This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] Fix bo_CN and bo_IN locales
- From: Rich Felker <dalias at aerifal dot cx>
- To: Andreas Jaeger <aj at suse dot com>
- Cc: Aurelien Jarno <aurelien at aurel32 dot net>, libc-alpha at sourceware dot org
- Date: Mon, 13 May 2013 12:29:37 -0400
- Subject: Re: [PATCH] Fix bo_CN and bo_IN locales
- References: <20130513013446 dot GA21407 at hall dot aurel32 dot net> <51908772 dot 4060906 at suse dot com>
On Mon, May 13, 2013 at 08:25:54AM +0200, Andreas Jaeger wrote:
> On 05/13/2013 03:34 AM, Aurelien Jarno wrote:
> >bo_CN and bo_IN have an empty name_fmt field, causing localedef to fail:
> >
> >$ localedef -c -f UTF-8 -i bo_IN bo_IN.UTF-8
> >LC_NAME: field `name_fmt' must not be empty
> >
> >While I don't know what are the correct values for this locale, I think
> >this should be fixed, to not break localedef and the tools using it. I
> >therefore suggest to change the value to "???" as it is already done for
> >the az_AZ and tt_RU locales.
>
> Yes, this looks like the best approach. I wanted to suggest
> contacting the authors but there're none given in the file, so go
> ahead,
The whole model of LC_NAME is unsuitable to Tibetan names, which are
not composed of a given and family name, but usually two given names,
which are meant to be kept in a fixed order (first, second) when both
used, and it varies from person to person which name they use when
going by a single name. Thus, the choice of name_fmt depends on how
the user has coerced Tibetan names into a model that's unsuitable for
representing them. There does not seem to be any standard for this.
In a Western or Indian context, the first and second name seem usually
to be treated as first and last, but this is mildly problematic
because at least some Tibetans also have family names which were not
traditionally used as part of a name, but which increasingly many
people are choosing to adopt when living in places where having a
family name is expected.
In a Chinese context, it's a lot more ambiguous whether the first and
second name will be mapped to (first,family) or (family,first). I've
seen it done both ways, and I've also seen people's names turned
backwards because it was originally done one way then rendered the
opposite.
In any case, "???" does not seem like a good value for name_fmt. I
would probably just go with first,family until somebody opens a bug
report asking for it to be changed...
Rich