Re: [PATCH v12 0/6][BZ 10871] Month names in alternative grammatical case

[Rafal Luzynski <digitalfreak at lingonborough dot com>]

14.01.2018 04:53 Carlos O'Donell <carlos@redhat.com> wrote:
> [...]
> I'm going to install these patches and test it out and give my OK.

Thank you for your effort.  Please note that you need to update
some locale data and use the corresponding locale in order to see
any difference.  But it's probably also worth testing without the
locale data updated in order to verify that there is no difference
if the locale data is not changed.

You can also use some of my test programs. [1] Feel free to ask if
you need some hints about how to verify if the date is correct in
an inflected language.

> I want to just verify the static binary case that Florian asked about
> so we can document the result.

I made a simple test and it confirms that the static binary is unable
to load any new locale data and it just falls back to a default
builtin "C" locale.  My tests were not thorough enough but I guess
that this is because of some sanity tests in the file locale/loadlocale.c
which make sure that the locale data is not too short or too long.
Also I noticed that a binary segfaults if it assumes that the newlocale()
function call is always successful and does not verify that the result
is nonzero.  But this can be treated as an application bug.

A workaround for the problem would be to deliver the old locale data
and put their directory name in the LOCPATH environment variable.

A similar problem has been reported as a bug 19084 [2] and the answer
was that it cannot be fixed for statically linked binaries.

Regards,

Rafal

[1] https://github.com/rluzynski/months-test
[2] https://sourceware.org/bugzilla/show_bug.cgi?id=19084