Add `urlcodec', obsolete `url{en,de}code'
[s-mailx.git] / head.c
bloba525c5de6557690fbe355fdc28e3f6906e8051c9
1 /*@ S-nail - a mail user agent derived from Berkeley Mail.
2 *@ Routines for processing and detecting headlines.
4 * Copyright (c) 2000-2004 Gunnar Ritter, Freiburg i. Br., Germany.
5 * Copyright (c) 2012 - 2015 Steffen (Daode) Nurpmeso <sdaoden@users.sf.net>.
6 */
7 /*
8 * Copyright (c) 1980, 1993
9 * The Regents of the University of California. All rights reserved.
11 * Redistribution and use in source and binary forms, with or without
12 * modification, are permitted provided that the following conditions
13 * are met:
14 * 1. Redistributions of source code must retain the above copyright
15 * notice, this list of conditions and the following disclaimer.
16 * 2. Redistributions in binary form must reproduce the above copyright
17 * notice, this list of conditions and the following disclaimer in the
18 * documentation and/or other materials provided with the distribution.
19 * 3. Neither the name of the University nor the names of its contributors
20 * may be used to endorse or promote products derived from this software
21 * without specific prior written permission.
23 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
24 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
25 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
26 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
27 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
28 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
29 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
30 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
31 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
32 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33 * SUCH DAMAGE.
35 #undef n_FILE
36 #define n_FILE head
38 #ifndef HAVE_AMALGAMATION
39 # include "nail.h"
40 #endif
42 #ifdef HAVE_IDNA
43 # if HAVE_IDNA == HAVE_IDNA_LIBIDNA
44 # include <idna.h>
45 # include <idn-free.h>
46 # include <stringprep.h>
47 # elif HAVE_IDNA == HAVE_IDNA_IDNKIT
48 # include <idn/api.h>
49 # endif
50 #endif
52 struct cmatch_data {
53 size_t tlen; /* Length of .tdata */
54 char const *tdata; /* Template date - see _cmatch_data[] */
57 /* Template characters for cmatch_data.tdata:
58 * 'A' An upper case char
59 * 'a' A lower case char
60 * ' ' A space
61 * '0' A digit
62 * 'O' An optional digit or space
63 * ':' A colon
64 * '+' Either a plus or a minus sign */
65 static struct cmatch_data const _cmatch_data[] = {
66 { 24, "Aaa Aaa O0 00:00:00 0000" }, /* BSD/ISO C90 ctime */
67 { 28, "Aaa Aaa O0 00:00:00 AAA 0000" }, /* BSD tmz */
68 { 21, "Aaa Aaa O0 00:00 0000" }, /* SysV ctime */
69 { 25, "Aaa Aaa O0 00:00 AAA 0000" }, /* SysV tmz */
70 /* RFC 822-alike From_ lines do not conform to RFC 4155, but seem to be used
71 * in the wild (by UW-imap) */
72 { 30, "Aaa Aaa O0 00:00:00 0000 +0000" },
73 /* RFC 822 with zone spec; 1. military, 2. UT, 3. north america time
74 * zone strings; note that 1. is strictly speaking not correct as some
75 * letters are not used, and 2. is not because only "UT" is defined */
76 #define __reuse "Aaa Aaa O0 00:00:00 0000 AAA"
77 { 28 - 2, __reuse }, { 28 - 1, __reuse }, { 28 - 0, __reuse },
78 { 0, NULL }
80 #define _DATE_MINLEN 21
82 /* Skip over "word" as found in From_ line */
83 static char const * _from__skipword(char const *wp);
85 /* Match the date string against the date template (tp), return if match.
86 * See _cmatch_data[] for template character description */
87 static int _cmatch(size_t len, char const *date,
88 char const *tp);
90 /* Check wether date is a valid 'From_' date.
91 * (Rather ctime(3) generated dates, according to RFC 4155) */
92 static int _is_date(char const *date);
94 /* Convert the domain part of a skinned address to IDNA.
95 * If an error occurs before Unicode information is available, revert the IDNA
96 * error to a normal CHAR one so that the error message doesn't talk Unicode */
97 #ifdef HAVE_IDNA
98 static struct addrguts * _idna_apply(struct addrguts *agp);
99 #endif
101 /* Classify and check a (possibly skinned) header body according to RFC
102 * *addr-spec* rules; if it (is assumed to has been) skinned it may however be
103 * also a file or a pipe command, so check that first, then.
104 * Otherwise perform content checking and isolate the domain part (for IDNA) */
105 static int _addrspec_check(int doskin, struct addrguts *agp);
107 /* Return the next header field found in the given message.
108 * Return >= 0 if something found, < 0 elsewise.
109 * "colon" is set to point to the colon in the header.
110 * Must deal with \ continuations & other such fraud */
111 static int gethfield(FILE *f, char **linebuf, size_t *linesize,
112 int rem, char **colon);
114 static int msgidnextc(char const **cp, int *status);
116 /* Count the occurances of c in str */
117 static int charcount(char *str, int c);
119 static char const * nexttoken(char const *cp);
121 static char const *
122 _from__skipword(char const *wp)
124 char c = 0;
125 NYD2_ENTER;
127 if (wp != NULL) {
128 while ((c = *wp++) != '\0' && !blankchar(c)) {
129 if (c == '"') {
130 while ((c = *wp++) != '\0' && c != '"')
132 if (c != '"')
133 --wp;
136 for (; blankchar(c); c = *wp++)
139 NYD2_LEAVE;
140 return (c == 0 ? NULL : wp - 1);
143 static int
144 _cmatch(size_t len, char const *date, char const *tp)
146 int ret = 0;
147 NYD2_ENTER;
149 while (len--) {
150 char c = date[len];
151 switch (tp[len]) {
152 case 'a':
153 if (!lowerchar(c))
154 goto jleave;
155 break;
156 case 'A':
157 if (!upperchar(c))
158 goto jleave;
159 break;
160 case ' ':
161 if (c != ' ')
162 goto jleave;
163 break;
164 case '0':
165 if (!digitchar(c))
166 goto jleave;
167 break;
168 case 'O':
169 if (c != ' ' && !digitchar(c))
170 goto jleave;
171 break;
172 case ':':
173 if (c != ':')
174 goto jleave;
175 break;
176 case '+':
177 if (c != '+' && c != '-')
178 goto jleave;
179 break;
182 ret = 1;
183 jleave:
184 NYD2_LEAVE;
185 return ret;
188 static int
189 _is_date(char const *date)
191 struct cmatch_data const *cmdp;
192 size_t dl;
193 int rv = 0;
194 NYD2_ENTER;
196 if ((dl = strlen(date)) >= _DATE_MINLEN)
197 for (cmdp = _cmatch_data; cmdp->tdata != NULL; ++cmdp)
198 if (dl == cmdp->tlen && (rv = _cmatch(dl, date, cmdp->tdata)))
199 break;
200 NYD2_LEAVE;
201 return rv;
204 #ifdef HAVE_IDNA
205 # if HAVE_IDNA == HAVE_IDNA_LIBIDNA
206 static struct addrguts *
207 _idna_apply(struct addrguts *agp)
209 char *idna_utf8, *idna_ascii, *cs;
210 size_t sz, i;
211 NYD_ENTER;
213 sz = agp->ag_slen - agp->ag_sdom_start;
214 assert(sz > 0);
215 idna_utf8 = ac_alloc(sz +1);
216 memcpy(idna_utf8, agp->ag_skinned + agp->ag_sdom_start, sz);
217 idna_utf8[sz] = '\0';
219 /* GNU Libidn settles on top of iconv(3) without any fallback, so let's just
220 * let it perform the charset conversion, if any should be necessary */
221 if (!(options & OPT_UNICODE)) {
222 char const *tcs = charset_get_lc();
223 idna_ascii = idna_utf8;
224 idna_utf8 = stringprep_convert(idna_ascii, "UTF-8", tcs);
225 i = (idna_utf8 == NULL && errno == EINVAL);
226 ac_free(idna_ascii);
228 if (idna_utf8 == NULL) {
229 if (i)
230 n_err(_("Cannot convert from %s to %s\n"), tcs, "UTF-8");
231 agp->ag_n_flags ^= NAME_ADDRSPEC_ERR_IDNA | NAME_ADDRSPEC_ERR_CHAR;
232 goto jleave;
236 if (idna_to_ascii_8z(idna_utf8, &idna_ascii, 0) != IDNA_SUCCESS) {
237 agp->ag_n_flags ^= NAME_ADDRSPEC_ERR_IDNA | NAME_ADDRSPEC_ERR_CHAR;
238 goto jleave1;
241 /* Replace the domain part of .ag_skinned with IDNA version */
242 sz = strlen(idna_ascii);
243 i = agp->ag_sdom_start;
244 cs = salloc(agp->ag_slen - i + sz +1);
245 memcpy(cs, agp->ag_skinned, i);
246 memcpy(cs + i, idna_ascii, sz);
247 i += sz;
248 cs[i] = '\0';
250 agp->ag_skinned = cs;
251 agp->ag_slen = i;
252 NAME_ADDRSPEC_ERR_SET(agp->ag_n_flags,
253 NAME_NAME_SALLOC | NAME_SKINNED | NAME_IDNA, 0);
255 idn_free(idna_ascii);
256 jleave1:
257 if (options & OPT_UNICODE)
258 ac_free(idna_utf8);
259 else
260 idn_free(idna_utf8);
261 jleave:
262 NYD_LEAVE;
263 return agp;
266 # elif HAVE_IDNA == HAVE_IDNA_IDNKIT /* IDNA==LIBIDNA */
267 static struct addrguts *
268 _idna_apply(struct addrguts *agp)
270 char *idna_in, *idna_out, *cs;
271 size_t sz, i;
272 idn_result_t r;
273 NYD_ENTER;
275 sz = agp->ag_slen - agp->ag_sdom_start;
276 assert(sz > 0);
277 idna_in = ac_alloc(sz +1);
278 memcpy(idna_in, agp->ag_skinned + agp->ag_sdom_start, sz);
279 idna_in[sz] = '\0';
281 for (idna_out = NULL, sz = HOST_NAME_MAX +1;; sz += HOST_NAME_MAX) {
282 idna_out = ac_alloc(sz);
284 r = idn_encodename(IDN_ENCODE_APP, idna_in, idna_out, sz);
285 switch (r) {
286 case idn_success:
287 case idn_buffer_overflow:
288 break;
289 case idn_invalid_encoding:
290 n_err(_("Cannot convert from %s to %s\n"), charset_get_lc(), "UTF-8");
291 /* FALLTHRU */
292 default:
293 agp->ag_n_flags ^= NAME_ADDRSPEC_ERR_IDNA | NAME_ADDRSPEC_ERR_CHAR;
294 goto jleave;
297 if (r == idn_success)
298 break;
299 ac_free(idna_out);
302 /* Replace the domain part of .ag_skinned with IDNA version */
303 sz = strlen(idna_out);
304 i = agp->ag_sdom_start;
305 cs = salloc(agp->ag_slen - i + sz +1);
306 memcpy(cs, agp->ag_skinned, i);
307 memcpy(cs + i, idna_out, sz);
308 i += sz;
309 cs[i] = '\0';
311 agp->ag_skinned = cs;
312 agp->ag_slen = i;
313 NAME_ADDRSPEC_ERR_SET(agp->ag_n_flags,
314 NAME_NAME_SALLOC | NAME_SKINNED | NAME_IDNA, 0);
316 jleave:
317 ac_free(idna_out);
318 ac_free(idna_in);
319 NYD_LEAVE;
320 return agp;
322 # endif /* IDNA==IDNKIT */
323 #endif /* HAVE_IDNA */
325 static int
326 _addrspec_check(int skinned, struct addrguts *agp)
328 char *addr, *p;
329 bool_t in_quote;
330 ui8_t in_domain, hadat;
331 union {char c; unsigned char u;} c;
332 #ifdef HAVE_IDNA
333 ui8_t use_idna;
334 #endif
335 NYD_ENTER;
337 #ifdef HAVE_IDNA
338 use_idna = ok_blook(idna_disable) ? 0 : 1;
339 #endif
340 agp->ag_n_flags |= NAME_ADDRSPEC_CHECKED;
341 addr = agp->ag_skinned;
343 if (agp->ag_iaddr_aend - agp->ag_iaddr_start == 0) {
344 NAME_ADDRSPEC_ERR_SET(agp->ag_n_flags, NAME_ADDRSPEC_ERR_EMPTY, 0);
345 goto jleave;
348 /* If the field is not a recipient, it cannot be a file or a pipe */
349 if (!skinned)
350 goto jaddr_check;
352 /* When changing any of the following adjust any RECIPIENTADDRSPEC;
353 * grep the latter for the complete picture */
354 if (*addr == '|') {
355 agp->ag_n_flags |= NAME_ADDRSPEC_ISPIPE;
356 goto jleave;
358 if (addr[0] == '/' || (addr[0] == '.' && addr[1] == '/'))
359 goto jisfile;
360 if (memchr(addr, '@', agp->ag_slen) == NULL) {
361 if (*addr == '+')
362 goto jisfile;
363 for (p = addr; (c.c = *p); ++p) {
364 if (c.c == '!' || c.c == '%')
365 break;
366 if (c.c == '/') {
367 jisfile:
368 agp->ag_n_flags |= NAME_ADDRSPEC_ISFILE;
369 goto jleave;
374 jaddr_check:
375 in_quote = FAL0;
376 in_domain = hadat = 0;
378 for (p = addr; (c.c = *p++) != '\0';) {
379 if (c.c == '"') {
380 in_quote = !in_quote;
381 } else if (c.u < 040 || c.u >= 0177) { /* TODO no magics: !bodychar()? */
382 #ifdef HAVE_IDNA
383 if (in_domain && use_idna > 0) {
384 if (use_idna == 1)
385 NAME_ADDRSPEC_ERR_SET(agp->ag_n_flags, NAME_ADDRSPEC_ERR_IDNA,
386 c.u);
387 use_idna = 2;
388 } else
389 #endif
390 break;
391 } else if (in_domain == 2) {
392 if ((c.c == ']' && *p != '\0') || c.c == '\\' || whitechar(c.c))
393 break;
394 } else if (in_quote && in_domain == 0) {
395 /*EMPTY*/;
396 } else if (c.c == '\\' && *p != '\0') {
397 ++p;
398 } else if (c.c == '@') {
399 if (hadat++ > 0) {
400 NAME_ADDRSPEC_ERR_SET(agp->ag_n_flags, NAME_ADDRSPEC_ERR_ATSEQ,
401 c.u);
402 goto jleave;
404 agp->ag_sdom_start = PTR2SIZE(p - addr);
405 agp->ag_n_flags |= NAME_ADDRSPEC_ISADDR; /* TODO .. really? */
406 in_domain = (*p == '[') ? 2 : 1;
407 continue;
408 } else if (c.c == '(' || c.c == ')' || c.c == '<' || c.c == '>' ||
409 c.c == ',' || c.c == ';' || c.c == ':' || c.c == '\\' ||
410 c.c == '[' || c.c == ']')
411 break;
412 hadat = 0;
414 if (c.c != '\0') {
415 NAME_ADDRSPEC_ERR_SET(agp->ag_n_flags, NAME_ADDRSPEC_ERR_CHAR, c.u);
416 goto jleave;
419 if (!(agp->ag_n_flags & NAME_ADDRSPEC_ISADDR))
420 agp->ag_n_flags |= NAME_ADDRSPEC_ISNAME;
422 #ifdef HAVE_IDNA
423 if (use_idna == 2)
424 agp = _idna_apply(agp);
425 #endif
426 jleave:
427 NYD_LEAVE;
428 return ((agp->ag_n_flags & NAME_ADDRSPEC_INVALID) != 0);
431 static int
432 gethfield(FILE *f, char **linebuf, size_t *linesize, int rem, char **colon)
434 char *line2 = NULL, *cp, *cp2;
435 size_t line2size = 0;
436 int c, isenc;
437 NYD2_ENTER;
439 if (*linebuf == NULL)
440 *linebuf = srealloc(*linebuf, *linesize = 1);
441 **linebuf = '\0';
442 for (;;) {
443 if (--rem < 0) {
444 rem = -1;
445 break;
447 if ((c = readline_restart(f, linebuf, linesize, 0)) <= 0) {
448 rem = -1;
449 break;
451 for (cp = *linebuf; fieldnamechar(*cp); ++cp)
453 if (cp > *linebuf)
454 while (blankchar(*cp))
455 ++cp;
456 if (*cp != ':' || cp == *linebuf)
457 continue;
459 /* I guess we got a headline. Handle wraparound */
460 *colon = cp;
461 cp = *linebuf + c;
462 for (;;) {
463 isenc = 0;
464 while (PTRCMP(--cp, >=, *linebuf) && blankchar(*cp))
466 cp++;
467 if (rem <= 0)
468 break;
469 if (PTRCMP(cp - 8, >=, *linebuf) && cp[-1] == '=' && cp[-2] == '?')
470 isenc |= 1;
471 ungetc(c = getc(f), f);
472 if (!blankchar(c))
473 break;
474 c = readline_restart(f, &line2, &line2size, 0);
475 if (c < 0)
476 break;
477 --rem;
478 for (cp2 = line2; blankchar(*cp2); ++cp2)
480 c -= (int)PTR2SIZE(cp2 - line2);
481 if (cp2[0] == '=' && cp2[1] == '?' && c > 8)
482 isenc |= 2;
483 if (PTRCMP(cp + c, >=, *linebuf + *linesize - 2)) {
484 size_t diff = PTR2SIZE(cp - *linebuf),
485 colondiff = PTR2SIZE(*colon - *linebuf);
486 *linebuf = srealloc(*linebuf, *linesize += c + 2);
487 cp = &(*linebuf)[diff];
488 *colon = &(*linebuf)[colondiff];
490 if (isenc != 3)
491 *cp++ = ' ';
492 memcpy(cp, cp2, c);
493 cp += c;
495 *cp = '\0';
497 if (line2 != NULL)
498 free(line2);
499 break;
501 NYD2_LEAVE;
502 return rem;
505 static int
506 msgidnextc(char const **cp, int *status)
508 int c;
509 NYD2_ENTER;
511 assert(cp != NULL);
512 assert(*cp != NULL);
513 assert(status != NULL);
515 for (;;) {
516 if (*status & 01) {
517 if (**cp == '"') {
518 *status &= ~01;
519 (*cp)++;
520 continue;
522 if (**cp == '\\') {
523 (*cp)++;
524 if (**cp == '\0')
525 goto jeof;
527 goto jdfl;
529 switch (**cp) {
530 case '(':
531 *cp = skip_comment(&(*cp)[1]);
532 continue;
533 case '>':
534 case '\0':
535 jeof:
536 c = '\0';
537 goto jleave;
538 case '"':
539 (*cp)++;
540 *status |= 01;
541 continue;
542 case '@':
543 *status |= 02;
544 /*FALLTHRU*/
545 default:
546 jdfl:
547 c = *(*cp)++ & 0377;
548 c = (*status & 02) ? lowerconv(c) : c;
549 goto jleave;
552 jleave:
553 NYD2_LEAVE;
554 return c;
557 static int
558 charcount(char *str, int c)
560 char *cp;
561 int i;
562 NYD2_ENTER;
564 for (i = 0, cp = str; *cp; ++cp)
565 if (*cp == c)
566 ++i;
567 NYD2_LEAVE;
568 return i;
571 static char const *
572 nexttoken(char const *cp)
574 NYD2_ENTER;
575 for (;;) {
576 if (*cp == '\0') {
577 cp = NULL;
578 break;
581 if (*cp == '(') {
582 size_t nesting = 1;
584 do switch (*++cp) {
585 case '(':
586 ++nesting;
587 break;
588 case ')':
589 --nesting;
590 break;
591 } while (nesting > 0 && *cp != '\0'); /* XXX error? */
592 } else if (blankchar(*cp) || *cp == ',')
593 ++cp;
594 else
595 break;
597 NYD2_LEAVE;
598 return cp;
601 FL char const *
602 myaddrs(struct header *hp)
604 struct name *np;
605 char *rv;
606 NYD_ENTER;
608 if (hp != NULL && (np = hp->h_from) != NULL) {
609 if ((rv = np->n_fullname) != NULL)
610 goto jleave;
611 if ((rv = np->n_name) != NULL)
612 goto jleave;
615 if ((rv = ok_vlook(from)) != NULL)
616 goto jleave;
618 /* When invoking *sendmail* directly, it's its task to generate an otherwise
619 * undeterminable From: address. However, if the user sets *hostname*,
620 * accept his desire */
621 if (ok_vlook(smtp) != NULL || ok_vlook(hostname) != NULL) {
622 char *hn = nodename(1);
623 size_t sz = strlen(myname) + strlen(hn) + 1 +1;
624 rv = salloc(sz);
625 sstpcpy(sstpcpy(sstpcpy(rv, myname), "@"), hn);
627 jleave:
628 NYD_LEAVE;
629 return rv;
632 FL char const *
633 myorigin(struct header *hp)
635 char const *rv = NULL, *ccp;
636 struct name *np;
637 NYD_ENTER;
639 if ((ccp = myaddrs(hp)) != NULL &&
640 (np = lextract(ccp, GEXTRA | GFULL)) != NULL)
641 rv = (np->n_flink != NULL) ? ok_vlook(sender) : ccp;
642 NYD_LEAVE;
643 return rv;
646 FL bool_t
647 is_head(char const *linebuf, size_t linelen, bool_t check_rfc4155)
649 char date[FROM_DATEBUF];
650 bool_t rv;
651 NYD2_ENTER;
653 if ((rv = (linelen >= 5 && !memcmp(linebuf, "From ", 5))) && check_rfc4155 &&
654 (extract_date_from_from_(linebuf, linelen, date) <= 0 ||
655 !_is_date(date)))
656 rv = TRUM1;
657 NYD2_LEAVE;
658 return rv;
661 FL int
662 extract_date_from_from_(char const *line, size_t linelen,
663 char datebuf[FROM_DATEBUF])
665 int rv;
666 char const *cp = line;
667 NYD_ENTER;
669 rv = 1;
671 /* "From " */
672 cp = _from__skipword(cp);
673 if (cp == NULL)
674 goto jerr;
675 /* "addr-spec " */
676 cp = _from__skipword(cp);
677 if (cp == NULL)
678 goto jerr;
679 if (cp[0] == 't' && cp[1] == 't' && cp[2] == 'y') {
680 cp = _from__skipword(cp);
681 if (cp == NULL)
682 goto jerr;
684 /* It seems there are invalid MBOX archives in the wild, compare
685 * . http://bugs.debian.org/624111
686 * . [Mutt] #3868: mutt should error if the imported mailbox is invalid
687 * What they do is that they obfuscate the address to "name at host",
688 * and even "name at host dot dom dot dom. I think we should handle that */
689 else if(cp[0] == 'a' && cp[1] == 't' && cp[2] == ' '){
690 rv = -1;
691 cp += 3;
692 jat_dot:
693 cp = _from__skipword(cp);
694 if (cp == NULL)
695 goto jerr;
696 if(cp[0] == 'd' && cp[1] == 'o' && cp[2] == 't' && cp[3] == ' '){
697 cp += 4;
698 goto jat_dot;
702 linelen -= PTR2SIZE(cp - line);
703 if (linelen < _DATE_MINLEN)
704 goto jerr;
705 if (cp[linelen - 1] == '\n') {
706 --linelen;
707 /* (Rather IMAP/POP3 only) */
708 if (cp[linelen - 1] == '\r')
709 --linelen;
710 if (linelen < _DATE_MINLEN)
711 goto jerr;
713 if (linelen >= FROM_DATEBUF)
714 goto jerr;
716 jleave:
717 memcpy(datebuf, cp, linelen);
718 datebuf[linelen] = '\0';
719 NYD_LEAVE;
720 return rv;
721 jerr:
722 cp = _("<Unknown date>");
723 linelen = strlen(cp);
724 if (linelen >= FROM_DATEBUF)
725 linelen = FROM_DATEBUF;
726 rv = 0;
727 goto jleave;
730 FL void
731 extract_header(FILE *fp, struct header *hp, si8_t *checkaddr_err)
733 /* See the prototype declaration for the hairy relationship of
734 * options&OPT_t_FLAG and/or pstate&PS_t_FLAG in here */
735 struct header nh, *hq = &nh;
736 char *linebuf = NULL /* TODO line pool */, *colon;
737 size_t linesize = 0, seenfields = 0;
738 int lc, c;
739 char const *val, *cp;
740 NYD_ENTER;
742 memset(hq, 0, sizeof *hq);
743 if ((pstate & PS_t_FLAG) && (options & OPT_t_FLAG)) {
744 hq->h_to = hp->h_to;
745 hq->h_cc = hp->h_cc;
746 hq->h_bcc = hp->h_bcc;
749 for (lc = 0; readline_restart(fp, &linebuf, &linesize, 0) > 0; ++lc)
752 /* TODO yippieia, cat(check(lextract)) :-) */
753 rewind(fp);
754 while ((lc = gethfield(fp, &linebuf, &linesize, lc, &colon)) >= 0) {
755 struct name *np;
757 /* We explicitly allow EAF_NAME for some addressees since aliases are not
758 * yet expanded when we parse these! */
759 if ((val = thisfield(linebuf, "to")) != NULL) {
760 ++seenfields;
761 hq->h_to = cat(hq->h_to, checkaddrs(lextract(val, GTO | GFULL),
762 EACM_NORMAL | EAF_NAME, checkaddr_err));
763 } else if ((val = thisfield(linebuf, "cc")) != NULL) {
764 ++seenfields;
765 hq->h_cc = cat(hq->h_cc, checkaddrs(lextract(val, GCC | GFULL),
766 EACM_NORMAL | EAF_NAME, checkaddr_err));
767 } else if ((val = thisfield(linebuf, "bcc")) != NULL) {
768 ++seenfields;
769 hq->h_bcc = cat(hq->h_bcc, checkaddrs(lextract(val, GBCC | GFULL),
770 EACM_NORMAL | EAF_NAME, checkaddr_err));
771 } else if ((val = thisfield(linebuf, "from")) != NULL) {
772 if (!(pstate & PS_t_FLAG) || (options & OPT_t_FLAG)) {
773 ++seenfields;
774 hq->h_from = cat(hq->h_from,
775 checkaddrs(lextract(val, GEXTRA | GFULL | GFULLEXTRA),
776 EACM_STRICT, NULL));
778 } else if ((val = thisfield(linebuf, "reply-to")) != NULL) {
779 ++seenfields;
780 hq->h_replyto = cat(hq->h_replyto,
781 checkaddrs(lextract(val, GEXTRA | GFULL), EACM_STRICT, NULL));
782 } else if ((val = thisfield(linebuf, "sender")) != NULL) {
783 if (!(pstate & PS_t_FLAG) || (options & OPT_t_FLAG)) {
784 ++seenfields;
785 hq->h_sender = cat(hq->h_sender, /* TODO cat? check! */
786 checkaddrs(lextract(val, GEXTRA | GFULL | GFULLEXTRA),
787 EACM_STRICT, NULL));
789 } else if ((val = thisfield(linebuf, "organization")) != NULL) {
790 ++seenfields;
791 for (cp = val; blankchar(*cp); ++cp)
793 hq->h_organization = (hq->h_organization != NULL)
794 ? save2str(hq->h_organization, cp) : savestr(cp);
795 } else if ((val = thisfield(linebuf, "subject")) != NULL ||
796 (val = thisfield(linebuf, "subj")) != NULL) {
797 ++seenfields;
798 for (cp = val; blankchar(*cp); ++cp)
800 hq->h_subject = (hq->h_subject != NULL)
801 ? save2str(hq->h_subject, cp) : savestr(cp);
803 /* The remaining are mostly hacked in and thus TODO -- at least in
804 * TODO respect to their content checking */
805 else if ((pstate & PS_t_FLAG) &&
806 (val = thisfield(linebuf, "message-id")) != NULL) {
807 np = checkaddrs(lextract(val, GREF),
808 /*EACM_STRICT | TODO '/' valid!! */ EACM_NOLOG | EACM_NONAME,
809 NULL);
810 if (np == NULL || np->n_flink != NULL)
811 goto jebadhead;
812 ++seenfields;
813 hq->h_message_id = np;
814 } else if ((pstate & PS_t_FLAG) &&
815 (val = thisfield(linebuf, "in-reply-to")) != NULL) {
816 np = checkaddrs(lextract(val, GREF),
817 /*EACM_STRICT | TODO '/' valid!! */ EACM_NOLOG | EACM_NONAME,
818 NULL);
819 if (np == NULL || np->n_flink != NULL)
820 goto jebadhead;
821 ++seenfields;
822 hq->h_in_reply_to = np;
823 } else if ((pstate & PS_t_FLAG) &&
824 (val = thisfield(linebuf, "references")) != NULL) {
825 ++seenfields;
826 /* TODO Limit number of references TODO better on parser side */
827 hq->h_ref = cat(hq->h_ref, checkaddrs(extract(val, GREF),
828 /*EACM_STRICT | TODO '/' valid!! */ EACM_NOLOG | EACM_NONAME,
829 NULL));
830 /* and that is very hairy */
831 } else if ((pstate & PS_t_FLAG) &&
832 (val = thisfield(linebuf, "mail-followup-to")) != NULL) {
833 ++seenfields;
834 hq->h_mft = cat(hq->h_mft, checkaddrs(lextract(val, GEXTRA | GFULL),
835 /*EACM_STRICT | TODO '/' valid!! | EACM_NOLOG | */ EACM_NONAME,
836 checkaddr_err));
837 } else
838 jebadhead:
839 n_err(_("Ignoring header field \"%s\"\n"), linebuf);
842 /* In case the blank line after the header has been edited out. Otherwise,
843 * fetch the header separator */
844 if (linebuf != NULL) {
845 if (linebuf[0] != '\0') {
846 for (cp = linebuf; *(++cp) != '\0';)
848 fseek(fp, (long)-PTR2SIZE(1 + cp - linebuf), SEEK_CUR);
849 } else {
850 if ((c = getc(fp)) != '\n' && c != EOF)
851 ungetc(c, fp);
855 if (seenfields > 0) {
856 hp->h_to = hq->h_to;
857 hp->h_cc = hq->h_cc;
858 hp->h_bcc = hq->h_bcc;
859 hp->h_from = hq->h_from;
860 hp->h_replyto = hq->h_replyto;
861 hp->h_sender = hq->h_sender;
862 hp->h_organization = hq->h_organization;
863 if (hq->h_subject != NULL || !(pstate & PS_t_FLAG) ||
864 !(options & OPT_t_FLAG))
865 hp->h_subject = hq->h_subject;
867 if (pstate & PS_t_FLAG) {
868 hp->h_ref = hq->h_ref;
869 hp->h_message_id = hq->h_message_id;
870 hp->h_in_reply_to = hq->h_in_reply_to;
871 hp->h_mft = hq->h_mft;
873 /* And perform additional validity checks so that we don't bail later
874 * on TODO this is good and the place where this should occur,
875 * TODO unfortunately a lot of other places do again and blabla */
876 if (pstate & PS_t_FLAG) {
877 if (hp->h_from == NULL)
878 hp->h_from = option_r_arg;
879 else if (hp->h_from->n_flink != NULL && hp->h_sender == NULL)
880 hp->h_sender = lextract(ok_vlook(sender),
881 GEXTRA | GFULL | GFULLEXTRA);
884 } else
885 n_err(_("Restoring deleted header lines\n"));
887 if (linebuf != NULL)
888 free(linebuf);
889 NYD_LEAVE;
892 FL char *
893 hfield_mult(char const *field, struct message *mp, int mult)
895 FILE *ibuf;
896 int lc;
897 struct str hfs;
898 size_t linesize = 0; /* TODO line pool */
899 char *linebuf = NULL, *colon;
900 char const *hfield;
901 NYD_ENTER;
903 /* There are (spam) messages which have header bytes which are many KB when
904 * joined, so resize a single heap storage until we are done if we shall
905 * collect a field that may have multiple bodies; only otherwise use the
906 * string dope directly */
907 memset(&hfs, 0, sizeof hfs);
909 if ((ibuf = setinput(&mb, mp, NEED_HEADER)) == NULL)
910 goto jleave;
911 if ((lc = mp->m_lines - 1) < 0)
912 goto jleave;
914 if ((mp->m_flag & MNOFROM) == 0 &&
915 readline_restart(ibuf, &linebuf, &linesize, 0) < 0)
916 goto jleave;
917 while (lc > 0) {
918 if ((lc = gethfield(ibuf, &linebuf, &linesize, lc, &colon)) < 0)
919 break;
920 if ((hfield = thisfield(linebuf, field)) != NULL && *hfield != '\0') {
921 if (mult)
922 n_str_add_buf(&hfs, hfield, strlen(hfield));
923 else {
924 hfs.s = savestr(hfield);
925 break;
930 jleave:
931 if (linebuf != NULL)
932 free(linebuf);
933 if (mult && hfs.s != NULL) {
934 colon = savestrbuf(hfs.s, hfs.l);
935 free(hfs.s);
936 hfs.s = colon;
938 NYD_LEAVE;
939 return hfs.s;
942 FL char const *
943 thisfield(char const *linebuf, char const *field)
945 char const *rv = NULL;
946 NYD2_ENTER;
948 while (lowerconv(*linebuf) == lowerconv(*field)) {
949 ++linebuf;
950 ++field;
952 if (*field != '\0')
953 goto jleave;
955 while (blankchar(*linebuf))
956 ++linebuf;
957 if (*linebuf++ != ':')
958 goto jleave;
960 while (blankchar(*linebuf)) /* TODO header parser.. strip trailing WS?!? */
961 ++linebuf;
962 rv = linebuf;
963 jleave:
964 NYD2_LEAVE;
965 return rv;
968 FL char *
969 nameof(struct message *mp, int reptype)
971 char *cp, *cp2;
972 NYD_ENTER;
974 cp = skin(name1(mp, reptype));
975 if (reptype != 0 || charcount(cp, '!') < 2)
976 goto jleave;
977 cp2 = strrchr(cp, '!');
978 --cp2;
979 while (cp2 > cp && *cp2 != '!')
980 --cp2;
981 if (*cp2 == '!')
982 cp = cp2 + 1;
983 jleave:
984 NYD_LEAVE;
985 return cp;
988 FL char const *
989 skip_comment(char const *cp)
991 size_t nesting;
992 NYD_ENTER;
994 for (nesting = 1; nesting > 0 && *cp; ++cp) {
995 switch (*cp) {
996 case '\\':
997 if (cp[1])
998 ++cp;
999 break;
1000 case '(':
1001 ++nesting;
1002 break;
1003 case ')':
1004 --nesting;
1005 break;
1008 NYD_LEAVE;
1009 return cp;
1012 FL char const *
1013 routeaddr(char const *name)
1015 char const *np, *rp = NULL;
1016 NYD_ENTER;
1018 for (np = name; *np; np++) {
1019 switch (*np) {
1020 case '(':
1021 np = skip_comment(np + 1) - 1;
1022 break;
1023 case '"':
1024 while (*np) {
1025 if (*++np == '"')
1026 break;
1027 if (*np == '\\' && np[1])
1028 np++;
1030 break;
1031 case '<':
1032 rp = np;
1033 break;
1034 case '>':
1035 goto jleave;
1038 rp = NULL;
1039 jleave:
1040 NYD_LEAVE;
1041 return rp;
1044 FL enum expand_addr_flags
1045 expandaddr_to_eaf(void)
1047 struct eafdesc {
1048 char const *eafd_name;
1049 bool_t eafd_is_target;
1050 ui8_t eafd_andoff;
1051 ui8_t eafd_or;
1052 } const eafa[] = {
1053 {"restrict", FAL0, EAF_TARGET_MASK, EAF_RESTRICT | EAF_RESTRICT_TARGETS},
1054 {"fail", FAL0, EAF_NONE, EAF_FAIL},
1055 {"all", TRU1, EAF_NONE, EAF_TARGET_MASK},
1056 {"file", TRU1, EAF_NONE, EAF_FILE},
1057 {"pipe", TRU1, EAF_NONE, EAF_PIPE},
1058 {"name", TRU1, EAF_NONE, EAF_NAME},
1059 {"addr", TRU1, EAF_NONE, EAF_ADDR}
1060 }, *eafp;
1062 char *buf;
1063 enum expand_addr_flags rv;
1064 char const *cp;
1065 NYD2_ENTER;
1067 if ((cp = ok_vlook(expandaddr)) == NULL)
1068 rv = EAF_RESTRICT_TARGETS;
1069 else if (*cp == '\0')
1070 rv = EAF_TARGET_MASK;
1071 else {
1072 rv = EAF_TARGET_MASK;
1074 for (buf = savestr(cp); (cp = n_strsep(&buf, ',', TRU1)) != NULL;) {
1075 bool_t minus;
1077 if ((minus = (*cp == '-')) || *cp == '+')
1078 ++cp;
1079 for (eafp = eafa;; ++eafp) {
1080 if (eafp == eafa + NELEM(eafa)) {
1081 if (options & OPT_D_V)
1082 n_err(_("Unknown *expandaddr* value: \"%s\"\n"), cp);
1083 break;
1084 } else if (!asccasecmp(cp, eafp->eafd_name)) {
1085 if (!minus) {
1086 rv &= ~eafp->eafd_andoff;
1087 rv |= eafp->eafd_or;
1088 } else {
1089 if (eafp->eafd_is_target)
1090 rv &= ~eafp->eafd_or;
1091 else if (options & OPT_D_V)
1092 n_err(_("\"-\" prefix invalid for *expandaddr* value: "
1093 "\"%s\"\n"), --cp);
1095 break;
1096 } else if (!asccasecmp(cp, "noalias")) { /* TODO v15 OBSOLETE */
1097 OBSOLETE(_("*expandaddr*: \"noalias\" is henceforth \"-name\""));
1098 rv &= ~EAF_NAME;
1099 break;
1104 if ((rv & EAF_RESTRICT) && (options & (OPT_INTERACTIVE | OPT_TILDE_FLAG)))
1105 rv |= EAF_TARGET_MASK;
1106 else if (options & OPT_D_V) {
1107 if (!(rv & EAF_TARGET_MASK))
1108 n_err(_("*expandaddr* doesn't allow any addresses\n"));
1109 else if ((rv & EAF_FAIL) && (rv & EAF_TARGET_MASK) == EAF_TARGET_MASK)
1110 n_err(_("*expandaddr* with \"fail\" but no restrictions\n"));
1113 NYD2_LEAVE;
1114 return rv;
1117 FL si8_t
1118 is_addr_invalid(struct name *np, enum expand_addr_check_mode eacm)
1120 char cbuf[sizeof "'\\U12340'"];
1121 enum expand_addr_flags eaf;
1122 char const *cs;
1123 int f;
1124 si8_t rv;
1125 NYD_ENTER;
1127 f = np->n_flags;
1129 if ((rv = ((f & NAME_ADDRSPEC_INVALID) != 0))) {
1130 if ((eacm & EACM_NOLOG) || (f & NAME_ADDRSPEC_ERR_EMPTY)) {
1132 } else {
1133 ui32_t c;
1134 char const *fmt = "'\\x%02X'";
1135 bool_t ok8bit = TRU1;
1137 if (f & NAME_ADDRSPEC_ERR_IDNA) {
1138 cs = _("Invalid domain name: \"%s\", character %s\n");
1139 fmt = "'\\U%04X'";
1140 ok8bit = FAL0;
1141 } else if (f & NAME_ADDRSPEC_ERR_ATSEQ)
1142 cs = _("\"%s\" contains invalid %s sequence\n");
1143 else
1144 cs = _("\"%s\" contains invalid character %s\n");
1146 c = NAME_ADDRSPEC_ERR_GETWC(f);
1147 snprintf(cbuf, sizeof cbuf,
1148 (ok8bit && c >= 040 && c <= 0177 ? "'%c'" : fmt), c);
1149 goto jprint;
1151 goto jleave;
1154 /* *expandaddr* stuff */
1155 if (!(rv = ((eacm & EACM_MODE_MASK) != EACM_NONE)))
1156 goto jleave;
1158 eaf = expandaddr_to_eaf();
1160 if ((eacm & EACM_STRICT) && (f & NAME_ADDRSPEC_ISFILEORPIPE)) {
1161 if (eaf & EAF_FAIL)
1162 rv = -rv;
1163 cs = _("\"%s\"%s: file or pipe addressees not allowed here\n");
1164 if (eacm & EACM_NOLOG)
1165 goto jleave;
1166 else
1167 goto j0print;
1170 eaf |= (eacm & EAF_TARGET_MASK);
1171 if (eacm & EACM_NONAME)
1172 eaf &= ~EAF_NAME;
1174 if (eaf == EAF_NONE) {
1175 rv = FAL0;
1176 goto jleave;
1178 if (eaf & EAF_FAIL)
1179 rv = -rv;
1181 if (!(eaf & EAF_FILE) && (f & NAME_ADDRSPEC_ISFILE)) {
1182 cs = _("\"%s\"%s: *expandaddr* doesn't allow file target\n");
1183 if (eacm & EACM_NOLOG)
1184 goto jleave;
1185 } else if (!(eaf & EAF_PIPE) && (f & NAME_ADDRSPEC_ISPIPE)) {
1186 cs = _("\"%s\"%s: *expandaddr* doesn't allow command pipe target\n");
1187 if (eacm & EACM_NOLOG)
1188 goto jleave;
1189 } else if (!(eaf & EAF_NAME) && (f & NAME_ADDRSPEC_ISNAME)) {
1190 cs = _("\"%s\"%s: *expandaddr* doesn't allow user name target\n");
1191 if (eacm & EACM_NOLOG)
1192 goto jleave;
1193 } else if (!(eaf & EAF_ADDR) && (f & NAME_ADDRSPEC_ISADDR)) {
1194 cs = _("\"%s\"%s: *expandaddr* doesn't allow mail address target\n");
1195 if (eacm & EACM_NOLOG)
1196 goto jleave;
1197 } else {
1198 rv = FAL0;
1199 goto jleave;
1202 j0print:
1203 cbuf[0] = '\0';
1204 jprint:
1205 n_err(cs, np->n_name, cbuf);
1206 jleave:
1207 NYD_LEAVE;
1208 return rv;
1211 FL char *
1212 skin(char const *name)
1214 struct addrguts ag;
1215 char *ret = NULL;
1216 NYD_ENTER;
1218 if (name != NULL) {
1219 addrspec_with_guts(1, name, &ag);
1220 ret = ag.ag_skinned;
1221 if (!(ag.ag_n_flags & NAME_NAME_SALLOC))
1222 ret = savestrbuf(ret, ag.ag_slen);
1224 NYD_LEAVE;
1225 return ret;
1228 /* TODO addrspec_with_guts: RFC 5322
1229 * TODO addrspec_with_guts: trim whitespace ETC. ETC. ETC.!!! */
1230 FL int
1231 addrspec_with_guts(int doskin, char const *name, struct addrguts *agp)
1233 char const *cp;
1234 char *cp2, *bufend, *nbuf, c, gotlt, gotaddr, lastsp;
1235 int rv = 1;
1236 NYD_ENTER;
1238 memset(agp, 0, sizeof *agp);
1240 if ((agp->ag_input = name) == NULL || (agp->ag_ilen = strlen(name)) == 0) {
1241 agp->ag_skinned = UNCONST(""); /* ok: NAME_SALLOC is not set */
1242 agp->ag_slen = 0;
1243 agp->ag_n_flags |= NAME_ADDRSPEC_CHECKED;
1244 NAME_ADDRSPEC_ERR_SET(agp->ag_n_flags, NAME_ADDRSPEC_ERR_EMPTY, 0);
1245 goto jleave;
1248 if (!doskin || !anyof(name, "(< ")) {
1249 /*agp->ag_iaddr_start = 0;*/
1250 agp->ag_iaddr_aend = agp->ag_ilen;
1251 agp->ag_skinned = UNCONST(name); /* (NAME_SALLOC not set) */
1252 agp->ag_slen = agp->ag_ilen;
1253 agp->ag_n_flags = NAME_SKINNED;
1254 goto jcheck;
1257 /* Something makes us think we have to perform the skin operation */
1258 nbuf = ac_alloc(agp->ag_ilen + 1);
1259 /*agp->ag_iaddr_start = 0;*/
1260 cp2 = bufend = nbuf;
1261 gotlt = gotaddr = lastsp = 0;
1263 for (cp = name++; (c = *cp++) != '\0'; ) {
1264 switch (c) {
1265 case '(':
1266 cp = skip_comment(cp);
1267 lastsp = 0;
1268 break;
1269 case '"':
1270 /* Start of a "quoted-string".
1271 * Copy it in its entirety */
1272 /* XXX RFC: quotes are "semantically invisible"
1273 * XXX But it was explicitly added (Changelog.Heirloom,
1274 * XXX [9.23] released 11/15/00, "Do not remove quotes
1275 * XXX when skinning names"? No more info.. */
1276 *cp2++ = c;
1277 while ((c = *cp) != '\0') { /* TODO improve */
1278 cp++;
1279 if (c == '"') {
1280 *cp2++ = c;
1281 break;
1283 if (c != '\\')
1284 *cp2++ = c;
1285 else if ((c = *cp) != '\0') {
1286 *cp2++ = c;
1287 cp++;
1290 lastsp = 0;
1291 break;
1292 case ' ':
1293 case '\t':
1294 if (gotaddr == 1) {
1295 gotaddr = 2;
1296 agp->ag_iaddr_aend = PTR2SIZE(cp - name);
1298 if (cp[0] == 'a' && cp[1] == 't' && blankchar(cp[2]))
1299 cp += 3, *cp2++ = '@';
1300 else if (cp[0] == '@' && blankchar(cp[1]))
1301 cp += 2, *cp2++ = '@';
1302 else
1303 lastsp = 1;
1304 break;
1305 case '<':
1306 agp->ag_iaddr_start = PTR2SIZE(cp - (name - 1));
1307 cp2 = bufend;
1308 gotlt = gotaddr = 1;
1309 lastsp = 0;
1310 break;
1311 case '>':
1312 if (gotlt) {
1313 /* (_addrspec_check() verifies these later!) */
1314 agp->ag_iaddr_aend = PTR2SIZE(cp - name);
1315 gotlt = 0;
1316 while ((c = *cp) != '\0' && c != ',') {
1317 cp++;
1318 if (c == '(')
1319 cp = skip_comment(cp);
1320 else if (c == '"')
1321 while ((c = *cp) != '\0') {
1322 cp++;
1323 if (c == '"')
1324 break;
1325 if (c == '\\' && *cp != '\0')
1326 ++cp;
1329 lastsp = 0;
1330 break;
1332 /* FALLTRHOUGH */
1333 default:
1334 if (lastsp) {
1335 lastsp = 0;
1336 if (gotaddr)
1337 *cp2++ = ' ';
1339 *cp2++ = c;
1340 if (c == ',') {
1341 if (!gotlt) {
1342 *cp2++ = ' ';
1343 for (; blankchar(*cp); ++cp)
1345 lastsp = 0;
1346 bufend = cp2;
1348 } else if (!gotaddr) {
1349 gotaddr = 1;
1350 agp->ag_iaddr_start = PTR2SIZE(cp - name);
1354 agp->ag_slen = PTR2SIZE(cp2 - nbuf);
1355 if (agp->ag_iaddr_aend == 0)
1356 agp->ag_iaddr_aend = agp->ag_ilen;
1358 agp->ag_skinned = savestrbuf(nbuf, agp->ag_slen);
1359 ac_free(nbuf);
1360 agp->ag_n_flags = NAME_NAME_SALLOC | NAME_SKINNED;
1361 jcheck:
1362 rv = _addrspec_check(doskin, agp);
1363 jleave:
1364 NYD_LEAVE;
1365 return rv;
1368 FL char *
1369 realname(char const *name)
1371 char const *cp, *cq, *cstart = NULL, *cend = NULL;
1372 char *rname, *rp;
1373 struct str in, out;
1374 int quoted, good, nogood;
1375 NYD_ENTER;
1377 if ((cp = UNCONST(name)) == NULL)
1378 goto jleave;
1379 for (; *cp != '\0'; ++cp) {
1380 switch (*cp) {
1381 case '(':
1382 if (cstart != NULL) {
1383 /* More than one comment in address, doesn't make sense to display
1384 * it without context. Return the entire field */
1385 cp = mime_fromaddr(name);
1386 goto jleave;
1388 cstart = cp++;
1389 cp = skip_comment(cp);
1390 cend = cp--;
1391 if (cend <= cstart)
1392 cend = cstart = NULL;
1393 break;
1394 case '"':
1395 while (*cp) {
1396 if (*++cp == '"')
1397 break;
1398 if (*cp == '\\' && cp[1])
1399 ++cp;
1401 break;
1402 case '<':
1403 if (cp > name) {
1404 cstart = name;
1405 cend = cp;
1407 break;
1408 case ',':
1409 /* More than one address. Just use the first one */
1410 goto jbrk;
1414 jbrk:
1415 if (cstart == NULL) {
1416 if (*name == '<') {
1417 /* If name contains only a route-addr, the surrounding angle brackets
1418 * don't serve any useful purpose when displaying, so remove */
1419 cp = prstr(skin(name));
1420 } else
1421 cp = mime_fromaddr(name);
1422 goto jleave;
1425 /* Strip quotes. Note that quotes that appear within a MIME encoded word are
1426 * not stripped. The idea is to strip only syntactical relevant things (but
1427 * this is not necessarily the most sensible way in practice) */
1428 rp = rname = ac_alloc(PTR2SIZE(cend - cstart +1));
1429 quoted = 0;
1430 for (cp = cstart; cp < cend; ++cp) {
1431 if (*cp == '(' && !quoted) {
1432 cq = skip_comment(++cp);
1433 if (PTRCMP(--cq, >, cend))
1434 cq = cend;
1435 while (cp < cq) {
1436 if (*cp == '\\' && PTRCMP(cp + 1, <, cq))
1437 ++cp;
1438 *rp++ = *cp++;
1440 } else if (*cp == '\\' && PTRCMP(cp + 1, <, cend))
1441 *rp++ = *++cp;
1442 else if (*cp == '"') {
1443 quoted = !quoted;
1444 continue;
1445 } else
1446 *rp++ = *cp;
1448 *rp = '\0';
1449 in.s = rname;
1450 in.l = rp - rname;
1451 mime_fromhdr(&in, &out, TD_ISPR | TD_ICONV);
1452 ac_free(rname);
1453 rname = savestr(out.s);
1454 free(out.s);
1456 while (blankchar(*rname))
1457 ++rname;
1458 for (rp = rname; *rp != '\0'; ++rp)
1460 while (PTRCMP(--rp, >=, rname) && blankchar(*rp))
1461 *rp = '\0';
1462 if (rp == rname) {
1463 cp = mime_fromaddr(name);
1464 goto jleave;
1467 /* mime_fromhdr() has converted all nonprintable characters to question
1468 * marks now. These and blanks are considered uninteresting; if the
1469 * displayed part of the real name contains more than 25% of them, it is
1470 * probably better to display the plain email address instead */
1471 good = 0;
1472 nogood = 0;
1473 for (rp = rname; *rp != '\0' && PTRCMP(rp, <, rname + 20); ++rp)
1474 if (*rp == '?' || blankchar(*rp))
1475 ++nogood;
1476 else
1477 ++good;
1478 cp = (good * 3 < nogood) ? prstr(skin(name)) : rname;
1479 jleave:
1480 NYD_LEAVE;
1481 return UNCONST(cp);
1484 FL char *
1485 name1(struct message *mp, int reptype)
1487 char *namebuf, *cp, *cp2, *linebuf = NULL /* TODO line pool */;
1488 size_t namesize, linesize = 0;
1489 FILE *ibuf;
1490 int f1st = 1;
1491 NYD_ENTER;
1493 if ((cp = hfield1("from", mp)) != NULL && *cp != '\0')
1494 goto jleave;
1495 if (reptype == 0 && (cp = hfield1("sender", mp)) != NULL && *cp != '\0')
1496 goto jleave;
1498 namebuf = smalloc(namesize = 1);
1499 namebuf[0] = 0;
1500 if (mp->m_flag & MNOFROM)
1501 goto jout;
1502 if ((ibuf = setinput(&mb, mp, NEED_HEADER)) == NULL)
1503 goto jout;
1504 if (readline_restart(ibuf, &linebuf, &linesize, 0) < 0)
1505 goto jout;
1507 jnewname:
1508 if (namesize <= linesize)
1509 namebuf = srealloc(namebuf, namesize = linesize +1);
1510 for (cp = linebuf; *cp != '\0' && *cp != ' '; ++cp)
1512 for (; blankchar(*cp); ++cp)
1514 for (cp2 = namebuf + strlen(namebuf);
1515 *cp && !blankchar(*cp) && PTRCMP(cp2, <, namebuf + namesize -1);)
1516 *cp2++ = *cp++;
1517 *cp2 = '\0';
1519 if (readline_restart(ibuf, &linebuf, &linesize, 0) < 0)
1520 goto jout;
1521 if ((cp = strchr(linebuf, 'F')) == NULL)
1522 goto jout;
1523 if (strncmp(cp, "From", 4))
1524 goto jout;
1525 if (namesize <= linesize)
1526 namebuf = srealloc(namebuf, namesize = linesize + 1);
1528 while ((cp = strchr(cp, 'r')) != NULL) {
1529 if (!strncmp(cp, "remote", 6)) {
1530 if ((cp = strchr(cp, 'f')) == NULL)
1531 break;
1532 if (strncmp(cp, "from", 4) != 0)
1533 break;
1534 if ((cp = strchr(cp, ' ')) == NULL)
1535 break;
1536 cp++;
1537 if (f1st) {
1538 strncpy(namebuf, cp, namesize);
1539 f1st = 0;
1540 } else {
1541 cp2 = strrchr(namebuf, '!') + 1;
1542 strncpy(cp2, cp, PTR2SIZE(namebuf + namesize - cp2));
1544 namebuf[namesize - 2] = '!';
1545 namebuf[namesize - 1] = '\0';
1546 goto jnewname;
1548 cp++;
1550 jout:
1551 if (*namebuf != '\0' || ((cp = hfield1("return-path", mp))) == NULL ||
1552 *cp == '\0')
1553 cp = savestr(namebuf);
1555 if (linebuf != NULL)
1556 free(linebuf);
1557 free(namebuf);
1558 jleave:
1559 NYD_LEAVE;
1560 return cp;
1563 FL char *
1564 subject_re_trim(char *s)
1566 struct {
1567 ui8_t len;
1568 char dat[7];
1569 } const *pp, ignored[] = { /* Update *reply-strings* manual upon change! */
1570 { 3, "re:" },
1571 { 3, "aw:" }, { 5, "antw:" }, /* de */
1572 { 0, "" }
1575 bool_t any = FAL0;
1576 char *orig_s = s, *re_st = NULL, *re_st_x;
1577 size_t re_l = 0 /* pacify CC */;
1578 NYD_ENTER;
1580 if ((re_st_x = ok_vlook(reply_strings)) != NULL &&
1581 (re_l = strlen(re_st_x)) > 0) {
1582 re_st = ac_alloc(++re_l * 2);
1583 memcpy(re_st, re_st_x, re_l);
1586 jouter:
1587 while (*s != '\0') {
1588 while (spacechar(*s))
1589 ++s;
1591 for (pp = ignored; pp->len > 0; ++pp)
1592 if (is_asccaseprefix(s, pp->dat)) {
1593 s += pp->len;
1594 any = TRU1;
1595 goto jouter;
1598 if (re_st != NULL) {
1599 char *cp;
1601 memcpy(re_st_x = re_st + re_l, re_st, re_l);
1602 while ((cp = n_strsep(&re_st_x, ',', TRU1)) != NULL)
1603 if (is_asccaseprefix(s, cp)) {
1604 s += strlen(cp);
1605 any = TRU1;
1606 goto jouter;
1609 break;
1612 if (re_st != NULL)
1613 ac_free(re_st);
1614 NYD_LEAVE;
1615 return any ? s : orig_s;
1618 FL int
1619 msgidcmp(char const *s1, char const *s2)
1621 int q1 = 0, q2 = 0, c1, c2;
1622 NYD_ENTER;
1624 while(*s1 == '<')
1625 ++s1;
1626 while(*s2 == '<')
1627 ++s2;
1629 do {
1630 c1 = msgidnextc(&s1, &q1);
1631 c2 = msgidnextc(&s2, &q2);
1632 if (c1 != c2)
1633 break;
1634 } while (c1 && c2);
1635 NYD_LEAVE;
1636 return c1 - c2;
1639 FL int
1640 is_ign(char const *field, size_t fieldlen, struct ignoretab igta[2])
1642 char *realfld;
1643 int rv;
1644 NYD_ENTER;
1646 rv = 0;
1647 if (igta == NULL)
1648 goto jleave;
1649 rv = 1;
1650 if (igta == allignore)
1651 goto jleave;
1653 /* Lowercase it so that "Status" and "status" will hash to the same place */
1654 realfld = ac_alloc(fieldlen +1);
1655 i_strcpy(realfld, field, fieldlen +1);
1656 if (igta[1].i_count > 0)
1657 rv = !member(realfld, igta + 1);
1658 else
1659 rv = member(realfld, igta);
1660 ac_free(realfld);
1661 jleave:
1662 NYD_LEAVE;
1663 return rv;
1666 FL int
1667 member(char const *realfield, struct ignoretab *table)
1669 struct ignored *igp;
1670 int rv = 0;
1671 NYD_ENTER;
1673 for (igp = table->i_head[hash(realfield)]; igp != 0; igp = igp->i_link)
1674 if (*igp->i_field == *realfield && !strcmp(igp->i_field, realfield)) {
1675 rv = 1;
1676 break;
1678 NYD_LEAVE;
1679 return rv;
1682 FL char const *
1683 fakefrom(struct message *mp)
1685 char const *name;
1686 NYD_ENTER;
1688 if (((name = skin(hfield1("return-path", mp))) == NULL || *name == '\0' ) &&
1689 ((name = skin(hfield1("from", mp))) == NULL || *name == '\0'))
1690 /* XXX MAILER-DAEMON is what an old MBOX manual page says.
1691 * RFC 4155 however requires a RFC 5322 (2822) conforming
1692 * "addr-spec", but we simply can't provide that */
1693 name = "MAILER-DAEMON";
1694 NYD_LEAVE;
1695 return name;
1698 FL char const *
1699 fakedate(time_t t)
1701 char *cp, *cq;
1702 NYD_ENTER;
1704 cp = ctime(&t);
1705 for (cq = cp; *cq != '\0' && *cq != '\n'; ++cq)
1707 *cq = '\0';
1708 cp = savestr(cp);
1709 NYD_LEAVE;
1710 return cp;
1713 #if defined HAVE_IMAP_SEARCH || defined HAVE_IMAP
1714 FL time_t
1715 unixtime(char const *fromline)
1717 char const *fp;
1718 char *xp;
1719 time_t t;
1720 int i, year, month, day, hour, minute, second, tzdiff;
1721 struct tm *tmptr;
1722 NYD2_ENTER;
1724 for (fp = fromline; *fp != '\0' && *fp != '\n'; ++fp)
1726 fp -= 24;
1727 if (PTR2SIZE(fp - fromline) < 7)
1728 goto jinvalid;
1729 if (fp[3] != ' ')
1730 goto jinvalid;
1731 for (i = 0;;) {
1732 if (!strncmp(fp + 4, month_names[i], 3))
1733 break;
1734 if (month_names[++i][0] == '\0')
1735 goto jinvalid;
1737 month = i + 1;
1738 if (fp[7] != ' ')
1739 goto jinvalid;
1740 day = strtol(fp + 8, &xp, 10);
1741 if (*xp != ' ' || xp != fp + 10)
1742 goto jinvalid;
1743 hour = strtol(fp + 11, &xp, 10);
1744 if (*xp != ':' || xp != fp + 13)
1745 goto jinvalid;
1746 minute = strtol(fp + 14, &xp, 10);
1747 if (*xp != ':' || xp != fp + 16)
1748 goto jinvalid;
1749 second = strtol(fp + 17, &xp, 10);
1750 if (*xp != ' ' || xp != fp + 19)
1751 goto jinvalid;
1752 year = strtol(fp + 20, &xp, 10);
1753 if (xp != fp + 24)
1754 goto jinvalid;
1755 if ((t = combinetime(year, month, day, hour, minute, second)) == (time_t)-1)
1756 goto jinvalid;
1757 tzdiff = t - mktime(gmtime(&t));
1758 tmptr = localtime(&t);
1759 if (tmptr->tm_isdst > 0)
1760 tzdiff += 3600;
1761 t -= tzdiff;
1762 jleave:
1763 NYD2_LEAVE;
1764 return t;
1765 jinvalid:
1766 t = n_time_epoch();
1767 goto jleave;
1769 #endif /* HAVE_IMAP_SEARCH || defined HAVE_IMAP */
1771 FL time_t
1772 rfctime(char const *date)
1774 char const *cp = date;
1775 char *x;
1776 time_t t;
1777 int i, year, month, day, hour, minute, second;
1778 NYD2_ENTER;
1780 if ((cp = nexttoken(cp)) == NULL)
1781 goto jinvalid;
1782 if (alphachar(cp[0]) && alphachar(cp[1]) && alphachar(cp[2]) &&
1783 cp[3] == ',') {
1784 if ((cp = nexttoken(&cp[4])) == NULL)
1785 goto jinvalid;
1787 day = strtol(cp, &x, 10); /* XXX strtol */
1788 if ((cp = nexttoken(x)) == NULL)
1789 goto jinvalid;
1790 for (i = 0;;) {
1791 if (!strncmp(cp, month_names[i], 3))
1792 break;
1793 if (month_names[++i][0] == '\0')
1794 goto jinvalid;
1796 month = i + 1;
1797 if ((cp = nexttoken(&cp[3])) == NULL)
1798 goto jinvalid;
1799 /* RFC 5322, 4.3:
1800 * Where a two or three digit year occurs in a date, the year is to be
1801 * interpreted as follows: If a two digit year is encountered whose
1802 * value is between 00 and 49, the year is interpreted by adding 2000,
1803 * ending up with a value between 2000 and 2049. If a two digit year
1804 * is encountered with a value between 50 and 99, or any three digit
1805 * year is encountered, the year is interpreted by adding 1900 */
1806 year = strtol(cp, &x, 10); /* XXX strtol */
1807 i = (int)PTR2SIZE(x - cp);
1808 if (i == 2 && year >= 0 && year <= 49)
1809 year += 2000;
1810 else if (i == 3 || (i == 2 && year >= 50 && year <= 99))
1811 year += 1900;
1812 if ((cp = nexttoken(x)) == NULL)
1813 goto jinvalid;
1814 hour = strtol(cp, &x, 10); /* XXX strtol */
1815 if (*x != ':')
1816 goto jinvalid;
1817 cp = &x[1];
1818 minute = strtol(cp, &x, 10);
1819 if (*x == ':') {
1820 cp = x + 1;
1821 second = strtol(cp, &x, 10);
1822 } else
1823 second = 0;
1825 if ((t = combinetime(year, month, day, hour, minute, second)) == (time_t)-1)
1826 goto jinvalid;
1827 if ((cp = nexttoken(x)) != NULL) {
1828 char buf[3];
1829 int sign = 1;
1831 switch (*cp) {
1832 case '+':
1833 sign = -1;
1834 /* FALLTHRU */
1835 case '-':
1836 ++cp;
1837 break;
1839 if (digitchar(cp[0]) && digitchar(cp[1]) && digitchar(cp[2]) &&
1840 digitchar(cp[3])) {
1841 long tadj;
1842 buf[2] = '\0';
1843 buf[0] = cp[0];
1844 buf[1] = cp[1];
1845 tadj = strtol(buf, NULL, 10) * 3600;/*XXX strtrol*/
1846 buf[0] = cp[2];
1847 buf[1] = cp[3];
1848 tadj += strtol(buf, NULL, 10) * 60; /* XXX strtol*/
1849 if (sign < 0)
1850 tadj = -tadj;
1851 t += tadj;
1853 /* TODO WE DO NOT YET PARSE (OBSOLETE) ZONE NAMES
1854 * TODO once again, Christos Zoulas and NetBSD Mail have done
1855 * TODO a really good job already, but using strptime(3), which
1856 * TODO is not portable. Nonetheless, WE must improve, not
1857 * TODO at last because we simply ignore obsolete timezones!!
1858 * TODO See RFC 5322, 4.3! */
1860 jleave:
1861 NYD2_LEAVE;
1862 return t;
1863 jinvalid:
1864 t = 0;
1865 goto jleave;
1868 #define is_leapyear(Y) ((((Y) % 100 ? (Y) : (Y) / 100) & 3) == 0)
1870 FL time_t
1871 combinetime(int year, int month, int day, int hour, int minute, int second)
1873 time_t t;
1874 NYD2_ENTER;
1876 if (second < 0 || minute < 0 || hour < 0 || day < 1) {
1877 t = (time_t)-1;
1878 goto jleave;
1881 t = second + minute * 60 + hour * 3600 + (day - 1) * 86400;
1882 if (month > 1)
1883 t += 86400 * 31;
1884 if (month > 2)
1885 t += 86400 * (is_leapyear(year) ? 29 : 28);
1886 if (month > 3)
1887 t += 86400 * 31;
1888 if (month > 4)
1889 t += 86400 * 30;
1890 if (month > 5)
1891 t += 86400 * 31;
1892 if (month > 6)
1893 t += 86400 * 30;
1894 if (month > 7)
1895 t += 86400 * 31;
1896 if (month > 8)
1897 t += 86400 * 31;
1898 if (month > 9)
1899 t += 86400 * 30;
1900 if (month > 10)
1901 t += 86400 * 31;
1902 if (month > 11)
1903 t += 86400 * 30;
1904 year -= 1900;
1905 t += (year - 70) * 31536000 + ((year - 69) / 4) * 86400 -
1906 ((year - 1) / 100) * 86400 + ((year + 299) / 400) * 86400;
1907 jleave:
1908 NYD2_LEAVE;
1909 return t;
1912 FL void
1913 substdate(struct message *m)
1915 char const *cp;
1916 NYD_ENTER;
1918 /* Determine the date to print in faked 'From ' lines. This is traditionally
1919 * the date the message was written to the mail file. Try to determine this
1920 * using RFC message header fields, or fall back to current time */
1921 if ((cp = hfield1("received", m)) != NULL) {
1922 while ((cp = nexttoken(cp)) != NULL && *cp != ';') {
1924 ++cp;
1925 while (alnumchar(*cp));
1927 if (cp && *++cp)
1928 m->m_time = rfctime(cp);
1930 if (m->m_time == 0 || m->m_time > time_current.tc_time) {
1931 if ((cp = hfield1("date", m)) != NULL)
1932 m->m_time = rfctime(cp);
1934 if (m->m_time == 0 || m->m_time > time_current.tc_time)
1935 m->m_time = time_current.tc_time;
1936 NYD_LEAVE;
1939 FL struct name const *
1940 check_from_and_sender(struct name const *fromfield,
1941 struct name const *senderfield)
1943 struct name const *rv = NULL;
1944 NYD_ENTER;
1946 if (senderfield != NULL) {
1947 if (senderfield->n_flink != NULL) {
1948 n_err(_("The \"Sender:\" field may contain only one address\n"));
1949 goto jleave;
1951 rv = senderfield;
1954 if (fromfield != NULL) {
1955 if (fromfield->n_flink != NULL && senderfield == NULL) {
1956 n_err(_("A \"Sender:\" field is required with multiple "
1957 "addresses in \"From:\" field\n"));
1958 goto jleave;
1960 if (rv == NULL)
1961 rv = fromfield;
1964 if (rv == NULL)
1965 rv = (struct name*)0x1;
1966 jleave:
1967 NYD_LEAVE;
1968 return rv;
1971 #ifdef HAVE_OPENSSL
1972 FL char *
1973 getsender(struct message *mp)
1975 char *cp;
1976 struct name *np;
1977 NYD_ENTER;
1979 if ((cp = hfield1("from", mp)) == NULL ||
1980 (np = lextract(cp, GEXTRA | GSKIN)) == NULL)
1981 cp = NULL;
1982 else
1983 cp = (np->n_flink != NULL) ? skin(hfield1("sender", mp)) : np->n_name;
1984 NYD_LEAVE;
1985 return cp;
1987 #endif
1989 FL int
1990 grab_headers(struct header *hp, enum gfield gflags, int subjfirst)
1992 /* TODO grab_headers: again, check counts etc. against RFC;
1993 * TODO (now assumes check_from_and_sender() is called afterwards ++ */
1994 int errs;
1995 int volatile comma;
1996 NYD_ENTER;
1998 errs = 0;
1999 comma = (ok_blook(bsdcompat) || ok_blook(bsdmsgs)) ? 0 : GCOMMA;
2001 if (gflags & GTO)
2002 hp->h_to = grab_names("To: ", hp->h_to, comma, GTO | GFULL);
2003 if (subjfirst && (gflags & GSUBJECT))
2004 hp->h_subject = n_input_cp_addhist("Subject: ", hp->h_subject, TRU1);
2005 if (gflags & GCC)
2006 hp->h_cc = grab_names("Cc: ", hp->h_cc, comma, GCC | GFULL);
2007 if (gflags & GBCC)
2008 hp->h_bcc = grab_names("Bcc: ", hp->h_bcc, comma, GBCC | GFULL);
2010 if (gflags & GEXTRA) {
2011 if (hp->h_from == NULL)
2012 hp->h_from = lextract(myaddrs(hp), GEXTRA | GFULL | GFULLEXTRA);
2013 hp->h_from = grab_names("From: ", hp->h_from, comma,
2014 GEXTRA | GFULL | GFULLEXTRA);
2015 if (hp->h_replyto == NULL)
2016 hp->h_replyto = lextract(ok_vlook(replyto), GEXTRA | GFULL);
2017 hp->h_replyto = grab_names("Reply-To: ", hp->h_replyto, comma,
2018 GEXTRA | GFULL);
2019 if (hp->h_sender == NULL)
2020 hp->h_sender = extract(ok_vlook(sender), GEXTRA | GFULL);
2021 hp->h_sender = grab_names("Sender: ", hp->h_sender, comma,
2022 GEXTRA | GFULL);
2023 if (hp->h_organization == NULL)
2024 hp->h_organization = ok_vlook(ORGANIZATION);
2025 hp->h_organization = n_input_cp_addhist("Organization: ",
2026 hp->h_organization, TRU1);
2029 if (!subjfirst && (gflags & GSUBJECT))
2030 hp->h_subject = n_input_cp_addhist("Subject: ", hp->h_subject, TRU1);
2032 NYD_LEAVE;
2033 return errs;
2036 FL bool_t
2037 header_match(struct message *mp, struct search_expr const *sep)
2039 struct str in, out;
2040 FILE *ibuf;
2041 int lc;
2042 size_t linesize = 0; /* TODO line pool */
2043 char *linebuf = NULL, *colon;
2044 bool_t rv = FAL0;
2045 NYD_ENTER;
2047 if ((ibuf = setinput(&mb, mp, NEED_HEADER)) == NULL)
2048 goto jleave;
2049 if ((lc = mp->m_lines - 1) < 0)
2050 goto jleave;
2052 if ((mp->m_flag & MNOFROM) == 0 &&
2053 readline_restart(ibuf, &linebuf, &linesize, 0) < 0)
2054 goto jleave;
2055 while (lc > 0) {
2056 if (gethfield(ibuf, &linebuf, &linesize, lc, &colon) <= 0)
2057 break;
2058 if (blankchar(*++colon))
2059 ++colon;
2060 in.l = strlen(in.s = colon);
2061 mime_fromhdr(&in, &out, TD_ICONV);
2062 #ifdef HAVE_REGEX
2063 if (sep->ss_sexpr == NULL)
2064 rv = (regexec(&sep->ss_regex, out.s, 0,NULL, 0) != REG_NOMATCH);
2065 else
2066 #endif
2067 rv = substr(out.s, sep->ss_sexpr);
2068 free(out.s);
2069 if (rv)
2070 break;
2073 jleave:
2074 if (linebuf != NULL)
2075 free(linebuf);
2076 NYD_LEAVE;
2077 return rv;
2080 /* s-it-mode */