1 /* $OpenBSD: main.c,v 1.77 2009/10/14 17:19:47 sthen 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. Neither the name of the University nor the names of its contributors
20 * may be used to endorse or promote products derived from this software
21 * without specific prior written permission.
22 *
23 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
24 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
25 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
26 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
27 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
28 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
29 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
30 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
31 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
32 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33 * SUCH DAMAGE.
34 */
35
36 /*
37 * main.c
38 * Facility: m4 macro processor
39 * by: oz
40 */
41
42 #include <assert.h>
43 #include <signal.h>
44 #include <err.h>
45 #include <errno.h>
46 #include <unistd.h>
47 #include <stdio.h>
48 #include <ctype.h>
49 #include <string.h>
50 #include <stddef.h>
51 #include <stdint.h>
52 #include <stdlib.h>
53 #include <ohash.h>
54 #include "mdef.h"
55 #include "stdd.h"
56 #include "extern.h"
57 #include "pathnames.h"
58
59 ndptr hashtab[HASHSIZE]; /* hash table for macros etc. */
60 stae *mstack; /* stack of m4 machine */
61 char *sstack; /* shadow stack, for string space extension */
62 static size_t STACKMAX; /* current maximum size of stack */
63 int sp; /* current m4 stack pointer */
64 int fp; /* m4 call frame pointer */
65 struct input_file infile[MAXINP];/* input file stack (0=stdin) */
66 FILE **outfile; /* diversion array(0=bitbucket)*/
67 int maxout;
68 FILE *active; /* active output file pointer */
69 int ilevel = 0; /* input file stack pointer */
70 int oindex = 0; /* diversion index.. */
71 char *null = ""; /* as it says.. just a null.. */
72 char **m4wraps = NULL; /* m4wraps array. */
73 int maxwraps = 0; /* size of m4wraps array */
74 int wrapindex = 0; /* current offset in m4wraps */
75 char lquote[MAXCCHARS+1] = {LQUOTE}; /* left quote character (`) */
76 char rquote[MAXCCHARS+1] = {RQUOTE}; /* right quote character (') */
77 char scommt[MAXCCHARS+1] = {SCOMMT}; /* start character for comment */
78 char ecommt[MAXCCHARS+1] = {ECOMMT}; /* end character for comment */
79 int synch_lines = 0; /* line synchronisation for C preprocessor */
80 int prefix_builtins = 0; /* -P option to prefix builtin keywords */
81
82 struct keyblk {
83 char *knam; /* keyword name */
84 int ktyp; /* keyword type */
85 };
86
87 struct keyblk keywrds[] = { /* m4 keywords to be installed */
88 { "include", INCLTYPE },
89 { "sinclude", SINCTYPE },
90 { "define", DEFITYPE },
91 { "defn", DEFNTYPE },
92 { "divert", DIVRTYPE | NOARGS },
93 { "expr", EXPRTYPE },
94 { "eval", EXPRTYPE },
95 { "substr", SUBSTYPE },
96 { "ifelse", IFELTYPE },
97 { "ifdef", IFDFTYPE },
98 { "len", LENGTYPE },
99 { "incr", INCRTYPE },
100 { "decr", DECRTYPE },
101 { "dnl", DNLNTYPE | NOARGS },
102 { "changequote", CHNQTYPE | NOARGS },
103 { "changecom", CHNCTYPE | NOARGS },
104 { "index", INDXTYPE },
105 #ifdef EXTENDED
106 { "paste", PASTTYPE },
107 { "spaste", SPASTYPE },
108 /* Newer extensions, needed to handle gnu-m4 scripts */
109 { "indir", INDIRTYPE},
110 { "builtin", BUILTINTYPE},
111 { "patsubst", PATSTYPE},
112 { "regexp", REGEXPTYPE},
113 { "esyscmd", ESYSCMDTYPE},
114 { "__file__", FILENAMETYPE | NOARGS},
115 { "__line__", LINETYPE | NOARGS},
116 #endif
117 { "popdef", POPDTYPE },
118 { "pushdef", PUSDTYPE },
119 { "dumpdef", DUMPTYPE | NOARGS },
120 { "shift", SHIFTYPE | NOARGS },
121 { "translit", TRNLTYPE },
122 { "undefine", UNDFTYPE },
123 { "undivert", UNDVTYPE | NOARGS },
124 { "divnum", DIVNTYPE | NOARGS },
125 { "maketemp", MKTMTYPE },
126 { "errprint", ERRPTYPE | NOARGS },
127 { "m4wrap", M4WRTYPE | NOARGS },
128 { "m4exit", EXITTYPE | NOARGS },
129 { "syscmd", SYSCTYPE },
130 { "sysval", SYSVTYPE | NOARGS },
131 { "traceon", TRACEONTYPE | NOARGS },
132 { "traceoff", TRACEOFFTYPE | NOARGS },
133
134 #if defined(unix) || defined(__unix__)
135 { "unix", SELFTYPE | NOARGS },
136 #else
137 #ifdef vms
138 { "vms", SELFTYPE | NOARGS },
139 #endif
140 #endif
141 };
142
143 #define MAXKEYS (sizeof(keywrds)/sizeof(struct keyblk))
144
145 extern int optind;
146 extern char *optarg;
147
148 #define MAXRECORD 50
149 static struct position {
150 char *name;
151 unsigned long line;
152 } quotes[MAXRECORD], paren[MAXRECORD];
153
154 static void record(struct position *, int);
155 static void dump_stack(struct position *, int);
156
157 static void macro(void);
158 static void initkwds(void);
159 static ndptr inspect(int, char *);
160 static int do_look_ahead(int, const char *);
161 static void reallyoutputstr(const char *);
162 static void reallyputchar(int);
163
164 static void enlarge_stack(void);
165
166 int main(int, char *[]);
167
168 int
main(int argc,char * argv[])169 main(int argc, char *argv[])
170 {
171 int c;
172 int n;
173 char *p;
174
175 if (signal(SIGINT, SIG_IGN) != SIG_IGN)
176 signal(SIGINT, onintr);
177
178 init_macros();
179 initspaces();
180 STACKMAX = INITSTACKMAX;
181
182 mstack = (stae *)xalloc(sizeof(stae) * STACKMAX, NULL);
183 sstack = (char *)xalloc(STACKMAX, NULL);
184
185 maxout = 0;
186 outfile = NULL;
187 resizedivs(MAXOUT);
188
189 while ((c = getopt(argc, argv, "gst:d:D:U:o:I:P")) != -1)
190 switch(c) {
191
192 case 'D': /* define something..*/
193 for (p = optarg; *p; p++)
194 if (*p == '=')
195 break;
196 if (*p)
197 *p++ = EOS;
198 dodefine(optarg, p);
199 break;
200 case 'I':
201 addtoincludepath(optarg);
202 break;
203 case 'P':
204 prefix_builtins = 1;
205 break;
206 case 'U': /* undefine... */
207 macro_popdef(optarg);
208 break;
209 case 'g':
210 mimic_gnu = 1;
211 break;
212 case 'd':
213 set_trace_flags(optarg);
214 break;
215 case 's':
216 synch_lines = 1;
217 break;
218 case 't':
219 mark_traced(optarg, 1);
220 break;
221 case 'o':
222 trace_file(optarg);
223 break;
224 case '?':
225 usage();
226 }
227
228 argc -= optind;
229 argv += optind;
230
231 initkwds();
232 if (mimic_gnu)
233 setup_builtin("format", FORMATTYPE);
234
235 active = stdout; /* default active output */
236 bbase[0] = bufbase;
237 if (!argc) {
238 sp = -1; /* stack pointer initialized */
239 fp = 0; /* frame pointer initialized */
240 set_input(infile+0, stdin, "stdin");
241 /* default input (naturally) */
242 macro();
243 } else
244 for (; argc--; ++argv) {
245 p = *argv;
246 if (p[0] == '-' && p[1] == EOS)
247 set_input(infile, stdin, "stdin");
248 else if (fopen_trypath(infile, p) == NULL)
249 err(1, "%s", p);
250 sp = -1;
251 fp = 0;
252 macro();
253 release_input(infile);
254 }
255
256 if (wrapindex) {
257 int i;
258
259 ilevel = 0; /* in case m4wrap includes.. */
260 bufbase = bp = buf; /* use the entire buffer */
261 if (mimic_gnu) {
262 while (wrapindex != 0) {
263 for (i = 0; i < wrapindex; i++)
264 pbstr(m4wraps[i]);
265 wrapindex =0;
266 macro();
267 }
268 } else {
269 for (i = 0; i < wrapindex; i++) {
270 pbstr(m4wraps[i]);
271 macro();
272 }
273 }
274 }
275
276 if (active != stdout)
277 active = stdout; /* reset output just in case */
278 for (n = 1; n < maxout; n++) /* default wrap-up: undivert */
279 if (outfile[n] != NULL)
280 getdiv(n);
281 /* remove bitbucket if used */
282 if (outfile[0] != NULL) {
283 (void) fclose(outfile[0]);
284 }
285
286 return 0;
287 }
288
289 /*
290 * Look ahead for `token'.
291 * (on input `t == token[0]')
292 * Used for comment and quoting delimiters.
293 * Returns 1 if `token' present; copied to output.
294 * 0 if `token' not found; all characters pushed back
295 */
296 static int
do_look_ahead(int t,const char * token)297 do_look_ahead(int t, const char *token)
298 {
299 int i;
300
301 assert((unsigned char)t == (unsigned char)token[0]);
302
303 for (i = 1; *++token; i++) {
304 t = gpbc();
305 if (t == EOF || (unsigned char)t != (unsigned char)*token) {
306 pushback(t);
307 while (--i)
308 pushback(*--token);
309 return 0;
310 }
311 }
312 return 1;
313 }
314
315 #define LOOK_AHEAD(t, token) (t != EOF && \
316 (unsigned char)(t)==(unsigned char)(token)[0] && \
317 do_look_ahead(t,token))
318
319 /*
320 * macro - the work horse..
321 */
322 static void
macro(void)323 macro(void)
324 {
325 char token[MAXTOK+1];
326 int t, l;
327 ndptr p;
328 int nlpar;
329
330 cycle {
331 t = gpbc();
332
333 if (LOOK_AHEAD(t,lquote)) { /* strip quotes */
334 nlpar = 0;
335 record(quotes, nlpar++);
336 /*
337 * Opening quote: scan forward until matching
338 * closing quote has been found.
339 */
340 do {
341
342 l = gpbc();
343 if (LOOK_AHEAD(l,rquote)) {
344 if (--nlpar > 0)
345 outputstr(rquote);
346 } else if (LOOK_AHEAD(l,lquote)) {
347 record(quotes, nlpar++);
348 outputstr(lquote);
349 } else if (l == EOF) {
350 if (nlpar == 1)
351 warnx("unclosed quote:");
352 else
353 warnx("%d unclosed quotes:", nlpar);
354 dump_stack(quotes, nlpar);
355 exit(1);
356 } else {
357 if (nlpar > 0) {
358 if (sp < 0)
359 reallyputchar(l);
360 else
361 CHRSAVE(l);
362 }
363 }
364 }
365 while (nlpar != 0);
366 } else if (sp < 0 && LOOK_AHEAD(t, scommt)) {
367 reallyoutputstr(scommt);
368
369 for(;;) {
370 t = gpbc();
371 if (LOOK_AHEAD(t, ecommt)) {
372 reallyoutputstr(ecommt);
373 break;
374 }
375 if (t == EOF)
376 break;
377 reallyputchar(t);
378 }
379 } else if (t == '_' || isalpha(t)) {
380 p = inspect(t, token);
381 if (p != NULL)
382 pushback(l = gpbc());
383 if (p == NULL || (l != LPAREN &&
384 (macro_getdef(p)->type & NEEDARGS) != 0))
385 outputstr(token);
386 else {
387 /*
388 * real thing.. First build a call frame:
389 */
390 pushf(fp); /* previous call frm */
391 pushf(macro_getdef(p)->type); /* type of the call */
392 pushf(is_traced(p));
393 pushf(0); /* parenthesis level */
394 fp = sp; /* new frame pointer */
395 /*
396 * now push the string arguments:
397 */
398 pushs1(macro_getdef(p)->defn); /* defn string */
399 pushs1((char *)macro_name(p)); /* macro name */
400 pushs(ep); /* start next..*/
401
402 if (l != LPAREN && PARLEV == 0) {
403 /* no bracks */
404 chrsave(EOS);
405
406 if (sp == STACKMAX)
407 errx(1, "internal stack overflow");
408 eval((const char **) mstack+fp+1, 2,
409 CALTYP, TRACESTATUS);
410
411 ep = PREVEP; /* flush strspace */
412 sp = PREVSP; /* previous sp.. */
413 fp = PREVFP; /* rewind stack...*/
414 }
415 }
416 } else if (t == EOF) {
417 if (sp > -1 && ilevel <= 0) {
418 warnx( "unexpected end of input, unclosed parenthesis:");
419 dump_stack(paren, PARLEV);
420 exit(1);
421 }
422 if (ilevel <= 0)
423 break; /* all done thanks.. */
424 release_input(infile+ilevel--);
425 emit_synchline();
426 bufbase = bbase[ilevel];
427 continue;
428 } else if (sp < 0) { /* not in a macro at all */
429 reallyputchar(t); /* output directly.. */
430 }
431
432 else switch(t) {
433
434 case LPAREN:
435 if (PARLEV > 0)
436 chrsave(t);
437 while (isspace(l = gpbc())) /* skip blank, tab, nl.. */
438 if (PARLEV > 0)
439 chrsave(l);
440 pushback(l);
441 record(paren, PARLEV++);
442 break;
443
444 case RPAREN:
445 if (--PARLEV > 0)
446 chrsave(t);
447 else { /* end of argument list */
448 chrsave(EOS);
449
450 if (sp == STACKMAX)
451 errx(1, "internal stack overflow");
452
453 eval((const char **) mstack+fp+1, sp-fp,
454 CALTYP, TRACESTATUS);
455
456 ep = PREVEP; /* flush strspace */
457 sp = PREVSP; /* previous sp.. */
458 fp = PREVFP; /* rewind stack...*/
459 }
460 break;
461
462 case COMMA:
463 if (PARLEV == 1) {
464 chrsave(EOS); /* new argument */
465 while (isspace(l = gpbc()))
466 ;
467 pushback(l);
468 pushs(ep);
469 } else
470 chrsave(t);
471 break;
472
473 default:
474 if (LOOK_AHEAD(t, scommt)) {
475 char *p;
476 for (p = scommt; *p; p++)
477 chrsave(*p);
478 for(;;) {
479 t = gpbc();
480 if (LOOK_AHEAD(t, ecommt)) {
481 for (p = ecommt; *p; p++)
482 chrsave(*p);
483 break;
484 }
485 if (t == EOF)
486 break;
487 CHRSAVE(t);
488 }
489 } else
490 CHRSAVE(t); /* stack the char */
491 break;
492 }
493 }
494 }
495
496 /*
497 * output string directly, without pushing it for reparses.
498 */
499 void
outputstr(const char * s)500 outputstr(const char *s)
501 {
502 if (sp < 0)
503 reallyoutputstr(s);
504 else
505 while (*s)
506 CHRSAVE(*s++);
507 }
508
509 void
reallyoutputstr(const char * s)510 reallyoutputstr(const char *s)
511 {
512 if (synch_lines) {
513 while (*s) {
514 fputc(*s, active);
515 if (*s++ == '\n') {
516 infile[ilevel].synch_lineno++;
517 if (infile[ilevel].synch_lineno !=
518 infile[ilevel].lineno)
519 do_emit_synchline();
520 }
521 }
522 } else
523 fputs(s, active);
524 }
525
526 void
reallyputchar(int c)527 reallyputchar(int c)
528 {
529 putc(c, active);
530 if (synch_lines && c == '\n') {
531 infile[ilevel].synch_lineno++;
532 if (infile[ilevel].synch_lineno != infile[ilevel].lineno)
533 do_emit_synchline();
534 }
535 }
536
537 /*
538 * build an input token..
539 * consider only those starting with _ or A-Za-z.
540 */
541 static ndptr
inspect(int c,char * tp)542 inspect(int c, char *tp)
543 {
544 char *name = tp;
545 char *etp = tp+MAXTOK;
546 ndptr p;
547
548 *tp++ = c;
549
550 while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
551 *tp++ = c;
552 if (c != EOF)
553 PUSHBACK(c);
554 *tp = EOS;
555 /* token is too long, it won't match anything, but it can still
556 * be output. */
557 if (tp == ep) {
558 outputstr(name);
559 while (isalnum(c = gpbc()) || c == '_') {
560 if (sp < 0)
561 reallyputchar(c);
562 else
563 CHRSAVE(c);
564 }
565 *name = EOS;
566 return NULL;
567 }
568
569 p = ohash_find(¯os, ohash_qlookupi(¯os, name, (const char **)&tp));
570 if (p == NULL)
571 return NULL;
572 if (macro_getdef(p) == NULL)
573 return NULL;
574 return p;
575 }
576
577 /*
578 * initkwds - initialise m4 keywords as fast as possible.
579 * This very similar to install, but without certain overheads,
580 * such as calling lookup. Malloc is not used for storing the
581 * keyword strings, since we simply use the static pointers
582 * within keywrds block.
583 */
584 static void
initkwds(void)585 initkwds(void)
586 {
587 unsigned int type;
588 int i;
589
590 for (i = 0; i < MAXKEYS; i++) {
591 type = keywrds[i].ktyp & TYPEMASK;
592 if ((keywrds[i].ktyp & NOARGS) == 0)
593 type |= NEEDARGS;
594 setup_builtin(keywrds[i].knam, type);
595 }
596 }
597
598 static void
record(struct position * t,int lev)599 record(struct position *t, int lev)
600 {
601 if (lev < MAXRECORD) {
602 t[lev].name = CURRENT_NAME;
603 t[lev].line = CURRENT_LINE;
604 }
605 }
606
607 static void
dump_stack(struct position * t,int lev)608 dump_stack(struct position *t, int lev)
609 {
610 int i;
611
612 for (i = 0; i < lev; i++) {
613 if (i == MAXRECORD) {
614 fprintf(stderr, " ...\n");
615 break;
616 }
617 fprintf(stderr, " %s at line %lu\n",
618 t[i].name, t[i].line);
619 }
620 }
621
622
623 static void
enlarge_stack(void)624 enlarge_stack(void)
625 {
626 STACKMAX += STACKMAX/2;
627 mstack = xrealloc(mstack, sizeof(stae) * STACKMAX,
628 "Evaluation stack overflow (%lu)",
629 (unsigned long)STACKMAX);
630 sstack = xrealloc(sstack, STACKMAX,
631 "Evaluation stack overflow (%lu)",
632 (unsigned long)STACKMAX);
633 }
634