1 /*
2 * Copyright (C) 1999-2002,2007 Thomas Roessler <roessler@does-not-exist.org>
3 *
4 * This program is free software; you can redistribute it
5 * and/or modify it under the terms of the GNU General Public
6 * License as published by the Free Software Foundation; either
7 * version 2 of the License, or (at your option) any later
8 * version.
9 *
10 * This program is distributed in the hope that it will be
11 * useful, but WITHOUT ANY WARRANTY; without even the implied
12 * warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
13 * PURPOSE. See the GNU General Public License for more
14 * details.
15 *
16 * You should have received a copy of the GNU General Public
17 * License along with this program; if not, write to the Free
18 * Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
19 * Boston, MA 02110-1301, USA.
20 */
21
22 #if HAVE_CONFIG_H
23 # include "config.h"
24 #endif
25
26 #include <string.h>
27 #include <stdio.h>
28 #include <stdlib.h>
29
30 #include <ctype.h>
31
32 #include <sys/types.h>
33 #include <dirent.h>
34 #include <unistd.h>
35 #include <errno.h>
36
37 #include "mutt.h"
38 #include "charset.h"
39
40 #ifndef EILSEQ
41 # define EILSEQ EINVAL
42 #endif
43
44 /*
45 * The following list has been created manually from the data under:
46 * http://www.isi.edu/in-notes/iana/assignments/character-sets
47 * Last update: 2000-09-07
48 *
49 * Note that it includes only the subset of character sets for which
50 * a preferred MIME name is given.
51 */
52
53 static struct
54 {
55 char *key;
56 char *pref;
57 }
58 PreferredMIMENames[] =
59 {
60 { "ansi_x3.4-1968", "us-ascii" },
61 { "iso-ir-6", "us-ascii" },
62 { "iso_646.irv:1991", "us-ascii" },
63 { "ascii", "us-ascii" },
64 { "iso646-us", "us-ascii" },
65 { "us", "us-ascii" },
66 { "ibm367", "us-ascii" },
67 { "cp367", "us-ascii" },
68 { "csASCII", "us-ascii" },
69
70 { "csISO2022KR", "iso-2022-kr" },
71 { "csEUCKR", "euc-kr" },
72 { "csISO2022JP", "iso-2022-jp" },
73 { "csISO2022JP2", "iso-2022-jp-2" },
74
75 { "ISO_8859-1:1987", "iso-8859-1" },
76 { "iso-ir-100", "iso-8859-1" },
77 { "iso_8859-1", "iso-8859-1" },
78 { "latin1", "iso-8859-1" },
79 { "l1", "iso-8859-1" },
80 { "IBM819", "iso-8859-1" },
81 { "CP819", "iso-8859-1" },
82 { "csISOLatin1", "iso-8859-1" },
83
84 { "ISO_8859-2:1987", "iso-8859-2" },
85 { "iso-ir-101", "iso-8859-2" },
86 { "iso_8859-2", "iso-8859-2" },
87 { "latin2", "iso-8859-2" },
88 { "l2", "iso-8859-2" },
89 { "csISOLatin2", "iso-8859-2" },
90
91 { "ISO_8859-3:1988", "iso-8859-3" },
92 { "iso-ir-109", "iso-8859-3" },
93 { "ISO_8859-3", "iso-8859-3" },
94 { "latin3", "iso-8859-3" },
95 { "l3", "iso-8859-3" },
96 { "csISOLatin3", "iso-8859-3" },
97
98 { "ISO_8859-4:1988", "iso-8859-4" },
99 { "iso-ir-110", "iso-8859-4" },
100 { "ISO_8859-4", "iso-8859-4" },
101 { "latin4", "iso-8859-4" },
102 { "l4", "iso-8859-4" },
103 { "csISOLatin4", "iso-8859-4" },
104
105 { "ISO_8859-6:1987", "iso-8859-6" },
106 { "iso-ir-127", "iso-8859-6" },
107 { "iso_8859-6", "iso-8859-6" },
108 { "ECMA-114", "iso-8859-6" },
109 { "ASMO-708", "iso-8859-6" },
110 { "arabic", "iso-8859-6" },
111 { "csISOLatinArabic", "iso-8859-6" },
112
113 { "ISO_8859-7:1987", "iso-8859-7" },
114 { "iso-ir-126", "iso-8859-7" },
115 { "ISO_8859-7", "iso-8859-7" },
116 { "ELOT_928", "iso-8859-7" },
117 { "ECMA-118", "iso-8859-7" },
118 { "greek", "iso-8859-7" },
119 { "greek8", "iso-8859-7" },
120 { "csISOLatinGreek", "iso-8859-7" },
121
122 { "ISO_8859-8:1988", "iso-8859-8" },
123 { "iso-ir-138", "iso-8859-8" },
124 { "ISO_8859-8", "iso-8859-8" },
125 { "hebrew", "iso-8859-8" },
126 { "csISOLatinHebrew", "iso-8859-8" },
127
128 { "ISO_8859-5:1988", "iso-8859-5" },
129 { "iso-ir-144", "iso-8859-5" },
130 { "ISO_8859-5", "iso-8859-5" },
131 { "cyrillic", "iso-8859-5" },
132 { "csISOLatinCyrillic", "iso-8859-5" },
133
134 { "ISO_8859-9:1989", "iso-8859-9" },
135 { "iso-ir-148", "iso-8859-9" },
136 { "ISO_8859-9", "iso-8859-9" },
137 { "latin5", "iso-8859-9" }, /* this is not a bug */
138 { "l5", "iso-8859-9" },
139 { "csISOLatin5", "iso-8859-9" },
140
141 { "ISO_8859-10:1992", "iso-8859-10" },
142 { "iso-ir-157", "iso-8859-10" },
143 { "latin6", "iso-8859-10" }, /* this is not a bug */
144 { "l6", "iso-8859-10" },
145 { "csISOLatin6", "iso-8859-10" },
146
147 { "csKOI8r", "koi8-r" },
148
149 { "MS_Kanji", "Shift_JIS" }, /* Note the underscore! */
150 { "csShiftJis", "Shift_JIS" },
151
152 { "Extended_UNIX_Code_Packed_Format_for_Japanese",
153 "euc-jp" },
154 { "csEUCPkdFmtJapanese",
155 "euc-jp" },
156
157 { "csGB2312", "gb2312" },
158 { "csbig5", "big5" },
159
160 /*
161 * End of official brain damage. What follows has been taken
162 * from glibc's localedata files.
163 */
164
165 { "iso_8859-13", "iso-8859-13" },
166 { "iso-ir-179", "iso-8859-13" },
167 { "latin7", "iso-8859-13" }, /* this is not a bug */
168 { "l7", "iso-8859-13" },
169
170 { "iso_8859-14", "iso-8859-14" },
171 { "latin8", "iso-8859-14" }, /* this is not a bug */
172 { "l8", "iso-8859-14" },
173
174 { "iso_8859-15", "iso-8859-15" },
175 { "latin9", "iso-8859-15" }, /* this is not a bug */
176
177 /* Suggested by Ionel Mugurel Ciobica <tgakic@sg10.chem.tue.nl> */
178 { "latin0", "iso-8859-15" }, /* this is not a bug */
179
180 { "iso_8859-16", "iso-8859-16" },
181 { "latin10", "iso-8859-16" }, /* this is not a bug */
182
183 /*
184 * David Champion <dgc@uchicago.edu> has observed this with
185 * nl_langinfo under SunOS 5.8.
186 */
187
188 { "646", "us-ascii" },
189
190 /*
191 * http://www.sun.com/software/white-papers/wp-unicode/
192 */
193
194 { "eucJP", "euc-jp" },
195 { "PCK", "Shift_JIS" },
196 { "ko_KR-euc", "euc-kr" },
197 { "zh_TW-big5", "big5" },
198
199 /* seems to be common on some systems */
200
201 { "sjis", "Shift_JIS" },
202 { "euc-jp-ms", "eucJP-ms" },
203
204
205 /*
206 * If you happen to encounter system-specific brain-damage with
207 * respect to character set naming, please add it above this
208 * comment, and submit a patch to <mutt-dev@mutt.org>.
209 */
210
211 /* End of aliases. Please keep this line last. */
212
213 { NULL, NULL }
214 };
215
216 #ifdef HAVE_LANGINFO_CODESET
217 # include <langinfo.h>
218
219
mutt_set_langinfo_charset(void)220 void mutt_set_langinfo_charset (void)
221 {
222 char buff[LONG_STRING];
223 char buff2[LONG_STRING];
224
225 strfcpy (buff, nl_langinfo (CODESET), sizeof (buff));
226 mutt_canonical_charset (buff2, sizeof (buff2), buff);
227
228 /* finally, set $charset */
229 if (!(Charset = safe_strdup (buff2)))
230 Charset = safe_strdup ("iso-8859-1");
231 }
232
233 #else
234
mutt_set_langinfo_charset(void)235 void mutt_set_langinfo_charset (void)
236 {
237 Charset = safe_strdup ("iso-8859-1");
238 }
239
240 #endif
241
242 /* this first ties off any charset extension such as //TRANSLIT,
243 canonicalizes the charset and re-adds the extension */
mutt_canonical_charset(char * dest,size_t dlen,const char * name)244 void mutt_canonical_charset (char *dest, size_t dlen, const char *name)
245 {
246 size_t i;
247 char *p, *ext;
248 char in[LONG_STRING], scratch[LONG_STRING];
249
250 strfcpy (in, name, sizeof (in));
251 if ((ext = strchr (in, '/')))
252 *ext++ = 0;
253
254 if (!ascii_strcasecmp (in, "utf-8") || !ascii_strcasecmp (in, "utf8"))
255 {
256 strfcpy (dest, "utf-8", dlen);
257 goto out;
258 }
259
260 /* catch some common iso-8859-something misspellings */
261 if (!ascii_strncasecmp (in, "8859", 4) && in[4] != '-')
262 snprintf (scratch, sizeof (scratch), "iso-8859-%s", in +4);
263 else if (!ascii_strncasecmp (in, "8859-", 5))
264 snprintf (scratch, sizeof (scratch), "iso-8859-%s", in + 5);
265 else if (!ascii_strncasecmp (in, "iso8859", 7) && in[7] != '-')
266 snprintf (scratch, sizeof (scratch), "iso_8859-%s", in + 7);
267 else if (!ascii_strncasecmp (in, "iso8859-", 8))
268 snprintf (scratch, sizeof (scratch), "iso_8859-%s", in + 8);
269 else
270 strfcpy (scratch, in, sizeof (scratch));
271
272 for (i = 0; PreferredMIMENames[i].key; i++)
273 if (!ascii_strcasecmp (scratch, PreferredMIMENames[i].key) ||
274 !mutt_strcasecmp (scratch, PreferredMIMENames[i].key))
275 {
276 strfcpy (dest, PreferredMIMENames[i].pref, dlen);
277 goto out;
278 }
279
280 strfcpy (dest, scratch, dlen);
281
282 /* for cosmetics' sake, transform to lowercase. */
283 for (p = dest; *p; p++)
284 *p = ascii_tolower (*p);
285
286 out:
287 if (ext && *ext)
288 {
289 safe_strcat (dest, dlen, "/");
290 safe_strcat (dest, dlen, ext);
291 }
292 }
293
mutt_chscmp(const char * s,const char * chs)294 int mutt_chscmp (const char *s, const char *chs)
295 {
296 char buffer[STRING];
297 int a, b;
298
299 if (!s) return 0;
300
301 /* charsets may have extensions mutt_canonical_charset()
302 leaves intact; we expect `chs' to originate from mutt
303 code, not user input (i.e. `chs' does _not_ have any
304 extension)
305 we simply check if the shorter string is a prefix for
306 the longer */
307 mutt_canonical_charset (buffer, sizeof (buffer), s);
308 a = mutt_strlen (buffer);
309 b = mutt_strlen (chs);
310 return !ascii_strncasecmp (a > b ? buffer : chs,
311 a > b ? chs : buffer, MIN(a,b));
312 }
313
mutt_get_default_charset()314 char *mutt_get_default_charset ()
315 {
316 static char fcharset[SHORT_STRING];
317 const char *c = AssumedCharset;
318 const char *c1;
319
320 if (c && *c) {
321 c1 = strchr (c, ':');
322 strfcpy (fcharset, c, c1 ? (c1 - c + 1) : sizeof (fcharset));
323 return fcharset;
324 }
325 return strcpy (fcharset, "us-ascii"); /* __STRCPY_CHECKED__ */
326 }
327
328 #ifndef HAVE_ICONV
329
iconv_open(const char * tocode,const char * fromcode)330 iconv_t iconv_open (const char *tocode, const char *fromcode)
331 {
332 return (iconv_t)(-1);
333 }
334
iconv(iconv_t cd,ICONV_CONST char ** inbuf,size_t * inbytesleft,char ** outbuf,size_t * outbytesleft)335 size_t iconv (iconv_t cd, ICONV_CONST char **inbuf, size_t *inbytesleft,
336 char **outbuf, size_t *outbytesleft)
337 {
338 return 0;
339 }
340
iconv_close(iconv_t cd)341 int iconv_close (iconv_t cd)
342 {
343 return 0;
344 }
345
346 #endif /* !HAVE_ICONV */
347
348
349 /*
350 * Like iconv_open, but canonicalises the charsets, applies
351 * charset-hooks, recanonicalises, and finally applies iconv-hooks.
352 * Parameter flags=0 skips charset-hooks, while M_ICONV_HOOK_FROM
353 * applies them to fromcode. Callers should use flags=0 when fromcode
354 * can safely be considered true, either some constant, or some value
355 * provided by the user; M_ICONV_HOOK_FROM should be used only when
356 * fromcode is unsure, taken from a possibly wrong incoming MIME label,
357 * or such. Misusing M_ICONV_HOOK_FROM leads to unwanted interactions
358 * in some setups. Note: By design charset-hooks should never be, and
359 * are never, applied to tocode. Highlight note: The top-well-named
360 * M_ICONV_HOOK_FROM acts on charset-hooks, not at all on iconv-hooks.
361 */
362
mutt_iconv_open(const char * tocode,const char * fromcode,int flags)363 iconv_t mutt_iconv_open (const char *tocode, const char *fromcode, int flags)
364 {
365 char tocode1[SHORT_STRING];
366 char fromcode1[SHORT_STRING];
367 char *tocode2, *fromcode2;
368 char *tmp;
369
370 iconv_t cd;
371
372 /* transform to MIME preferred charset names */
373 mutt_canonical_charset (tocode1, sizeof (tocode1), tocode);
374 mutt_canonical_charset (fromcode1, sizeof (fromcode1), fromcode);
375
376 /* maybe apply charset-hooks and recanonicalise fromcode,
377 * but only when caller asked us to sanitize a potentialy wrong
378 * charset name incoming from the wild exterior. */
379 if ((flags & M_ICONV_HOOK_FROM) && (tmp = mutt_charset_hook (fromcode1)))
380 mutt_canonical_charset (fromcode1, sizeof (fromcode1), tmp);
381
382 /* always apply iconv-hooks to suit system's iconv tastes */
383 tocode2 = mutt_iconv_hook (tocode1);
384 tocode2 = (tocode2) ? tocode2 : tocode1;
385 fromcode2 = mutt_iconv_hook (fromcode1);
386 fromcode2 = (fromcode2) ? fromcode2 : fromcode1;
387
388 /* call system iconv with names it appreciates */
389 if ((cd = iconv_open (tocode2, fromcode2)) != (iconv_t) -1)
390 return cd;
391
392 return (iconv_t) -1;
393 }
394
395
396 /*
397 * Like iconv, but keeps going even when the input is invalid
398 * If you're supplying inrepls, the source charset should be stateless;
399 * if you're supplying an outrepl, the target charset should be.
400 */
401
mutt_iconv(iconv_t cd,ICONV_CONST char ** inbuf,size_t * inbytesleft,char ** outbuf,size_t * outbytesleft,ICONV_CONST char ** inrepls,const char * outrepl)402 size_t mutt_iconv (iconv_t cd, ICONV_CONST char **inbuf, size_t *inbytesleft,
403 char **outbuf, size_t *outbytesleft,
404 ICONV_CONST char **inrepls, const char *outrepl)
405 {
406 size_t ret = 0, ret1;
407 ICONV_CONST char *ib = *inbuf;
408 size_t ibl = *inbytesleft;
409 char *ob = *outbuf;
410 size_t obl = *outbytesleft;
411
412 for (;;)
413 {
414 ret1 = iconv (cd, &ib, &ibl, &ob, &obl);
415 if (ret1 != (size_t)-1)
416 ret += ret1;
417 if (ibl && obl && errno == EILSEQ)
418 {
419 if (inrepls)
420 {
421 /* Try replacing the input */
422 ICONV_CONST char **t;
423 for (t = inrepls; *t; t++)
424 {
425 ICONV_CONST char *ib1 = *t;
426 size_t ibl1 = strlen (*t);
427 char *ob1 = ob;
428 size_t obl1 = obl;
429 iconv (cd, &ib1, &ibl1, &ob1, &obl1);
430 if (!ibl1)
431 {
432 ++ib, --ibl;
433 ob = ob1, obl = obl1;
434 ++ret;
435 break;
436 }
437 }
438 if (*t)
439 continue;
440 }
441 /* Replace the output */
442 if (!outrepl)
443 outrepl = "?";
444 iconv (cd, 0, 0, &ob, &obl);
445 if (obl)
446 {
447 int n = strlen (outrepl);
448 if (n > obl)
449 {
450 outrepl = "?";
451 n = 1;
452 }
453 memcpy (ob, outrepl, n);
454 ++ib, --ibl;
455 ob += n, obl -= n;
456 ++ret;
457 iconv (cd, 0, 0, 0, 0); /* for good measure */
458 continue;
459 }
460 }
461 *inbuf = ib, *inbytesleft = ibl;
462 *outbuf = ob, *outbytesleft = obl;
463 return ret;
464 }
465 }
466
467
468 /*
469 * Convert a string
470 * Used in rfc2047.c, rfc2231.c, crypt-gpgme.c, mutt_idna.c, and more.
471 * Parameter flags is given as-is to mutt_iconv_open(). See there
472 * for its meaning and usage policy.
473 */
474
mutt_convert_string(char ** ps,const char * from,const char * to,int flags)475 int mutt_convert_string (char **ps, const char *from, const char *to, int flags)
476 {
477 iconv_t cd;
478 ICONV_CONST char *repls[] = { "\357\277\275", "?", 0 };
479 char *s = *ps;
480
481 if (!s || !*s)
482 return 0;
483
484 if (option (OPTSANITIZEJACHARS) && !ascii_strncasecmp (from, "iso-2022-jp", 11))
485 mutt_sanitize_ja_chars (s, mutt_strlen(s), 0);
486
487 if (to && from && (cd = mutt_iconv_open (to, from, flags)) != (iconv_t)-1)
488 {
489 int len;
490 ICONV_CONST char *ib;
491 char *buf, *ob;
492 size_t ibl, obl;
493 ICONV_CONST char **inrepls = 0;
494 char *outrepl = 0;
495
496 if (mutt_is_utf8 (to))
497 outrepl = "\357\277\275";
498 else if (mutt_is_utf8 (from))
499 inrepls = repls;
500 else
501 outrepl = "?";
502
503 len = strlen (s);
504 ib = s, ibl = len + 1;
505 obl = MB_LEN_MAX * ibl;
506 ob = buf = safe_malloc (obl + 1);
507
508 mutt_iconv (cd, &ib, &ibl, &ob, &obl, inrepls, outrepl);
509 iconv_close (cd);
510
511 *ob = '\0';
512
513 FREE (ps); /* __FREE_CHECKED__ */
514 *ps = buf;
515
516 mutt_str_adjust (ps);
517 return 0;
518 }
519 else
520 return -1;
521 }
522
523
524 /*
525 * FGETCONV stuff for converting a file while reading it
526 * Used in sendlib.c for converting from mutt's Charset
527 */
528
529 struct fgetconv_s
530 {
531 FILE *file;
532 iconv_t cd;
533 char bufi[512];
534 char bufo[512];
535 char *p;
536 char *ob;
537 char *ib;
538 size_t ibl;
539 ICONV_CONST char **inrepls;
540 };
541
542 struct fgetconv_not
543 {
544 FILE *file;
545 iconv_t cd;
546 };
547
548 /*
549 * Parameter flags is given as-is to mutt_iconv_open(). See there
550 * for its meaning and usage policy.
551 */
fgetconv_open(FILE * file,const char * from,const char * to,int flags)552 FGETCONV *fgetconv_open (FILE *file, const char *from, const char *to, int flags)
553 {
554 struct fgetconv_s *fc;
555 iconv_t cd = (iconv_t)-1;
556 static ICONV_CONST char *repls[] = { "\357\277\275", "?", 0 };
557
558 if (from && to)
559 cd = mutt_iconv_open (to, from, flags);
560
561 if (cd != (iconv_t)-1)
562 {
563 fc = safe_malloc (sizeof (struct fgetconv_s));
564 fc->p = fc->ob = fc->bufo;
565 fc->ib = fc->bufi;
566 fc->ibl = 0;
567 fc->inrepls = mutt_is_utf8 (to) ? repls : repls + 1;
568 }
569 else
570 fc = safe_malloc (sizeof (struct fgetconv_not));
571 fc->file = file;
572 fc->cd = cd;
573 return (FGETCONV *)fc;
574 }
575
fgetconvs(char * buf,size_t l,FGETCONV * _fc)576 char *fgetconvs (char *buf, size_t l, FGETCONV *_fc)
577 {
578 int c;
579 size_t r;
580
581 for (r = 0; r + 1 < l;)
582 {
583 if ((c = fgetconv (_fc)) == EOF)
584 break;
585 buf[r++] = (char) c;
586 if (c == '\n')
587 break;
588 }
589 buf[r] = '\0';
590
591 if (r)
592 return buf;
593 else
594 return NULL;
595 }
596
fgetconv(FGETCONV * _fc)597 int fgetconv (FGETCONV *_fc)
598 {
599 struct fgetconv_s *fc = (struct fgetconv_s *)_fc;
600
601 if (!fc)
602 return EOF;
603 if (fc->cd == (iconv_t)-1)
604 return fgetc (fc->file);
605 if (!fc->p)
606 return EOF;
607 if (fc->p < fc->ob)
608 return (unsigned char)*(fc->p)++;
609
610 /* Try to convert some more */
611 fc->p = fc->ob = fc->bufo;
612 if (fc->ibl)
613 {
614 size_t obl = sizeof (fc->bufo);
615 iconv (fc->cd, (ICONV_CONST char **)&fc->ib, &fc->ibl, &fc->ob, &obl);
616 if (fc->p < fc->ob)
617 return (unsigned char)*(fc->p)++;
618 }
619
620 /* If we trusted iconv a bit more, we would at this point
621 * ask why it had stopped converting ... */
622
623 /* Try to read some more */
624 if (fc->ibl == sizeof (fc->bufi) ||
625 (fc->ibl && fc->ib + fc->ibl < fc->bufi + sizeof (fc->bufi)))
626 {
627 fc->p = 0;
628 return EOF;
629 }
630 if (fc->ibl)
631 memcpy (fc->bufi, fc->ib, fc->ibl);
632 fc->ib = fc->bufi;
633 fc->ibl += fread (fc->ib + fc->ibl, 1, sizeof (fc->bufi) - fc->ibl, fc->file);
634
635 /* Try harder this time to convert some */
636 if (fc->ibl)
637 {
638 size_t obl = sizeof (fc->bufo);
639 mutt_iconv (fc->cd, (ICONV_CONST char **)&fc->ib, &fc->ibl, &fc->ob, &obl,
640 fc->inrepls, 0);
641 if (fc->p < fc->ob)
642 return (unsigned char)*(fc->p)++;
643 }
644
645 /* Either the file has finished or one of the buffers is too small */
646 fc->p = 0;
647 return EOF;
648 }
649
fgetconv_close(FGETCONV ** _fc)650 void fgetconv_close (FGETCONV **_fc)
651 {
652 struct fgetconv_s *fc = (struct fgetconv_s *) *_fc;
653
654 if (fc->cd != (iconv_t)-1)
655 iconv_close (fc->cd);
656 FREE (_fc); /* __FREE_CHECKED__ */
657 }
658
mutt_check_charset(const char * s,int strict)659 int mutt_check_charset (const char *s, int strict)
660 {
661 int i;
662 iconv_t cd;
663
664 if (mutt_is_utf8 (s))
665 return 0;
666
667 if (!strict)
668 for (i = 0; PreferredMIMENames[i].key; i++)
669 {
670 if (ascii_strcasecmp (PreferredMIMENames[i].key, s) == 0 ||
671 ascii_strcasecmp (PreferredMIMENames[i].pref, s) == 0)
672 return 0;
673 }
674
675 if ((cd = mutt_iconv_open (s, s, 0)) != (iconv_t)(-1))
676 {
677 iconv_close (cd);
678 return 0;
679 }
680
681 return -1;
682 }
683
684 /*
685 * mutt_sanitize_ja_chars()
686 * Adapted by TAKIZAWA Takashi <taki@cyber.email.ne.jp>
687 *
688 * - It replaces undefined KANJI characters to GETA mark.
689 * - It replaces character of 'JIS X 0201 kana' to '?'.
690 * - If $charset is EUC-JP, it replaces third character 'J' of
691 * escape sequence switching to 'JIS X 0201 latin' to 'B' indicating
692 * 'US-ASCII'.
693 * - If $charset is Shift_JIS, it replaces third character 'B' of
694 * escape sequence switching to 'US-ASCII' to 'J' indicating
695 * 'JIS X 0201 latin'.
696 */
697
698 #define ASCII 0
699 #define JISX0201LATIN 1
700 #define JISX0201KANA 2
701 #define JISX0208 3
702 #define OTHER_CS 4
703
mutt_sanitize_ja_chars(char * s,size_t len,int keep_state)704 void mutt_sanitize_ja_chars(char *s, size_t len, int keep_state)
705 {
706 static int cs = ASCII;
707 static int kanji_cont = 0;
708 static int illegal_kanji = 0;
709 static int es = 0;
710 static char pes = '\0';
711 static char ascii_3rd_char = 'B';
712 static char jisx0201_3rd_char = 'J';
713
714 char *p = s;
715 char *p1 = NULL;
716 unsigned char c;
717
718 if (!keep_state || *p == 0x1b) /* consideration about mbstate's buffer */
719 {
720 if (!ascii_strcasecmp (Charset, "euc-jp"))
721 jisx0201_3rd_char = 'B';
722 else if (!ascii_strcasecmp (Charset, "shift_jis"))
723 ascii_3rd_char = 'J';
724 cs = ASCII;
725 kanji_cont = 0;
726 illegal_kanji = 0;
727 es = 0;
728 pes = '\0';
729 }
730
731 for (;p - s < len;p++)
732 {
733 if (es == 0)
734 {
735 if (*p == 0x1b)
736 es++;
737 else
738 {
739 switch (cs)
740 {
741 case ASCII:
742 case JISX0201LATIN:
743 break;
744 case JISX0201KANA:
745 *p = '?';
746 break;
747 case JISX0208:
748 /* replace ku-ten code from 9 to 15 and 85 or more to "GETA MARK" */
749 c = (unsigned char)*p;
750 if (! kanji_cont)
751 {
752 if ((size_t)(p - s + 1) == len)
753 return; /* the last character is a primary byte of KANJI */
754 if (c <= 0x20 || (c >= 0x29 && c <= 0x2f)
755 || (c >= 0x75 && c <= 0xa0))
756 illegal_kanji = 1;
757 kanji_cont = 1;
758 p1 = p;
759 }
760 else
761 {
762 if (c <= 0x20 || c >= 0x7f)
763 illegal_kanji = 1;
764 if (illegal_kanji && p1)
765 *p1 = 0x22, *p = 0x2e;
766 kanji_cont = 0;
767 illegal_kanji = 0;
768 }
769 break;
770 }
771 }
772 }
773 else if (es == 1)
774 {
775 if (*p == '$' || (*p >= '(' && *p <= '/' && *p != ','))
776 {
777 es++;
778 pes = *p;
779 }
780 else
781 {
782 es = 0;
783 return; /* broken */
784 }
785 }
786 else if (es == 2)
787 {
788 if (pes == '(')
789 {
790 switch (*p)
791 {
792 case 'B':
793 cs = ASCII, *p = ascii_3rd_char;
794 break;
795 case 'J':
796 cs = JISX0201LATIN, *p = jisx0201_3rd_char;
797 break;
798 case 'I':
799 /* ready to replace character to '?' */
800 cs = JISX0201KANA, *p = ascii_3rd_char;
801 break;
802 default:
803 cs = OTHER_CS;
804 }
805 es = 0;
806 }
807 else if (pes == '$')
808 {
809 switch (*p)
810 {
811 case '@': /* JIS X 0208-1978 */
812 case 'B': /* JIS X 0208-1983 */
813 cs = JISX0208;
814 es = 0;
815 break;
816 case 'A':
817 cs = OTHER_CS; /* GB 2312 */
818 es = 0;
819 break;
820 case '(':
821 case ')':
822 case '*':
823 case '+':
824 case '-':
825 case '.':
826 case '/':
827 es++;
828 break;
829 default:
830 es = 0;
831 return; /* broken */
832 }
833 }
834 else
835 {
836 cs = OTHER_CS;
837 es = 0;
838 }
839 }
840 else /* es == 3 */
841 {
842 cs = OTHER_CS;
843 es = 0;
844 }
845 }
846 }
847
mutt_copy_bytes_sanitize_ja(FILE * in,FILE * out,size_t size)848 int mutt_copy_bytes_sanitize_ja (FILE *in, FILE *out, size_t size)
849 {
850 char buf[2048];
851 size_t chunk;
852
853 mutt_sanitize_ja_chars (NULL, 0, 0);
854 while (size > 0)
855 {
856 chunk = (size > sizeof (buf)) ? sizeof (buf) : size;
857 if ((chunk = fread (buf, 1, chunk, in)) < 1)
858 break;
859 mutt_sanitize_ja_chars (buf, chunk, 1);
860 if (fwrite (buf, 1, chunk, out) != chunk)
861 return (-1);
862 size -= chunk;
863 }
864
865 return 0;
866 }
867
868