2 /* This work is part of OpenLDAP Software <http://www.openldap.org/>.
4 * Copyright 1998-2004 The OpenLDAP Foundation.
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted only as authorized by the OpenLDAP
11 * A copy of this license is available in the file LICENSE in the
12 * top-level directory of the distribution or, alternatively, at
13 * <http://www.OpenLDAP.org/license.html>.
15 /* Portions Copyright (C) 1999, 2000 Novell, Inc. All Rights Reserved.
17 * THIS WORK IS SUBJECT TO U.S. AND INTERNATIONAL COPYRIGHT LAWS AND
18 * TREATIES. USE, MODIFICATION, AND REDISTRIBUTION OF THIS WORK IS SUBJECT
19 * TO VERSION 2.0.1 OF THE OPENLDAP PUBLIC LICENSE, A COPY OF WHICH IS
20 * AVAILABLE AT HTTP://WWW.OPENLDAP.ORG/LICENSE.HTML OR IN THE FILE "LICENSE"
21 * IN THE TOP-LEVEL DIRECTORY OF THE DISTRIBUTION. ANY USE OR EXPLOITATION
22 * OF THIS WORK OTHER THAN AS AUTHORIZED IN VERSION 2.0.1 OF THE OPENLDAP
23 * PUBLIC LICENSE, OR OTHER PRIOR WRITTEN CONSENT FROM NOVELL, COULD SUBJECT
24 * THE PERPETRATOR TO CRIMINAL AND CIVIL LIABILITY.
26 * Note: A verbatim copy of version 2.0.1 of the OpenLDAP Public License
27 * can be found in the file "build/LICENSE-2.0.1" in this distribution
28 * of OpenLDAP Software.
32 * UTF-8 Conversion Routines
34 * These routines convert between Wide Character and UTF-8,
35 * or between MultiByte and UTF-8 encodings.
37 * Both single character and string versions of the functions are provided.
38 * All functions return -1 if the character or string cannot be converted.
43 #if SIZEOF_WCHAR_T >= 4
44 /* These routines assume ( sizeof(wchar_t) >= 4 ) */
47 #include <ac/stdlib.h> /* For wctomb, wcstombs, mbtowc, mbstowcs */
48 #include <ac/string.h>
49 #include <ac/time.h> /* for time_t */
53 #include <ldap_utf8.h>
55 static unsigned char mask[] = { 0, 0x7f, 0x1f, 0x0f, 0x07, 0x03, 0x01 };
58 /*-----------------------------------------------------------------------------
64 2-character UTF-8 sequence: 11 bits
67 3-character UTF-8 16 bits
68 1110xxxx 10xxxxxx 10xxxxxx
71 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
74 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
77 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
79 Unicode address space (0 - 0x10FFFF) 21 bits
80 ISO-10646 address space (0 - 0x7FFFFFFF) 31 bits
82 Note: This code does not prevent UTF-8 sequences which are longer than
83 necessary from being decoded.
86 /*-----------------------------------------------------------------------------
87 Convert a UTF-8 character to a wide char.
88 Return the length of the UTF-8 input character in bytes.
91 ldap_x_utf8_to_wc ( wchar_t *wchar, const char *utf8char )
96 /* If input ptr is NULL, treat it as empty string. */
100 /* Get UTF-8 sequence length from 1st byte */
101 utflen = LDAP_UTF8_CHARLEN2(utf8char, utflen);
103 if( utflen==0 || utflen > (int)LDAP_MAX_UTF8_LEN )
104 return -1; /* Invalid input */
106 /* First byte minus length tag */
107 ch = (wchar_t)(utf8char[0] & mask[utflen]);
109 for(i=1; i < utflen; i++)
111 /* Subsequent bytes must start with 10 */
112 if ((utf8char[i] & 0xc0) != 0x80)
115 ch <<= 6; /* 6 bits of data in each subsequent byte */
116 ch |= (wchar_t)(utf8char[i] & 0x3f);
125 /*-----------------------------------------------------------------------------
126 Convert a UTF-8 string to a wide char string.
127 No more than 'count' wide chars will be written to the output buffer.
128 Return the size of the converted string in wide chars, excl null terminator.
131 ldap_x_utf8s_to_wcs ( wchar_t *wcstr, const char *utf8str, size_t count )
138 /* If input ptr is NULL, treat it as empty string. */
142 /* Examine next UTF-8 character. If output buffer is NULL, ignore count */
143 while ( *utf8str && (wcstr==NULL || wclen<count) )
145 /* Get UTF-8 sequence length from 1st byte */
146 utflen = LDAP_UTF8_CHARLEN2(utf8str, utflen);
148 if( utflen==0 || utflen > (int)LDAP_MAX_UTF8_LEN )
149 return -1; /* Invalid input */
151 /* First byte minus length tag */
152 ch = (wchar_t)(utf8str[0] & mask[utflen]);
154 for(i=1; i < utflen; i++)
156 /* Subsequent bytes must start with 10 */
157 if ((utf8str[i] & 0xc0) != 0x80)
160 ch <<= 6; /* 6 bits of data in each subsequent byte */
161 ch |= (wchar_t)(utf8str[i] & 0x3f);
167 utf8str += utflen; /* Move to next UTF-8 character */
168 wclen++; /* Count number of wide chars stored/required */
171 /* Add null terminator if there's room in the buffer. */
172 if (wcstr && wclen < count)
179 /*-----------------------------------------------------------------------------
180 Convert one wide char to a UTF-8 character.
181 Return the length of the converted UTF-8 character in bytes.
182 No more than 'count' bytes will be written to the output buffer.
185 ldap_x_wc_to_utf8 ( char *utf8char, wchar_t wchar, size_t count )
189 if (utf8char == NULL) /* Just determine the required UTF-8 char length. */
197 if( wchar < 0x10000 )
199 if( wchar < 0x200000 )
201 if( wchar < 0x4000000 )
203 if( wchar < 0x80000000 )
209 if ( wchar < 0 ) { /* Invalid wide character */
212 } else if( wchar < 0x80 ) {
214 utf8char[len++] = (char)wchar;
217 } else if( wchar < 0x800 ) {
219 utf8char[len++] = 0xc0 | ( wchar >> 6 );
220 utf8char[len++] = 0x80 | ( wchar & 0x3f );
223 } else if( wchar < 0x10000 ) {
225 utf8char[len++] = 0xe0 | ( wchar >> 12 );
226 utf8char[len++] = 0x80 | ( (wchar >> 6) & 0x3f );
227 utf8char[len++] = 0x80 | ( wchar & 0x3f );
230 } else if( wchar < 0x200000 ) {
232 utf8char[len++] = 0xf0 | ( wchar >> 18 );
233 utf8char[len++] = 0x80 | ( (wchar >> 12) & 0x3f );
234 utf8char[len++] = 0x80 | ( (wchar >> 6) & 0x3f );
235 utf8char[len++] = 0x80 | ( wchar & 0x3f );
238 } else if( wchar < 0x4000000 ) {
240 utf8char[len++] = 0xf8 | ( wchar >> 24 );
241 utf8char[len++] = 0x80 | ( (wchar >> 18) & 0x3f );
242 utf8char[len++] = 0x80 | ( (wchar >> 12) & 0x3f );
243 utf8char[len++] = 0x80 | ( (wchar >> 6) & 0x3f );
244 utf8char[len++] = 0x80 | ( wchar & 0x3f );
247 } else if( wchar < 0x80000000 ) {
249 utf8char[len++] = 0xfc | ( wchar >> 30 );
250 utf8char[len++] = 0x80 | ( (wchar >> 24) & 0x3f );
251 utf8char[len++] = 0x80 | ( (wchar >> 18) & 0x3f );
252 utf8char[len++] = 0x80 | ( (wchar >> 12) & 0x3f );
253 utf8char[len++] = 0x80 | ( (wchar >> 6) & 0x3f );
254 utf8char[len++] = 0x80 | ( wchar & 0x3f );
265 /*-----------------------------------------------------------------------------
266 Convert a wide char string to a UTF-8 string.
267 No more than 'count' bytes will be written to the output buffer.
268 Return the # of bytes written to the output buffer, excl null terminator.
271 ldap_x_wcs_to_utf8s ( char *utf8str, const wchar_t *wcstr, size_t count )
276 wchar_t empty = 0; /* To avoid use of L"" construct */
278 if (wcstr == NULL) /* Treat input ptr NULL as an empty string */
281 if (utf8str == NULL) /* Just compute size of output, excl null */
285 /* Get UTF-8 size of next wide char */
286 n = ldap_x_wc_to_utf8( NULL, *wcstr++, LDAP_MAX_UTF8_LEN);
296 /* Do the actual conversion. */
298 n = 1; /* In case of empty wcstr */
301 n = ldap_x_wc_to_utf8( p, *wcstr++, count);
303 if (n <= 0) /* If encoding error (-1) or won't fit (0), quit */
307 count -= n; /* Space left in output buffer */
310 /* If not enough room for last character, pad remainder with null
311 so that return value = original count, indicating buffer full. */
318 /* Add a null terminator if there's room. */
322 if (n == -1) /* Conversion encountered invalid wide char. */
325 /* Return the number of bytes written to output buffer, excl null. */
326 return (p - utf8str);
330 /*-----------------------------------------------------------------------------
331 Convert a UTF-8 character to a MultiByte character.
332 Return the size of the converted character in bytes.
335 ldap_x_utf8_to_mb ( char *mbchar, const char *utf8char,
336 int (*f_wctomb)(char *mbchar, wchar_t wchar) )
340 char tmp[6]; /* Large enough for biggest multibyte char */
342 if (f_wctomb == NULL) /* If no conversion function was given... */
343 f_wctomb = wctomb; /* use the local ANSI C function */
345 /* First convert UTF-8 char to a wide char */
346 n = ldap_x_utf8_to_wc( &wchar, utf8char);
349 return -1; /* Invalid UTF-8 character */
352 n = f_wctomb( tmp, wchar );
354 n = f_wctomb( mbchar, wchar);
359 /*-----------------------------------------------------------------------------
360 Convert a UTF-8 string to a MultiByte string.
361 No more than 'count' bytes will be written to the output buffer.
362 Return the size of the converted string in bytes, excl null terminator.
365 ldap_x_utf8s_to_mbs ( char *mbstr, const char *utf8str, size_t count,
366 size_t (*f_wcstombs)(char *mbstr, const wchar_t *wcstr, size_t count) )
372 if (f_wcstombs == NULL) /* If no conversion function was given... */
373 f_wcstombs = wcstombs; /* use the local ANSI C function */
375 if (utf8str == NULL || *utf8str == 0) /* NULL or empty input string */
382 /* Allocate memory for the maximum size wchar string that we could get. */
383 wcsize = strlen(utf8str) + 1;
384 wcs = (wchar_t *)LDAP_MALLOC(wcsize * sizeof(wchar_t));
386 return -1; /* Memory allocation failure. */
388 /* First convert the UTF-8 string to a wide char string */
389 n = ldap_x_utf8s_to_wcs( wcs, utf8str, wcsize);
391 /* Then convert wide char string to multi-byte string */
394 n = f_wcstombs(mbstr, wcs, count);
402 /*-----------------------------------------------------------------------------
403 Convert a MultiByte character to a UTF-8 character.
404 'mbsize' indicates the number of bytes of 'mbchar' to check.
405 Returns the number of bytes written to the output character.
408 ldap_x_mb_to_utf8 ( char *utf8char, const char *mbchar, size_t mbsize,
409 int (*f_mbtowc)(wchar_t *wchar, const char *mbchar, size_t count) )
414 if (f_mbtowc == NULL) /* If no conversion function was given... */
415 f_mbtowc = mbtowc; /* use the local ANSI C function */
417 if (mbsize == 0) /* 0 is not valid. */
420 if (mbchar == NULL || *mbchar == 0)
427 /* First convert the MB char to a Wide Char */
428 n = f_mbtowc( &wchar, mbchar, mbsize);
433 /* Convert the Wide Char to a UTF-8 character. */
434 n = ldap_x_wc_to_utf8( utf8char, wchar, LDAP_MAX_UTF8_LEN);
440 /*-----------------------------------------------------------------------------
441 Convert a MultiByte string to a UTF-8 string.
442 No more than 'count' bytes will be written to the output buffer.
443 Return the size of the converted string in bytes, excl null terminator.
446 ldap_x_mbs_to_utf8s ( char *utf8str, const char *mbstr, size_t count,
447 size_t (*f_mbstowcs)(wchar_t *wcstr, const char *mbstr, size_t count) )
453 if (mbstr == NULL) /* Treat NULL input string as an empty string */
456 if (f_mbstowcs == NULL) /* If no conversion function was given... */
457 f_mbstowcs = mbstowcs; /* use the local ANSI C function */
459 /* Allocate memory for the maximum size wchar string that we could get. */
460 wcsize = strlen(mbstr) + 1;
461 wcs = (wchar_t *)LDAP_MALLOC( wcsize * sizeof(wchar_t) );
465 /* First convert multi-byte string to a wide char string */
466 n = f_mbstowcs(wcs, mbstr, wcsize);
468 /* Convert wide char string to UTF-8 string */
471 n = ldap_x_wcs_to_utf8s( utf8str, wcs, count);