Re: PR libc/5023: New locales for li_NL, nds_NL and nds_DE

[Petter Reinholdtsen <pere at hungry dot com>]

Sorry for my late reply.  I've been busy.

[Kenneth Rohde Christiansen]
> Apparently I never got the mail - maybe my gnu address is bouncing. If
> so please cc: kenneth@krn.dk

I'll try the gnu address and see if it work.

You might want so subscribe to the libc-locales mailing list.  See the
bottom of <URL:http://www.student.uit.no/~pere/linux/glibc/> for
instructions.

> Since it is a long time ago I will send you my latest version. Can you
> look and comment on those? 

In the mean time, the glibc project started using bugzilla.  You
should submit your proposed locales into
<URL:http://sources.redhat.com/bugzilla/>.

> I will probably need some help

I'll do my best.

> Is there a tool to "convert" the string to <unicode numner>'s ?

Yes.  I got this perl script capable of doing that.  But it is a bit
stupid, and converts everything, and not only those part that need to
be converted.

#!/usr/bin/perl -w
use utf8;

while(<>)
{
  #Encode all chars in quoted strings
  s/\"([^\"]+)\"/'"'.encode_string($1).'"'/ge;
  
  #Encode all national letters everywhere
  s/([\x{0080}-\x{FFFF}])/"<U".sprintf("%04X",unpack('U',$1)+0).">"/ge;
  
  print;
}

sub encode_string
{
  my $str=shift;

  #Decode <Uxxxx> back to unicode char to prevent double encoding
  $str=~s/<U([0-9A-Za-z]{4})>/pack('U',hex($1))/ge;
 
  #Encode char into <Uxxxx> form
  $str=~s/(.)/"<U".sprintf("%04X",unpack('U',$1)+0).">"/ge;
  return $str;
}


> Can you explain me exactly what I should set lang_term, lang_lib and
> lang_ab, country_ab2, country_ab3 to?

I'm not sure.  I'll try to find time to look at it when the locale is
in bugzilla.  In the mean time, I spend my time follow up on the bugs
in bugzilla, and try to write a small guide on how to write locales.