main.c revision 97296 
1/*	$OpenBSD: main.c,v 1.53 2002/04/26 16:15:16 espie Exp $	*/
2/*	$NetBSD: main.c,v 1.12 1997/02/08 23:54:49 cgd Exp $	*/
3
4/*-
5 * Copyright (c) 1989, 1993
6 *	The Regents of the University of California.  All rights reserved.
7 *
8 * This code is derived from software contributed to Berkeley by
9 * Ozan Yigit at York University.
10 *
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. All advertising materials mentioning features or use of this software
20 *    must display the following acknowledgement:
21 *	This product includes software developed by the University of
22 *	California, Berkeley and its contributors.
23 * 4. Neither the name of the University nor the names of its contributors
24 *    may be used to endorse or promote products derived from this software
25 *    without specific prior written permission.
26 *
27 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
28 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
29 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
30 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
31 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
32 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
33 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
34 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
35 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
36 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
37 * SUCH DAMAGE.
38 */
39
40#ifndef lint
41static char copyright[] =
42"@(#) Copyright (c) 1989, 1993\n\
43	The Regents of the University of California.  All rights reserved.\n";
44#endif /* not lint */
45
46#ifndef lint
47#if 0
48static char sccsid[] = "@(#)main.c	8.1 (Berkeley) 6/6/93";
49#else
50#if 0
51static char rcsid[] = "$OpenBSD: main.c,v 1.53 2002/04/26 16:15:16 espie Exp $";
52#endif
53#endif
54#endif /* not lint */
55
56#include <sys/cdefs.h>
57__FBSDID("$FreeBSD: head/usr.bin/m4/main.c 97296 2002-05-26 06:15:15Z tjr $");
58
59/*
60 * main.c
61 * Facility: m4 macro processor
62 * by: oz
63 */
64
65#include <sys/types.h>
66#include <assert.h>
67#include <signal.h>
68#include <errno.h>
69#include <unistd.h>
70#include <stdio.h>
71#include <ctype.h>
72#include <string.h>
73#include <stddef.h>
74#include <stdlib.h>
75#include <err.h>
76#include "mdef.h"
77#include "stdd.h"
78#include "extern.h"
79#include "pathnames.h"
80
81ndptr hashtab[HASHSIZE];	/* hash table for macros etc.  */
82stae *mstack;		 	/* stack of m4 machine         */
83char *sstack;		 	/* shadow stack, for string space extension */
84static size_t STACKMAX;		/* current maximum size of stack */
85int sp; 			/* current m4  stack pointer   */
86int fp; 			/* m4 call frame pointer       */
87struct input_file infile[MAXINP];/* input file stack (0=stdin)  */
88char *inname[MAXINP];		/* names of these input files */
89int inlineno[MAXINP];		/* current number in each input file */
90FILE **outfile;			/* diversion array(0=bitbucket)*/
91int maxout;
92FILE *active;			/* active output file pointer  */
93int ilevel = 0; 		/* input file stack pointer    */
94int oindex = 0; 		/* diversion index..	       */
95const char *null = "";          /* as it says.. just a null..  */
96const char *m4wraps = "";       /* m4wrap string default..     */
97char lquote[MAXCCHARS+1] = {LQUOTE};	/* left quote character  (`)   */
98char rquote[MAXCCHARS+1] = {RQUOTE};	/* right quote character (')   */
99char scommt[MAXCCHARS+1] = {SCOMMT};	/* start character for comment */
100char ecommt[MAXCCHARS+1] = {ECOMMT};	/* end character for comment   */
101int synccpp;			/* Line synchronisation for C preprocessor */
102
103struct keyblk keywrds[] = {	/* m4 keywords to be installed */
104	{ "include",      INCLTYPE },
105	{ "sinclude",     SINCTYPE },
106	{ "define",       DEFITYPE },
107	{ "defn",         DEFNTYPE },
108	{ "divert",       DIVRTYPE | NOARGS },
109	{ "expr",         EXPRTYPE },
110	{ "eval",         EXPRTYPE },
111	{ "substr",       SUBSTYPE },
112	{ "ifelse",       IFELTYPE },
113	{ "ifdef",        IFDFTYPE },
114	{ "len",          LENGTYPE },
115	{ "incr",         INCRTYPE },
116	{ "decr",         DECRTYPE },
117	{ "dnl",          DNLNTYPE | NOARGS },
118	{ "changequote",  CHNQTYPE | NOARGS },
119	{ "changecom",    CHNCTYPE | NOARGS },
120	{ "index",        INDXTYPE },
121#ifdef EXTENDED
122	{ "paste",        PASTTYPE },
123	{ "spaste",       SPASTYPE },
124    	/* Newer extensions, needed to handle gnu-m4 scripts */
125	{ "indir",        INDIRTYPE},
126	{ "builtin",      BUILTINTYPE},
127	{ "patsubst",	  PATSTYPE},
128	{ "regexp",	  REGEXPTYPE},
129	{ "esyscmd",	  ESYSCMDTYPE},
130	{ "__file__",	  FILENAMETYPE | NOARGS},
131	{ "__line__",	  LINETYPE | NOARGS},
132#endif
133	{ "popdef",       POPDTYPE },
134	{ "pushdef",      PUSDTYPE },
135	{ "dumpdef",      DUMPTYPE | NOARGS },
136	{ "shift",        SHIFTYPE | NOARGS },
137	{ "translit",     TRNLTYPE },
138	{ "undefine",     UNDFTYPE },
139	{ "undivert",     UNDVTYPE | NOARGS },
140	{ "divnum",       DIVNTYPE | NOARGS },
141	{ "maketemp",     MKTMTYPE },
142	{ "errprint",     ERRPTYPE | NOARGS },
143	{ "m4wrap",       M4WRTYPE | NOARGS },
144	{ "m4exit",       EXITTYPE | NOARGS },
145	{ "syscmd",       SYSCTYPE },
146	{ "sysval",       SYSVTYPE | NOARGS },
147	{ "traceon",	  TRACEONTYPE | NOARGS },
148	{ "traceoff",	  TRACEOFFTYPE | NOARGS },
149
150#if defined(unix) || defined(__unix__)
151	{ "unix",         SELFTYPE | NOARGS },
152#else
153#ifdef vms
154	{ "vms",          SELFTYPE | NOARGS },
155#endif
156#endif
157};
158
159#define MAXKEYS	(sizeof(keywrds)/sizeof(struct keyblk))
160
161#define MAXRECORD 50
162static struct position {
163	char *name;
164	unsigned long line;
165} quotes[MAXRECORD], paren[MAXRECORD];
166
167static void record(struct position *, int);
168static void dump_stack(struct position *, int);
169
170static void macro(void);
171static void initkwds(void);
172static ndptr inspect(int, char *);
173static int do_look_ahead(int, const char *);
174
175static void enlarge_stack(void);
176
177int main(int, char *[]);
178
179int
180main(int argc, char *argv[])
181{
182	int c;
183	int n;
184	int rval;
185	char *p;
186
187	traceout = stderr;
188
189	if (signal(SIGINT, SIG_IGN) != SIG_IGN)
190		signal(SIGINT, onintr);
191
192	initkwds();
193	initspaces();
194	STACKMAX = INITSTACKMAX;
195
196	mstack = (stae *)xalloc(sizeof(stae) * STACKMAX);
197	sstack = (char *)xalloc(STACKMAX);
198
199	maxout = 0;
200	outfile = NULL;
201	resizedivs(MAXOUT);
202
203	while ((c = getopt(argc, argv, "gst:d:D:U:o:I:")) != -1)
204		switch(c) {
205
206		case 'D':               /* define something..*/
207			for (p = optarg; *p; p++)
208				if (*p == '=')
209					break;
210			if (*p)
211				*p++ = EOS;
212			dodefine(optarg, p);
213			break;
214		case 'I':
215			addtoincludepath(optarg);
216			break;
217		case 'U':               /* undefine...       */
218			remhash(optarg, TOP);
219			break;
220		case 'g':
221			mimic_gnu = 1;
222			break;
223		case 'd':
224			set_trace_flags(optarg);
225			break;
226		case 's':
227			synccpp = 1;
228			break;
229		case 't':
230			mark_traced(optarg, 1);
231			break;
232		case 'o':
233			trace_file(optarg);
234                        break;
235		case '?':
236			usage();
237		}
238
239        argc -= optind;
240        argv += optind;
241
242	rval = 0;
243	active = stdout;		/* default active output     */
244	bbase[0] = bufbase;
245        if (!argc) {
246 		sp = -1;		/* stack pointer initialized */
247		fp = 0; 		/* frame pointer initialized */
248		set_input(infile+0, stdin, "stdin");
249					/* default input (naturally) */
250		if ((inname[0] = strdup("-")) == NULL)
251			err(1, NULL);
252		inlineno[0] = 1;
253		emitline();
254		macro();
255	} else
256		for (; argc--; ++argv) {
257			p = *argv;
258			if (p[0] == '-' && p[1] == EOS)
259				set_input(infile, stdin, "stdin");
260			else if (fopen_trypath(infile, p) == NULL) {
261				warn("%s", p);
262				rval = 1;
263				continue;
264			}
265			sp = -1;
266			fp = 0;
267			if ((inname[0] = strdup(p)) == NULL)
268				err(1, NULL);
269			inlineno[0] = 1;
270			emitline();
271			macro();
272		    	release_input(infile);
273		}
274
275	if (*m4wraps) { 		/* anything for rundown ??   */
276		ilevel = 0;		/* in case m4wrap includes.. */
277		bufbase = bp = buf;	/* use the entire buffer   */
278		pbstr(m4wraps); 	/* user-defined wrapup act   */
279		macro();		/* last will and testament   */
280	}
281
282	if (active != stdout)
283		active = stdout;	/* reset output just in case */
284	for (n = 1; n < maxout; n++)	/* default wrap-up: undivert */
285		if (outfile[n] != NULL)
286			getdiv(n);
287					/* remove bitbucket if used  */
288	if (outfile[0] != NULL) {
289		(void) fclose(outfile[0]);
290	}
291
292	exit(rval);
293}
294
295/*
296 * Look ahead for `token'.
297 * (on input `t == token[0]')
298 * Used for comment and quoting delimiters.
299 * Returns 1 if `token' present; copied to output.
300 *         0 if `token' not found; all characters pushed back
301 */
302static int
303do_look_ahead(int t, const char *token)
304{
305	int i;
306
307	assert((unsigned char)t == (unsigned char)token[0]);
308
309	for (i = 1; *++token; i++) {
310		t = gpbc();
311		if (t == EOF || (unsigned char)t != (unsigned char)*token) {
312			putback(t);
313			while (--i)
314				putback(*--token);
315			return 0;
316		}
317	}
318	return 1;
319}
320
321#define LOOK_AHEAD(t, token) (t != EOF && 		\
322    (unsigned char)(t)==(unsigned char)(token)[0] && 	\
323    do_look_ahead(t,token))
324
325/*
326 * macro - the work horse..
327 */
328static void
329macro()
330{
331	char token[MAXTOK+1];
332	int t, l;
333	ndptr p;
334	int  nlpar;
335
336	cycle {
337		t = gpbc();
338		if (t == '_' || isalpha(t)) {
339			p = inspect(t, token);
340			if (p != nil)
341				putback(l = gpbc());
342			if (p == nil || (l != LPAREN &&
343			    (p->type & NEEDARGS) != 0))
344				outputstr(token);
345			else {
346		/*
347		 * real thing.. First build a call frame:
348		 */
349				pushf(fp);	/* previous call frm */
350				pushf(p->type); /* type of the call  */
351				pushf(0);	/* parenthesis level */
352				fp = sp;	/* new frame pointer */
353		/*
354		 * now push the string arguments:
355		 */
356				pushs1(p->defn);	/* defn string */
357				pushs1(p->name);	/* macro name  */
358				pushs(ep);	      	/* start next..*/
359
360				if (l != LPAREN && PARLEV == 0)  {
361				    /* no bracks  */
362					chrsave(EOS);
363
364					if (sp == STACKMAX)
365						errx(1, "internal stack overflow");
366					eval((const char **) mstack+fp+1, 2,
367					    CALTYP);
368
369					ep = PREVEP;	/* flush strspace */
370					sp = PREVSP;	/* previous sp..  */
371					fp = PREVFP;	/* rewind stack...*/
372				}
373			}
374		} else if (t == EOF) {
375			if (sp > -1) {
376				warnx( "unexpected end of input, unclosed parenthesis:");
377				dump_stack(paren, PARLEV);
378				exit(1);
379			}
380			if (ilevel <= 0)
381				break;			/* all done thanks.. */
382			release_input(infile+ilevel--);
383			free(inname[ilevel+1]);
384			bufbase = bbase[ilevel];
385			emitline();
386			continue;
387		}
388	/*
389	 * non-alpha token possibly seen..
390	 * [the order of else if .. stmts is important.]
391	 */
392		else if (LOOK_AHEAD(t,lquote)) {	/* strip quotes */
393			nlpar = 0;
394			record(quotes, nlpar++);
395			/*
396			 * Opening quote: scan forward until matching
397			 * closing quote has been found.
398			 */
399			do {
400
401				l = gpbc();
402				if (LOOK_AHEAD(l,rquote)) {
403					if (--nlpar > 0)
404						outputstr(rquote);
405				} else if (LOOK_AHEAD(l,lquote)) {
406					record(quotes, nlpar++);
407					outputstr(lquote);
408				} else if (l == EOF) {
409					if (nlpar == 1)
410						warnx("unclosed quote:");
411					else
412						warnx("%d unclosed quotes:", nlpar);
413					dump_stack(quotes, nlpar);
414					exit(1);
415				} else {
416					if (nlpar > 0) {
417						if (sp < 0)
418							putc(l, active);
419						else
420							CHRSAVE(l);
421					}
422				}
423			}
424			while (nlpar != 0);
425		}
426
427		else if (sp < 0 && LOOK_AHEAD(t, scommt)) {
428			fputs(scommt, active);
429
430			for(;;) {
431				t = gpbc();
432				if (LOOK_AHEAD(t, ecommt)) {
433					fputs(ecommt, active);
434					break;
435				}
436				if (t == EOF)
437					break;
438				putc(t, active);
439			}
440		}
441
442		else if (sp < 0) {		/* not in a macro at all */
443			putc(t, active);	/* output directly..	 */
444		}
445
446		else switch(t) {
447
448		case LPAREN:
449			if (PARLEV > 0)
450				chrsave(t);
451			while (isspace(l = gpbc()))
452				;		/* skip blank, tab, nl.. */
453			putback(l);
454			record(paren, PARLEV++);
455			break;
456
457		case RPAREN:
458			if (--PARLEV > 0)
459				chrsave(t);
460			else {			/* end of argument list */
461				chrsave(EOS);
462
463				if (sp == STACKMAX)
464					errx(1, "internal stack overflow");
465
466				eval((const char **) mstack+fp+1, sp-fp,
467				    CALTYP);
468
469				ep = PREVEP;	/* flush strspace */
470				sp = PREVSP;	/* previous sp..  */
471				fp = PREVFP;	/* rewind stack...*/
472			}
473			break;
474
475		case COMMA:
476			if (PARLEV == 1) {
477				chrsave(EOS);		/* new argument   */
478				while (isspace(l = gpbc()))
479					;
480				putback(l);
481				pushs(ep);
482			} else
483				chrsave(t);
484			break;
485
486		default:
487			if (LOOK_AHEAD(t, scommt)) {
488				char *pc;
489				for (pc = scommt; *pc; pc++)
490					chrsave(*pc);
491				for(;;) {
492					t = gpbc();
493					if (LOOK_AHEAD(t, ecommt)) {
494						for (pc = ecommt; *pc; pc++)
495							chrsave(*pc);
496						break;
497					}
498					if (t == EOF)
499					    break;
500					CHRSAVE(t);
501				}
502			} else
503				CHRSAVE(t);		/* stack the char */
504			break;
505		}
506	}
507}
508
509/*
510 * output string directly, without pushing it for reparses.
511 */
512void
513outputstr(const char *s)
514{
515	if (sp < 0)
516		while (*s)
517			putc(*s++, active);
518	else
519		while (*s)
520			CHRSAVE(*s++);
521}
522
523/*
524 * build an input token..
525 * consider only those starting with _ or A-Za-z. This is a
526 * combo with lookup to speed things up.
527 */
528static ndptr
529inspect(int c, char *tp)
530{
531	char *name = tp;
532	char *etp = tp+MAXTOK;
533	ndptr p;
534	unsigned int h;
535
536	h = *tp++ = c;
537
538	while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
539		h = (h << 5) + h + (*tp++ = c);
540	if (c != EOF)
541		PUTBACK(c);
542	*tp = EOS;
543	/* token is too long, it won't match anything, but it can still
544	 * be output. */
545	if (tp == ep) {
546		outputstr(name);
547		while (isalnum(c = gpbc()) || c == '_') {
548			if (sp < 0)
549				putc(c, active);
550			else
551				CHRSAVE(c);
552		}
553		*name = EOS;
554		return nil;
555	}
556
557	for (p = hashtab[h % HASHSIZE]; p != nil; p = p->nxtptr)
558		if (h == p->hv && STREQ(name, p->name))
559			break;
560	return p;
561}
562
563/*
564 * initkwds - initialise m4 keywords as fast as possible.
565 * This very similar to install, but without certain overheads,
566 * such as calling lookup. Malloc is not used for storing the
567 * keyword strings, since we simply use the static pointers
568 * within keywrds block.
569 */
570static void
571initkwds()
572{
573	size_t i;
574	unsigned int h;
575	ndptr p;
576
577	for (i = 0; i < MAXKEYS; i++) {
578		h = hash(keywrds[i].knam);
579		p = (ndptr) xalloc(sizeof(struct ndblock));
580		p->nxtptr = hashtab[h % HASHSIZE];
581		hashtab[h % HASHSIZE] = p;
582		p->name = xstrdup(keywrds[i].knam);
583		p->defn = xstrdup(null);
584		p->hv = h;
585		p->type = keywrds[i].ktyp & TYPEMASK;
586		if ((keywrds[i].ktyp & NOARGS) == 0)
587			p->type |= NEEDARGS;
588	}
589}
590
591/* Look up a builtin type, even if overridden by the user */
592int
593builtin_type(const char *key)
594{
595	int i;
596
597	for (i = 0; i != MAXKEYS; i++)
598		if (STREQ(keywrds[i].knam, key))
599			return keywrds[i].ktyp;
600	return -1;
601}
602
603const char *
604builtin_realname(int n)
605{
606	int i;
607
608	for (i = 0; i != MAXKEYS; i++)
609		if (((keywrds[i].ktyp ^ n) & TYPEMASK) == 0)
610			return keywrds[i].knam;
611	return NULL;
612}
613
614static void
615record(struct position *t, int lev)
616{
617	if (lev < MAXRECORD) {
618		t[lev].name = CURRENT_NAME;
619		t[lev].line = CURRENT_LINE;
620	}
621}
622
623static void
624dump_stack(struct position *t, int lev)
625{
626	int i;
627
628	for (i = 0; i < lev; i++) {
629		if (i == MAXRECORD) {
630			fprintf(stderr, "   ...\n");
631			break;
632		}
633		fprintf(stderr, "   %s at line %lu\n",
634			t[i].name, t[i].line);
635	}
636}
637
638
639static void
640enlarge_stack()
641{
642	STACKMAX *= 2;
643	mstack = realloc(mstack, sizeof(stae) * STACKMAX);
644	sstack = realloc(sstack, STACKMAX);
645	if (mstack == NULL || sstack == NULL)
646		errx(1, "Evaluation stack overflow (%lu)",
647		    (unsigned long)STACKMAX);
648}
649
650/* Emit preprocessor #line directive if -s option used. */
651void
652emitline(void)
653{
654
655	if (synccpp)
656		fprintf(active, "#line %d \"%s\"\n", inlineno[ilevel],
657			inname[ilevel]);
658}
659