/* PSPP - a program for statistical analysis.
- Copyright (C) 2006 Free Software Foundation, Inc.
+ Copyright (C) 2006, 2009 Free Software Foundation, Inc.
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
+#include <libintl.h>
#include <iconv.h>
#include <errno.h>
+#include <relocatable.h>
#include "assertion.h"
+#include "hmapx.h"
+#include "hash-functions.h"
+#include "pool.h"
#include "i18n.h"
+#include "version.h"
+
#include <localcharset.h>
#include "xstrndup.h"
+#if HAVE_NL_LANGINFO
+#include <langinfo.h>
+#endif
-static char *locale = 0;
-static const char *charset;
-
-
-static iconv_t convertor[n_CONV];
+struct converter
+ {
+ char *tocode;
+ char *fromcode;
+ iconv_t conv;
+ };
+static char *default_encoding;
+static struct hmapx map;
/* A wrapper around iconv_open */
static iconv_t
create_iconv (const char* tocode, const char* fromcode)
{
- iconv_t conv = iconv_open (tocode, fromcode);
+ size_t hash;
+ struct hmapx_node *node;
+ struct converter *converter;
+ assert (fromcode);
+
+ hash = hash_string (tocode, hash_string (fromcode, 0));
+ HMAPX_FOR_EACH_WITH_HASH (converter, node, hash, &map)
+ if (!strcmp (tocode, converter->tocode)
+ && !strcmp (fromcode, converter->fromcode))
+ return converter->conv;
+
+ converter = xmalloc (sizeof *converter);
+ converter->tocode = xstrdup (tocode);
+ converter->fromcode = xstrdup (fromcode);
+ converter->conv = iconv_open (tocode, fromcode);
+ hmapx_insert (&map, converter, hash);
/* I don't think it's safe to translate this string or to use messaging
- as the convertors have not yet been set up */
- if ( (iconv_t) -1 == conv && 0 != strcmp (tocode, fromcode))
+ as the converters have not yet been set up */
+ if ( (iconv_t) -1 == converter->conv && 0 != strcmp (tocode, fromcode))
{
const int err = errno;
fprintf (stderr,
- "Warning: cannot create a convertor for \"%s\" to \"%s\": %s\n",
- fromcode, tocode, strerror (err));
+ "Warning: "
+ "cannot create a converter for \"%s\" to \"%s\": %s\n",
+ fromcode, tocode, strerror (err));
}
- return conv;
+ return converter->conv;
}
-/* Return a string based on TEXT converted according to HOW.
- If length is not -1, then it must be the number of bytes in TEXT.
- The returned string must be freed when no longer required.
+
+/* Similar to recode_string_pool, but allocates the returned value on the heap instead of
+ in a pool. It is the caller's responsibility to free the returned value. */
+char *
+recode_string (const char *to, const char *from,
+ const char *text, int length)
+{
+ return recode_string_pool (to, from, text, length, NULL);
+}
+
+
+/*
+Converts the string TEXT, which should be encoded in FROM-encoding, to a
+dynamically allocated string in TO-encoding. Any characters which cannot
+be converted will be represented by '?'.
+
+LENGTH should be the length of the string or -1, if null terminated.
+
+The returned string will be allocated on POOL.
+
+This function's behaviour differs from that of g_convert_with_fallback provided
+by GLib. The GLib function will fail (returns NULL) if any part of the input
+string is not valid in the declared input encoding. This function however perseveres
+even in the presence of badly encoded input.
*/
char *
-recode_string (enum conv_id how, const char *text, int length)
+recode_string_pool (const char *to, const char *from,
+ const char *text, int length, struct pool *pool)
{
char *outbuf = 0;
size_t outbufferlength;
char *op ;
size_t inbytes = 0;
size_t outbytes ;
+ iconv_t conv ;
/* FIXME: Need to ensure that this char is valid in the target encoding */
const char fallbackchar = '?';
if ( length == -1 )
length = strlen(text);
- assert (how < n_CONV);
+ if (to == NULL)
+ to = default_encoding;
- if (convertor[how] == (iconv_t) -1)
- return xstrndup (text, length);
+ if (from == NULL)
+ from = default_encoding;
for ( outbufferlength = 1 ; outbufferlength != 0; outbufferlength <<= 1 )
if ( outbufferlength > length)
break;
- outbuf = xmalloc(outbufferlength);
+ outbuf = pool_malloc (pool, outbufferlength);
op = outbuf;
outbytes = outbufferlength;
inbytes = length;
+
+ conv = create_iconv (to, from);
+
+ if ( (iconv_t) -1 == conv )
+ return xstrdup (text);
+
do {
const char *ip = text;
- result = iconv (convertor[how], (ICONV_CONST char **) &text, &inbytes,
+ result = iconv (conv, (ICONV_CONST char **) &text, &inbytes,
&op, &outbytes);
if ( -1 == result )
}
/* Fall through */
case E2BIG:
- free (outbuf);
+ pool_free (pool, outbuf);
outbufferlength <<= 1;
- outbuf = xmalloc (outbufferlength);
+ outbuf = pool_malloc (pool, outbufferlength);
op = outbuf;
outbytes = outbufferlength;
inbytes = length;
break;
default:
/* should never happen */
+ fprintf (stderr, "Character conversion error: %s\n", strerror (the_error));
NOT_REACHED ();
break;
}
if (outbytes == 0 )
{
char *const oldaddr = outbuf;
- outbuf = xrealloc (outbuf, outbufferlength + 1);
+ outbuf = pool_realloc (pool, outbuf, outbufferlength + 1);
op += (outbuf - oldaddr) ;
}
}
-/* Returns the current PSPP locale */
+void
+i18n_init (void)
+{
+#if ENABLE_NLS
+ setlocale (LC_CTYPE, "");
+#ifdef LC_MESSAGES
+ setlocale (LC_MESSAGES, "");
+#endif
+#if HAVE_LC_PAPER
+ setlocale (LC_PAPER, "");
+#endif
+ bindtextdomain (PACKAGE, relocate(locale_dir));
+ textdomain (PACKAGE);
+#endif /* ENABLE_NLS */
+
+ assert (default_encoding == NULL);
+ default_encoding = xstrdup (locale_charset ());
+
+ hmapx_init (&map);
+}
+
+
const char *
-get_pspp_locale (void)
+get_default_encoding (void)
{
- assert (locale);
- return locale;
+ return default_encoding;
}
-/* Set the PSPP locale */
void
-set_pspp_locale (const char *l)
+set_default_encoding (const char *enc)
{
- char *current_locale;
- const char *current_charset;
+ free (default_encoding);
+ default_encoding = xstrdup (enc);
+}
- free(locale);
- locale = strdup(l);
- current_locale = setlocale (LC_CTYPE, 0);
- current_charset = locale_charset ();
- setlocale (LC_CTYPE, locale);
+/* Attempts to set the encoding from a locale name
+ returns true if successfull.
+ This function does not (should not!) alter the current locale.
+*/
+bool
+set_encoding_from_locale (const char *loc)
+{
+ bool ok = true;
+ char *c_encoding;
+ char *loc_encoding;
+ char *tmp = xstrdup (setlocale (LC_CTYPE, NULL));
- charset = locale_charset ();
- setlocale (LC_CTYPE, current_locale);
+ setlocale (LC_CTYPE, "C");
+ c_encoding = xstrdup (locale_charset ());
- iconv_close (convertor[CONV_PSPP_TO_UTF8]);
- convertor[CONV_PSPP_TO_UTF8] = create_iconv ("UTF-8", charset);
+ setlocale (LC_CTYPE, loc);
+ loc_encoding = xstrdup (locale_charset ());
- iconv_close (convertor[CONV_SYSTEM_TO_PSPP]);
- convertor[CONV_SYSTEM_TO_PSPP] = create_iconv (charset, current_charset);
- iconv_close (convertor[CONV_UTF8_TO_PSPP]);
- convertor[CONV_UTF8_TO_PSPP] = create_iconv (charset, "UTF-8");
-}
+ if ( 0 == strcmp (loc_encoding, c_encoding))
+ {
+ ok = false;
+ }
-void
-i18n_init (void)
-{
- assert (!locale) ;
- locale = strdup (setlocale (LC_CTYPE, NULL));
- setlocale (LC_CTYPE, locale);
- charset = locale_charset ();
+ setlocale (LC_CTYPE, tmp);
- convertor[CONV_PSPP_TO_UTF8] = create_iconv ("UTF-8", charset);
- convertor[CONV_SYSTEM_TO_PSPP] = create_iconv (charset, charset);
- convertor[CONV_UTF8_TO_PSPP] = create_iconv (charset, "UTF-8");
-}
+ free (tmp);
+ if (ok)
+ {
+ free (default_encoding);
+ default_encoding = loc_encoding;
+ }
+ else
+ free (loc_encoding);
+
+ free (c_encoding);
+
+ return ok;
+}
void
i18n_done (void)
{
- int i;
- free (locale);
- locale = 0;
+ struct hmapx_node *node;
+ struct converter *cvtr;
- for(i = 0 ; i < n_CONV; ++i )
+ HMAPX_FOR_EACH (cvtr, node, &map)
{
- if ( (iconv_t) -1 == convertor[i] )
- continue;
- iconv_close (convertor[i]);
+ free (cvtr->tocode);
+ free (cvtr->fromcode);
+ iconv_close (cvtr->conv);
+ free (cvtr);
}
+
+ hmapx_destroy (&map);
+
+ free (default_encoding);
+ default_encoding = NULL;
+}
+
+
+
+bool
+valid_encoding (const char *enc)
+{
+ iconv_t conv = iconv_open ("UTF8", enc);
+
+ if ( conv == (iconv_t) -1)
+ return false;
+
+ iconv_close (conv);
+
+ return true;
+}
+
+
+/* Return the system local's idea of the
+ decimal seperator character */
+char
+get_system_decimal (void)
+{
+ char radix_char;
+
+ char *ol = xstrdup (setlocale (LC_NUMERIC, NULL));
+ setlocale (LC_NUMERIC, "");
+
+#if HAVE_NL_LANGINFO
+ radix_char = nl_langinfo (RADIXCHAR)[0];
+#else
+ {
+ char buf[10];
+ snprintf (buf, sizeof buf, "%f", 2.5);
+ radix_char = buf[1];
+ }
+#endif
+
+ /* We MUST leave LC_NUMERIC untouched, since it would
+ otherwise interfere with data_{in,out} */
+ setlocale (LC_NUMERIC, ol);
+ free (ol);
+ return radix_char;
}