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.
44 #include <ac/stdlib.h> /* For wctomb, wcstombs, mbtowc, mbstowcs */
45 #include <ac/string.h>
46 #include <ac/time.h> /* for time_t */
50 #include <ldap_utf8.h>
52 static unsigned char mask[] = { 0, 0x7f, 0x1f, 0x0f, 0x07, 0x03, 0x01 };
55 /*-----------------------------------------------------------------------------
61 2-character UTF-8 sequence: 11 bits
64 3-character UTF-8 16 bits
65 1110xxxx 10xxxxxx 10xxxxxx
68 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
71 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
74 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
76 Unicode address space (0 - 0x10FFFF) 21 bits
77 ISO-10646 address space (0 - 0x7FFFFFFF) 31 bits
79 Note: This code does not prevent UTF-8 sequences which are longer than
80 necessary from being decoded.
83 /*-----------------------------------------------------------------------------
84 Convert a UTF-8 character to a wide char.
85 Return the length of the UTF-8 input character in bytes.
88 ldap_x_utf8_to_wc ( wchar_t *wchar, const char *utf8char )
93 /* If input ptr is NULL, treat it as empty string. */
97 /* Get UTF-8 sequence length from 1st byte */
98 utflen = LDAP_UTF8_CHARLEN2(utf8char, utflen);
100 if( utflen==0 || utflen > (int)LDAP_MAX_UTF8_LEN )
101 return -1; /* Invalid input */
103 /* First byte minus length tag */
104 ch = (wchar_t)(utf8char[0] & mask[utflen]);
106 for(i=1; i < utflen; i++)
108 /* Subsequent bytes must start with 10 */
109 if ((utf8char[i] & 0xc0) != 0x80)
112 ch <<= 6; /* 6 bits of data in each subsequent byte */
113 ch |= (wchar_t)(utf8char[i] & 0x3f);
122 /*-----------------------------------------------------------------------------
123 Convert a UTF-8 string to a wide char string.
124 No more than 'count' wide chars will be written to the output buffer.
125 Return the size of the converted string in wide chars, excl null terminator.
128 ldap_x_utf8s_to_wcs ( wchar_t *wcstr, const char *utf8str, size_t count )
135 /* If input ptr is NULL, treat it as empty string. */
139 /* Examine next UTF-8 character. If output buffer is NULL, ignore count */
140 while ( *utf8str && (wcstr==NULL || wclen<count) )
142 /* Get UTF-8 sequence length from 1st byte */
143 utflen = LDAP_UTF8_CHARLEN2(utf8str, utflen);
145 if( utflen==0 || utflen > (int)LDAP_MAX_UTF8_LEN )
146 return -1; /* Invalid input */
148 /* First byte minus length tag */
149 ch = (wchar_t)(utf8str[0] & mask[utflen]);
151 for(i=1; i < utflen; i++)
153 /* Subsequent bytes must start with 10 */
154 if ((utf8str[i] & 0xc0) != 0x80)
157 ch <<= 6; /* 6 bits of data in each subsequent byte */
158 ch |= (wchar_t)(utf8str[i] & 0x3f);
164 utf8str += utflen; /* Move to next UTF-8 character */
165 wclen++; /* Count number of wide chars stored/required */
168 /* Add null terminator if there's room in the buffer. */
169 if (wcstr && wclen < count)
176 /*-----------------------------------------------------------------------------
177 Convert one wide char to a UTF-8 character.
178 Return the length of the converted UTF-8 character in bytes.
179 No more than 'count' bytes will be written to the output buffer.
182 ldap_x_wc_to_utf8 ( char *utf8char, wchar_t wchar, size_t count )
186 if (utf8char == NULL) /* Just determine the required UTF-8 char length. */
194 if( wchar < 0x10000 )
196 if( wchar < 0x200000 )
198 if( wchar < 0x4000000 )
200 if( wchar < 0x80000000 )
206 if ( wchar < 0 ) { /* Invalid wide character */
209 } else if( wchar < 0x80 ) {
211 utf8char[len++] = (char)wchar;
214 } else if( wchar < 0x800 ) {
216 utf8char[len++] = 0xc0 | ( wchar >> 6 );
217 utf8char[len++] = 0x80 | ( wchar & 0x3f );
220 } else if( wchar < 0x10000 ) {
222 utf8char[len++] = 0xe0 | ( wchar >> 12 );
223 utf8char[len++] = 0x80 | ( (wchar >> 6) & 0x3f );
224 utf8char[len++] = 0x80 | ( wchar & 0x3f );
227 } else if( wchar < 0x200000 ) {
229 utf8char[len++] = 0xf0 | ( wchar >> 18 );
230 utf8char[len++] = 0x80 | ( (wchar >> 12) & 0x3f );
231 utf8char[len++] = 0x80 | ( (wchar >> 6) & 0x3f );
232 utf8char[len++] = 0x80 | ( wchar & 0x3f );
235 } else if( wchar < 0x4000000 ) {
237 utf8char[len++] = 0xf8 | ( wchar >> 24 );
238 utf8char[len++] = 0x80 | ( (wchar >> 18) & 0x3f );
239 utf8char[len++] = 0x80 | ( (wchar >> 12) & 0x3f );
240 utf8char[len++] = 0x80 | ( (wchar >> 6) & 0x3f );
241 utf8char[len++] = 0x80 | ( wchar & 0x3f );
244 } else if( wchar < 0x80000000 ) {
246 utf8char[len++] = 0xfc | ( wchar >> 30 );
247 utf8char[len++] = 0x80 | ( (wchar >> 24) & 0x3f );
248 utf8char[len++] = 0x80 | ( (wchar >> 18) & 0x3f );
249 utf8char[len++] = 0x80 | ( (wchar >> 12) & 0x3f );
250 utf8char[len++] = 0x80 | ( (wchar >> 6) & 0x3f );
251 utf8char[len++] = 0x80 | ( wchar & 0x3f );
262 /*-----------------------------------------------------------------------------
263 Convert a wide char string to a UTF-8 string.
264 No more than 'count' bytes will be written to the output buffer.
265 Return the # of bytes written to the output buffer, excl null terminator.
268 ldap_x_wcs_to_utf8s ( char *utf8str, const wchar_t *wcstr, size_t count )
273 wchar_t empty = 0; /* To avoid use of L"" construct */
275 if (wcstr == NULL) /* Treat input ptr NULL as an empty string */
278 if (utf8str == NULL) /* Just compute size of output, excl null */
282 /* Get UTF-8 size of next wide char */
283 n = ldap_x_wc_to_utf8( NULL, *wcstr++, LDAP_MAX_UTF8_LEN);
293 /* Do the actual conversion. */
295 n = 1; /* In case of empty wcstr */
298 n = ldap_x_wc_to_utf8( p, *wcstr++, count);
300 if (n <= 0) /* If encoding error (-1) or won't fit (0), quit */
304 count -= n; /* Space left in output buffer */
307 /* If not enough room for last character, pad remainder with null
308 so that return value = original count, indicating buffer full. */
315 /* Add a null terminator if there's room. */
319 if (n == -1) /* Conversion encountered invalid wide char. */
322 /* Return the number of bytes written to output buffer, excl null. */
323 return (p - utf8str);
327 /*-----------------------------------------------------------------------------
328 Convert a UTF-8 character to a MultiByte character.
329 Return the size of the converted character in bytes.
332 ldap_x_utf8_to_mb ( char *mbchar, const char *utf8char,
333 int (*f_wctomb)(char *mbchar, wchar_t wchar) )
337 char tmp[6]; /* Large enough for biggest multibyte char */
339 if (f_wctomb == NULL) /* If no conversion function was given... */
340 f_wctomb = wctomb; /* use the local ANSI C function */
342 /* First convert UTF-8 char to a wide char */
343 n = ldap_x_utf8_to_wc( &wchar, utf8char);
346 return -1; /* Invalid UTF-8 character */
349 n = f_wctomb( tmp, wchar );
351 n = f_wctomb( mbchar, wchar);
356 /*-----------------------------------------------------------------------------
357 Convert a UTF-8 string to a MultiByte string.
358 No more than 'count' bytes will be written to the output buffer.
359 Return the size of the converted string in bytes, excl null terminator.
362 ldap_x_utf8s_to_mbs ( char *mbstr, const char *utf8str, size_t count,
363 size_t (*f_wcstombs)(char *mbstr, const wchar_t *wcstr, size_t count) )
369 if (f_wcstombs == NULL) /* If no conversion function was given... */
370 f_wcstombs = wcstombs; /* use the local ANSI C function */
372 if (utf8str == NULL || *utf8str == 0) /* NULL or empty input string */
379 /* Allocate memory for the maximum size wchar string that we could get. */
380 wcsize = strlen(utf8str) + 1;
381 wcs = (wchar_t *)LDAP_MALLOC(wcsize * sizeof(wchar_t));
383 return -1; /* Memory allocation failure. */
385 /* First convert the UTF-8 string to a wide char string */
386 n = ldap_x_utf8s_to_wcs( wcs, utf8str, wcsize);
388 /* Then convert wide char string to multi-byte string */
391 n = f_wcstombs(mbstr, wcs, count);
399 /*-----------------------------------------------------------------------------
400 Convert a MultiByte character to a UTF-8 character.
401 'mbsize' indicates the number of bytes of 'mbchar' to check.
402 Returns the number of bytes written to the output character.
405 ldap_x_mb_to_utf8 ( char *utf8char, const char *mbchar, size_t mbsize,
406 int (*f_mbtowc)(wchar_t *wchar, const char *mbchar, size_t count) )
411 if (f_mbtowc == NULL) /* If no conversion function was given... */
412 f_mbtowc = mbtowc; /* use the local ANSI C function */
414 if (mbsize == 0) /* 0 is not valid. */
417 if (mbchar == NULL || *mbchar == 0)
424 /* First convert the MB char to a Wide Char */
425 n = f_mbtowc( &wchar, mbchar, mbsize);
430 /* Convert the Wide Char to a UTF-8 character. */
431 n = ldap_x_wc_to_utf8( utf8char, wchar, LDAP_MAX_UTF8_LEN);
437 /*-----------------------------------------------------------------------------
438 Convert a MultiByte string to a UTF-8 string.
439 No more than 'count' bytes will be written to the output buffer.
440 Return the size of the converted string in bytes, excl null terminator.
443 ldap_x_mbs_to_utf8s ( char *utf8str, const char *mbstr, size_t count,
444 size_t (*f_mbstowcs)(wchar_t *wcstr, const char *mbstr, size_t count) )
450 if (mbstr == NULL) /* Treat NULL input string as an empty string */
453 if (f_mbstowcs == NULL) /* If no conversion function was given... */
454 f_mbstowcs = mbstowcs; /* use the local ANSI C function */
456 /* Allocate memory for the maximum size wchar string that we could get. */
457 wcsize = strlen(mbstr) + 1;
458 wcs = (wchar_t *)LDAP_MALLOC( wcsize * sizeof(wchar_t) );
462 /* First convert multi-byte string to a wide char string */
463 n = f_mbstowcs(wcs, mbstr, wcsize);
465 /* Convert wide char string to UTF-8 string */
468 n = ldap_x_wcs_to_utf8s( utf8str, wcs, count);