linux/glibc/howto.html

   1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
   2 <html lang="en">
   3   <head>
   4     <title>How to write a GNU libc locale</title>
   5     <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
   6     <meta http-equiv="Content-Language" content="en">
   7     <link rel="stylesheet" type="text/css" href="http://i18n.skolelinux.no/stilsett.css" id="nn1">
   8     <link rel="stylesheet" type="text/css" href="http://i18n.skolelinux.no/utskrift.css" media="print" id="nn2">
   9
  10   </head>
  11
  12   <body>
  13     <div class="topp">
  14     <h1>How to write a GNU libc locale</h1>
  15     </div>
  16
  17     <div class="meny">
  18       <a href="./">Back</a>
  19     </div>
  20
  21     <div class="hovuddel">
  22
  23     <p>This is a draft document explaining how to write locale files
  24       for GNU libc.  It will not go into details, but reference
  25       specifications.  It will on the other hand mention some of the
  26       pitfalls, and try to document the current practice.</p>
  27
  28     <h2>How to choose the locale file name</h2>
  29
  30       <p>Locale names consist of three parts.  The language code, the
  31         country/region code, and the optional modifier.  The format is
  32         language_REGION@modifier.  The language code is a code from
  33         ISO 639.  The two-letter code is prefered, but a three letter
  34         code is accepted if no two-letter code is available.  The
  35         country/region code is a code from ISO 3166.  If the language
  36         or region in question is missing in the ISO standard, one need
  37         to get the ISO standard updated before the locale will be
  38         included in glibc.  If one can't convince the ISO 639
  39         maintainers that your language exists (and thus need a
  40         language code), the glibc maintainers will refuse to add the
  41         locale.  In addition, the glibc maintainers seem to refuse
  42         "artificial languages" like Esperanto and Lojban, even if they
  43         got a ISO 639 code.</p>
  44
  45       <p>Little is known about the requirements for the naming of
  46         modifiers.  The following modifiers are currently used:
  47         abegede, cyrillic, euro and saaho.  This might indicate that
  48         lower case letters are prefered in modifier names.</p>
  49
  50       <p>It is recommended to follow RFC 3066 when selecting locale
  51         names.</p>
  52
  53       <ul>
  54
  55         <li><a href="http://www.unicode.org/onlinedat/countries.html">ISO
  56             3166</a></li>
  57
  58         <li><a href="http://www.loc.gov/standards/iso639-2/">ISO 639</a></li>
  59
  60         <li><a href="http://rfc.sunsite.dk/rfc/rfc3066.html"> RFC 3066
  61             - Tags for the Identification of Languages</a></li>
  62
  63       </ul>
  64
  65     <h2>Category order</h2>
  66
  67     <p>To make it easier to compare locales with each other, I
  68       recommend using the same order for the categories in all
  69       locales.  Any order will do, so I picked the order used in most
  70       locales, and decided to recommend this order:</p>
  71
  72     <ol>
  73       <li>LC_IDENTIFICATION
  74       <li>LC_CTYPE
  75       <li>LC_COLLATE
  76       <li>LC_MONETARY
  77       <li>LC_NUMERIC
  78       <li>LC_TIME
  79       <li>LC_MESSAGES
  80       <li>LC_PAPER
  81       <li>LC_NAME
  82       <li>LC_ADDRESS
  83       <li>LC_TELEPHONE
  84       <li>LC_MEASUREMENT
  85     </ol>
  86
  87     <h2>Reuse when possible</h2>
  88
  89       <p>One should avoid cut-n-paste when possible, and instead use
  90         the <tt>copy</tt> statement to include sections from locales
  91         with identical content.</p>
  92
  93     <h2>LD_IDENTIFICATION</h2>
  94
  95       <p>The category entries are references to the standard used when
  96         writing the given section.  The standard refs should have
  97         quotes around them, and should not use the &lt;U#&gt;
  98         notation.  They should normally look something like this:</p>
  99
 100       <blockquote><pre>
 101 category  "i18n:1997";LC_IDENTIFICATION
 102         </pre></blockquote>
 103
 104     <h2>LC_MESSAGES</h2>
 105
 106       <p>Then yesexpr and noexpr entries should have the form
 107         <tt>^[yY&lt;extra&gt;]</tt> and <tt>^[nN&lt;extra&gt;]</tt>,
 108         without 0 and 1 and without trailing "<tt>.*</tt>".  The
 109         reason is to make sure the expressions have the same form as
 110         the expressions used in the C/POSIX locale (<tt>^[yY]</tt> and
 111         <tt>^[nN]</tt>).</p>
 112
 113     <h2>Standard documents and specifications</h2>
 114
 115     <h2>Testing the new locale file</h2>
 116
 117         <p>To test a new locale on a test machine, do the
 118         following:</p>
 119
 120         <ul>
 121
 122           <li>Copy the new locale to
 123             <tt>/usr/share/i18n/locales/<em>filename</em></tt></li>
 124
 125           <li>Run <tt>localedef -i <em>inputfile</em> -c -f
 126               <em>charset<em> <em>locale</em></tt> to generate a
 127             binary locale file in
 128             <tt>/usr/lib/locale/<em>locale</em>/</tt></li>
 129
 130           <li>Test it using LANG=<em>locale</em>, for example by
 131             running <tt>date</tt></li>
 132
 133         </ul>
 134
 135         <p>Example, generating a new <tt>de_DE@euro</tt> locale using
 136           the ISO-8859-15 charset and save it as 'de_DE':</p>
 137
 138         <pre>
 139           cp de_DE@euro /usr/share/i18n/locales/de_DE@euro
 140           localedef -i de_DE@euro -c -f ISO-8859-15 de_DE
 141           LANG=de_DE date
 142         </pre>
 143
 144       <p>I've made a small tool <a href="check-locale">check-locale</a>
 145         capable of detecting a few common mistakes with locales</p>
 146
 147     </div>
 148
 149     <hr>
 150     <address><a href="mailto:pere@hungry.com">Petter Reinholdtsen</a></address>
 151 <!-- Created: Sun Mar 21 18:14:42 CET 2004 -->
 152 <!-- hhmts start -->
 153 Last modified: Mon Dec 20 20:17:59 CET 2004
 154 <!-- hhmts end -->
 155   </body>
 156 </html>