2 ** mf.c -- mail filter subroutines
4 ** This code is Copyright (c) 2002, by the authors of nmh. See the
5 ** COPYRIGHT file in the root directory of the nmh distribution for
6 ** complete copyright information.
17 static char *getcpy(char *);
18 static int isat(char *);
19 static int parse_address(void);
20 static int phrase(char *);
21 static int route_addr(char *);
22 static int local_part(char *);
23 static int domain(char *);
24 static int route(char *);
25 static int my_lex(char *);
35 ** causes compiles to blow up because the symbol _cleanup
36 ** is undefined where did this ever come from?
43 p = mh_xmalloc((size_t) (strlen(s) + 2));
49 #define CHKADR 0 /* undertermined address style */
50 #define UNIXDR 1 /* UNIX-style address */
51 #define ARPADR 2 /* ARPAnet-style address */
57 return (strncmp(p, " AT ", 4)
58 && strncmp(p, " At ", 4)
59 && strncmp(p, " aT ", 4)
60 && strncmp(p, " at ", 4) ? FALSE : TRUE);
66 ** getadrx() implements a partial 822-style address parser. The parser
67 ** is neither complete nor correct. It does however recognize nearly all
68 ** of the 822 address syntax. In addition it handles the majority of the
69 ** 733 syntax as well. Most problems arise from trying to accomodate both.
71 ** In terms of 822, the route-specification in
73 ** "<" [route] local-part "@" domain ">"
75 ** is parsed and returned unchanged. Multiple at-signs are compressed
76 ** via source-routing. Recursive groups are not allowed as per the
79 ** In terms of 733, " at " is recognized as equivalent to "@".
81 ** In terms of both the parser will not complain about missing hosts.
85 ** We should not allow addresses like
87 ** Marshall T. Rose <MRose@UCI>
89 ** but should insist on
91 ** "Marshall T. Rose" <MRose@UCI>
93 ** Unfortunately, a lot of mailers stupidly let people get away with this.
97 ** We should not allow addresses like
101 ** but should insist on
105 ** Unfortunately, a lot of mailers stupidly let people's UAs get away with
110 ** We should not allow addresses like
112 ** @UCI:MRose@UCI-750a
114 ** but should insist on
116 ** Marshall Rose <@UCI:MRose@UCI-750a>
118 ** Unfortunately, a lot of mailers stupidly do this.
142 static struct specials special[] = {
159 static int glevel = 0;
160 static int ingrp = 0;
161 static int last_lex = LX_END;
163 static char *dp = NULL;
164 static unsigned char *cp = NULL;
165 static unsigned char *ap = NULL;
166 static char *pers = NULL;
167 static char *mbox = NULL;
168 static char *host = NULL;
169 static char *path = NULL;
170 static char *grp = NULL;
171 static char *note = NULL;
172 static char err[BUFSIZ];
173 static char adr[BUFSIZ];
175 static struct adrx adrxs2;
182 register struct adrx *adrxp = &adrxs2;
196 pers = mbox = host = path = grp = note = NULL;
200 dp = cp = getcpy(addrs ? addrs : "");
202 } else if (cp == NULL) {
208 switch (parse_address()) {
220 default: /* catch trailing comments */
248 sprintf(adr, "%.*s", (int)(cp - ap), ap);
251 bp = adr + strlen(adr) - 1;
252 if (*bp == ',' || *bp == ';' || *bp == '\n')
261 adrxp->ingrp = ingrp;
263 adrxp->err = err[0] ? err : NULL;
276 switch (my_lex(buffer)) {
279 pers = getcpy(buffer);
284 strcpy(err, "extraneous semi-colon");
297 case LX_LBRK: /* sigh (2) */
300 case LX_AT: /* sigh (3) */
302 if (route_addr(buffer) == NOTOK)
304 return OK; /* why be choosy? */
307 sprintf(err, "illegal address construct (%s)", buffer);
311 switch (my_lex(buffer)) {
314 pers = add(buffer, add(" ", pers));
315 more_phrase: ; /* sigh (1) */
316 if (phrase(buffer) == NOTOK)
322 if (route_addr(buffer) == NOTOK)
324 if (last_lex == LX_RBRK)
326 sprintf(err, "missing right-bracket (%s)", buffer);
332 sprintf(err, "nested groups not allowed (%s)", pers);
335 grp = add(": ", pers);
341 switch (my_lex(buffer)) {
343 case LX_END: /* tsk, tsk */
352 return parse_address();
356 case LX_DOT: /* sigh (1) */
357 pers = add(".", pers);
361 sprintf(err, "no mailbox in address, only a phrase (%s%s)", pers, buffer);
372 mbox = add(buffer, pers);
374 if (route_addr(buffer) == NOTOK)
382 if (domain(buffer) == NOTOK)
388 strcpy(err, "extraneous semi-colon");
396 sprintf(err, "junk after local@domain (%s)", buffer);
400 case LX_SEMI: /* no host */
404 if (last_lex == LX_SEMI && glevel-- <= 0) {
405 strcpy(err, "extraneous semi-colon");
413 sprintf(err, "missing mailbox (%s)", buffer);
423 switch (my_lex(buffer)) {
426 pers = add(buffer, add(" ", pers));
436 route_addr(char *buffer)
438 register char *pp = cp;
440 if (my_lex(buffer) == LX_AT) {
441 if (route(buffer) == NOTOK)
447 if (local_part(buffer) == NOTOK)
452 return domain(buffer);
454 case LX_SEMI: /* if in group */
455 case LX_RBRK: /* no host */
461 sprintf(err, "no at-sign after local-part (%s)", buffer);
468 local_part(char *buffer)
473 switch (my_lex(buffer)) {
476 mbox = add(buffer, mbox);
480 sprintf(err, "no mailbox in local-part (%s)", buffer);
484 switch (my_lex(buffer)) {
486 mbox = add(buffer, mbox);
500 switch (my_lex(buffer)) {
503 host = add(buffer, host);
507 sprintf(err, "no sub-domain in domain-part of address (%s)", buffer);
511 switch (my_lex(buffer)) {
513 host = add(buffer, host);
516 case LX_AT: /* sigh (0) */
517 mbox = add(host, add("%", mbox));
535 switch (my_lex(buffer)) {
538 path = add(buffer, path);
542 sprintf(err, "no sub-domain in domain-part of address (%s)", buffer);
545 switch (my_lex(buffer)) {
547 path = add(buffer, path);
549 switch (my_lex(buffer)) {
554 path = add(buffer, path);
558 sprintf(err, "no at-sign found for next domain in route (%s)",
565 case LX_AT: /* XXX */
567 path = add(buffer, path);
571 path = add(buffer, path);
575 sprintf(err, "no colon found to terminate route (%s)", buffer);
585 /* buffer should be at least BUFSIZ bytes long */
587 register unsigned char c;
591 ** Add C to the buffer bp. After use of this macro *bp is guaranteed
592 ** to be within the buffer.
597 if ((bp - buffer) == (BUFSIZ-1)) \
598 goto my_lex_buffull; \
604 return (last_lex = LX_END);
612 return (last_lex = LX_END);
621 return (last_lex = LX_ERR);
624 if ((c = *cp++) == 0) {
626 return (last_lex = LX_ERR);
639 note = note ? add(buffer, add(" ", note)) : getcpy(buffer);
640 return my_lex(buffer);
651 return (last_lex = LX_ERR);
654 if ((c = *cp++) == 0) {
656 return (last_lex = LX_ERR);
664 return (last_lex = LX_QSTR);
674 return (last_lex = LX_ERR);
677 if ((c = *cp++) == 0) {
679 return (last_lex = LX_ERR);
687 return (last_lex = LX_DLIT);
693 for (i = 0; special[i].lx_chr != 0; i++)
694 if (c == special[i].lx_chr)
695 return (last_lex = special[i].lx_val);
698 return (last_lex = LX_ERR);
701 if ((c = *cp++) == 0)
703 for (i = 0; special[i].lx_chr != 0; i++)
704 if (c == special[i].lx_chr)
706 if (iscntrl(c) || isspace(c))
716 last_lex = !gotat || cp == NULL || strchr(cp, '<') != NULL
721 /* Out of buffer space. *bp is the last byte in the buffer */
723 return (last_lex = LX_ERR);
728 legal_person(char *p)
732 static char buffer[BUFSIZ];
736 for (cp = p; *cp; cp++)
737 for (i = 0; special[i].lx_chr; i++)
738 if (*cp == special[i].lx_chr) {
739 sprintf(buffer, "\"%s\"", p);