2 * Copyright (c) 1993, 1994
3 * The Regents of the University of California. All rights reserved.
4 * Copyright (c) 1993, 1994, 1995, 1996
5 * Keith Bostic. All rights reserved.
6 * Copyright (c) 2011, 2012
7 * Zhihao Yuan. All rights reserved.
9 * See the LICENSE file for redistribution information.
15 static const char sccsid[] = "$Id: conv.c,v 2.39 2013/07/01 23:28:13 zy Exp $";
18 #include <sys/types.h>
19 #include <sys/queue.h>
22 #include <bitstring.h>
37 * Get the locale encoding.
39 * PUBLIC: char * codeset __P((void));
46 cs = nl_langinfo(CODESET);
52 raw2int(SCR *sp, const char * str, ssize_t len, CONVWIN *cw,
53 size_t *tolen, CHAR_T **dst)
56 CHAR_T **tostr = &cw->bp1.wc;
57 size_t *blen = &cw->blen1;
59 BINC_RETW(NULL, *tostr, *blen, len);
62 for (i = 0; i < len; ++i)
63 (*tostr)[i] = (u_char) str[i];
70 #define CONV_BUFFER_SIZE 512
71 /* fill the buffer with codeset encoding of string pointed to by str
72 * left has the number of bytes left in str and is adjusted
73 * len contains the number of bytes put in the buffer
76 #define CONVERT(str, left, src, len) \
80 outleft = CONV_BUFFER_SIZE; \
82 if (iconv(id, (iconv_src_t)&str, &left, &bp, &outleft) == -1 && \
85 if ((len = CONV_BUFFER_SIZE - outleft) == 0) { \
94 if (id != (iconv_t)-1) \
95 iconv(id, NULL, NULL, NULL, NULL); \
98 #define CONVERT(str, left, src, len)
103 default_char2int(SCR *sp, const char * str, ssize_t len, CONVWIN *cw,
104 size_t *tolen, CHAR_T **dst, iconv_t id)
107 CHAR_T **tostr = &cw->bp1.wc;
108 size_t *blen = &cw->blen1;
112 char *src = (char *)str;
114 char buffer[CONV_BUFFER_SIZE];
120 BINC_RETW(NULL, *tostr, *blen, nlen);
123 if (id != (iconv_t)-1)
124 CONVERT(str, left, src, len);
127 for (i = 0, j = 0; j < len; ) {
128 n = mbrtowc((*tostr)+i, src+j, len-j, &mbs);
129 /* NULL character converted */
130 if (n == -2) error = -(len-j);
131 if (n == -1 || n == -2) goto err;
136 BINC_RETW(NULL, *tostr, *blen, nlen);
138 if (id != (iconv_t)-1 && j == len && left) {
139 CONVERT(str, left, src, len);
154 fe_char2int(SCR *sp, const char * str, ssize_t len, CONVWIN *cw,
155 size_t *tolen, CHAR_T **dst)
157 return default_char2int(sp, str, len, cw, tolen, dst,
158 sp->conv.id[IC_FE_CHAR2INT]);
162 ie_char2int(SCR *sp, const char * str, ssize_t len, CONVWIN *cw,
163 size_t *tolen, CHAR_T **dst)
165 return default_char2int(sp, str, len, cw, tolen, dst,
166 sp->conv.id[IC_IE_CHAR2INT]);
170 cs_char2int(SCR *sp, const char * str, ssize_t len, CONVWIN *cw,
171 size_t *tolen, CHAR_T **dst)
173 return default_char2int(sp, str, len, cw, tolen, dst,
178 int2raw(SCR *sp, const CHAR_T * str, ssize_t len, CONVWIN *cw,
179 size_t *tolen, char **dst)
182 char **tostr = &cw->bp1.c;
183 size_t *blen = &cw->blen1;
185 BINC_RETC(NULL, *tostr, *blen, len);
188 for (i = 0; i < len; ++i)
189 (*tostr)[i] = str[i];
197 default_int2char(SCR *sp, const CHAR_T * str, ssize_t len, CONVWIN *cw,
198 size_t *tolen, char **pdst, iconv_t id)
200 size_t i, j, offset = 0;
201 char **tostr = &cw->bp1.c;
202 size_t *blen = &cw->blen1;
205 ssize_t nlen = len + MB_CUR_MAX;
209 char buffer[CONV_BUFFER_SIZE];
213 /* convert first len bytes of buffer and append it to cw->bp
214 * len is adjusted => 0
215 * offset contains the offset in cw->bp and is adjusted
216 * cw->bp is grown as required
219 #define CONVERT2(_buffer, lenp, cw, offset) \
221 char *bp = _buffer; \
224 size_t outleft = cw->blen1 - offset; \
225 char *obp = cw->bp1.c + offset; \
226 if (cw->blen1 < offset + MB_CUR_MAX) { \
228 BINC_RETC(NULL, cw->bp1.c, cw->blen1, nlen); \
231 ret = iconv(id, (iconv_src_t)&bp, lenp, &obp, &outleft); \
232 if (ret == -1 && errno != E2BIG) \
234 offset = cw->blen1 - outleft; \
235 } while (ret != 0); \
238 #define CONVERT2(_buffer, lenp, cw, offset)
243 BINC_RETC(NULL, *tostr, *blen, nlen);
244 dst = *tostr; buflen = *blen;
247 if (id != (iconv_t)-1) {
248 dst = buffer; buflen = CONV_BUFFER_SIZE;
252 for (i = 0, j = 0; i < len; ++i) {
253 n = wcrtomb(dst+j, str[i], &mbs);
254 if (n == -1) goto err;
256 if (buflen < j + MB_CUR_MAX) {
257 if (id != (iconv_t)-1) {
258 CONVERT2(buffer, &j, cw, offset);
261 BINC_RETC(NULL, *tostr, *blen, nlen);
262 dst = *tostr; buflen = *blen;
267 n = wcrtomb(dst+j, L'\0', &mbs);
268 j += n - 1; /* don't count NUL at the end */
271 if (id != (iconv_t)-1) {
272 CONVERT2(buffer, &j, cw, offset);
273 CONVERT2(NULL, NULL, cw, offset); /* back to the initial state */
288 fe_int2char(SCR *sp, const CHAR_T * str, ssize_t len, CONVWIN *cw,
289 size_t *tolen, char **dst)
291 return default_int2char(sp, str, len, cw, tolen, dst,
292 sp->conv.id[IC_FE_INT2CHAR]);
296 cs_int2char(SCR *sp, const CHAR_T * str, ssize_t len, CONVWIN *cw,
297 size_t *tolen, char **dst)
299 return default_int2char(sp, str, len, cw, tolen, dst,
307 * Initialize the iconv environment.
309 * PUBLIC: void conv_init __P((SCR *, SCR *));
312 conv_init(SCR *orig, SCR *sp)
317 setlocale(LC_ALL, "");
319 BCOPY(&orig->conv, &sp->conv, 1);
322 char *ctype = setlocale(LC_CTYPE, NULL);
326 * This hack fixes the libncursesw issue on FreeBSD.
328 if (!strcmp(ctype, "ko_KR.CP949"))
329 setlocale(LC_CTYPE, "ko_KR.eucKR");
330 else if (!strcmp(ctype, "zh_CN.GB2312"))
331 setlocale(LC_CTYPE, "zh_CN.eucCN");
332 else if (!strcmp(ctype, "zh_CN.GBK"))
333 setlocale(LC_CTYPE, "zh_CN.GB18030");
336 * Switch to 8bit mode if locale is C;
337 * LC_CTYPE should be reseted to C if unmatched.
339 if (!strcmp(ctype, "C") || !strcmp(ctype, "POSIX")) {
340 sp->conv.sys2int = sp->conv.file2int = raw2int;
341 sp->conv.int2sys = sp->conv.int2file = int2raw;
342 sp->conv.input2int = raw2int;
344 sp->conv.sys2int = cs_char2int;
345 sp->conv.int2sys = cs_int2char;
346 sp->conv.file2int = fe_char2int;
347 sp->conv.int2file = fe_int2char;
348 sp->conv.input2int = ie_char2int;
351 o_set(sp, O_INPUTENCODING, OS_STRDUP, codeset(), 0);
356 /* iconv descriptors must be distinct to screens. */
357 for (i = 0; i <= IC_IE_TO_UTF16; ++i)
358 sp->conv.id[i] = (iconv_t)-1;
360 conv_enc(sp, O_INPUTENCODING, 0);
366 * Convert file/input encoding.
368 * PUBLIC: int conv_enc __P((SCR *, int, char *));
371 conv_enc(SCR *sp, int option, char *enc)
373 #if defined(USE_WIDECHAR) && defined(USE_ICONV)
378 c2w = sp->conv.id + IC_FE_CHAR2INT;
379 w2c = sp->conv.id + IC_FE_INT2CHAR;
380 if (!enc) enc = O_STR(sp, O_FILEENCODING);
381 if (*c2w != (iconv_t)-1)
383 if (*w2c != (iconv_t)-1)
385 if (strcasecmp(codeset(), enc)) {
386 if ((*c2w = iconv_open(codeset(), enc)) == (iconv_t)-1)
388 if ((*w2c = iconv_open(enc, codeset())) == (iconv_t)-1)
390 } else *c2w = *w2c = (iconv_t)-1;
392 case O_INPUTENCODING:
393 c2w = sp->conv.id + IC_IE_CHAR2INT;
394 w2c = sp->conv.id + IC_IE_TO_UTF16;
395 if (!enc) enc = O_STR(sp, O_INPUTENCODING);
396 if (*c2w != (iconv_t)-1)
398 if (*w2c != (iconv_t)-1)
400 if (strcasecmp(codeset(), enc)) {
401 if ((*c2w = iconv_open(codeset(), enc)) == (iconv_t)-1)
403 } else *c2w = (iconv_t)-1;
404 /* UTF-16 can not be locale and can not be inputed. */
405 if ((*w2c = iconv_open("utf-16be", enc)) == (iconv_t)-1)
410 F_CLR(sp, SC_CONV_ERROR);
411 F_SET(sp, SC_SCR_REFORMAT);
419 "321|File encoding conversion not supported");
421 case O_INPUTENCODING:
423 "322|Input encoding conversion not supported");
431 * Close the iconv descriptors, release the buffer.
433 * PUBLIC: void conv_end __P((SCR *));
438 #if defined(USE_WIDECHAR) && defined(USE_ICONV)
440 for (i = 0; i <= IC_IE_TO_UTF16; ++i)
441 if (sp->conv.id[i] != (iconv_t)-1)
442 iconv_close(sp->conv.id[i]);
443 if (sp->cw.bp1.c != NULL)