13312Seric # include "sendmail.h"
2297Seric 
3*9585Seric SCCSID(@(#)parseaddr.c	3.68		12/09/82);
4407Seric 
5297Seric /*
6297Seric **  PARSE -- Parse an address
7297Seric **
8297Seric **	Parses an address and breaks it up into three parts: a
9297Seric **	net to transmit the message on, the host to transmit it
10297Seric **	to, and a user on that host.  These are loaded into an
112973Seric **	ADDRESS header with the values squirreled away if necessary.
12297Seric **	The "user" part may not be a real user; the process may
13297Seric **	just reoccur on that machine.  For example, on a machine
14297Seric **	with an arpanet connection, the address
15297Seric **		csvax.bill@berkeley
16297Seric **	will break up to a "user" of 'csvax.bill' and a host
17297Seric **	of 'berkeley' -- to be transmitted over the arpanet.
18297Seric **
19297Seric **	Parameters:
20297Seric **		addr -- the address to parse.
21297Seric **		a -- a pointer to the address descriptor buffer.
22297Seric **			If NULL, a header will be created.
23297Seric **		copyf -- determines what shall be copied:
24297Seric **			-1 -- don't copy anything.  The printname
25297Seric **				(q_paddr) is just addr, and the
26297Seric **				user & host are allocated internally
27297Seric **				to parse.
28297Seric **			0 -- copy out the parsed user & host, but
29297Seric **				don't copy the printname.
30297Seric **			+1 -- copy everything.
31297Seric **
32297Seric **	Returns:
33297Seric **		A pointer to the address descriptor header (`a' if
34297Seric **			`a' is non-NULL).
35297Seric **		NULL on error.
36297Seric **
37297Seric **	Side Effects:
38297Seric **		none
39297Seric */
40297Seric 
419374Seric /* following delimiters are inherent to the internal algorithms */
423380Seric # define DELIMCHARS	"$()<>,;\\\"\r\n"	/* word delimiters */
432091Seric 
442973Seric ADDRESS *
45297Seric parse(addr, a, copyf)
46297Seric 	char *addr;
472973Seric 	register ADDRESS *a;
48297Seric 	int copyf;
49297Seric {
503149Seric 	register char **pvp;
513149Seric 	register struct mailer *m;
523149Seric 	extern char **prescan();
533149Seric 	extern ADDRESS *buildaddr();
54297Seric 
55297Seric 	/*
56297Seric 	**  Initialize and prescan address.
57297Seric 	*/
58297Seric 
596903Seric 	CurEnv->e_to = addr;
603188Seric # ifdef DEBUG
617675Seric 	if (tTd(20, 1))
623188Seric 		printf("\n--parse(%s)\n", addr);
633188Seric # endif DEBUG
643188Seric 
658078Seric 	pvp = prescan(addr, ',');
663149Seric 	if (pvp == NULL)
67297Seric 		return (NULL);
68297Seric 
69297Seric 	/*
703149Seric 	**  Apply rewriting rules.
717889Seric 	**	Ruleset 0 does basic parsing.  It must resolve.
72297Seric 	*/
73297Seric 
748181Seric 	rewrite(pvp, 3);
754070Seric 	rewrite(pvp, 0);
76297Seric 
773149Seric 	/*
783149Seric 	**  See if we resolved to a real mailer.
793149Seric 	*/
80297Seric 
813149Seric 	if (pvp[0][0] != CANONNET)
823149Seric 	{
833149Seric 		setstat(EX_USAGE);
843149Seric 		usrerr("cannot resolve name");
853149Seric 		return (NULL);
86297Seric 	}
87297Seric 
88297Seric 	/*
893149Seric 	**  Build canonical address from pvp.
90297Seric 	*/
91297Seric 
923149Seric 	a = buildaddr(pvp, a);
934279Seric 	if (a == NULL)
944279Seric 		return (NULL);
954598Seric 	m = a->q_mailer;
96297Seric 
97297Seric 	/*
983149Seric 	**  Make local copies of the host & user and then
993149Seric 	**  transport them out.
100297Seric 	*/
101297Seric 
102297Seric 	if (copyf > 0)
1038078Seric 	{
1048078Seric 		extern char *DelimChar;
1058078Seric 		char savec = *DelimChar;
1068078Seric 
1078078Seric 		*DelimChar = '\0';
1082973Seric 		a->q_paddr = newstr(addr);
1098078Seric 		*DelimChar = savec;
1108078Seric 	}
111297Seric 	else
112297Seric 		a->q_paddr = addr;
1133149Seric 	if (copyf >= 0)
114297Seric 	{
1153149Seric 		if (a->q_host != NULL)
1163149Seric 			a->q_host = newstr(a->q_host);
117297Seric 		else
1183149Seric 			a->q_host = "";
1193149Seric 		if (a->q_user != a->q_paddr)
1203149Seric 			a->q_user = newstr(a->q_user);
121297Seric 	}
122297Seric 
123297Seric 	/*
124297Seric 	**  Do UPPER->lower case mapping unless inhibited.
125297Seric 	*/
126297Seric 
1273149Seric 	if (!bitset(M_HST_UPPER, m->m_flags))
128297Seric 		makelower(a->q_host);
1293149Seric 	if (!bitset(M_USR_UPPER, m->m_flags))
130297Seric 		makelower(a->q_user);
131297Seric 
132297Seric 	/*
133297Seric 	**  Compute return value.
134297Seric 	*/
135297Seric 
136297Seric # ifdef DEBUG
1377675Seric 	if (tTd(20, 1))
1384443Seric 	{
1394443Seric 		printf("parse-->");
1404443Seric 		printaddr(a, FALSE);
1414443Seric 	}
142297Seric # endif DEBUG
143297Seric 
144297Seric 	return (a);
145297Seric }
146297Seric /*
147297Seric **  PRESCAN -- Prescan name and make it canonical
148297Seric **
1499374Seric **	Scans a name and turns it into a set of tokens.  This process
1509374Seric **	deletes blanks and comments (in parentheses).
151297Seric **
152297Seric **	This routine knows about quoted strings and angle brackets.
153297Seric **
154297Seric **	There are certain subtleties to this routine.  The one that
155297Seric **	comes to mind now is that backslashes on the ends of names
156297Seric **	are silently stripped off; this is intentional.  The problem
157297Seric **	is that some versions of sndmsg (like at LBL) set the kill
158297Seric **	character to something other than @ when reading addresses;
159297Seric **	so people type "csvax.eric\@berkeley" -- which screws up the
160297Seric **	berknet mailer.
161297Seric **
162297Seric **	Parameters:
163297Seric **		addr -- the name to chomp.
164297Seric **		delim -- the delimiter for the address, normally
165297Seric **			'\0' or ','; \0 is accepted in any case.
166297Seric **
167297Seric **	Returns:
1683149Seric **		A pointer to a vector of tokens.
169297Seric **		NULL on error.
170297Seric **
171297Seric **	Side Effects:
1723149Seric **		none.
173297Seric */
174297Seric 
1758078Seric /* states and character types */
1768078Seric # define OPR		0	/* operator */
1778078Seric # define ATM		1	/* atom */
1788078Seric # define QST		2	/* in quoted string */
1798078Seric # define SPC		3	/* chewing up spaces */
1808078Seric # define ONE		4	/* pick up one character */
1813149Seric 
1828078Seric # define NSTATES	5	/* number of states */
1838078Seric # define TYPE		017	/* mask to select state type */
1848078Seric 
1858078Seric /* meta bits for table */
1868078Seric # define M		020	/* meta character; don't pass through */
1878078Seric # define B		040	/* cause a break */
1888078Seric # define MB		M|B	/* meta-break */
1898078Seric 
1908078Seric static short StateTab[NSTATES][NSTATES] =
1918078Seric {
1928087Seric    /*	oldst	chtype>	OPR	ATM	QST	SPC	ONE	*/
1939051Seric 	/*OPR*/		OPR|B,	ATM|B,	QST|B,	SPC|MB,	ONE|B,
1949051Seric 	/*ATM*/		OPR|B,	ATM,	QST|B,	SPC|MB,	ONE|B,
1959051Seric 	/*QST*/		QST,	QST,	OPR,	QST,	QST,
1968078Seric 	/*SPC*/		OPR,	ATM,	QST,	SPC|M,	ONE,
1978078Seric 	/*ONE*/		OPR,	OPR,	OPR,	OPR,	OPR,
1988078Seric };
1998078Seric 
2008078Seric # define NOCHAR		-1	/* signal nothing in lookahead token */
2018078Seric 
2028078Seric char	*DelimChar;		/* set to point to the delimiter */
2038078Seric 
2043149Seric char **
2053149Seric prescan(addr, delim)
206297Seric 	char *addr;
207297Seric 	char delim;
208297Seric {
209297Seric 	register char *p;
2108078Seric 	register char *q;
2119346Seric 	register int c;
2123149Seric 	char **avp;
213297Seric 	bool bslashmode;
214297Seric 	int cmntcnt;
2158423Seric 	int anglecnt;
2163149Seric 	char *tok;
2178078Seric 	int state;
2188078Seric 	int newstate;
2198078Seric 	static char buf[MAXNAME+MAXATOM];
2208078Seric 	static char *av[MAXATOM+1];
221297Seric 
222297Seric 	q = buf;
2233149Seric 	bslashmode = FALSE;
2247800Seric 	cmntcnt = 0;
2258423Seric 	anglecnt = 0;
2263149Seric 	avp = av;
2278078Seric 	state = OPR;
2288078Seric 	c = NOCHAR;
2298078Seric 	p = addr;
2308078Seric # ifdef DEBUG
2318078Seric 	if (tTd(22, 45))
232297Seric 	{
2338078Seric 		printf("prescan: ");
2348078Seric 		xputs(p);
2358078Seric 		putchar('\n');
2368078Seric 	}
2378078Seric # endif DEBUG
2388078Seric 
2398078Seric 	do
2408078Seric 	{
2413149Seric 		/* read a token */
2423149Seric 		tok = q;
2438078Seric 		for (;;)
244297Seric 		{
2458078Seric 			/* store away any old lookahead character */
2468078Seric 			if (c != NOCHAR)
2478078Seric 			{
2488078Seric 				/* squirrel it away */
2498078Seric 				if (q >= &buf[sizeof buf - 5])
2508078Seric 				{
2518078Seric 					usrerr("Address too long");
2528078Seric 					DelimChar = p;
2538078Seric 					return (NULL);
2548078Seric 				}
2558078Seric 				*q++ = c;
2568078Seric 			}
2578078Seric 
2588078Seric 			/* read a new input character */
2598078Seric 			c = *p++;
2608078Seric 			if (c == '\0')
2618078Seric 				break;
2628078Seric # ifdef DEBUG
2638078Seric 			if (tTd(22, 101))
2648078Seric 				printf("c=%c, s=%d; ", c, state);
2658078Seric # endif DEBUG
2668078Seric 
2673149Seric 			/* chew up special characters */
2684100Seric 			c &= ~0200;
2693149Seric 			*q = '\0';
2703149Seric 			if (bslashmode)
2713149Seric 			{
2723149Seric 				c |= 0200;
2733149Seric 				bslashmode = FALSE;
2743149Seric 			}
2753149Seric 			else if (c == '\\')
2763149Seric 			{
2773149Seric 				bslashmode = TRUE;
2788078Seric 				c = NOCHAR;
2793149Seric 			}
2808514Seric 			else if (state == QST)
2818514Seric 			{
2828514Seric 				/* do nothing, just avoid next clauses */
2838514Seric 			}
2848078Seric 			else if (c == '(')
2854100Seric 			{
2868078Seric 				cmntcnt++;
2878078Seric 				c = NOCHAR;
2884100Seric 			}
2898078Seric 			else if (c == ')')
2903149Seric 			{
2918078Seric 				if (cmntcnt <= 0)
2923149Seric 				{
2938078Seric 					usrerr("Unbalanced ')'");
2948078Seric 					DelimChar = p;
2958078Seric 					return (NULL);
2963149Seric 				}
2978078Seric 				else
2988078Seric 					cmntcnt--;
2998078Seric 			}
3008078Seric 			else if (cmntcnt > 0)
3018078Seric 				c = NOCHAR;
3028423Seric 			else if (c == '<')
3038423Seric 				anglecnt++;
3048423Seric 			else if (c == '>')
3058423Seric 			{
3068423Seric 				if (anglecnt <= 0)
3078423Seric 				{
3088423Seric 					usrerr("Unbalanced '>'");
3098423Seric 					DelimChar = p;
3108423Seric 					return (NULL);
3118423Seric 				}
3128423Seric 				anglecnt--;
3138423Seric 			}
3143149Seric 
3158078Seric 			if (c == NOCHAR)
3168078Seric 				continue;
3173149Seric 
3188078Seric 			/* see if this is end of input */
3198423Seric 			if (c == delim && anglecnt <= 0)
3203149Seric 				break;
3213149Seric 
3228078Seric 			newstate = StateTab[state][toktype(c)];
3238078Seric # ifdef DEBUG
3248078Seric 			if (tTd(22, 101))
3258078Seric 				printf("ns=%02o\n", newstate);
3268078Seric # endif DEBUG
3278078Seric 			state = newstate & TYPE;
3288078Seric 			if (bitset(M, newstate))
3298078Seric 				c = NOCHAR;
3308078Seric 			if (bitset(B, newstate))
3314228Seric 				break;
332297Seric 		}
3333149Seric 
3343149Seric 		/* new token */
3358078Seric 		if (tok != q)
3361378Seric 		{
3378078Seric 			*q++ = '\0';
3388078Seric # ifdef DEBUG
3398078Seric 			if (tTd(22, 36))
340297Seric 			{
3418078Seric 				printf("tok=");
3428078Seric 				xputs(tok);
3438078Seric 				putchar('\n');
344297Seric 			}
3458078Seric # endif DEBUG
3468078Seric 			if (avp >= &av[MAXATOM])
347297Seric 			{
3488078Seric 				syserr("prescan: too many tokens");
3498078Seric 				DelimChar = p;
3508078Seric 				return (NULL);
351297Seric 			}
3528078Seric 			*avp++ = tok;
353297Seric 		}
3548423Seric 	} while (c != '\0' && (c != delim || anglecnt > 0));
3553149Seric 	*avp = NULL;
3568078Seric 	DelimChar = --p;
3573149Seric 	if (cmntcnt > 0)
3583149Seric 		usrerr("Unbalanced '('");
3598423Seric 	else if (anglecnt > 0)
3608423Seric 		usrerr("Unbalanced '<'");
3618078Seric 	else if (state == QST)
3623149Seric 		usrerr("Unbalanced '\"'");
3633149Seric 	else if (av[0] != NULL)
3643149Seric 		return (av);
3653149Seric 	return (NULL);
3663149Seric }
3673149Seric /*
3683149Seric **  TOKTYPE -- return token type
3693149Seric **
3703149Seric **	Parameters:
3713149Seric **		c -- the character in question.
3723149Seric **
3733149Seric **	Returns:
3743149Seric **		Its type.
3753149Seric **
3763149Seric **	Side Effects:
3773149Seric **		none.
3783149Seric */
379297Seric 
3803149Seric toktype(c)
3813149Seric 	register char c;
3823149Seric {
3833380Seric 	static char buf[50];
3843382Seric 	static bool firstime = TRUE;
3853380Seric 
3863382Seric 	if (firstime)
3873380Seric 	{
3883382Seric 		firstime = FALSE;
3896977Seric 		expand("$o", buf, &buf[sizeof buf - 1], CurEnv);
3907005Seric 		(void) strcat(buf, DELIMCHARS);
3913380Seric 	}
392*9585Seric 	if (c == MATCHCLASS || c == MATCHREPL || c == MATCHNCLASS)
3938078Seric 		return (ONE);
3948078Seric 	if (c == '"')
3958078Seric 		return (QST);
3964100Seric 	if (!isascii(c))
3978078Seric 		return (ATM);
3988078Seric 	if (isspace(c) || c == ')')
3998078Seric 		return (SPC);
4003380Seric 	if (iscntrl(c) || index(buf, c) != NULL)
4018078Seric 		return (OPR);
4028078Seric 	return (ATM);
4033149Seric }
4043149Seric /*
4053149Seric **  REWRITE -- apply rewrite rules to token vector.
4063149Seric **
4074476Seric **	This routine is an ordered production system.  Each rewrite
4084476Seric **	rule has a LHS (called the pattern) and a RHS (called the
4094476Seric **	rewrite); 'rwr' points the the current rewrite rule.
4104476Seric **
4114476Seric **	For each rewrite rule, 'avp' points the address vector we
4124476Seric **	are trying to match against, and 'pvp' points to the pattern.
4138058Seric **	If pvp points to a special match value (MATCHZANY, MATCHANY,
414*9585Seric **	MATCHONE, MATCHCLASS, MATCHNCLASS) then the address in avp
415*9585Seric **	matched is saved away in the match vector (pointed to by 'mvp').
4164476Seric **
4174476Seric **	When a match between avp & pvp does not match, we try to
418*9585Seric **	back out.  If we back up over MATCHONE, MATCHCLASS, or MATCHNCLASS
4194476Seric **	we must also back out the match in mvp.  If we reach a
4208058Seric **	MATCHANY or MATCHZANY we just extend the match and start
4218058Seric **	over again.
4224476Seric **
4234476Seric **	When we finally match, we rewrite the address vector
4244476Seric **	and try over again.
4254476Seric **
4263149Seric **	Parameters:
4273149Seric **		pvp -- pointer to token vector.
4283149Seric **
4293149Seric **	Returns:
4303149Seric **		none.
4313149Seric **
4323149Seric **	Side Effects:
4333149Seric **		pvp is modified.
4343149Seric */
4352091Seric 
4363149Seric struct match
4373149Seric {
4384468Seric 	char	**first;	/* first token matched */
4394468Seric 	char	**last;		/* last token matched */
4403149Seric };
4413149Seric 
4424468Seric # define MAXMATCH	9	/* max params per rewrite */
4433149Seric 
4443149Seric 
4454070Seric rewrite(pvp, ruleset)
4463149Seric 	char **pvp;
4474070Seric 	int ruleset;
4483149Seric {
4493149Seric 	register char *ap;		/* address pointer */
4503149Seric 	register char *rp;		/* rewrite pointer */
4513149Seric 	register char **avp;		/* address vector pointer */
4523149Seric 	register char **rvp;		/* rewrite vector pointer */
4538058Seric 	register struct match *mlp;	/* cur ptr into mlist */
4548058Seric 	register struct rewrite *rwr;	/* pointer to current rewrite rule */
4554468Seric 	struct match mlist[MAXMATCH];	/* stores match on LHS */
4563149Seric 	char *npvp[MAXATOM+1];		/* temporary space for rebuild */
4574060Seric 	extern bool sameword();
4583149Seric 
4599279Seric 	if (OpMode == MD_TEST || tTd(21, 2))
4603149Seric 	{
4618959Seric 		printf("rewrite: ruleset %2d   input:", ruleset);
4623149Seric 		printav(pvp);
4633149Seric 	}
4648423Seric 	if (pvp == NULL)
4658423Seric 		return;
4663149Seric 
4673149Seric 	/*
4683149Seric 	**  Run through the list of rewrite rules, applying
4693149Seric 	**	any that match.
4703149Seric 	*/
4713149Seric 
4724070Seric 	for (rwr = RewriteRules[ruleset]; rwr != NULL; )
4733149Seric 	{
4744100Seric # ifdef DEBUG
4757675Seric 		if (tTd(21, 12))
476297Seric 		{
4778069Seric 			printf("-----trying rule:");
4783149Seric 			printav(rwr->r_lhs);
4793149Seric 		}
4804100Seric # endif DEBUG
4813149Seric 
4823149Seric 		/* try to match on this rule */
4834468Seric 		mlp = mlist;
4848058Seric 		rvp = rwr->r_lhs;
4858058Seric 		avp = pvp;
4868058Seric 		while ((ap = *avp) != NULL || *rvp != NULL)
4873149Seric 		{
4883149Seric 			rp = *rvp;
4898058Seric # ifdef DEBUG
4908058Seric 			if (tTd(21, 35))
4918058Seric 			{
4928069Seric 				printf("ap=");
4938058Seric 				xputs(ap);
4948069Seric 				printf(", rp=");
4958058Seric 				xputs(rp);
4968069Seric 				printf("\n");
4978058Seric 			}
4988058Seric # endif DEBUG
4993149Seric 			if (rp == NULL)
500297Seric 			{
5013149Seric 				/* end-of-pattern before end-of-address */
5028058Seric 				goto backup;
503297Seric 			}
5048058Seric 			if (ap == NULL && *rp != MATCHZANY)
5058058Seric 			{
5068058Seric 				/* end-of-input */
5078058Seric 				break;
5088058Seric 			}
5093149Seric 
5103149Seric 			switch (*rp)
5113149Seric 			{
5124060Seric 				register STAB *s;
5134060Seric 				register int class;
5144060Seric 
5154060Seric 			  case MATCHCLASS:
516*9585Seric 			  case MATCHNCLASS:
517*9585Seric 				/* match any token in (not in) a class */
5184060Seric 				class = rp[1];
5194060Seric 				if (!isalpha(class))
5208058Seric 					goto backup;
5214060Seric 				if (isupper(class))
5224060Seric 					class -= 'A';
5234060Seric 				else
5244060Seric 					class -= 'a';
5254100Seric 				s = stab(ap, ST_CLASS, ST_FIND);
5266273Seric 				if (s == NULL || (s->s_class & (1L << class)) == 0)
527*9585Seric 				{
528*9585Seric 					if (*rp == MATCHCLASS)
529*9585Seric 						goto backup;
530*9585Seric 				}
531*9585Seric 				else if (*rp == MATCHNCLASS)
5328058Seric 					goto backup;
5334468Seric 
5344476Seric 				/* explicit fall-through */
5354476Seric 
5364476Seric 			  case MATCHONE:
5374476Seric 			  case MATCHANY:
5384476Seric 				/* match exactly one token */
5398058Seric 				mlp->first = avp;
5408058Seric 				mlp->last = avp++;
5414468Seric 				mlp++;
5424060Seric 				break;
5434060Seric 
5448058Seric 			  case MATCHZANY:
5458058Seric 				/* match zero or more tokens */
5468058Seric 				mlp->first = avp;
5478058Seric 				mlp->last = avp - 1;
5488058Seric 				mlp++;
5498058Seric 				break;
5508058Seric 
5513149Seric 			  default:
5523149Seric 				/* must have exact match */
5534060Seric 				if (!sameword(rp, ap))
5548058Seric 					goto backup;
5554468Seric 				avp++;
5563149Seric 				break;
5573149Seric 			}
5583149Seric 
5593149Seric 			/* successful match on this token */
5603149Seric 			rvp++;
5613149Seric 			continue;
5623149Seric 
5638058Seric 		  backup:
5643149Seric 			/* match failed -- back up */
5653149Seric 			while (--rvp >= rwr->r_lhs)
5663149Seric 			{
5673149Seric 				rp = *rvp;
5688058Seric 				if (*rp == MATCHANY || *rp == MATCHZANY)
5694468Seric 				{
5704476Seric 					/* extend binding and continue */
5718058Seric 					avp = ++mlp[-1].last;
5728058Seric 					avp++;
5734476Seric 					rvp++;
5743149Seric 					break;
5754468Seric 				}
5764476Seric 				avp--;
577*9585Seric 				if (*rp == MATCHONE || *rp == MATCHCLASS ||
578*9585Seric 				    *rp == MATCHNCLASS)
5793149Seric 				{
5804468Seric 					/* back out binding */
5814468Seric 					mlp--;
5823149Seric 				}
5833149Seric 			}
5843149Seric 
5853149Seric 			if (rvp < rwr->r_lhs)
5863149Seric 			{
5873149Seric 				/* total failure to match */
5883149Seric 				break;
5893149Seric 			}
590297Seric 		}
5913149Seric 
5923149Seric 		/*
5933149Seric 		**  See if we successfully matched
5943149Seric 		*/
5953149Seric 
5969374Seric 		if (rvp < rwr->r_lhs || *rvp != NULL)
5973149Seric 		{
5984100Seric # ifdef DEBUG
5999374Seric 			if (tTd(21, 10))
6009374Seric 				printf("----- rule fails\n");
6014100Seric # endif DEBUG
6029374Seric 			rwr = rwr->r_next;
6039374Seric 			continue;
6049374Seric 		}
6053149Seric 
6069374Seric 		rvp = rwr->r_rhs;
6079374Seric # ifdef DEBUG
6089374Seric 		if (tTd(21, 12))
6099374Seric 		{
6109374Seric 			printf("-----rule matches:");
6119374Seric 			printav(rvp);
6129374Seric 		}
6139374Seric # endif DEBUG
6149374Seric 
6159374Seric 		rp = *rvp;
6169374Seric 		if (*rp == CANONUSER)
6179374Seric 		{
6189374Seric 			rvp++;
6199374Seric 			rwr = rwr->r_next;
6209374Seric 		}
6219374Seric 		else if (*rp == CANONHOST)
6229374Seric 		{
6239374Seric 			rvp++;
6249374Seric 			rwr = NULL;
6259374Seric 		}
6269374Seric 		else if (*rp == CANONNET)
6279374Seric 			rwr = NULL;
6289374Seric 
6299374Seric 		/* substitute */
6309374Seric 		for (avp = npvp; *rvp != NULL; rvp++)
6319374Seric 		{
6329374Seric 			register struct match *m;
6339374Seric 			register char **pp;
6349374Seric 
6358058Seric 			rp = *rvp;
6369374Seric 			if (*rp != MATCHREPL)
6378058Seric 			{
6389374Seric 				if (avp >= &npvp[MAXATOM])
6399374Seric 				{
6409374Seric 					syserr("rewrite: expansion too long");
6419374Seric 					return;
6429374Seric 				}
6439374Seric 				*avp++ = rp;
6449374Seric 				continue;
6458069Seric 			}
6468058Seric 
6479374Seric 			/* substitute from LHS */
6489374Seric 			m = &mlist[rp[1] - '1'];
6499374Seric # ifdef DEBUG
6509374Seric 			if (tTd(21, 15))
6513149Seric 			{
6529374Seric 				printf("$%c:", rp[1]);
6539374Seric 				pp = m->first;
6549374Seric 				while (pp <= m->last)
6553149Seric 				{
6569374Seric 					printf(" %x=\"", *pp);
6579374Seric 					(void) fflush(stdout);
6589374Seric 					printf("%s\"", *pp++);
6593149Seric 				}
6609374Seric 				printf("\n");
6613149Seric 			}
6628226Seric # endif DEBUG
6639374Seric 			pp = m->first;
6649374Seric 			while (pp <= m->last)
6658226Seric 			{
6669374Seric 				if (avp >= &npvp[MAXATOM])
6679374Seric 				{
6689374Seric 					syserr("rewrite: expansion too long");
6699374Seric 					return;
6709374Seric 				}
6719374Seric 				*avp++ = *pp++;
6728226Seric 			}
6739374Seric 		}
6749374Seric 		*avp++ = NULL;
6759374Seric 		if (**npvp == CALLSUBR)
6769374Seric 		{
6779374Seric 			bmove((char *) &npvp[2], (char *) pvp,
6789374Seric 				(avp - npvp - 2) * sizeof *avp);
6798226Seric # ifdef DEBUG
6809374Seric 			if (tTd(21, 3))
6819374Seric 				printf("-----callsubr %s\n", npvp[1]);
6823149Seric # endif DEBUG
6839374Seric 			rewrite(pvp, atoi(npvp[1]));
6843149Seric 		}
6853149Seric 		else
6863149Seric 		{
6879374Seric 			bmove((char *) npvp, (char *) pvp,
6889374Seric 				(avp - npvp) * sizeof *avp);
6899374Seric 		}
6904100Seric # ifdef DEBUG
6919374Seric 		if (tTd(21, 4))
6929374Seric 		{
6939374Seric 			printf("rewritten as:");
6949374Seric 			printav(pvp);
6959374Seric 		}
6964100Seric # endif DEBUG
697297Seric 	}
6988069Seric 
6999279Seric 	if (OpMode == MD_TEST || tTd(21, 2))
7008069Seric 	{
7018959Seric 		printf("rewrite: ruleset %2d returns:", ruleset);
7028069Seric 		printav(pvp);
7038069Seric 	}
7043149Seric }
7053149Seric /*
7063149Seric **  BUILDADDR -- build address from token vector.
7073149Seric **
7083149Seric **	Parameters:
7093149Seric **		tv -- token vector.
7103149Seric **		a -- pointer to address descriptor to fill.
7113149Seric **			If NULL, one will be allocated.
7123149Seric **
7133149Seric **	Returns:
7144279Seric **		NULL if there was an error.
7154279Seric **		'a' otherwise.
7163149Seric **
7173149Seric **	Side Effects:
7183149Seric **		fills in 'a'
7193149Seric */
7203149Seric 
7213149Seric ADDRESS *
7223149Seric buildaddr(tv, a)
7233149Seric 	register char **tv;
7243149Seric 	register ADDRESS *a;
7253149Seric {
7263149Seric 	static char buf[MAXNAME];
7273149Seric 	struct mailer **mp;
7283149Seric 	register struct mailer *m;
7294635Seric 	extern bool sameword();
7303149Seric 
7313149Seric 	if (a == NULL)
7323149Seric 		a = (ADDRESS *) xalloc(sizeof *a);
7334988Seric 	clear((char *) a, sizeof *a);
7343149Seric 
7353149Seric 	/* figure out what net/mailer to use */
7363149Seric 	if (**tv != CANONNET)
7374279Seric 	{
7383149Seric 		syserr("buildaddr: no net");
7394279Seric 		return (NULL);
7404279Seric 	}
7413149Seric 	tv++;
7424635Seric 	if (sameword(*tv, "error"))
7434279Seric 	{
7444279Seric 		if (**++tv != CANONUSER)
7454279Seric 			syserr("buildaddr: error: no user");
7464279Seric 		buf[0] = '\0';
7474279Seric 		while (*++tv != NULL)
7484279Seric 		{
7494279Seric 			if (buf[0] != '\0')
7507005Seric 				(void) strcat(buf, " ");
7517005Seric 			(void) strcat(buf, *tv);
7524279Seric 		}
7534279Seric 		usrerr(buf);
7544279Seric 		return (NULL);
7554279Seric 	}
7564598Seric 	for (mp = Mailer; (m = *mp++) != NULL; )
7573149Seric 	{
7584635Seric 		if (sameword(m->m_name, *tv))
7593149Seric 			break;
7603149Seric 	}
7613149Seric 	if (m == NULL)
7624279Seric 	{
7633149Seric 		syserr("buildaddr: unknown net %s", *tv);
7644279Seric 		return (NULL);
7654279Seric 	}
7664598Seric 	a->q_mailer = m;
7673149Seric 
7683149Seric 	/* figure out what host (if any) */
7693149Seric 	tv++;
7704195Seric 	if (!bitset(M_LOCAL, m->m_flags))
7713149Seric 	{
7725704Seric 		if (**tv++ != CANONHOST)
7734279Seric 		{
7743149Seric 			syserr("buildaddr: no host");
7754279Seric 			return (NULL);
7764279Seric 		}
7775704Seric 		buf[0] = '\0';
7785704Seric 		while (*tv != NULL && **tv != CANONUSER)
7797005Seric 			(void) strcat(buf, *tv++);
7805704Seric 		a->q_host = newstr(buf);
7813149Seric 	}
7823149Seric 	else
7833149Seric 		a->q_host = NULL;
7843149Seric 
7853149Seric 	/* figure out the user */
7863149Seric 	if (**tv != CANONUSER)
7874279Seric 	{
7883149Seric 		syserr("buildaddr: no user");
7894279Seric 		return (NULL);
7904279Seric 	}
7914228Seric 	cataddr(++tv, buf, sizeof buf);
7923149Seric 	a->q_user = buf;
7933149Seric 
7943149Seric 	return (a);
7953149Seric }
7963188Seric /*
7974228Seric **  CATADDR -- concatenate pieces of addresses (putting in <LWSP> subs)
7984228Seric **
7994228Seric **	Parameters:
8004228Seric **		pvp -- parameter vector to rebuild.
8014228Seric **		buf -- buffer to build the string into.
8024228Seric **		sz -- size of buf.
8034228Seric **
8044228Seric **	Returns:
8054228Seric **		none.
8064228Seric **
8074228Seric **	Side Effects:
8084228Seric **		Destroys buf.
8094228Seric */
8104228Seric 
8114228Seric cataddr(pvp, buf, sz)
8124228Seric 	char **pvp;
8134228Seric 	char *buf;
8144228Seric 	register int sz;
8154228Seric {
8164228Seric 	bool oatomtok = FALSE;
8174228Seric 	bool natomtok = FALSE;
8184228Seric 	register int i;
8194228Seric 	register char *p;
8204228Seric 
8218423Seric 	if (pvp == NULL)
8228423Seric 	{
8238423Seric 		strcpy(buf, "");
8248423Seric 		return;
8258423Seric 	}
8264228Seric 	p = buf;
8274228Seric 	sz--;
8284228Seric 	while (*pvp != NULL && (i = strlen(*pvp)) < sz)
8294228Seric 	{
8308078Seric 		natomtok = (toktype(**pvp) == ATM);
8314228Seric 		if (oatomtok && natomtok)
8329042Seric 			*p++ = SpaceSub;
8334228Seric 		(void) strcpy(p, *pvp);
8344228Seric 		oatomtok = natomtok;
8354228Seric 		p += i;
8364228Seric 		sz -= i;
8374228Seric 		pvp++;
8384228Seric 	}
8394228Seric 	*p = '\0';
8404228Seric }
8414228Seric /*
8423188Seric **  SAMEADDR -- Determine if two addresses are the same
8433188Seric **
8443188Seric **	This is not just a straight comparison -- if the mailer doesn't
8453188Seric **	care about the host we just ignore it, etc.
8463188Seric **
8473188Seric **	Parameters:
8483188Seric **		a, b -- pointers to the internal forms to compare.
8493188Seric **
8503188Seric **	Returns:
8513188Seric **		TRUE -- they represent the same mailbox.
8523188Seric **		FALSE -- they don't.
8533188Seric **
8543188Seric **	Side Effects:
8553188Seric **		none.
8563188Seric */
8573188Seric 
8583188Seric bool
8599374Seric sameaddr(a, b)
8603188Seric 	register ADDRESS *a;
8613188Seric 	register ADDRESS *b;
8623188Seric {
8633188Seric 	/* if they don't have the same mailer, forget it */
8643188Seric 	if (a->q_mailer != b->q_mailer)
8653188Seric 		return (FALSE);
8663188Seric 
8673188Seric 	/* if the user isn't the same, we can drop out */
8689374Seric 	if (strcmp(a->q_user, b->q_user) != 0)
8693188Seric 		return (FALSE);
8703188Seric 
8713188Seric 	/* if the mailer ignores hosts, we have succeeded! */
8724598Seric 	if (bitset(M_LOCAL, a->q_mailer->m_flags))
8733188Seric 		return (TRUE);
8743188Seric 
8753188Seric 	/* otherwise compare hosts (but be careful for NULL ptrs) */
8763188Seric 	if (a->q_host == NULL || b->q_host == NULL)
8773188Seric 		return (FALSE);
8783188Seric 	if (strcmp(a->q_host, b->q_host) != 0)
8793188Seric 		return (FALSE);
8803188Seric 
8813188Seric 	return (TRUE);
8823188Seric }
8833234Seric /*
8843234Seric **  PRINTADDR -- print address (for debugging)
8853234Seric **
8863234Seric **	Parameters:
8873234Seric **		a -- the address to print
8883234Seric **		follow -- follow the q_next chain.
8893234Seric **
8903234Seric **	Returns:
8913234Seric **		none.
8923234Seric **
8933234Seric **	Side Effects:
8943234Seric **		none.
8953234Seric */
8963234Seric 
8974317Seric # ifdef DEBUG
8984317Seric 
8993234Seric printaddr(a, follow)
9003234Seric 	register ADDRESS *a;
9013234Seric 	bool follow;
9023234Seric {
9035001Seric 	bool first = TRUE;
9045001Seric 
9053234Seric 	while (a != NULL)
9063234Seric 	{
9075001Seric 		first = FALSE;
9084443Seric 		printf("%x=", a);
9094085Seric 		(void) fflush(stdout);
9103234Seric 		printf("%s: mailer %d (%s), host `%s', user `%s'\n", a->q_paddr,
9118181Seric 		       a->q_mailer->m_mno, a->q_mailer->m_name, a->q_host,
9128181Seric 		       a->q_user);
9138181Seric 		printf("\tnext=%x, flags=%o, alias %x\n", a->q_next, a->q_flags,
9148181Seric 		       a->q_alias);
9158181Seric 		printf("\thome=\"%s\", fullname=\"%s\"\n", a->q_home,
9168181Seric 		       a->q_fullname);
9174996Seric 
9183234Seric 		if (!follow)
9193234Seric 			return;
9204996Seric 		a = a->q_next;
9213234Seric 	}
9225001Seric 	if (first)
9234443Seric 		printf("[NULL]\n");
9243234Seric }
9254317Seric 
9264317Seric # endif DEBUG
9277682Seric /*
9287682Seric **  REMOTENAME -- return the name relative to the current mailer
9297682Seric **
9307682Seric **	Parameters:
9317682Seric **		name -- the name to translate.
9328069Seric **		m -- the mailer that we want to do rewriting relative
9338069Seric **			to.
9348069Seric **		senderaddress -- if set, uses the sender rewriting rules
9358069Seric **			rather than the recipient rewriting rules.
9367682Seric **
9377682Seric **	Returns:
9387682Seric **		the text string representing this address relative to
9397682Seric **			the receiving mailer.
9407682Seric **
9417682Seric **	Side Effects:
9427682Seric **		none.
9437682Seric **
9447682Seric **	Warnings:
9457682Seric **		The text string returned is tucked away locally;
9467682Seric **			copy it if you intend to save it.
9477682Seric */
9487682Seric 
9497682Seric char *
9508069Seric remotename(name, m, senderaddress)
9517682Seric 	char *name;
9527682Seric 	struct mailer *m;
9538069Seric 	bool senderaddress;
9547682Seric {
9558069Seric 	register char **pvp;
9568069Seric 	char *fancy;
9578069Seric 	extern char *macvalue();
9588181Seric 	char *oldg = macvalue('g', CurEnv);
9597682Seric 	static char buf[MAXNAME];
9607682Seric 	char lbuf[MAXNAME];
9617682Seric 	extern char **prescan();
9627889Seric 	extern char *crackaddr();
9637682Seric 
9647755Seric # ifdef DEBUG
9657755Seric 	if (tTd(12, 1))
9667755Seric 		printf("remotename(%s)\n", name);
9677755Seric # endif DEBUG
9687755Seric 
9697682Seric 	/*
9708181Seric 	**  Do a heuristic crack of this name to extract any comment info.
9718181Seric 	**	This will leave the name as a comment and a $g macro.
9727889Seric 	*/
9737889Seric 
9747889Seric 	fancy = crackaddr(name);
9757889Seric 
9768181Seric 	/*
9778181Seric 	**  Turn the name into canonical form.
9788181Seric 	**	Normally this will be RFC 822 style, i.e., "user@domain".
9798181Seric 	**	If this only resolves to "user", and the "C" flag is
9808181Seric 	**	specified in the sending mailer, then the sender's
9818181Seric 	**	domain will be appended.
9828181Seric 	*/
9838181Seric 
9847889Seric 	pvp = prescan(name, '\0');
9857889Seric 	if (pvp == NULL)
9867889Seric 		return (name);
9878181Seric 	rewrite(pvp, 3);
9888181Seric 	if (CurEnv->e_fromdomain != NULL)
9898181Seric 	{
9908181Seric 		/* append from domain to this address */
9918181Seric 		register char **pxp = pvp;
9928181Seric 
9938181Seric 		while (*pxp != NULL && strcmp(*pxp, "@") != 0)
9948181Seric 			pxp++;
9958181Seric 		if (*pxp == NULL)
9968181Seric 		{
9978181Seric 			register char **qxq = CurEnv->e_fromdomain;
9988181Seric 
9998181Seric 			while (*qxq != NULL)
10008181Seric 				*pxp++ = *qxq++;
10018181Seric 		}
10028181Seric 	}
10038181Seric 
10048181Seric 	/*
10058959Seric 	**  Do more specific rewriting.
10068181Seric 	**	Rewrite using ruleset 1 or 2 depending on whether this is
10078181Seric 	**		a sender address or not.
10088181Seric 	**	Then run it through any receiving-mailer-specific rulesets.
10098181Seric 	*/
10108181Seric 
10118069Seric 	if (senderaddress)
10127755Seric 	{
10137889Seric 		rewrite(pvp, 1);
10148069Seric 		if (m->m_s_rwset > 0)
10158069Seric 			rewrite(pvp, m->m_s_rwset);
10168069Seric 	}
10178069Seric 	else
10188069Seric 	{
10197889Seric 		rewrite(pvp, 2);
10208069Seric 		if (m->m_r_rwset > 0)
10218069Seric 			rewrite(pvp, m->m_r_rwset);
10227682Seric 	}
10237682Seric 
10248181Seric 	/*
10258959Seric 	**  Do any final sanitation the address may require.
10268959Seric 	**	This will normally be used to turn internal forms
10278959Seric 	**	(e.g., user@host.LOCAL) into external form.  This
10288959Seric 	**	may be used as a default to the above rules.
10298959Seric 	*/
10308959Seric 
10318959Seric 	rewrite(pvp, 4);
10328959Seric 
10338959Seric 	/*
10348181Seric 	**  Now restore the comment information we had at the beginning.
10358181Seric 	*/
10368181Seric 
10377682Seric 	cataddr(pvp, lbuf, sizeof lbuf);
10389374Seric 	define('g', lbuf, CurEnv);
10397889Seric 	expand(fancy, buf, &buf[sizeof buf - 1], CurEnv);
10409374Seric 	define('g', oldg, CurEnv);
10417682Seric 
10427682Seric # ifdef DEBUG
10437682Seric 	if (tTd(12, 1))
10447755Seric 		printf("remotename => `%s'\n", buf);
10457682Seric # endif DEBUG
10467682Seric 	return (buf);
10477682Seric }
10487682Seric /*
10497682Seric **  CANONNAME -- make name canonical
10507682Seric **
10517682Seric **	This is used for SMTP and misc. printing.  Given a print
10528181Seric **	address, it strips out comments, etc.
10537682Seric **
10547682Seric **	Parameters:
10557682Seric **		name -- the name to make canonical.
10568353Seric **		ruleset -- the canonicalizing ruleset.
10577682Seric **
10587682Seric **	Returns:
10597682Seric **		pointer to canonical name.
10607682Seric **
10617682Seric **	Side Effects:
10627682Seric **		none.
10637682Seric **
10647682Seric **	Warning:
10657682Seric **		result is saved in static buf; future calls will trash it.
10667682Seric */
10677682Seric 
10687682Seric char *
10698353Seric canonname(name, ruleset)
10707682Seric 	char *name;
10718353Seric 	int ruleset;
10727682Seric {
10738069Seric 	static char nbuf[MAXNAME];
10747940Seric 	register char **pvp;
10757682Seric 
10767940Seric 	pvp = prescan(name, '\0');
10778069Seric 	rewrite(pvp, 3);
10788353Seric 	rewrite(pvp, ruleset);
10798069Seric 	cataddr(pvp, nbuf, sizeof nbuf);
10807682Seric 	return (nbuf);
10817682Seric }
1082