[RFC][PATCH v5 00/13][BZ 10871] Month names in genitive case

[Rafal Luzynski]

13.01.2017 10:01 Florian Weimer <fweimer@redhat.com> wrote:
>
>
> On 01/03/2017 03:10 AM, Rafal Luzynski wrote:
> > I think it would be pretty destructive to the users if we initially
> > implemented %OB as genitive and %B nominative and then swapped them.
>
> This is quite confusing. *You* want to retroactively change the meaning
> of %B and the MON_* values, breaking applications such as Thunderbird
> and cal.
>
> Thanks,
> Florian

Yes, my proposed change is kinda destructive but I think it's
easier to explain that we change the meaning of some format
specifiers because we add new ones than to explain that we
change the meaning of some format specifiers because we have
changed our minds.  Without any way for the application
developers to tell which version is on the computer where their
application works.

I think that the current consensus, as stated in [1], is to:

- add ALTMON_x and %OB format specifier;
- document it as a new experimental feature but not to specify
  which is nominative, genitive, full-date format, standalone, etc.

Note that there is no difference in the implementation between:

- adding %OB and making it a genitive case,

 and

- changing %B to the genitive case and adding %OB for the
  nominative case. [2]

Now I can imagine it could be left to the language community
(translators) to decide whether they want %B / MON_x to be
nominative and %OB / ALTMON_x to be genitive or the other way
round.

Regarding the applications that will be broken: yes, cal will
be broken and will need a fix.  Any other applications with
the same problem?  I use Thunderbird so it should be easy
for me to analyze the problem.  I see it using only numbers
to format months in dates, can you point me a specific place
in Thunderbird where it may use months in nominative/genitive
case?  Does Thunderbird use strftime() at all, doesn't it
use ICU which does not have this bug?

Regards,

Rafal

[1] https://sourceware.org/ml/libc-alpha/2016-12/msg01103.html
[2] https://sourceware.org/ml/libc-alpha/2016-12/msg01079.html