c03d30daa32994362023f777624b10d
Author: Victor Stinner <vstinner@redhat.com>
Date:   Wed Aug 29 11:25:15 2018 +0200

    bpo-34485, Windows: LC_CTYPE set to user preference (GH-8988)
    
    On Windows, the LC_CTYPE is now set to the user preferred locale at
    startup: _Py_SetLocaleFromEnv(LC_CTYPE) is now called during the
    Python initialization. Previously, the LC_CTYPE locale was "C" at
    startup, but changed when calling setlocale(LC_CTYPE, "") or
    setlocale(LC_ALL, "").
    
    pymain_read_conf() now also calls _Py_SetLocaleFromEnv(LC_CTYPE) to
    behave as _Py_InitializeCore(). Moreover, it doesn't save/restore the
    LC_ALL anymore.
    
    On Windows, standard streams like sys.stdout now always use
    surrogateescape error handler by default (ignore the locale).

test_c_locale_coerce tests invalid locales like LANG="invalid.ascii". It seems like Python supported such configuration before my changes.

My recent work:

* bpo-34523: filesystem encoding/errors
* bpo-34485: stdio encoding/errors