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.
14 #include <sys/types.h>
15 #include <sys/queue.h>
18 #include <bitstring.h>
33 * Get the locale encoding.
35 * PUBLIC: char * codeset(void);
43 cs
= nl_langinfo(CODESET
);
50 raw2int(SCR
*sp
, const char * str
, ssize_t len
, CONVWIN
*cw
, size_t *tolen
,
54 CHAR_T
**tostr
= &cw
->bp1
.wc
;
55 size_t *blen
= &cw
->blen1
;
57 BINC_RETW(NULL
, *tostr
, *blen
, len
);
60 for (i
= 0; i
< len
; ++i
)
61 (*tostr
)[i
] = (u_char
) str
[i
];
68 #define CONV_BUFFER_SIZE 512
69 /* fill the buffer with codeset encoding of string pointed to by str
70 * left has the number of bytes left in str and is adjusted
71 * len contains the number of bytes put in the buffer
74 #define CONVERT(str, left, src, len) \
78 outleft = CONV_BUFFER_SIZE; \
80 if (iconv(id, (iconv_src_t)&str, &left, &bp, &outleft) \
81 == -1 && errno != E2BIG) \
83 if ((len = CONV_BUFFER_SIZE - outleft) == 0) { \
92 if (id != (iconv_t)-1) \
93 iconv(id, NULL, NULL, NULL, NULL); \
96 #define CONVERT(str, left, src, len)
101 default_char2int(SCR
*sp
, const char * str
, ssize_t len
, CONVWIN
*cw
,
102 size_t *tolen
, CHAR_T
**dst
, iconv_t id
)
105 CHAR_T
**tostr
= &cw
->bp1
.wc
;
106 size_t *blen
= &cw
->blen1
;
110 char *src
= (char *)str
;
112 char buffer
[CONV_BUFFER_SIZE
];
117 memset(&mbs
, 0, sizeof(mbs
));
118 BINC_RETW(NULL
, *tostr
, *blen
, nlen
);
121 if (id
!= (iconv_t
)-1)
122 CONVERT(str
, left
, src
, len
);
125 for (i
= 0, j
= 0; j
< len
; ) {
126 n
= mbrtowc((*tostr
)+i
, src
+j
, len
-j
, &mbs
);
127 /* NULL character converted */
130 if (n
== -1 || n
== -2)
137 BINC_RETW(NULL
, *tostr
, *blen
, nlen
);
139 if (id
!= (iconv_t
)-1 && j
== len
&& left
) {
140 CONVERT(str
, left
, src
, len
);
155 fe_char2int(SCR
*sp
, const char * str
, ssize_t len
, CONVWIN
*cw
, size_t *tolen
,
158 return default_char2int(sp
, str
, len
, cw
, tolen
, dst
,
159 sp
->conv
.id
[IC_FE_CHAR2INT
]);
163 ie_char2int(SCR
*sp
, const char * str
, ssize_t len
, CONVWIN
*cw
, size_t *tolen
,
166 return default_char2int(sp
, str
, len
, cw
, tolen
, dst
,
167 sp
->conv
.id
[IC_IE_CHAR2INT
]);
171 cs_char2int(SCR
*sp
, const char * str
, ssize_t len
, CONVWIN
*cw
, size_t *tolen
,
174 return default_char2int(sp
, str
, len
, cw
, tolen
, dst
, (iconv_t
)-1);
178 int2raw(SCR
*sp
, const CHAR_T
* str
, ssize_t len
, CONVWIN
*cw
, size_t *tolen
,
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, \
232 ret = iconv(id, (iconv_src_t)&bp, lenp, &obp, \
234 if (ret == -1 && errno != E2BIG) \
236 offset = cw->blen1 - outleft; \
237 } while (ret != 0); \
240 #define CONVERT2(_buffer, lenp, cw, offset)
244 memset(&mbs
, 0, sizeof(mbs
));
245 BINC_RETC(NULL
, *tostr
, *blen
, nlen
);
246 dst
= *tostr
; buflen
= *blen
;
249 if (id
!= (iconv_t
)-1) {
250 dst
= buffer
; buflen
= CONV_BUFFER_SIZE
;
254 for (i
= 0, j
= 0; i
< len
; ++i
) {
255 n
= wcrtomb(dst
+j
, str
[i
], &mbs
);
259 if (buflen
< j
+ MB_CUR_MAX
) {
260 if (id
!= (iconv_t
)-1) {
261 CONVERT2(buffer
, &j
, cw
, offset
);
264 BINC_RETC(NULL
, *tostr
, *blen
, nlen
);
265 dst
= *tostr
; buflen
= *blen
;
270 n
= wcrtomb(dst
+j
, L
'\0', &mbs
);
271 j
+= n
- 1; /* don't count NUL at the end */
274 if (id
!= (iconv_t
)-1) {
275 CONVERT2(buffer
, &j
, cw
, offset
);
276 /* back to the initial state */
277 CONVERT2(NULL
, NULL
, cw
, offset
);
292 fe_int2char(SCR
*sp
, const CHAR_T
* str
, ssize_t len
, CONVWIN
*cw
,
293 size_t *tolen
, char **dst
)
295 return default_int2char(sp
, str
, len
, cw
, tolen
, dst
,
296 sp
->conv
.id
[IC_FE_INT2CHAR
]);
300 cs_int2char(SCR
*sp
, const CHAR_T
* str
, ssize_t len
, CONVWIN
*cw
,
301 size_t *tolen
, char **dst
)
303 return default_int2char(sp
, str
, len
, cw
, tolen
, dst
, (iconv_t
)-1);
310 * Initialize the iconv environment.
312 * PUBLIC: void conv_init(SCR *, SCR *);
315 conv_init(SCR
*orig
, SCR
*sp
)
320 setlocale(LC_ALL
, "");
322 memmove(&sp
->conv
, &orig
->conv
, sizeof(CONV
));
325 char *ctype
= setlocale(LC_CTYPE
, NULL
);
329 * This hack fixes the libncursesw issue on FreeBSD.
331 if (!strcmp(ctype
, "ko_KR.CP949"))
332 setlocale(LC_CTYPE
, "ko_KR.eucKR");
333 else if (!strcmp(ctype
, "zh_CN.GB2312"))
334 setlocale(LC_CTYPE
, "zh_CN.eucCN");
335 else if (!strcmp(ctype
, "zh_CN.GBK"))
336 setlocale(LC_CTYPE
, "zh_CN.GB18030");
337 else if (!strcmp(ctype
, "zh_Hans_CN.GB2312"))
338 setlocale(LC_CTYPE
, "zh_Hans_CN.eucCN");
339 else if (!strcmp(ctype
, "zh_Hans_CN.GBK"))
340 setlocale(LC_CTYPE
, "zh_Hans_CN.GB18030");
343 * Switch to 8bit mode if locale is C;
344 * LC_CTYPE should be reseted to C if unmatched.
346 if (!strcmp(ctype
, "C") || !strcmp(ctype
, "POSIX")) {
347 sp
->conv
.sys2int
= sp
->conv
.file2int
= raw2int
;
348 sp
->conv
.int2sys
= sp
->conv
.int2file
= int2raw
;
349 sp
->conv
.input2int
= raw2int
;
351 sp
->conv
.sys2int
= cs_char2int
;
352 sp
->conv
.int2sys
= cs_int2char
;
353 sp
->conv
.file2int
= fe_char2int
;
354 sp
->conv
.int2file
= fe_int2char
;
355 sp
->conv
.input2int
= ie_char2int
;
358 o_set(sp
, O_INPUTENCODING
, OS_STRDUP
, codeset(), 0);
363 /* iconv descriptors must be distinct to screens. */
364 for (i
= 0; i
<= IC_IE_TO_UTF16
; ++i
)
365 sp
->conv
.id
[i
] = (iconv_t
)-1;
367 conv_enc(sp
, O_INPUTENCODING
, 0);
373 * Convert file/input encoding.
375 * PUBLIC: int conv_enc(SCR *, int, char *);
378 conv_enc(SCR
*sp
, int option
, char *enc
)
380 #if defined(USE_WIDECHAR) && defined(USE_ICONV)
382 iconv_t id_c2w
, id_w2c
;
386 c2w
= sp
->conv
.id
+ IC_FE_CHAR2INT
;
387 w2c
= sp
->conv
.id
+ IC_FE_INT2CHAR
;
389 enc
= O_STR(sp
, O_FILEENCODING
);
391 if (strcasecmp(codeset(), enc
)) {
392 if ((id_c2w
= iconv_open(codeset(), enc
)) ==
395 if ((id_w2c
= iconv_open(enc
, codeset())) ==
399 id_c2w
= (iconv_t
)-1;
400 id_w2c
= (iconv_t
)-1;
405 case O_INPUTENCODING
:
406 c2w
= sp
->conv
.id
+ IC_IE_CHAR2INT
;
407 w2c
= sp
->conv
.id
+ IC_IE_TO_UTF16
;
409 enc
= O_STR(sp
, O_INPUTENCODING
);
411 if (strcasecmp(codeset(), enc
)) {
412 if ((id_c2w
= iconv_open(codeset(), enc
)) ==
416 id_c2w
= (iconv_t
)-1;
418 /* UTF-16 can not be locale and can not be inputed. */
419 if ((id_w2c
= iconv_open("utf-16be", enc
)) == (iconv_t
)-1)
428 if (*c2w
!= (iconv_t
)-1)
430 if (*w2c
!= (iconv_t
)-1)
436 F_CLR(sp
, SC_CONV_ERROR
);
437 F_SET(sp
, SC_SCR_REFORMAT
);
444 msgq(sp
, M_ERR
, "321|File encoding conversion not supported");
446 case O_INPUTENCODING
:
447 msgq(sp
, M_ERR
, "322|Input encoding conversion not supported");
455 * Close the iconv descriptors, release the buffer.
457 * PUBLIC: void conv_end(SCR *);
462 #if defined(USE_WIDECHAR) && defined(USE_ICONV)
464 for (i
= 0; i
<= IC_IE_TO_UTF16
; ++i
)
465 if (sp
->conv
.id
[i
] != (iconv_t
)-1)
466 iconv_close(sp
->conv
.id
[i
]);