Re: [PATCH] Extra decimal point and thousands-sep

[Hamed Malek <hamed at bamdad dot org>]

I've slightly changed my patch.

Hamed

On Wed, 2004-02-04 at 17:56, Hamed Malek wrote:
> Hi,
> 
> Some languages like Persian need extra decimal point and thousands-sep. 
> Attached is a patch which adds these new elements to glibc.
> 
> I wrote it based on Ulrich suggestion, which defines a new map
> (to_outpunct) for those locales, and call it for I18n numbers.
> The patch also contains the necessary changes in fa_IR.
> 
> Hamed
> 

--- glibc-2.3.2-200310271512.orig/include/wctype.h	2004-01-05 18:07:20.000000000 +0330
+++ glibc-2.3.2-200310271512/include/wctype.h	2004-02-03 17:52:08.000000000 +0330
@@ -49,6 +49,7 @@
 extern int __iswctype (wint_t __wc, wctype_t __desc);
 extern int __iswctype_internal (wint_t __wc, wctype_t __desc) attribute_hidden;
 extern wctype_t __wctype (__const char *__property);
+extern wctrans_t __wctrans (const char *__property);
 extern wint_t __towctrans (wint_t __wc, wctrans_t __desc);
 
 extern __typeof (iswalnum_l) __iswalnum_l;
--- glibc-2.3.2-200310271512.orig/localedata/locales/fa_IR	2004-01-05 18:07:26.000000000 +0330
+++ glibc-2.3.2-200310271512/localedata/locales/fa_IR	2004-02-03 18:03:35.000000000 +0330
@@ -49,6 +49,10 @@
 LC_CTYPE
 copy "i18n"
 
+map to_outpunct; /
+  (<U002E>,<U066B>); /
+  (<U002C>,<U066C>)
+
 outdigit <U06F0>..<U06F9>
 END LC_CTYPE
 
@@ -269,8 +273,8 @@
 END LC_MONETARY
 
 LC_NUMERIC
-decimal_point     "<U066B>"
-thousands_sep     "<U066C>"
+decimal_point     "<U002E>"
+thousands_sep     "<U002C>"
 grouping          3
 END LC_NUMERIC
 
--- glibc-2.3.2-200310271512.orig/stdio-common/_i18n_number.h	2004-01-05 18:07:41.000000000 +0330
+++ glibc-2.3.2-200310271512/stdio-common/_i18n_number.h	2004-02-08 14:23:33.000000000 +0330
@@ -17,12 +17,54 @@
    Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
    02111-1307 USA.  */
 
+#include <wctype.h>
+#include <wchar.h>
+
 #include "../locale/outdigits.h"
 #include "../locale/outdigitswc.h"
 
+static char *
+_put_outpunct (char *s, const char *outpunct)
+{
+  size_t dlen;
+  dlen = strlen (outpunct);
+
+  s -= dlen;
+  while (dlen-- > 0)
+    s[dlen] = outpunct[dlen];
+
+  return s;
+}
+
 static CHAR_T *
 _i18n_number_rewrite (CHAR_T *w, CHAR_T *rear_ptr)
 {
+  /* "to_outpunct" is a map from ASCII decimal point and thousands-sep
+     to their equivalent in locale. This is defined for locales which 
+     use extra decimal point and thousands-sep.  */
+
+  wctrans_t map = __wctrans ("to_outpunct");
+
+  wint_t wdecimal = (map) ? __towctrans ('.', map) : L'.';
+  wint_t wthousands = (map) ?  __towctrans (',', map) : L'.';
+  
+  char decimal[MB_LEN_MAX];
+  char thousands[MB_LEN_MAX];
+ 
+  mbstate_t state;
+  memset (&state, '\0', sizeof (state));
+  wint_t n;
+ 
+  n = (map) ? __wcrtomb (decimal, wdecimal, &state) : (size_t)-1;
+  if (n == (size_t)-1)
+    memcpy (decimal, ".", 2);
+ 
+  memset (&state, '\0', sizeof (state));
+
+  n = (map) ? __wcrtomb (thousands, wthousands, &state) : (size_t)-1;
+  if (n == (size_t)-1)
+    memcpy (thousands, ",", 2);
+
   CHAR_T *src, *s;
 
   /* Copy existing string so that nothing gets overwritten.  */
@@ -41,6 +83,20 @@
 	  else
 	    *--w = (CHAR_T) outdigitwc_value (*s - '0');
 	}
+      else if (map && (*s == '.'))
+        {
+	  if (sizeof (CHAR_T) == 1)
+	    w = _put_outpunct (w, decimal);
+	  else
+	    *--w = (CHAR_T) wdecimal;
+        }
+      else if (map && (*s == ','))
+        {
+	  if (sizeof (CHAR_T) == 1)
+	    w = _put_outpunct (w, thousands);
+	  else
+	    *--w = (CHAR_T) wthousands;
+        }
       else
 	*--w = *s;
     }
--- glibc-2.3.2-200310271512.orig/wctype/wctrans.c	2004-01-05 18:08:12.000000000 +0330
+++ glibc-2.3.2-200310271512/wctype/wctrans.c	2004-02-03 17:56:57.000000000 +0330
@@ -1,4 +1,4 @@
-/* Copyright (C) 1996,97,99,2000,02 Free Software Foundation, Inc.
+/* Copyright (C) 1996,97,99,2000,02,04 Free Software Foundation, Inc.
    This file is part of the GNU C Library.
    Contributed by Ulrich Drepper <drepper@gnu.org>, 1996.
 
@@ -23,7 +23,7 @@
 #include "../locale/localeinfo.h"
 
 wctrans_t
-wctrans (const char *property)
+__wctrans (const char *property)
 {
   const char *names;
   size_t cnt;
@@ -46,3 +46,4 @@
   i = _NL_CURRENT_WORD (LC_CTYPE, _NL_CTYPE_MAP_OFFSET) + cnt;
   return (wctrans_t) _NL_CURRENT_DATA (LC_CTYPE)->values[i].string;
 }
+weak_alias (__wctrans, wctrans)