Annotation of src/usr.bin/m4/main.c, Revision 1.41
1.41 ! espie 1: /* $OpenBSD: main.c,v 1.40 2001/07/28 05:36:18 pvalchev Exp $ */
1.7 deraadt 2: /* $NetBSD: main.c,v 1.12 1997/02/08 23:54:49 cgd Exp $ */
1.1 deraadt 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
41: static 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
48: static char sccsid[] = "@(#)main.c 8.1 (Berkeley) 6/6/93";
49: #else
1.41 ! espie 50: static char rcsid[] = "$OpenBSD: main.c,v 1.40 2001/07/28 05:36:18 pvalchev Exp $";
1.1 deraadt 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>
1.21 espie 61: #include <assert.h>
1.1 deraadt 62: #include <signal.h>
63: #include <errno.h>
64: #include <unistd.h>
65: #include <stdio.h>
66: #include <ctype.h>
67: #include <string.h>
1.13 espie 68: #include <stddef.h>
1.34 espie 69: #include <stdlib.h>
1.11 espie 70: #include <err.h>
1.1 deraadt 71: #include "mdef.h"
72: #include "stdd.h"
73: #include "extern.h"
74: #include "pathnames.h"
75:
76: ndptr hashtab[HASHSIZE]; /* hash table for macros etc. */
1.34 espie 77: stae *mstack; /* stack of m4 machine */
78: char *sstack; /* shadow stack, for string space extension */
79: static size_t STACKMAX; /* current maximum size of stack */
1.1 deraadt 80: int sp; /* current m4 stack pointer */
81: int fp; /* m4 call frame pointer */
1.26 espie 82: struct input_file infile[MAXINP];/* input file stack (0=stdin) */
1.36 espie 83: FILE **outfile; /* diversion array(0=bitbucket)*/
84: int maxout;
1.1 deraadt 85: FILE *active; /* active output file pointer */
86: int ilevel = 0; /* input file stack pointer */
87: int oindex = 0; /* diversion index.. */
88: char *null = ""; /* as it says.. just a null.. */
89: char *m4wraps = ""; /* m4wrap string default.. */
1.2 deraadt 90: char lquote[MAXCCHARS+1] = {LQUOTE}; /* left quote character (`) */
91: char rquote[MAXCCHARS+1] = {RQUOTE}; /* right quote character (') */
92: char scommt[MAXCCHARS+1] = {SCOMMT}; /* start character for comment */
93: char ecommt[MAXCCHARS+1] = {ECOMMT}; /* end character for comment */
1.1 deraadt 94:
95: struct keyblk keywrds[] = { /* m4 keywords to be installed */
1.8 millert 96: { "include", INCLTYPE },
97: { "sinclude", SINCTYPE },
98: { "define", DEFITYPE },
99: { "defn", DEFNTYPE },
1.24 espie 100: { "divert", DIVRTYPE | NOARGS },
1.8 millert 101: { "expr", EXPRTYPE },
102: { "eval", EXPRTYPE },
103: { "substr", SUBSTYPE },
104: { "ifelse", IFELTYPE },
105: { "ifdef", IFDFTYPE },
106: { "len", LENGTYPE },
107: { "incr", INCRTYPE },
108: { "decr", DECRTYPE },
1.24 espie 109: { "dnl", DNLNTYPE | NOARGS },
110: { "changequote", CHNQTYPE | NOARGS },
111: { "changecom", CHNCTYPE | NOARGS },
1.8 millert 112: { "index", INDXTYPE },
1.1 deraadt 113: #ifdef EXTENDED
1.8 millert 114: { "paste", PASTTYPE },
115: { "spaste", SPASTYPE },
1.31 espie 116: /* Newer extensions, needed to handle gnu-m4 scripts */
117: { "indir", INDIRTYPE},
118: { "builtin", BUILTINTYPE},
119: { "patsubst", PATSTYPE},
120: { "regexp", REGEXPTYPE},
1.35 espie 121: { "esyscmd", ESYSCMDTYPE},
1.31 espie 122: { "__file__", FILENAMETYPE | NOARGS},
123: { "__line__", LINETYPE | NOARGS},
1.1 deraadt 124: #endif
1.8 millert 125: { "popdef", POPDTYPE },
126: { "pushdef", PUSDTYPE },
1.24 espie 127: { "dumpdef", DUMPTYPE | NOARGS },
128: { "shift", SHIFTYPE | NOARGS },
1.8 millert 129: { "translit", TRNLTYPE },
130: { "undefine", UNDFTYPE },
1.24 espie 131: { "undivert", UNDVTYPE | NOARGS },
132: { "divnum", DIVNTYPE | NOARGS },
1.8 millert 133: { "maketemp", MKTMTYPE },
1.24 espie 134: { "errprint", ERRPTYPE | NOARGS },
135: { "m4wrap", M4WRTYPE | NOARGS },
136: { "m4exit", EXITTYPE | NOARGS },
1.8 millert 137: { "syscmd", SYSCTYPE },
1.24 espie 138: { "sysval", SYSVTYPE | NOARGS },
1.1 deraadt 139:
1.24 espie 140: #if defined(unix) || defined(__unix__)
141: { "unix", SELFTYPE | NOARGS },
1.1 deraadt 142: #else
143: #ifdef vms
1.24 espie 144: { "vms", SELFTYPE | NOARGS },
1.1 deraadt 145: #endif
146: #endif
147: };
148:
149: #define MAXKEYS (sizeof(keywrds)/sizeof(struct keyblk))
150:
151: extern int optind;
152: extern char *optarg;
153:
1.27 espie 154: #define MAXRECORD 50
155: static struct position {
156: char *name;
157: unsigned long line;
158: } quotes[MAXRECORD], paren[MAXRECORD];
159:
160: static void record __P((struct position *, int));
161: static void dump_stack __P((struct position *, int));
162:
1.18 espie 163: static void macro __P((void));
164: static void initkwds __P((void));
1.39 espie 165: static ndptr inspect __P((int, char *));
1.18 espie 166: static int do_look_ahead __P((int, const char *));
167:
1.34 espie 168: static void enlarge_stack __P((void));
169:
1.18 espie 170: int main __P((int, char *[]));
1.1 deraadt 171:
172: int
173: main(argc,argv)
174: int argc;
175: char *argv[];
176: {
1.17 espie 177: int c;
178: int n;
1.1 deraadt 179: char *p;
180:
181: if (signal(SIGINT, SIG_IGN) != SIG_IGN)
182: signal(SIGINT, onintr);
183:
184: initkwds();
1.14 espie 185: initspaces();
1.34 espie 186: STACKMAX = INITSTACKMAX;
187:
188: mstack = (stae *)xalloc(sizeof(stae) * STACKMAX);
189: sstack = (char *)xalloc(STACKMAX);
1.1 deraadt 190:
1.36 espie 191: maxout = 0;
192: outfile = NULL;
193: resizedivs(MAXOUT);
194:
1.32 espie 195: while ((c = getopt(argc, argv, "gtD:U:o:I:")) != -1)
1.1 deraadt 196: switch(c) {
197:
198: case 'D': /* define something..*/
199: for (p = optarg; *p; p++)
200: if (*p == '=')
201: break;
202: if (*p)
203: *p++ = EOS;
204: dodefine(optarg, p);
1.16 espie 205: break;
206: case 'I':
207: addtoincludepath(optarg);
1.1 deraadt 208: break;
209: case 'U': /* undefine... */
210: remhash(optarg, TOP);
1.32 espie 211: break;
212: case 'g':
213: mimic_gnu = 1;
1.1 deraadt 214: break;
1.38 aaron 215: case 'o':
216: /* XXX accept -o for compatibility */
217: break;
1.1 deraadt 218: case '?':
219: usage();
220: }
221:
222: argc -= optind;
223: argv += optind;
224:
225: active = stdout; /* default active output */
226: bbase[0] = bufbase;
227: if (!argc) {
228: sp = -1; /* stack pointer initialized */
229: fp = 0; /* frame pointer initialized */
1.26 espie 230: set_input(infile+0, stdin, "stdin");
231: /* default input (naturally) */
1.1 deraadt 232: macro();
233: } else
234: for (; argc--; ++argv) {
235: p = *argv;
1.13 espie 236: if (p[0] == '-' && p[1] == EOS)
1.26 espie 237: set_input(infile, stdin, "stdin");
238: else if (fopen_trypath(infile, p) == NULL)
1.11 espie 239: err(1, "%s", p);
1.1 deraadt 240: sp = -1;
241: fp = 0;
242: macro();
1.26 espie 243: release_input(infile);
1.1 deraadt 244: }
245:
246: if (*m4wraps) { /* anything for rundown ?? */
247: ilevel = 0; /* in case m4wrap includes.. */
248: bufbase = bp = buf; /* use the entire buffer */
249: pbstr(m4wraps); /* user-defined wrapup act */
250: macro(); /* last will and testament */
251: }
252:
253: if (active != stdout)
254: active = stdout; /* reset output just in case */
1.36 espie 255: for (n = 1; n < maxout; n++) /* default wrap-up: undivert */
1.1 deraadt 256: if (outfile[n] != NULL)
257: getdiv(n);
258: /* remove bitbucket if used */
259: if (outfile[0] != NULL) {
260: (void) fclose(outfile[0]);
261: }
262:
263: return 0;
264: }
265:
266: /*
1.21 espie 267: * Look ahead for `token'.
1.2 deraadt 268: * (on input `t == token[0]')
269: * Used for comment and quoting delimiters.
270: * Returns 1 if `token' present; copied to output.
271: * 0 if `token' not found; all characters pushed back
272: */
1.18 espie 273: static int
1.2 deraadt 274: do_look_ahead(t, token)
275: int t;
1.18 espie 276: const char *token;
1.2 deraadt 277: {
278: int i;
279:
1.21 espie 280: assert(t == token[0]);
1.2 deraadt 281:
282: for (i = 1; *++token; i++) {
283: t = gpbc();
284: if (t == EOF || t != *token) {
1.28 espie 285: putback(t);
1.2 deraadt 286: while (--i)
287: putback(*--token);
288: return 0;
289: }
290: }
291: return 1;
292: }
293:
294: #define LOOK_AHEAD(t, token) ((t)==(token)[0] && do_look_ahead(t,token))
295:
296: /*
1.1 deraadt 297: * macro - the work horse..
298: */
1.18 espie 299: static void
1.17 espie 300: macro()
301: {
1.34 espie 302: char token[MAXTOK+1];
1.17 espie 303: int t, l;
304: ndptr p;
305: int nlpar;
1.1 deraadt 306:
307: cycle {
1.2 deraadt 308: t = gpbc();
309: if (t == '_' || isalpha(t)) {
1.29 espie 310: p = inspect(t, token);
1.24 espie 311: if (p != nil)
312: putback(l = gpbc());
313: if (p == nil || (l != LPAREN &&
314: (p->type & NEEDARGS) != 0))
1.29 espie 315: outputstr(token);
1.1 deraadt 316: else {
317: /*
318: * real thing.. First build a call frame:
319: */
320: pushf(fp); /* previous call frm */
321: pushf(p->type); /* type of the call */
322: pushf(0); /* parenthesis level */
323: fp = sp; /* new frame pointer */
324: /*
325: * now push the string arguments:
326: */
1.34 espie 327: pushs1(p->defn); /* defn string */
328: pushs1(p->name); /* macro name */
329: pushs(ep); /* start next..*/
1.1 deraadt 330:
1.41 ! espie 331: if (l != LPAREN && PARLEV == 0) {
! 332: /* no bracks */
! 333: chrsave(EOS);
! 334:
! 335: if (sp == STACKMAX)
! 336: errx(1, "internal stack overflow");
! 337: if (CALTYP == MACRTYPE)
! 338: expand((const char **) mstack+fp+1, 2);
! 339: else
! 340: eval((const char **) mstack+fp+1, 2, CALTYP);
! 341:
! 342: ep = PREVEP; /* flush strspace */
! 343: sp = PREVSP; /* previous sp.. */
! 344: fp = PREVFP; /* rewind stack...*/
1.1 deraadt 345: }
346: }
1.41 ! espie 347: } else if (t == EOF) {
1.27 espie 348: if (sp > -1) {
349: warnx( "unexpected end of input, unclosed parenthesis:");
350: dump_stack(paren, PARLEV);
351: exit(1);
352: }
1.1 deraadt 353: if (ilevel <= 0)
354: break; /* all done thanks.. */
1.26 espie 355: release_input(infile+ilevel--);
1.1 deraadt 356: bufbase = bbase[ilevel];
357: continue;
358: }
359: /*
1.7 deraadt 360: * non-alpha token possibly seen..
1.1 deraadt 361: * [the order of else if .. stmts is important.]
362: */
1.2 deraadt 363: else if (LOOK_AHEAD(t,lquote)) { /* strip quotes */
1.27 espie 364: nlpar = 0;
365: record(quotes, nlpar++);
1.30 espie 366: /*
367: * Opening quote: scan forward until matching
368: * closing quote has been found.
369: */
1.1 deraadt 370: do {
1.7 deraadt 371:
1.2 deraadt 372: l = gpbc();
1.7 deraadt 373: if (LOOK_AHEAD(l,rquote)) {
1.29 espie 374: if (--nlpar > 0)
375: outputstr(rquote);
1.7 deraadt 376: } else if (LOOK_AHEAD(l,lquote)) {
1.27 espie 377: record(quotes, nlpar++);
1.29 espie 378: outputstr(lquote);
1.17 espie 379: } else if (l == EOF) {
380: if (nlpar == 1)
1.27 espie 381: warnx("unclosed quote:");
1.17 espie 382: else
1.27 espie 383: warnx("%d unclosed quotes:", nlpar);
384: dump_stack(quotes, nlpar);
385: exit(1);
1.17 espie 386: } else {
1.29 espie 387: if (nlpar > 0) {
388: if (sp < 0)
389: putc(l, active);
390: else
391: chrsave(l);
392: }
1.7 deraadt 393: }
1.1 deraadt 394: }
395: while (nlpar != 0);
396: }
397:
1.2 deraadt 398: else if (sp < 0 && LOOK_AHEAD(t, scommt)) {
1.21 espie 399: fputs(scommt, active);
1.2 deraadt 400:
401: for(;;) {
402: t = gpbc();
403: if (LOOK_AHEAD(t, ecommt)) {
1.21 espie 404: fputs(ecommt, active);
1.2 deraadt 405: break;
406: }
407: if (t == EOF)
408: break;
1.1 deraadt 409: putc(t, active);
410: }
1.2 deraadt 411: }
412:
413: else if (sp < 0) { /* not in a macro at all */
1.1 deraadt 414: putc(t, active); /* output directly.. */
415: }
416:
417: else switch(t) {
418:
419: case LPAREN:
420: if (PARLEV > 0)
421: chrsave(t);
422: while (isspace(l = gpbc()))
423: ; /* skip blank, tab, nl.. */
424: putback(l);
1.27 espie 425: record(paren, PARLEV++);
1.1 deraadt 426: break;
427:
428: case RPAREN:
429: if (--PARLEV > 0)
430: chrsave(t);
431: else { /* end of argument list */
432: chrsave(EOS);
433:
434: if (sp == STACKMAX)
1.11 espie 435: errx(1, "internal stack overflow");
1.1 deraadt 436:
437: if (CALTYP == MACRTYPE)
1.18 espie 438: expand((const char **) mstack+fp+1, sp-fp);
1.1 deraadt 439: else
1.18 espie 440: eval((const char **) mstack+fp+1, sp-fp, CALTYP);
1.1 deraadt 441:
442: ep = PREVEP; /* flush strspace */
443: sp = PREVSP; /* previous sp.. */
444: fp = PREVFP; /* rewind stack...*/
445: }
446: break;
447:
448: case COMMA:
449: if (PARLEV == 1) {
450: chrsave(EOS); /* new argument */
451: while (isspace(l = gpbc()))
452: ;
453: putback(l);
454: pushs(ep);
455: } else
456: chrsave(t);
457: break;
458:
459: default:
1.22 espie 460: if (LOOK_AHEAD(t, scommt)) {
461: char *p;
462: for (p = scommt; *p; p++)
463: chrsave(*p);
464: for(;;) {
465: t = gpbc();
466: if (LOOK_AHEAD(t, ecommt)) {
467: for (p = ecommt; *p; p++)
468: chrsave(*p);
469: break;
470: }
471: if (t == EOF)
472: break;
473: chrsave(t);
474: }
475: } else
476: chrsave(t); /* stack the char */
1.1 deraadt 477: break;
478: }
479: }
480: }
481:
1.24 espie 482: /*
483: * output string directly, without pushing it for reparses.
484: */
485: void
486: outputstr(s)
487: const char *s;
488: {
489: if (sp < 0)
490: while (*s)
491: putc(*s++, active);
492: else
493: while (*s)
494: chrsave(*s++);
495: }
496:
1.1 deraadt 497: /*
498: * build an input token..
499: * consider only those starting with _ or A-Za-z. This is a
500: * combo with lookup to speed things up.
501: */
1.18 espie 502: static ndptr
1.25 espie 503: inspect(c, tp)
1.39 espie 504: int c;
1.17 espie 505: char *tp;
1.1 deraadt 506: {
1.17 espie 507: char *name = tp;
508: char *etp = tp+MAXTOK;
509: ndptr p;
1.25 espie 510: unsigned int h;
511:
512: h = *tp++ = c;
1.1 deraadt 513:
514: while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
515: h = (h << 5) + h + (*tp++ = c);
516: putback(c);
517: *tp = EOS;
1.33 espie 518: /* token is too long, it won't match anything, but it can still
519: * be output. */
520: if (tp == ep) {
521: outputstr(name);
522: while (isalnum(c = gpbc()) || c == '_') {
523: if (sp < 0)
524: putc(c, active);
525: else
526: chrsave(c);
527: }
528: *name = EOS;
529: return nil;
530: }
1.1 deraadt 531:
1.19 espie 532: for (p = hashtab[h % HASHSIZE]; p != nil; p = p->nxtptr)
533: if (h == p->hv && STREQ(name, p->name))
1.1 deraadt 534: break;
535: return p;
536: }
537:
538: /*
539: * initkwds - initialise m4 keywords as fast as possible.
540: * This very similar to install, but without certain overheads,
541: * such as calling lookup. Malloc is not used for storing the
1.17 espie 542: * keyword strings, since we simply use the static pointers
1.1 deraadt 543: * within keywrds block.
544: */
1.18 espie 545: static void
1.17 espie 546: initkwds()
547: {
548: size_t i;
1.20 millert 549: unsigned int h;
1.17 espie 550: ndptr p;
1.1 deraadt 551:
552: for (i = 0; i < MAXKEYS; i++) {
553: h = hash(keywrds[i].knam);
554: p = (ndptr) xalloc(sizeof(struct ndblock));
1.19 espie 555: p->nxtptr = hashtab[h % HASHSIZE];
556: hashtab[h % HASHSIZE] = p;
1.1 deraadt 557: p->name = keywrds[i].knam;
558: p->defn = null;
1.19 espie 559: p->hv = h;
1.24 espie 560: p->type = (keywrds[i].ktyp & TYPEMASK) | STATIC;
561: if ((keywrds[i].ktyp & NOARGS) == 0)
562: p->type |= NEEDARGS;
1.1 deraadt 563: }
564: }
1.31 espie 565:
566: /* Look up a builtin type, even if overridden by the user */
567: int
568: builtin_type(key)
569: const char *key;
570: {
571: int i;
572:
573: for (i = 0; i != MAXKEYS; i++)
574: if (STREQ(keywrds[i].knam, key))
575: return keywrds[i].ktyp;
576: return -1;
577: }
578:
1.17 espie 579:
1.27 espie 580: static void
581: record(t, lev)
582: struct position *t;
583: int lev;
584: {
585: if (lev < MAXRECORD) {
586: t[lev].name = CURRENT_NAME;
587: t[lev].line = CURRENT_LINE;
588: }
589: }
590:
591: static void
592: dump_stack(t, lev)
593: struct position *t;
594: int lev;
595: {
596: int i;
597:
598: for (i = 0; i < lev; i++) {
599: if (i == MAXRECORD) {
600: fprintf(stderr, " ...\n");
601: break;
602: }
603: fprintf(stderr, " %s at line %lu\n",
604: t[i].name, t[i].line);
605: }
1.34 espie 606: }
607:
608:
609: static void
610: enlarge_stack()
611: {
612: STACKMAX *= 2;
1.37 deraadt 613: fprintf(stderr, "%lu\n", (unsigned long)STACKMAX);
1.34 espie 614: mstack = realloc(mstack, sizeof(stae) * STACKMAX);
615: sstack = realloc(sstack, STACKMAX);
616: if (mstack == NULL || sstack == NULL)
617: errx(1, "Evaluation stack overflow (%lu)",
618: (unsigned long)STACKMAX);
1.27 espie 619: }