3 * mf.c -- mail filter subroutines
5 * This code is Copyright (c) 2002, by the authors of nmh. See the
6 * COPYRIGHT file in the root directory of the nmh distribution for
7 * complete copyright information.
18 static char *getcpy (char *);
19 static void compress (char *, unsigned char *);
20 static int isat (char *);
21 static int parse_address (void);
22 static int phrase (char *);
23 static int route_addr (char *);
24 static int local_part (char *);
25 static int domain (char *);
26 static int route (char *);
27 static int my_lex (char *);
36 /* causes compiles to blow up because the symbol _cleanup is undefined
37 where did this ever come from? */
43 p = mh_xmalloc ((size_t) (strlen (s) + 2));
52 return (strncmp (string, "From ", 5) == 0
53 || strncmp (string, ">From ", 6) == 0);
58 lequal (unsigned char *a, unsigned char *b)
64 char c1 = islower (*a) ? toupper (*a) : *a;
65 char c2 = islower (*b) ? toupper (*b) : *b;
75 * seekadrx() is tricky. We want to cover both UUCP-style and ARPA-style
76 * addresses, so for each list of addresses we see if we can find some
77 * character to give us a hint.
81 #define CHKADR 0 /* undertermined address style */
82 #define UNIXDR 1 /* UNIX-style address */
83 #define ARPADR 2 /* ARPAnet-style address */
86 static char *punctuators = ";<>.()[]";
87 static char *vp = NULL;
88 static char *tp = NULL;
90 static struct adrx adrxs1;
94 seekadrx (char *addrs)
96 static int state = CHKADR;
98 register struct adrx *adrxp;
101 for (state = UNIXDR, cp = addrs; *cp; cp++)
102 if (strchr(punctuators, *cp)) {
109 adrxp = uucpadrx (addrs);
114 adrxp = getadrx (addrs);
126 * uucpadrx() implements a partial UUCP-style address parser. It's based
127 * on the UUCP notion that addresses are separated by spaces or commas.
132 uucpadrx (char *addrs)
134 register unsigned char *cp, *wp, *xp, *yp;
136 register struct adrx *adrxp = &adrxs1;
139 vp = tp = getcpy (addrs);
140 compress (addrs, vp);
149 for (cp = tp; isspace (*cp); cp++)
157 if ((wp = strchr(cp, ',')) == NULL) {
158 if ((wp = strchr(cp, ' ')) != NULL) {
160 while (isspace (*xp))
162 if (*xp != 0 && isat (--xp)) {
164 while (isspace (*yp))
167 if ((zp = strchr(yp, ' ')) != NULL)
186 adrxp->text = getcpy (cp);
188 adrxp->host = adrxp->path = NULL;
189 if ((wp = strrchr(cp, '@')) != NULL) {
191 adrxp->host = *wp ? wp : NULL;
194 for (wp = cp + strlen (cp) - 4; wp >= cp; wp--)
197 adrxp->host = wp + 3;
200 adrxp->pers = adrxp->grp = adrxp->note = adrxp->err = NULL;
208 compress (char *fp, unsigned char *tp)
211 register unsigned char *cp;
213 for (c = ' ', cp = tp; (*tp = *fp++) != 0;)
221 if (c == ' ' && cp < tp)
229 return (strncmp (p, " AT ", 4)
230 && strncmp (p, " At ", 4)
231 && strncmp (p, " aT ", 4)
232 && strncmp (p, " at ", 4) ? FALSE : TRUE);
238 * getadrx() implements a partial 822-style address parser. The parser
239 * is neither complete nor correct. It does however recognize nearly all
240 * of the 822 address syntax. In addition it handles the majority of the
241 * 733 syntax as well. Most problems arise from trying to accomodate both.
243 * In terms of 822, the route-specification in
245 * "<" [route] local-part "@" domain ">"
247 * is parsed and returned unchanged. Multiple at-signs are compressed
248 * via source-routing. Recursive groups are not allowed as per the
251 * In terms of 733, " at " is recognized as equivalent to "@".
253 * In terms of both the parser will not complain about missing hosts.
257 * We should not allow addresses like
259 * Marshall T. Rose <MRose@UCI>
261 * but should insist on
263 * "Marshall T. Rose" <MRose@UCI>
265 * Unfortunately, a lot of mailers stupidly let people get away with this.
269 * We should not allow addresses like
273 * but should insist on
277 * Unfortunately, a lot of mailers stupidly let people's UAs get away with
282 * We should not allow addresses like
284 * @UCI:MRose@UCI-750a
286 * but should insist on
288 * Marshall Rose <@UCI:MRose@UCI-750a>
290 * Unfortunately, a lot of mailers stupidly do this.
314 static struct specials special[] = {
331 static int glevel = 0;
332 static int ingrp = 0;
333 static int last_lex = LX_END;
335 static char *dp = NULL;
336 static unsigned char *cp = NULL;
337 static unsigned char *ap = NULL;
338 static char *pers = NULL;
339 static char *mbox = NULL;
340 static char *host = NULL;
341 static char *path = NULL;
342 static char *grp = NULL;
343 static char *note = NULL;
344 static char err[BUFSIZ];
345 static char adr[BUFSIZ];
347 static struct adrx adrxs2;
351 getadrx (char *addrs)
354 register struct adrx *adrxp = &adrxs2;
368 pers = mbox = host = path = grp = note = NULL;
372 dp = cp = getcpy (addrs ? addrs : "");
382 switch (parse_address ()) {
394 default: /* catch trailing comments */
419 while (isspace (*ap))
422 sprintf (adr, "%.*s", (int)(cp - ap), ap);
425 bp = adr + strlen (adr) - 1;
426 if (*bp == ',' || *bp == ';' || *bp == '\n')
435 adrxp->ingrp = ingrp;
437 adrxp->err = err[0] ? err : NULL;
450 switch (my_lex (buffer)) {
453 pers = getcpy (buffer);
458 strcpy (err, "extraneous semi-colon");
471 case LX_LBRK: /* sigh (2) */
474 case LX_AT: /* sigh (3) */
476 if (route_addr (buffer) == NOTOK)
478 return OK; /* why be choosy? */
481 sprintf (err, "illegal address construct (%s)", buffer);
485 switch (my_lex (buffer)) {
488 pers = add (buffer, add (" ", pers));
489 more_phrase: ; /* sigh (1) */
490 if (phrase (buffer) == NOTOK)
496 if (route_addr (buffer) == NOTOK)
498 if (last_lex == LX_RBRK)
500 sprintf (err, "missing right-bracket (%s)", buffer);
506 sprintf (err, "nested groups not allowed (%s)", pers);
509 grp = add (": ", pers);
515 switch (my_lex (buffer)) {
517 case LX_END: /* tsk, tsk */
526 return parse_address ();
530 case LX_DOT: /* sigh (1) */
531 pers = add (".", pers);
535 sprintf (err, "no mailbox in address, only a phrase (%s%s)",
547 mbox = add (buffer, pers);
549 if (route_addr (buffer) == NOTOK)
557 if (domain (buffer) == NOTOK)
563 strcpy (err, "extraneous semi-colon");
571 sprintf (err, "junk after local@domain (%s)", buffer);
575 case LX_SEMI: /* no host */
579 if (last_lex == LX_SEMI && glevel-- <= 0) {
580 strcpy (err, "extraneous semi-colon");
588 sprintf (err, "missing mailbox (%s)", buffer);
595 phrase (char *buffer)
598 switch (my_lex (buffer)) {
601 pers = add (buffer, add (" ", pers));
611 route_addr (char *buffer)
613 register char *pp = cp;
615 if (my_lex (buffer) == LX_AT) {
616 if (route (buffer) == NOTOK)
622 if (local_part (buffer) == NOTOK)
627 return domain (buffer);
629 case LX_SEMI: /* if in group */
630 case LX_RBRK: /* no host */
636 sprintf (err, "no at-sign after local-part (%s)", buffer);
643 local_part (char *buffer)
648 switch (my_lex (buffer)) {
651 mbox = add (buffer, mbox);
655 sprintf (err, "no mailbox in local-part (%s)", buffer);
659 switch (my_lex (buffer)) {
661 mbox = add (buffer, mbox);
672 domain (char *buffer)
675 switch (my_lex (buffer)) {
678 host = add (buffer, host);
682 sprintf (err, "no sub-domain in domain-part of address (%s)", buffer);
686 switch (my_lex (buffer)) {
688 host = add (buffer, host);
691 case LX_AT: /* sigh (0) */
692 mbox = add (host, add ("%", mbox));
710 switch (my_lex (buffer)) {
713 path = add (buffer, path);
717 sprintf (err, "no sub-domain in domain-part of address (%s)", buffer);
720 switch (my_lex (buffer)) {
722 path = add (buffer, path);
724 switch (my_lex (buffer)) {
729 path = add (buffer, path);
733 sprintf (err, "no at-sign found for next domain in route (%s)",
740 case LX_AT: /* XXX */
742 path = add (buffer, path);
746 path = add (buffer, path);
750 sprintf (err, "no colon found to terminate route (%s)", buffer);
758 my_lex (char *buffer)
760 /* buffer should be at least BUFSIZ bytes long */
762 register unsigned char c;
765 /* Add C to the buffer bp. After use of this macro *bp is guaranteed to be within the buffer. */
766 #define ADDCHR(C) do { *bp++ = (C); if ((bp - buffer) == (BUFSIZ-1)) goto my_lex_buffull; } while (0)
771 return (last_lex = LX_END);
779 return (last_lex = LX_END);
788 return (last_lex = LX_ERR);
791 if ((c = *cp++) == 0) {
793 return (last_lex = LX_ERR);
806 note = note ? add (buffer, add (" ", note))
808 return my_lex (buffer);
819 return (last_lex = LX_ERR);
822 if ((c = *cp++) == 0) {
824 return (last_lex = LX_ERR);
832 return (last_lex = LX_QSTR);
842 return (last_lex = LX_ERR);
845 if ((c = *cp++) == 0) {
847 return (last_lex = LX_ERR);
855 return (last_lex = LX_DLIT);
861 for (i = 0; special[i].lx_chr != 0; i++)
862 if (c == special[i].lx_chr)
863 return (last_lex = special[i].lx_val);
866 return (last_lex = LX_ERR);
869 if ((c = *cp++) == 0)
871 for (i = 0; special[i].lx_chr != 0; i++)
872 if (c == special[i].lx_chr)
874 if (iscntrl (c) || isspace (c))
884 last_lex = !gotat || cp == NULL || strchr(cp, '<') != NULL
889 /* Out of buffer space. *bp is the last byte in the buffer */
891 return (last_lex = LX_ERR);
896 legal_person (char *p)
900 static char buffer[BUFSIZ];
904 for (cp = p; *cp; cp++)
905 for (i = 0; special[i].lx_chr; i++)
906 if (*cp == special[i].lx_chr) {
907 sprintf (buffer, "\"%s\"", p);
916 mfgets (FILE *in, char **bp)
919 register char *cp, *dp, *ep;
921 static char *pp = NULL;
924 pp = mh_xmalloc ((size_t) (len = BUFSIZ));
926 for (ep = (cp = pp) + len - 2;;) {
927 switch (i = getc (in)) {
945 if (cp == pp) /* end of headers, gobble it */
947 switch (i = getc (in)) {
948 default: /* end of line */
949 case '\n': /* end of headers, save for next call */
953 case ' ': /* continue headers */
957 } /* fall into default case */
964 dp = mh_xrealloc (pp, (size_t) (len += BUFSIZ));
965 cp += dp - pp, ep = (pp = cp) + len - 2;