about summary refs log tree commit diff
path: root/libidn/toutf8.c
diff options
context:
space:
mode:
Diffstat (limited to 'libidn/toutf8.c')
-rw-r--r--libidn/toutf8.c150
1 files changed, 0 insertions, 150 deletions
diff --git a/libidn/toutf8.c b/libidn/toutf8.c
deleted file mode 100644
index c7e67ca563..0000000000
--- a/libidn/toutf8.c
+++ /dev/null
@@ -1,150 +0,0 @@
-/* toutf8.c --- Convert strings from system locale into UTF-8.
- * Copyright (C) 2002, 2003, 2004, 2005  Simon Josefsson
- *
- * This file is part of GNU Libidn.
- *
- * GNU Libidn is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * GNU Libidn is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with GNU Libidn; if not, see <http://www.gnu.org/licenses/>.
- */
-
-#if HAVE_CONFIG_H
-# include "config.h"
-#endif
-
-/* Get prototypes. */
-#include "stringprep.h"
-
-/* Get fprintf. */
-#include <stdio.h>
-
-/* Get getenv. */
-#include <stdlib.h>
-
-/* Get strlen. */
-#include <string.h>
-
-/* Get iconv_string. */
-#include "iconvme.h"
-
-#ifdef _LIBC
-# define HAVE_ICONV 1
-# define LOCALE_WORKS 1
-#endif
-
-#if LOCALE_WORKS
-# include <langinfo.h>
-# include <locale.h>
-#endif
-
-#ifdef _LIBC
-# define stringprep_locale_charset() nl_langinfo (CODESET)
-#else
-/**
- * stringprep_locale_charset - return charset used in current locale
- *
- * Find out current locale charset.  The function respect the CHARSET
- * environment variable, but typically uses nl_langinfo(CODESET) when
- * it is supported.  It fall back on "ASCII" if CHARSET isn't set and
- * nl_langinfo isn't supported or return anything.
- *
- * Note that this function return the application's locale's preferred
- * charset (or thread's locale's preffered charset, if your system
- * support thread-specific locales).  It does not return what the
- * system may be using.  Thus, if you receive data from external
- * sources you cannot in general use this function to guess what
- * charset it is encoded in.  Use stringprep_convert from the external
- * representation into the charset returned by this function, to have
- * data in the locale encoding.
- *
- * Return value: Return the character set used by the current locale.
- *   It will never return NULL, but use "ASCII" as a fallback.
- **/
-const char *
-stringprep_locale_charset (void)
-{
-  const char *charset = getenv ("CHARSET");	/* flawfinder: ignore */
-
-  if (charset && *charset)
-    return charset;
-
-# ifdef LOCALE_WORKS
-  charset = nl_langinfo (CODESET);
-
-  if (charset && *charset)
-    return charset;
-# endif
-
-  return "ASCII";
-}
-#endif
-
-/**
- * stringprep_convert - encode string using new character set
- * @str: input zero-terminated string.
- * @to_codeset: name of destination character set.
- * @from_codeset: name of origin character set, as used by @str.
- *
- * Convert the string from one character set to another using the
- * system's iconv() function.
- *
- * Return value: Returns newly allocated zero-terminated string which
- *   is @str transcoded into to_codeset.
- **/
-char *
-stringprep_convert (const char *str,
-		    const char *to_codeset, const char *from_codeset)
-{
-#if HAVE_ICONV
-  return iconv_string (str, from_codeset, to_codeset);
-#else
-  char *p;
-  fprintf (stderr, "libidn: warning: libiconv not installed, cannot "
-	   "convert data to UTF-8\n");
-  p = malloc (strlen (str) + 1);
-  if (!p)
-    return NULL;
-  return strcpy (p, str);
-#endif
-}
-
-/**
- * stringprep_locale_to_utf8 - convert locale encoded string to UTF-8
- * @str: input zero terminated string.
- *
- * Convert string encoded in the locale's character set into UTF-8 by
- * using stringprep_convert().
- *
- * Return value: Returns newly allocated zero-terminated string which
- *   is @str transcoded into UTF-8.
- **/
-char *
-stringprep_locale_to_utf8 (const char *str)
-{
-  return stringprep_convert (str, "UTF-8", stringprep_locale_charset ());
-}
-
-/**
- * stringprep_utf8_to_locale - encode UTF-8 string to locale encoding
- * @str: input zero terminated string.
- *
- * Convert string encoded in UTF-8 into the locale's character set by
- * using stringprep_convert().
- *
- * Return value: Returns newly allocated zero-terminated string which
- *   is @str transcoded into the locale's character set.
- **/
-char *
-stringprep_utf8_to_locale (const char *str)
-{
-  return stringprep_convert (str, stringprep_locale_charset (), "UTF-8");
-}