2 * Copyright (C) 2001 Edmund Grimley Evans <edmundo@rano.org>
4 * This program is free software; you can redistribute it and/or modify
5 * it under the terms of the GNU General Public License as published by
6 * the Free Software Foundation; either version 2 of the License, or
7 * (at your option) any later version.
9 * This program is distributed in the hope that it will be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 * GNU General Public License for more details.
14 * You should have received a copy of the GNU General Public License
15 * along with this program; if not, write to the Free Software
16 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
32 #include "share/alloc.h"
35 * Convert data from one encoding to another. Return:
37 * -2 : memory allocation failed
38 * -1 : unknown encoding
39 * 0 : data was converted exactly
40 * 1 : data was converted inexactly
41 * 2 : data was invalid (but still converted)
43 * We convert in two steps, via UTF-8, as this is the only
44 * reliable way of distinguishing between invalid input
45 * and valid input which iconv refuses to transliterate.
46 * We convert from UTF-8 twice, because we have no way of
47 * knowing whether the conversion was exact if iconv returns
48 * E2BIG (due to a bug in the specification of iconv).
49 * An alternative approach is to assume that the output of
50 * iconv is never more than 4 times as long as the input,
51 * but I prefer to avoid that assumption if possible.
54 int iconvert(const char *fromcode, const char *tocode,
55 const char *from, size_t fromlen,
56 char **to, size_t *tolen)
62 char *utfbuf = 0, *outbuf, *newbuf;
63 size_t utflen, outlen, ibl, obl, k;
66 cd1 = iconv_open("UTF-8", fromcode);
67 if (cd1 == (iconv_t)(-1))
71 /* Don't use strcasecmp() as it's locale-dependent. */
72 if (!strchr("Uu", tocode[0]) ||
73 !strchr("Tt", tocode[1]) ||
74 !strchr("Ff", tocode[2]) ||
81 * Try using this non-standard feature of glibc and libiconv.
82 * This is deliberately not a config option as people often
83 * change their iconv library without rebuilding applications.
85 tocode1 = (char *)safe_malloc_add_2op_(strlen(tocode), /*+*/11);
89 strcpy(tocode1, tocode);
90 strcat(tocode1, "//TRANSLIT");
91 cd2 = iconv_open(tocode1, "UTF-8");
94 if (cd2 == (iconv_t)(-1))
95 cd2 = iconv_open(tocode, fromcode);
97 if (cd2 == (iconv_t)(-1)) {
103 utflen = 1; /*fromlen * 2 + 1; XXX */
104 utfbuf = (char *)malloc(utflen);
108 /* Convert to UTF-8 */
114 k = iconv(cd1, &ib, &ibl, &ob, &obl);
115 assert((!k && !ibl) ||
116 (k == (size_t)(-1) && errno == E2BIG && ibl && obl < 6) ||
117 (k == (size_t)(-1) &&
118 (errno == EILSEQ || errno == EINVAL) && ibl));
122 /* Enlarge the buffer */
123 if(utflen*2 < utflen) /* overflow check */
126 newbuf = (char *)realloc(utfbuf, utflen);
129 ob = (ob - utfbuf) + newbuf;
130 obl = utflen - (ob - newbuf);
138 iconv(cd1, 0, 0, 0, 0);
142 if (cd2 == (iconv_t)(-1)) {
143 /* The target encoding was UTF-8 */
145 *tolen = ob - utfbuf;
151 newbuf = (char *)safe_realloc_add_2op_(utfbuf, (ob - utfbuf), /*+*/1);
154 ob = (ob - utfbuf) + newbuf;
161 /* Truncate the buffer to be tidy */
162 utflen = ob - utfbuf;
163 newbuf = (char *)realloc(utfbuf, utflen);
168 /* Convert from UTF-8 to discover how long the output is */
175 k = iconv(cd2, &ib, &ibl, &ob, &obl);
176 assert((k != (size_t)(-1) && !ibl) ||
177 (k == (size_t)(-1) && errno == E2BIG && ibl) ||
178 (k == (size_t)(-1) && errno == EILSEQ && ibl));
179 if (ibl && !(k == (size_t)(-1) && errno == E2BIG)) {
180 /* Replace one character */
187 k = iconv(cd2, &tb, &tbl, &ob, &obl);
188 assert((!k && !tbl) ||
189 (k == (size_t)(-1) && errno == EILSEQ && tbl));
190 for (++ib, --ibl; ibl && (*ib & 0x80); ib++, ibl--)
197 k = iconv(cd2, 0, 0, &ob, &obl);
201 /* Convert from UTF-8 for real */
202 outbuf = (char *)safe_malloc_add_2op_(outlen, /*+*/1);
210 k = iconv(cd2, &ib, &ibl, &ob, &obl);
211 assert((k != (size_t)(-1) && !ibl) ||
212 (k == (size_t)(-1) && errno == EILSEQ && ibl));
215 if (ibl && !(k == (size_t)(-1) && errno == E2BIG)) {
216 /* Replace one character */
220 k = iconv(cd2, &tb, &tbl, &ob, &obl);
221 assert((!k && !tbl) ||
222 (k == (size_t)(-1) && errno == EILSEQ && tbl));
223 for (++ib, --ibl; ibl && (*ib & 0x80); ib++, ibl--)
227 k = iconv(cd2, 0, 0, &ob, &obl);
248 if (cd2 != (iconv_t)(-1))
253 #endif /* HAVE_ICONV */