imapext-2007
view src/c-client/utf8aux.c @ 0:ada5e610ab86
imap-2007e
| author | yuuji@gentei.org |
|---|---|
| date | Mon, 14 Sep 2009 15:17:45 +0900 |
| parents | |
| children |
line source
1 /* ========================================================================
2 * Copyright 1988-2007 University of Washington
3 *
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
7 *
8 * http://www.apache.org/licenses/LICENSE-2.0
9 *
10 *
11 * ========================================================================
12 */
14 /*
15 * Program: UTF-8 auxillary routines (c-client and MIME2 support)
16 *
17 * Author: Mark Crispin
18 * Networks and Distributed Computing
19 * Computing & Communications
20 * University of Washington
21 * Administration Building, AG-44
22 * Seattle, WA 98195
23 * Internet: MRC@CAC.Washington.EDU
24 *
25 * Date: 11 June 1997
26 * Last Edited: 12 October 2007
27 */
30 #include <stdio.h>
31 #include <ctype.h>
32 #include "c-client.h"
34 /* Convert charset labelled stringlist to UTF-8 in place
35 * Accepts: string list
36 * charset
37 */
39 static void utf8_stringlist (STRINGLIST *st,char *charset)
40 {
41 SIZEDTEXT txt;
42 /* convert entire stringstruct */
43 if (st) do if (utf8_text (&st->text,charset,&txt,U8T_CANONICAL)) {
44 fs_give ((void **) &st->text.data);
45 st->text.data = txt.data; /* transfer this text */
46 st->text.size = txt.size;
47 } while (st = st->next);
48 }
51 /* Convert charset labelled searchpgm to UTF-8 in place
52 * Accepts: search program
53 * charset
54 */
56 void utf8_searchpgm (SEARCHPGM *pgm,char *charset)
57 {
58 SIZEDTEXT txt;
59 SEARCHHEADER *hl;
60 SEARCHOR *ol;
61 SEARCHPGMLIST *pl;
62 if (pgm) { /* must have a search program */
63 utf8_stringlist (pgm->bcc,charset);
64 utf8_stringlist (pgm->cc,charset);
65 utf8_stringlist (pgm->from,charset);
66 utf8_stringlist (pgm->to,charset);
67 utf8_stringlist (pgm->subject,charset);
68 for (hl = pgm->header; hl; hl = hl->next) {
69 if (utf8_text (&hl->line,charset,&txt,U8T_CANONICAL)) {
70 fs_give ((void **) &hl->line.data);
71 hl->line.data = txt.data;
72 hl->line.size = txt.size;
73 }
74 if (utf8_text (&hl->text,charset,&txt,U8T_CANONICAL)) {
75 fs_give ((void **) &hl->text.data);
76 hl->text.data = txt.data;
77 hl->text.size = txt.size;
78 }
79 }
80 utf8_stringlist (pgm->body,charset);
81 utf8_stringlist (pgm->text,charset);
82 for (ol = pgm->or; ol; ol = ol->next) {
83 utf8_searchpgm (ol->first,charset);
84 utf8_searchpgm (ol->second,charset);
85 }
86 for (pl = pgm->not; pl; pl = pl->next) utf8_searchpgm (pl->pgm,charset);
87 utf8_stringlist (pgm->return_path,charset);
88 utf8_stringlist (pgm->sender,charset);
89 utf8_stringlist (pgm->reply_to,charset);
90 utf8_stringlist (pgm->in_reply_to,charset);
91 utf8_stringlist (pgm->message_id,charset);
92 utf8_stringlist (pgm->newsgroups,charset);
93 utf8_stringlist (pgm->followup_to,charset);
94 utf8_stringlist (pgm->references,charset);
95 }
96 }
98 /* Convert MIME-2 sized text to UTF-8
99 * Accepts: source sized text
100 * charset
101 * flags (same as utf8_text())
102 * Returns: T if successful, NIL if failure
103 */
105 #define MINENCWORD 9
106 #define MAXENCWORD 75
108 /* This resizing algorithm is stupid, but hopefully it should never be triggered
109 * except for a pathological header. The main concern is that we don't get a
110 * buffer overflow.
111 */
113 #define DSIZE 65536 /* real headers should never be this big */
114 #define FUZZ 10 /* paranoia fuzz */
116 long utf8_mime2text (SIZEDTEXT *src,SIZEDTEXT *dst,long flags)
117 {
118 unsigned char *s,*se,*e,*ee,*t,*te;
119 char *cs,*ce,*ls;
120 SIZEDTEXT txt,rtxt;
121 unsigned long i;
122 size_t dsize = min (DSIZE,((src->size / 4) + 1) * 9);
123 /* always create buffer if canonicalizing */
124 dst->data = (flags & U8T_CANONICAL) ?
125 (unsigned char *) fs_get ((size_t) dsize) : NIL;
126 dst->size = 0; /* nothing written yet */
127 /* look for encoded words */
128 for (s = src->data, se = src->data + src->size; s < se; s++) {
129 if (((se - s) > MINENCWORD) && (*s == '=') && (s[1] == '?') &&
130 (cs = (char *) mime2_token (s+2,se,(unsigned char **) &ce)) &&
131 (e = mime2_token ((unsigned char *) ce+1,se,&ee)) &&
132 (te = mime2_text (t = e+2,se)) && (ee == e + 1) &&
133 ((te - s) < MAXENCWORD)) {
134 if (mime2_decode (e,t,te,&txt)) {
135 *ce = '\0'; /* temporarily tie off charset */
136 if (ls = strchr (cs,'*')) *ls = '\0';
137 /* convert to UTF-8 as best we can */
138 if (!utf8_text (&txt,cs,&rtxt,flags)) utf8_text (&txt,NIL,&rtxt,flags);
139 if (dst->data) { /* make sure existing buffer fits */
140 while (dsize <= (dst->size + rtxt.size + FUZZ)) {
141 dsize += DSIZE; /* kick it up */
142 fs_resize ((void **) &dst->data,dsize);
143 }
144 }
145 else { /* make a new buffer */
146 while (dsize <= (dst->size + rtxt.size)) dsize += DSIZE;
147 memcpy (dst->data = (unsigned char *) fs_get (dsize),src->data,
148 dst->size = s - src->data);
149 }
150 for (i = 0; i < rtxt.size; i++) dst->data[dst->size++] = rtxt.data[i];
152 /* all done with converted text */
153 if (rtxt.data != txt.data) fs_give ((void **) &rtxt.data);
154 if (ls) *ls = '*'; /* restore language tag delimiter */
155 *ce = '?'; /* restore charset delimiter */
156 /* all done with decoded text */
157 fs_give ((void **) &txt.data);
158 s = te+1; /* continue scan after encoded word */
159 /* skip leading whitespace */
160 for (t = s + 1; (t < se) && ((*t == ' ') || (*t == '\t')); t++);
161 /* see if likely continuation encoded word */
162 if (t < (se - MINENCWORD)) switch (*t) {
163 case '=': /* possible encoded word? */
164 if (t[1] == '?') s = t - 1;
165 break;
166 case '\015': /* CR, eat a following LF */
167 if (t[1] == '\012') t++;
168 case '\012': /* possible end of logical line */
169 if ((t[1] == ' ') || (t[1] == '\t')) {
170 do t++;
171 while ((t < (se - MINENCWORD)) && ((t[1] == ' ')||(t[1] == '\t')));
172 if ((t < (se - MINENCWORD)) && (t[1] == '=') && (t[2] == '?'))
173 s = t; /* definitely looks like continuation */
174 }
175 }
176 }
177 else { /* restore original text */
178 if (dst->data) fs_give ((void **) &dst->data);
179 dst->data = src->data;
180 dst->size = src->size;
181 return NIL; /* syntax error: MIME-2 decoding failure */
182 }
183 }
184 else do if (dst->data) { /* stash ASCII characters until LWSP */
185 if (dsize < (dst->size + FUZZ)) {
186 dsize += DSIZE; /* kick it up */
187 fs_resize ((void **) &dst->data,dsize);
188 }
189 /* kludge: assumes ASCII doesn't decompose and titlecases to one byte */
190 dst->data[dst->size++] = (flags & U8T_CASECANON) ?
191 (unsigned char) ucs4_titlecase (*s) : *s;
192 }
193 while ((*s != ' ') && (*s != '\t') && (*s != '\015') && (*s != '\012') &&
194 (++s < se));
195 }
196 if (dst->data) dst->data[dst->size] = '\0';
197 else { /* nothing converted, return identity */
198 dst->data = src->data;
199 dst->size = src->size;
200 }
201 return T; /* success */
202 }
204 /* Decode MIME-2 text
205 * Accepts: Encoding
206 * text
207 * text end
208 * destination sized text
209 * Returns: T if successful, else NIL
210 */
212 long mime2_decode (unsigned char *e,unsigned char *t,unsigned char *te,
213 SIZEDTEXT *txt)
214 {
215 unsigned char *q;
216 txt->data = NIL; /* initially no returned data */
217 switch (*e) { /* dispatch based upon encoding */
218 case 'Q': case 'q': /* sort-of QUOTED-PRINTABLE */
219 txt->data = (unsigned char *) fs_get ((size_t) (te - t) + 1);
220 for (q = t,txt->size = 0; q < te; q++) switch (*q) {
221 case '=': /* quoted character */
222 /* both must be hex */
223 if (!isxdigit (q[1]) || !isxdigit (q[2])) {
224 fs_give ((void **) &txt->data);
225 return NIL; /* syntax error: bad quoted character */
226 }
227 /* assemble character */
228 txt->data[txt->size++] = hex2byte (q[1],q[2]);
229 q += 2; /* advance past quoted character */
230 break;
231 case '_': /* convert to space */
232 txt->data[txt->size++] = ' ';
233 break;
234 default: /* ordinary character */
235 txt->data[txt->size++] = *q;
236 break;
237 }
238 txt->data[txt->size] = '\0';
239 break;
240 case 'B': case 'b': /* BASE64 */
241 if (txt->data = (unsigned char *) rfc822_base64 (t,te - t,&txt->size))
242 break;
243 default: /* any other encoding is unknown */
244 return NIL; /* syntax error: unknown encoding */
245 }
246 return T;
247 }
249 /* Get MIME-2 token from encoded word
250 * Accepts: current text pointer
251 * text limit pointer
252 * pointer to returned end pointer
253 * Returns: current text pointer & end pointer if success, else NIL
254 */
256 unsigned char *mime2_token (unsigned char *s,unsigned char *se,
257 unsigned char **t)
258 {
259 for (*t = s; **t != '?'; ++*t) {
260 if ((*t < se) && isgraph (**t)) switch (**t) {
261 case '(': case ')': case '<': case '>': case '@': case ',': case ';':
262 case ':': case '\\': case '"': case '/': case '[': case ']': case '.':
263 case '=':
264 return NIL; /* none of these are valid in tokens */
265 }
266 else return NIL; /* out of text or CTL or space */
267 }
268 return s;
269 }
272 /* Get MIME-2 text from encoded word
273 * Accepts: current text pointer
274 * text limit pointer
275 * pointer to returned end pointer
276 * Returns: end pointer if success, else NIL
277 */
279 unsigned char *mime2_text (unsigned char *s,unsigned char *se)
280 {
281 unsigned char *t = se - 1;
282 /* search for closing ?, make sure valid */
283 while ((s < t) && (*s != '?') && isgraph (*s++));
284 return ((s < t) && (*s == '?') && (s[1] == '=') &&
285 ((se == (s + 2)) || (s[2] == ' ') || (s[2] == '\t') ||
286 (s[2] == '\015') || (s[2] == '\012'))) ? s : NIL;
287 }
289 /* Convert UTF-16 string to Modified Base64
290 * Accepts: destination pointer
291 * source string
292 * source length in octets
293 * Returns: updated destination pointer
294 */
296 static unsigned char *utf16_to_mbase64 (unsigned char *t,unsigned char *s,
297 size_t i)
298 {
299 char *v = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+,";
300 *t++ = '&'; /* write shift-in */
301 while (i >= 3) { /* process tuplets */
302 *t++ = v[s[0] >> 2]; /* byte 1: high 6 bits (1) */
303 /* byte 2: low 2 bits (1), high 4 bits (2) */
304 *t++ = v[((s[0] << 4) + (s[1] >> 4)) & 0x3f];
305 /* byte 3: low 4 bits (2), high 2 bits (3) */
306 *t++ = v[((s[1] << 2) + (s[2] >> 6)) & 0x3f];
307 *t++ = v[s[2] & 0x3f]; /* byte 4: low 6 bits (3) */
308 s += 3;
309 i -= 3;
310 }
311 if (i) {
312 *t++ = v[s[0] >> 2]; /* byte 1: high 6 bits (1) */
313 /* byte 2: low 2 bits (1), high 4 bits (2) */
314 *t++ = v[((s[0] << 4) + (--i ? (s[1] >> 4) : 0)) & 0x3f];
315 /* byte 3: low 4 bits (2) */
316 if (i) *t++ = v[(s[1] << 2) & 0x3f];
317 }
318 *t++ = '-'; /* write shift-out */
319 return t;
320 }
323 /* Poot a UTF-16 value to a buffer
324 * Accepts: buffer pointer
325 * value
326 * Returns: updated pointer
327 */
329 static unsigned char *utf16_poot (unsigned char *s,unsigned long c)
330 {
331 *s++ = (unsigned char) (c >> 8);
332 *s++ = (unsigned char) (c & 0xff);
333 return s;
334 }
336 /* Convert UTF-8 to Modified UTF-7
337 * Accepts: UTF-8 string
338 * Returns: Modified UTF-7 string on success, NIL if invalid UTF-8
339 */
341 #define MAXUNIUTF8 4 /* maximum length of Unicode UTF-8 sequence */
343 unsigned char *utf8_to_mutf7 (unsigned char *src)
344 {
345 unsigned char *u16buf,*utf16;
346 unsigned char *ret,*t;
347 unsigned long j,c;
348 unsigned char *s = src;
349 unsigned long i = 0;
350 int nonascii = 0;
351 while (*s) { /* pass one: count destination octets */
352 if (*s & 0x80) { /* non-ASCII character? */
353 j = MAXUNIUTF8; /* get single UCS-4 codepoint */
354 if ((c = utf8_get (&s,&j)) & U8G_ERROR) return NIL;
355 /* tally number of UTF-16 octets */
356 nonascii += (c & U8GM_NONBMP) ? 4 : 2;
357 }
358 else { /* ASCII character */
359 if (nonascii) { /* add pending Modified BASE64 size + shifts */
360 i += ((nonascii / 3) * 4) + ((j = nonascii % 3) ? j + 1 : 0) + 2;
361 nonascii = 0; /* back to ASCII */
362 }
363 if (*s == '&') i += 2; /* two octets if the escape */
364 else ++i; /* otherwise just count another octet */
365 ++s; /* advance to next source octet */
366 }
367 }
368 if (nonascii) /* add pending Modified BASE64 size + shifts */
369 i += ((nonascii / 3) * 4) + ((j = nonascii % 3) ? j + 1 : 0) + 2;
371 /* create return buffer */
372 t = ret = (unsigned char *) fs_get (i + 1);
373 /* and scratch buffer */
374 utf16 = u16buf = (unsigned char *) fs_get (i + 1);
375 for (s = src; *s;) { /* pass two: copy destination octets */
376 if (*s & 0x80) { /* non-ASCII character? */
377 j = MAXUNIUTF8; /* get single UCS-4 codepoint */
378 if ((c = utf8_get (&s,&j)) & U8G_ERROR) return NIL;
379 if (c & U8GM_NONBMP) { /* non-BMP? */
380 c -= UTF16_BASE; /* yes, convert to surrogate */
381 utf16 = utf16_poot (utf16_poot (utf16,(c >> UTF16_SHIFT)+UTF16_SURRH),
382 (c & UTF16_MASK) + UTF16_SURRL);
383 }
384 else utf16 = utf16_poot (utf16,c);
385 }
386 else { /* ASCII character */
387 if (utf16 != u16buf) { /* add pending Modified BASE64 size + shifts */
388 t = utf16_to_mbase64 (t,u16buf,utf16 - u16buf);
389 utf16 = u16buf; /* reset buffer */
390 }
391 *t++ = *s; /* copy the character */
392 if (*s == '&') *t++ = '-';/* special sequence if the escape */
393 ++s; /* advance to next source octet */
394 }
395 }
396 /* add pending Modified BASE64 size + shifts */
397 if (utf16 != u16buf) t = utf16_to_mbase64 (t,u16buf,utf16 - u16buf);
398 *t = '\0'; /* tie off destination */
399 if (i != (t - ret)) fatal ("utf8_to_mutf7 botch");
400 fs_give ((void **) &u16buf);
401 return ret;
402 }
404 /* Convert Modified UTF-7 to UTF-8
405 * Accepts: Modified UTF-7 string
406 * Returns: UTF-8 string on success, NIL if invalid Modified UTF-7
407 */
409 unsigned char *utf8_from_mutf7 (unsigned char *src)
410 {
411 SIZEDTEXT utf8,utf7;
412 unsigned char *s;
413 int mbase64 = 0;
414 /* disallow bogus strings */
415 if (mail_utf7_valid (src)) return NIL;
416 /* initialize SIZEDTEXTs */
417 memset (&utf7,0,sizeof (SIZEDTEXT));
418 memset (&utf8,0,sizeof (SIZEDTEXT));
419 /* make copy of source */
420 for (s = cpytxt (&utf7,src,strlen (src)); *s; ++s) switch (*s) {
421 case '&': /* Modified UTF-7 uses & instead of + */
422 *s = '+';
423 mbase64 = T; /* note that we are in Modified BASE64 */
424 break;
425 case '+': /* temporarily swap text + to & */
426 if (!mbase64) *s = '&';
427 break;
428 case '-': /* shift back to ASCII */
429 mbase64 = NIL;
430 break;
431 case ',': /* Modified UTF-7 uses , instead of / ... */
432 if (mbase64) *s = '/'; /* ...in Modified BASE64 */
433 break;
434 }
435 /* do the conversion */
436 utf8_text_utf7 (&utf7,&utf8,NIL,NIL);
437 /* no longer need copy of source */
438 fs_give ((void **) &utf7.data);
439 /* post-process: switch & and + */
440 for (s = utf8.data; *s; ++s) switch (*s) {
441 case '&':
442 *s = '+';
443 break;
444 case '+':
445 *s = '&';
446 break;
447 }
448 return utf8.data;
449 }