]>
diplodocus.org Git - nmh/blob - zotnet/mf/mf.c
3 * mf.c -- mail filter subroutines
15 static char *getcpy (char *);
16 static char *add (char *, char *);
17 static void compress (char *, 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 *);
39 if ((p
= malloc ((size_t) (strlen (s
) + 2))))
46 add (char *s1
, char *s2
)
53 if ((p
= malloc ((size_t) (strlen (s1
) + strlen (s2
) + 2))))
54 sprintf (p
, "%s%s", s2
, s1
);
62 return (strncmp (string
, "From ", 5) == 0
63 || strncmp (string
, ">From ", 6) == 0);
68 lequal (char *a
, char *b
)
74 char c1
= islower (*a
) ? toupper (*a
) : *a
;
75 char c2
= islower (*b
) ? toupper (*b
) : *b
;
85 * seekadrx() is tricky. We want to cover both UUCP-style and ARPA-style
86 * addresses, so for each list of addresses we see if we can find some
87 * character to give us a hint.
91 #define CHKADR 0 /* undertermined address style */
92 #define UNIXDR 1 /* UNIX-style address */
93 #define ARPADR 2 /* ARPAnet-style address */
96 static char *punctuators
= ";<>.()[]";
97 static char *vp
= NULL
;
98 static char *tp
= NULL
;
100 static struct adrx adrxs1
;
104 seekadrx (char *addrs
)
106 static int state
= CHKADR
;
108 register struct adrx
*adrxp
;
111 for (state
= UNIXDR
, cp
= addrs
; *cp
; cp
++)
112 if (strchr(punctuators
, *cp
)) {
119 adrxp
= uucpadrx (addrs
);
124 adrxp
= getadrx (addrs
);
136 * uucpadrx() implements a partial UUCP-style address parser. It's based
137 * on the UUCP notion that addresses are separated by spaces or commas.
142 uucpadrx (char *addrs
)
144 register char *cp
, *wp
, *xp
, *yp
, *zp
;
145 register struct adrx
*adrxp
= &adrxs1
;
148 vp
= tp
= getcpy (addrs
);
149 compress (addrs
, vp
);
158 for (cp
= tp
; isspace (*cp
); cp
++)
166 if ((wp
= strchr(cp
, ',')) == NULL
) {
167 if ((wp
= strchr(cp
, ' ')) != NULL
) {
169 while (isspace (*xp
))
171 if (*xp
!= 0 && isat (--xp
)) {
173 while (isspace (*yp
))
176 if ((zp
= strchr(yp
, ' ')) != NULL
)
195 adrxp
->text
= getcpy (cp
);
197 adrxp
->host
= adrxp
->path
= NULL
;
198 if ((wp
= strrchr(cp
, '@')) != NULL
) {
200 adrxp
->host
= *wp
? wp
: NULL
;
203 for (wp
= cp
+ strlen (cp
) - 4; wp
>= cp
; wp
--)
206 adrxp
->host
= wp
+ 3;
209 adrxp
->pers
= adrxp
->grp
= adrxp
->note
= adrxp
->err
= NULL
;
217 compress (char *fp
, char *tp
)
219 register char c
, *cp
;
221 for (c
= ' ', cp
= tp
; (*tp
= *fp
++) != 0;)
229 if (c
== ' ' && cp
< tp
)
237 return (strncmp (p
, " AT ", 4)
238 && strncmp (p
, " At ", 4)
239 && strncmp (p
, " aT ", 4)
240 && strncmp (p
, " at ", 4) ? FALSE
: TRUE
);
246 * getadrx() implements a partial 822-style address parser. The parser
247 * is neither complete nor correct. It does however recognize nearly all
248 * of the 822 address syntax. In addition it handles the majority of the
249 * 733 syntax as well. Most problems arise from trying to accomodate both.
251 * In terms of 822, the route-specification in
253 * "<" [route] local-part "@" domain ">"
255 * is parsed and returned unchanged. Multiple at-signs are compressed
256 * via source-routing. Recursive groups are not allowed as per the
259 * In terms of 733, " at " is recognized as equivalent to "@".
261 * In terms of both the parser will not complain about missing hosts.
265 * We should not allow addresses like
267 * Marshall T. Rose <MRose@UCI>
269 * but should insist on
271 * "Marshall T. Rose" <MRose@UCI>
273 * Unfortunately, a lot of mailers stupidly let people get away with this.
277 * We should not allow addresses like
281 * but should insist on
285 * Unfortunately, a lot of mailers stupidly let people's UAs get away with
290 * We should not allow addresses like
292 * @UCI:MRose@UCI-750a
294 * but should insist on
296 * Marshall Rose <@UCI:MRose@UCI-750a>
298 * Unfortunately, a lot of mailers stupidly do this.
322 static struct specials special
[] = {
339 static int glevel
= 0;
340 static int ingrp
= 0;
341 static int last_lex
= LX_END
;
343 static char *dp
= NULL
;
344 static char *cp
= NULL
;
345 static char *ap
= NULL
;
346 static char *pers
= NULL
;
347 static char *mbox
= NULL
;
348 static char *host
= NULL
;
349 static char *path
= NULL
;
350 static char *grp
= NULL
;
351 static char *note
= NULL
;
352 static char err
[BUFSIZ
];
353 static char adr
[BUFSIZ
];
355 static struct adrx adrxs2
;
359 getadrx (char *addrs
)
362 register struct adrx
*adrxp
= &adrxs2
;
376 pers
= mbox
= host
= path
= grp
= note
= NULL
;
380 dp
= cp
= getcpy (addrs
? addrs
: "");
390 switch (parse_address ()) {
402 default: /* catch trailing comments */
427 while (isspace (*ap
))
430 sprintf (adr
, "%.*s", cp
- ap
, ap
);
433 bp
= adr
+ strlen (adr
) - 1;
434 if (*bp
== ',' || *bp
== ';' || *bp
== '\n')
443 adrxp
->ingrp
= ingrp
;
445 adrxp
->err
= err
[0] ? err
: NULL
;
458 switch (my_lex (buffer
)) {
461 pers
= getcpy (buffer
);
466 strcpy (err
, "extraneous semi-colon");
479 case LX_LBRK
: /* sigh (2) */
482 case LX_AT
: /* sigh (3) */
484 if (route_addr (buffer
) == NOTOK
)
486 return OK
; /* why be choosy? */
489 sprintf (err
, "illegal address construct (%s)", buffer
);
493 switch (my_lex (buffer
)) {
496 pers
= add (buffer
, add (" ", pers
));
497 more_phrase
: ; /* sigh (1) */
498 if (phrase (buffer
) == NOTOK
)
504 if (route_addr (buffer
) == NOTOK
)
506 if (last_lex
== LX_RBRK
)
508 sprintf (err
, "missing right-bracket (%s)", buffer
);
514 sprintf (err
, "nested groups not allowed (%s)", pers
);
517 grp
= add (": ", pers
);
523 switch (my_lex (buffer
)) {
525 case LX_END
: /* tsk, tsk */
534 return parse_address ();
538 case LX_DOT
: /* sigh (1) */
539 pers
= add (".", pers
);
543 sprintf (err
, "no mailbox in address, only a phrase (%s%s)",
555 mbox
= add (buffer
, pers
);
557 if (route_addr (buffer
) == NOTOK
)
565 if (domain (buffer
) == NOTOK
)
571 strcpy (err
, "extraneous semi-colon");
579 sprintf (err
, "junk after local@domain (%s)", buffer
);
583 case LX_SEMI
: /* no host */
587 if (last_lex
== LX_SEMI
&& glevel
-- <= 0) {
588 strcpy (err
, "extraneous semi-colon");
596 sprintf (err
, "missing mailbox (%s)", buffer
);
603 phrase (char *buffer
)
606 switch (my_lex (buffer
)) {
609 pers
= add (buffer
, add (" ", pers
));
619 route_addr (char *buffer
)
621 register char *pp
= cp
;
623 if (my_lex (buffer
) == LX_AT
) {
624 if (route (buffer
) == NOTOK
)
630 if (local_part (buffer
) == NOTOK
)
635 return domain (buffer
);
637 case LX_SEMI
: /* if in group */
638 case LX_RBRK
: /* no host */
644 sprintf (err
, "no at-sign after local-part (%s)", buffer
);
651 local_part (char *buffer
)
656 switch (my_lex (buffer
)) {
659 mbox
= add (buffer
, mbox
);
663 sprintf (err
, "no mailbox in local-part (%s)", buffer
);
667 switch (my_lex (buffer
)) {
669 mbox
= add (buffer
, mbox
);
680 domain (char *buffer
)
683 switch (my_lex (buffer
)) {
686 host
= add (buffer
, host
);
690 sprintf (err
, "no sub-domain in domain-part of address (%s)", buffer
);
694 switch (my_lex (buffer
)) {
696 host
= add (buffer
, host
);
699 case LX_AT
: /* sigh (0) */
700 mbox
= add (host
, add ("%", mbox
));
718 switch (my_lex (buffer
)) {
721 path
= add (buffer
, path
);
725 sprintf (err
, "no sub-domain in domain-part of address (%s)", buffer
);
728 switch (my_lex (buffer
)) {
730 path
= add (buffer
, path
);
732 switch (my_lex (buffer
)) {
737 path
= add (buffer
, path
);
741 sprintf (err
, "no at-sign found for next domain in route (%s)",
748 case LX_AT
: /* XXX */
750 path
= add (buffer
, path
);
754 path
= add (buffer
, path
);
758 sprintf (err
, "no colon found to terminate route (%s)", buffer
);
766 my_lex (char *buffer
)
769 register char c
, *bp
;
774 return (last_lex
= LX_END
);
782 return (last_lex
= LX_END
);
786 for (*bp
++ = c
, i
= 0;;)
790 return (last_lex
= LX_ERR
);
793 if ((c
= *cp
++) == 0) {
795 return (last_lex
= LX_ERR
);
808 note
= note
? add (buffer
, add (" ", note
))
810 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
);
840 return (last_lex
= LX_ERR
);
843 if ((c
= *cp
++) == 0) {
845 return (last_lex
= LX_ERR
);
853 return (last_lex
= LX_DLIT
);
858 for (i
= 0; special
[i
].lx_chr
!= 0; i
++)
859 if (c
== special
[i
].lx_chr
)
860 return (last_lex
= special
[i
].lx_val
);
863 return (last_lex
= LX_ERR
);
866 if ((c
= *cp
++) == 0)
868 for (i
= 0; special
[i
].lx_chr
!= 0; i
++)
869 if (c
== special
[i
].lx_chr
)
871 if (iscntrl (c
) || isspace (c
))
881 last_lex
= !gotat
|| cp
== NULL
|| strchr(cp
, '<') != NULL
888 legal_person (char *p
)
892 static char buffer
[BUFSIZ
];
896 for (cp
= p
; *cp
; cp
++)
897 for (i
= 0; special
[i
].lx_chr
; i
++)
898 if (*cp
== special
[i
].lx_chr
) {
899 sprintf (buffer
, "\"%s\"", p
);
908 mfgets (FILE *in
, char **bp
)
911 register char *cp
, *dp
, *ep
;
913 static char *pp
= NULL
;
916 if (!(pp
= malloc ((size_t) (len
= BUFSIZ
))))
919 for (ep
= (cp
= pp
) + len
- 2;;) {
920 switch (i
= getc (in
)) {
938 if (cp
== pp
) /* end of headers, gobble it */
940 switch (i
= getc (in
)) {
941 default: /* end of line */
942 case '\n': /* end of headers, save for next call */
946 case ' ': /* continue headers */
950 } /* fall into default case */
957 if (!(dp
= realloc (pp
, (size_t) (len
+= BUFSIZ
)))) {
963 cp
+= dp
- pp
, ep
= (pp
= cp
) + len
- 2;