main.c revision 1.32 1 /* $NetBSD: main.c,v 1.32 2002/03/05 15:37:35 tv Exp $ */
2 /* $OpenBSD: main.c,v 1.51 2001/10/06 10:52:25 espie 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 #include <sys/cdefs.h>
41 #if defined(__COPYRIGHT) && !defined(lint)
42 __COPYRIGHT("@(#) Copyright (c) 1989, 1993\n\
43 The Regents of the University of California. All rights reserved.\n");
44 #endif /* not lint */
45
46 #if defined(__RCSID) && !defined(lint)
47 #if 0
48 static char sccsid[] = "@(#)main.c 8.1 (Berkeley) 6/6/93";
49 #else
50 __RCSID("$NetBSD: main.c,v 1.32 2002/03/05 15:37:35 tv Exp $");
51 #endif
52 #endif /* not lint */
53
54 /*
55 * main.c
56 * Facility: m4 macro processor
57 * by: oz
58 */
59
60 #include <sys/types.h>
61 #include <assert.h>
62 #include <ctype.h>
63 #include <errno.h>
64 #include <signal.h>
65 #include <stddef.h>
66 #include <stdio.h>
67 #include <stdlib.h>
68 #include <string.h>
69 #include "mdef.h"
70 #include "stdd.h"
71 #include "extern.h"
72 #include "pathnames.h"
73
74 ndptr hashtab[HASHSIZE]; /* hash table for macros etc. */
75 stae *mstack; /* stack of m4 machine */
76 char *sstack; /* shadow stack, for string space extension */
77 static size_t STACKMAX; /* current maximum size of stack */
78 int sp; /* current m4 stack pointer */
79 int fp; /* m4 call frame pointer */
80 struct input_file infile[MAXINP];/* input file stack (0=stdin) */
81 FILE **outfile; /* diversion array(0=bitbucket)*/
82 int maxout;
83 FILE *active; /* active output file pointer */
84 int ilevel = 0; /* input file stack pointer */
85 int oindex = 0; /* diversion index.. */
86 char *null = ""; /* as it says.. just a null.. */
87 char *m4wraps = ""; /* m4wrap string default.. */
88 int m4prefix = 0; /* prefix keywords with m4_ */
89 char lquote[MAXCCHARS+1] = {LQUOTE}; /* left quote character (`) */
90 char rquote[MAXCCHARS+1] = {RQUOTE}; /* right quote character (') */
91 char scommt[MAXCCHARS+1] = {SCOMMT}; /* start character for comment */
92 char ecommt[MAXCCHARS+1] = {ECOMMT}; /* end character for comment */
93
94 struct keyblk keywrds[] = { /* m4 keywords to be installed */
95 { "include", INCLTYPE },
96 { "sinclude", SINCTYPE },
97 { "define", DEFITYPE },
98 { "defn", DEFNTYPE },
99 { "divert", DIVRTYPE | NOARGS },
100 { "expr", EXPRTYPE },
101 { "eval", EXPRTYPE },
102 { "substr", SUBSTYPE },
103 { "ifelse", IFELTYPE },
104 { "ifdef", IFDFTYPE },
105 { "len", LENGTYPE },
106 { "incr", INCRTYPE },
107 { "decr", DECRTYPE },
108 { "dnl", DNLNTYPE | NOARGS },
109 { "changequote", CHNQTYPE | NOARGS },
110 { "changecom", CHNCTYPE | NOARGS },
111 { "index", INDXTYPE },
112 #ifdef EXTENDED
113 { "paste", PASTTYPE },
114 { "spaste", SPASTYPE },
115 /* Newer extensions, needed to handle gnu-m4 scripts */
116 { "indir", INDIRTYPE},
117 { "builtin", BUILTINTYPE},
118 { "patsubst", PATSTYPE},
119 { "regexp", REGEXPTYPE},
120 { "esyscmd", ESYSCMDTYPE},
121 { "__file__", FILENAMETYPE | NOARGS},
122 { "__line__", LINETYPE | NOARGS},
123 #endif
124 { "popdef", POPDTYPE },
125 { "pushdef", PUSDTYPE },
126 { "dumpdef", DUMPTYPE | NOARGS },
127 { "shift", SHIFTYPE | NOARGS },
128 { "translit", TRNLTYPE },
129 { "undefine", UNDFTYPE },
130 { "undivert", UNDVTYPE | NOARGS },
131 { "divnum", DIVNTYPE | NOARGS },
132 { "maketemp", MKTMTYPE },
133 { "errprint", ERRPTYPE | NOARGS },
134 { "m4wrap", M4WRTYPE | NOARGS },
135 { "m4exit", EXITTYPE | NOARGS },
136 { "syscmd", SYSCTYPE },
137 { "sysval", SYSVTYPE | NOARGS },
138 { "traceon", TRACEONTYPE | NOARGS },
139 { "traceoff", TRACEOFFTYPE | NOARGS },
140
141 #if defined(unix) || defined(__unix__)
142 { "unix", SELFTYPE | NOARGS },
143 #else
144 #ifdef vms
145 { "vms", SELFTYPE | NOARGS },
146 #endif
147 #endif
148 };
149
150 #define MAXKEYS (sizeof(keywrds)/sizeof(struct keyblk))
151
152 extern int optind;
153 extern char *optarg;
154
155 #define MAXRECORD 50
156 static struct position {
157 char *name;
158 unsigned long line;
159 } quotes[MAXRECORD], paren[MAXRECORD];
160
161 static void record __P((struct position *, int));
162 static void dump_stack __P((struct position *, int));
163
164 static void macro __P((void));
165 static void initkwds __P((void));
166 static ndptr inspect __P((int, char *));
167 static int do_look_ahead __P((int, const char *));
168
169 static void enlarge_stack __P((void));
170
171 int main __P((int, char *[]));
172
173 int
174 main(argc,argv)
175 int argc;
176 char *argv[];
177 {
178 int c;
179 int n;
180 char *p;
181
182 if (signal(SIGINT, SIG_IGN) != SIG_IGN)
183 signal(SIGINT, onintr);
184
185 /*
186 * We need to know if -P is there before checking -D and -U.
187 */
188 while ((c = getopt(argc, argv, "D:I:PU:d:go:t:")) != -1)
189 if (c == 'P')
190 m4prefix = 1;
191 optind = 1;
192
193 initkwds();
194 initspaces();
195 STACKMAX = INITSTACKMAX;
196
197 mstack = (stae *)xalloc(sizeof(stae) * STACKMAX);
198 sstack = (char *)xalloc(STACKMAX);
199
200 maxout = 0;
201 outfile = NULL;
202 resizedivs(MAXOUT);
203
204 while ((c = getopt(argc, argv, "D:I:PU:d:go:t:")) != -1)
205 switch (c) {
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 'P':
218 break;
219 case 'U': /* undefine... */
220 remhash(optarg, TOP);
221 break;
222 case 'd':
223 set_trace_flags(optarg);
224 break;
225 case 'g':
226 mimic_gnu = 1;
227 break;
228 case 'o':
229 trace_file(optarg);
230 break;
231 case 't':
232 mark_traced(optarg, 1);
233 break;
234 case '?':
235 default:
236 usage(argv[0]);
237 }
238
239 argc -= optind;
240 argv += optind;
241
242 active = stdout; /* default active output */
243 bbase[0] = bufbase;
244 if (!argc) {
245 sp = -1; /* stack pointer initialized */
246 fp = 0; /* frame pointer initialized */
247 set_input(infile+0, stdin, "stdin");
248 /* default input (naturally) */
249 macro();
250 } else
251 for (; argc--; ++argv) {
252 p = *argv;
253 if (p[0] == '-' && p[1] == EOS)
254 set_input(infile, stdin, "stdin");
255 else if (fopen_trypath(infile, p) == NULL)
256 err(1, "%s", p);
257 sp = -1;
258 fp = 0;
259 macro();
260 release_input(infile);
261 }
262
263 if (*m4wraps) { /* anything for rundown ?? */
264 ilevel = 0; /* in case m4wrap includes.. */
265 bufbase = bp = buf; /* use the entire buffer */
266 pbstr(m4wraps); /* user-defined wrapup act */
267 macro(); /* last will and testament */
268 }
269
270 if (active != stdout)
271 active = stdout; /* reset output just in case */
272 for (n = 1; n < maxout; n++) /* default wrap-up: undivert */
273 if (outfile[n] != NULL)
274 getdiv(n);
275 /* remove bitbucket if used */
276 if (outfile[0] != NULL) {
277 (void) fclose(outfile[0]);
278 }
279
280 return 0;
281 }
282
283 /*
284 * Look ahead for `token'.
285 * (on input `t == token[0]')
286 * Used for comment and quoting delimiters.
287 * Returns 1 if `token' present; copied to output.
288 * 0 if `token' not found; all characters pushed back
289 */
290 static int
291 do_look_ahead(t, token)
292 int t;
293 const char *token;
294 {
295 int i;
296
297 assert((unsigned char)t == (unsigned char)token[0]);
298
299 for (i = 1; *++token; i++) {
300 t = gpbc();
301 if (t == EOF || (unsigned char)t != (unsigned char)*token) {
302 putback(t);
303 while (--i)
304 putback(*--token);
305 return 0;
306 }
307 }
308 return 1;
309 }
310
311 #define LOOK_AHEAD(t, token) (t != EOF && \
312 (unsigned char)(t)==(unsigned char)(token)[0] && \
313 do_look_ahead(t,token))
314
315 /*
316 * macro - the work horse..
317 */
318 static void
319 macro()
320 {
321 char token[MAXTOK+1];
322 int t, l;
323 ndptr p;
324 int nlpar;
325
326 cycle {
327 t = gpbc();
328 if (t == '_' || isalpha(t)) {
329 p = inspect(t, token);
330 if (p != nil)
331 putback(l = gpbc());
332 if (p == nil || (l != LPAREN &&
333 (p->type & NEEDARGS) != 0))
334 outputstr(token);
335 else {
336 /*
337 * real thing.. First build a call frame:
338 */
339 pushf(fp); /* previous call frm */
340 pushf(p->type); /* type of the call */
341 pushf(0); /* parenthesis level */
342 fp = sp; /* new frame pointer */
343 /*
344 * now push the string arguments:
345 */
346 pushs1(p->defn); /* defn string */
347 pushs1(p->name); /* macro name */
348 pushs(ep); /* start next..*/
349
350 if (l != LPAREN && PARLEV == 0) {
351 /* no bracks */
352 chrsave(EOS);
353
354 if (sp == STACKMAX)
355 errx(1, "internal stack overflow");
356 eval((const char **) mstack+fp+1, 2,
357 CALTYP);
358
359 ep = PREVEP; /* flush strspace */
360 sp = PREVSP; /* previous sp.. */
361 fp = PREVFP; /* rewind stack...*/
362 }
363 }
364 } else if (t == EOF) {
365 if (sp > -1) {
366 warnx( "unexpected end of input, unclosed parenthesis:");
367 dump_stack(paren, PARLEV);
368 exit(1);
369 }
370 if (ilevel <= 0)
371 break; /* all done thanks.. */
372 release_input(infile+ilevel--);
373 bufbase = bbase[ilevel];
374 continue;
375 }
376 /*
377 * non-alpha token possibly seen..
378 * [the order of else if .. stmts is important.]
379 */
380 else if (LOOK_AHEAD(t,lquote)) { /* strip quotes */
381 nlpar = 0;
382 record(quotes, nlpar++);
383 /*
384 * Opening quote: scan forward until matching
385 * closing quote has been found.
386 */
387 do {
388
389 l = gpbc();
390 if (LOOK_AHEAD(l,rquote)) {
391 if (--nlpar > 0)
392 outputstr(rquote);
393 } else if (LOOK_AHEAD(l,lquote)) {
394 record(quotes, nlpar++);
395 outputstr(lquote);
396 } else if (l == EOF) {
397 if (nlpar == 1)
398 warnx("unclosed quote:");
399 else
400 warnx("%d unclosed quotes:", nlpar);
401 dump_stack(quotes, nlpar);
402 exit(1);
403 } else {
404 if (nlpar > 0) {
405 if (sp < 0)
406 putc(l, active);
407 else
408 CHRSAVE(l);
409 }
410 }
411 }
412 while (nlpar != 0);
413 }
414
415 else if (sp < 0 && LOOK_AHEAD(t, scommt)) {
416 fputs(scommt, active);
417
418 for(;;) {
419 t = gpbc();
420 if (LOOK_AHEAD(t, ecommt)) {
421 fputs(ecommt, active);
422 break;
423 }
424 if (t == EOF)
425 break;
426 putc(t, active);
427 }
428 }
429
430 else if (sp < 0) { /* not in a macro at all */
431 putc(t, active); /* output directly.. */
432 }
433
434 else switch(t) {
435
436 case LPAREN:
437 if (PARLEV > 0)
438 chrsave(t);
439 while (isspace(l = gpbc()))
440 ; /* skip blank, tab, nl.. */
441 putback(l);
442 record(paren, PARLEV++);
443 break;
444
445 case RPAREN:
446 if (--PARLEV > 0)
447 chrsave(t);
448 else { /* end of argument list */
449 chrsave(EOS);
450
451 if (sp == STACKMAX)
452 errx(1, "internal stack overflow");
453
454 eval((const char **) mstack+fp+1, sp-fp,
455 CALTYP);
456
457 ep = PREVEP; /* flush strspace */
458 sp = PREVSP; /* previous sp.. */
459 fp = PREVFP; /* rewind stack...*/
460 }
461 break;
462
463 case COMMA:
464 if (PARLEV == 1) {
465 chrsave(EOS); /* new argument */
466 while (isspace(l = gpbc()))
467 ;
468 putback(l);
469 pushs(ep);
470 } else
471 chrsave(t);
472 break;
473
474 default:
475 if (LOOK_AHEAD(t, scommt)) {
476 char *p;
477 for (p = scommt; *p; p++)
478 chrsave(*p);
479 for(;;) {
480 t = gpbc();
481 if (LOOK_AHEAD(t, ecommt)) {
482 for (p = ecommt; *p; p++)
483 chrsave(*p);
484 break;
485 }
486 if (t == EOF)
487 break;
488 CHRSAVE(t);
489 }
490 } else
491 CHRSAVE(t); /* stack the char */
492 break;
493 }
494 }
495 }
496
497 /*
498 * output string directly, without pushing it for reparses.
499 */
500 void
501 outputstr(s)
502 const char *s;
503 {
504 if (sp < 0)
505 while (*s)
506 putc(*s++, active);
507 else
508 while (*s)
509 CHRSAVE(*s++);
510 }
511
512 /*
513 * build an input token..
514 * consider only those starting with _ or A-Za-z. This is a
515 * combo with lookup to speed things up.
516 */
517 static ndptr
518 inspect(c, tp)
519 int c;
520 char *tp;
521 {
522 char *name = tp;
523 char *etp = tp+MAXTOK;
524 ndptr p;
525 unsigned int h;
526
527 h = *tp++ = c;
528
529 while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
530 h = (h << 5) + h + (*tp++ = c);
531 if (c != EOF)
532 PUTBACK(c);
533 *tp = EOS;
534 /* token is too long, it won't match anything, but it can still
535 * be output. */
536 if (tp == ep) {
537 outputstr(name);
538 while (isalnum(c = gpbc()) || c == '_') {
539 if (sp < 0)
540 putc(c, active);
541 else
542 CHRSAVE(c);
543 }
544 *name = EOS;
545 return nil;
546 }
547
548 for (p = hashtab[h % HASHSIZE]; p != nil; p = p->nxtptr)
549 if (h == p->hv && STREQ(name, p->name))
550 break;
551 return p;
552 }
553
554 /*
555 * initkwds - initialise m4 keywords as fast as possible.
556 * This very similar to install, but without certain overheads,
557 * such as calling lookup. Malloc is not used for storing the
558 * keyword strings, since we simply use the static pointers
559 * within keywrds block.
560 */
561 static void
562 initkwds()
563 {
564 size_t i;
565 unsigned int h;
566 ndptr p;
567 char *k;
568
569 for (i = 0; i < MAXKEYS; i++) {
570 k = (char *)keywrds[i].knam;
571 if (m4prefix) {
572 size_t klen = strlen(k);
573 char *newk = malloc(klen + 4);
574
575 if (snprintf(newk, klen, "m4_%s", k) == -1)
576 err(1, "snprintf");
577 keywrds[i].knam = newk;
578 k = newk;
579 }
580 h = hash(k);
581 p = (ndptr) xalloc(sizeof(struct ndblock));
582 p->nxtptr = hashtab[h % HASHSIZE];
583 hashtab[h % HASHSIZE] = p;
584 p->name = xstrdup(keywrds[i].knam);
585 p->defn = null;
586 p->hv = h;
587 p->type = keywrds[i].ktyp & TYPEMASK;
588 if ((keywrds[i].ktyp & NOARGS) == 0)
589 p->type |= NEEDARGS;
590 }
591 }
592
593 /* Look up a builtin type, even if overridden by the user */
594 int
595 builtin_type(key)
596 const char *key;
597 {
598 int i;
599
600 for (i = 0; i != MAXKEYS; i++)
601 if (STREQ(keywrds[i].knam, key))
602 return keywrds[i].ktyp;
603 return -1;
604 }
605
606 const char *
607 builtin_realname(n)
608 int n;
609 {
610 int i;
611
612 for (i = 0; i != MAXKEYS; i++)
613 if (((keywrds[i].ktyp ^ n) & TYPEMASK) == 0)
614 return keywrds[i].knam;
615 return NULL;
616 }
617
618 static void
619 record(t, lev)
620 struct position *t;
621 int lev;
622 {
623 if (lev < MAXRECORD) {
624 t[lev].name = CURRENT_NAME;
625 t[lev].line = CURRENT_LINE;
626 }
627 }
628
629 static void
630 dump_stack(t, lev)
631 struct position *t;
632 int lev;
633 {
634 int i;
635
636 for (i = 0; i < lev; i++) {
637 if (i == MAXRECORD) {
638 fprintf(stderr, " ...\n");
639 break;
640 }
641 fprintf(stderr, " %s at line %lu\n",
642 t[i].name, t[i].line);
643 }
644 }
645
646
647 static void
648 enlarge_stack()
649 {
650 STACKMAX *= 2;
651 mstack = realloc(mstack, sizeof(stae) * STACKMAX);
652 sstack = realloc(sstack, STACKMAX);
653 if (mstack == NULL || sstack == NULL)
654 errx(1, "Evaluation stack overflow (%lu)",
655 (unsigned long)STACKMAX);
656 }
657