Annotation of src/usr.bin/m4/main.c, Revision 1.46
1.46 ! espie 1: /* $OpenBSD: main.c,v 1.45 2001/09/18 14:05:14 espie 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
1.46 ! espie 50: static char rcsid[] = "$OpenBSD: main.c,v 1.45 2001/09/18 14:05:14 espie 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:
1.46 ! espie 195: while ((c = getopt(argc, argv, "gt:d:D: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;
1.46 ! espie 215: case 'd':
! 216: set_trace_flags(optarg);
! 217: break;
1.38 aaron 218: case 'o':
1.46 ! espie 219: trace_file(optarg);
1.38 aaron 220: break;
1.1 deraadt 221: case '?':
222: usage();
223: }
224:
225: argc -= optind;
226: argv += optind;
227:
228: active = stdout; /* default active output */
229: bbase[0] = bufbase;
230: if (!argc) {
231: sp = -1; /* stack pointer initialized */
232: fp = 0; /* frame pointer initialized */
1.26 espie 233: set_input(infile+0, stdin, "stdin");
234: /* default input (naturally) */
1.1 deraadt 235: macro();
236: } else
237: for (; argc--; ++argv) {
238: p = *argv;
1.13 espie 239: if (p[0] == '-' && p[1] == EOS)
1.26 espie 240: set_input(infile, stdin, "stdin");
241: else if (fopen_trypath(infile, p) == NULL)
1.11 espie 242: err(1, "%s", p);
1.1 deraadt 243: sp = -1;
244: fp = 0;
245: macro();
1.26 espie 246: release_input(infile);
1.1 deraadt 247: }
248:
249: if (*m4wraps) { /* anything for rundown ?? */
250: ilevel = 0; /* in case m4wrap includes.. */
251: bufbase = bp = buf; /* use the entire buffer */
252: pbstr(m4wraps); /* user-defined wrapup act */
253: macro(); /* last will and testament */
254: }
255:
256: if (active != stdout)
257: active = stdout; /* reset output just in case */
1.36 espie 258: for (n = 1; n < maxout; n++) /* default wrap-up: undivert */
1.1 deraadt 259: if (outfile[n] != NULL)
260: getdiv(n);
261: /* remove bitbucket if used */
262: if (outfile[0] != NULL) {
263: (void) fclose(outfile[0]);
264: }
265:
266: return 0;
267: }
268:
269: /*
1.21 espie 270: * Look ahead for `token'.
1.2 deraadt 271: * (on input `t == token[0]')
272: * Used for comment and quoting delimiters.
273: * Returns 1 if `token' present; copied to output.
274: * 0 if `token' not found; all characters pushed back
275: */
1.18 espie 276: static int
1.2 deraadt 277: do_look_ahead(t, token)
278: int t;
1.18 espie 279: const char *token;
1.2 deraadt 280: {
281: int i;
282:
1.43 espie 283: assert((unsigned char)t == (unsigned char)token[0]);
1.2 deraadt 284:
285: for (i = 1; *++token; i++) {
286: t = gpbc();
1.43 espie 287: if (t == EOF || (unsigned char)t != (unsigned char)*token) {
1.28 espie 288: putback(t);
1.2 deraadt 289: while (--i)
290: putback(*--token);
291: return 0;
292: }
293: }
294: return 1;
295: }
296:
1.43 espie 297: #define LOOK_AHEAD(t, token) (t != EOF && \
298: (unsigned char)(t)==(unsigned char)(token)[0] && \
299: do_look_ahead(t,token))
1.2 deraadt 300:
301: /*
1.1 deraadt 302: * macro - the work horse..
303: */
1.18 espie 304: static void
1.17 espie 305: macro()
306: {
1.34 espie 307: char token[MAXTOK+1];
1.17 espie 308: int t, l;
309: ndptr p;
310: int nlpar;
1.1 deraadt 311:
312: cycle {
1.2 deraadt 313: t = gpbc();
314: if (t == '_' || isalpha(t)) {
1.29 espie 315: p = inspect(t, token);
1.24 espie 316: if (p != nil)
317: putback(l = gpbc());
318: if (p == nil || (l != LPAREN &&
319: (p->type & NEEDARGS) != 0))
1.29 espie 320: outputstr(token);
1.1 deraadt 321: else {
322: /*
323: * real thing.. First build a call frame:
324: */
325: pushf(fp); /* previous call frm */
326: pushf(p->type); /* type of the call */
327: pushf(0); /* parenthesis level */
328: fp = sp; /* new frame pointer */
329: /*
330: * now push the string arguments:
331: */
1.34 espie 332: pushs1(p->defn); /* defn string */
333: pushs1(p->name); /* macro name */
334: pushs(ep); /* start next..*/
1.1 deraadt 335:
1.41 espie 336: if (l != LPAREN && PARLEV == 0) {
337: /* no bracks */
338: chrsave(EOS);
339:
340: if (sp == STACKMAX)
341: errx(1, "internal stack overflow");
1.44 espie 342: eval((const char **) mstack+fp+1, 2,
343: CALTYP);
1.41 espie 344:
345: ep = PREVEP; /* flush strspace */
346: sp = PREVSP; /* previous sp.. */
347: fp = PREVFP; /* rewind stack...*/
1.1 deraadt 348: }
349: }
1.41 espie 350: } else if (t == EOF) {
1.27 espie 351: if (sp > -1) {
352: warnx( "unexpected end of input, unclosed parenthesis:");
353: dump_stack(paren, PARLEV);
354: exit(1);
355: }
1.1 deraadt 356: if (ilevel <= 0)
357: break; /* all done thanks.. */
1.26 espie 358: release_input(infile+ilevel--);
1.1 deraadt 359: bufbase = bbase[ilevel];
360: continue;
361: }
362: /*
1.7 deraadt 363: * non-alpha token possibly seen..
1.1 deraadt 364: * [the order of else if .. stmts is important.]
365: */
1.2 deraadt 366: else if (LOOK_AHEAD(t,lquote)) { /* strip quotes */
1.27 espie 367: nlpar = 0;
368: record(quotes, nlpar++);
1.30 espie 369: /*
370: * Opening quote: scan forward until matching
371: * closing quote has been found.
372: */
1.1 deraadt 373: do {
1.7 deraadt 374:
1.2 deraadt 375: l = gpbc();
1.7 deraadt 376: if (LOOK_AHEAD(l,rquote)) {
1.29 espie 377: if (--nlpar > 0)
378: outputstr(rquote);
1.7 deraadt 379: } else if (LOOK_AHEAD(l,lquote)) {
1.27 espie 380: record(quotes, nlpar++);
1.29 espie 381: outputstr(lquote);
1.17 espie 382: } else if (l == EOF) {
383: if (nlpar == 1)
1.27 espie 384: warnx("unclosed quote:");
1.17 espie 385: else
1.27 espie 386: warnx("%d unclosed quotes:", nlpar);
387: dump_stack(quotes, nlpar);
388: exit(1);
1.17 espie 389: } else {
1.29 espie 390: if (nlpar > 0) {
391: if (sp < 0)
392: putc(l, active);
393: else
394: chrsave(l);
395: }
1.7 deraadt 396: }
1.1 deraadt 397: }
398: while (nlpar != 0);
399: }
400:
1.2 deraadt 401: else if (sp < 0 && LOOK_AHEAD(t, scommt)) {
1.21 espie 402: fputs(scommt, active);
1.2 deraadt 403:
404: for(;;) {
405: t = gpbc();
406: if (LOOK_AHEAD(t, ecommt)) {
1.21 espie 407: fputs(ecommt, active);
1.2 deraadt 408: break;
409: }
410: if (t == EOF)
411: break;
1.1 deraadt 412: putc(t, active);
413: }
1.2 deraadt 414: }
415:
416: else if (sp < 0) { /* not in a macro at all */
1.1 deraadt 417: putc(t, active); /* output directly.. */
418: }
419:
420: else switch(t) {
421:
422: case LPAREN:
423: if (PARLEV > 0)
424: chrsave(t);
425: while (isspace(l = gpbc()))
426: ; /* skip blank, tab, nl.. */
427: putback(l);
1.27 espie 428: record(paren, PARLEV++);
1.1 deraadt 429: break;
430:
431: case RPAREN:
432: if (--PARLEV > 0)
433: chrsave(t);
434: else { /* end of argument list */
435: chrsave(EOS);
436:
437: if (sp == STACKMAX)
1.11 espie 438: errx(1, "internal stack overflow");
1.1 deraadt 439:
1.44 espie 440: eval((const char **) mstack+fp+1, sp-fp,
441: CALTYP);
1.1 deraadt 442:
443: ep = PREVEP; /* flush strspace */
444: sp = PREVSP; /* previous sp.. */
445: fp = PREVFP; /* rewind stack...*/
446: }
447: break;
448:
449: case COMMA:
450: if (PARLEV == 1) {
451: chrsave(EOS); /* new argument */
452: while (isspace(l = gpbc()))
453: ;
454: putback(l);
455: pushs(ep);
456: } else
457: chrsave(t);
458: break;
459:
460: default:
1.22 espie 461: if (LOOK_AHEAD(t, scommt)) {
462: char *p;
463: for (p = scommt; *p; p++)
464: chrsave(*p);
465: for(;;) {
466: t = gpbc();
467: if (LOOK_AHEAD(t, ecommt)) {
468: for (p = ecommt; *p; p++)
469: chrsave(*p);
470: break;
471: }
472: if (t == EOF)
473: break;
474: chrsave(t);
475: }
476: } else
477: chrsave(t); /* stack the char */
1.1 deraadt 478: break;
479: }
480: }
481: }
482:
1.24 espie 483: /*
484: * output string directly, without pushing it for reparses.
485: */
486: void
487: outputstr(s)
488: const char *s;
489: {
490: if (sp < 0)
491: while (*s)
492: putc(*s++, active);
493: else
494: while (*s)
495: chrsave(*s++);
496: }
497:
1.1 deraadt 498: /*
499: * build an input token..
500: * consider only those starting with _ or A-Za-z. This is a
501: * combo with lookup to speed things up.
502: */
1.18 espie 503: static ndptr
1.25 espie 504: inspect(c, tp)
1.39 espie 505: int c;
1.17 espie 506: char *tp;
1.1 deraadt 507: {
1.17 espie 508: char *name = tp;
509: char *etp = tp+MAXTOK;
510: ndptr p;
1.25 espie 511: unsigned int h;
512:
513: h = *tp++ = c;
1.1 deraadt 514:
515: while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
516: h = (h << 5) + h + (*tp++ = c);
517: putback(c);
518: *tp = EOS;
1.33 espie 519: /* token is too long, it won't match anything, but it can still
520: * be output. */
521: if (tp == ep) {
522: outputstr(name);
523: while (isalnum(c = gpbc()) || c == '_') {
524: if (sp < 0)
525: putc(c, active);
526: else
527: chrsave(c);
528: }
529: *name = EOS;
530: return nil;
531: }
1.1 deraadt 532:
1.19 espie 533: for (p = hashtab[h % HASHSIZE]; p != nil; p = p->nxtptr)
534: if (h == p->hv && STREQ(name, p->name))
1.1 deraadt 535: break;
536: return p;
537: }
538:
539: /*
540: * initkwds - initialise m4 keywords as fast as possible.
541: * This very similar to install, but without certain overheads,
542: * such as calling lookup. Malloc is not used for storing the
1.17 espie 543: * keyword strings, since we simply use the static pointers
1.1 deraadt 544: * within keywrds block.
545: */
1.18 espie 546: static void
1.17 espie 547: initkwds()
548: {
549: size_t i;
1.20 millert 550: unsigned int h;
1.17 espie 551: ndptr p;
1.1 deraadt 552:
553: for (i = 0; i < MAXKEYS; i++) {
554: h = hash(keywrds[i].knam);
555: p = (ndptr) xalloc(sizeof(struct ndblock));
1.19 espie 556: p->nxtptr = hashtab[h % HASHSIZE];
557: hashtab[h % HASHSIZE] = p;
1.42 espie 558: p->name = xstrdup(keywrds[i].knam);
1.1 deraadt 559: p->defn = null;
1.19 espie 560: p->hv = h;
1.42 espie 561: p->type = keywrds[i].ktyp & TYPEMASK;
1.24 espie 562: if ((keywrds[i].ktyp & NOARGS) == 0)
563: p->type |= NEEDARGS;
1.1 deraadt 564: }
565: }
1.31 espie 566:
567: /* Look up a builtin type, even if overridden by the user */
568: int
569: builtin_type(key)
570: const char *key;
571: {
572: int i;
573:
574: for (i = 0; i != MAXKEYS; i++)
575: if (STREQ(keywrds[i].knam, key))
576: return keywrds[i].ktyp;
577: return -1;
578: }
579:
1.45 espie 580: char *
581: builtin_realname(n)
582: int n;
583: {
584: int i;
585:
586: for (i = 0; i != MAXKEYS; i++)
587: if (((keywrds[i].ktyp ^ n) & TYPEMASK) == 0)
588: return keywrds[i].knam;
589: return NULL;
590: }
1.17 espie 591:
1.27 espie 592: static void
593: record(t, lev)
594: struct position *t;
595: int lev;
596: {
597: if (lev < MAXRECORD) {
598: t[lev].name = CURRENT_NAME;
599: t[lev].line = CURRENT_LINE;
600: }
601: }
602:
603: static void
604: dump_stack(t, lev)
605: struct position *t;
606: int lev;
607: {
608: int i;
609:
610: for (i = 0; i < lev; i++) {
611: if (i == MAXRECORD) {
612: fprintf(stderr, " ...\n");
613: break;
614: }
615: fprintf(stderr, " %s at line %lu\n",
616: t[i].name, t[i].line);
617: }
1.34 espie 618: }
619:
620:
621: static void
622: enlarge_stack()
623: {
624: STACKMAX *= 2;
1.37 deraadt 625: fprintf(stderr, "%lu\n", (unsigned long)STACKMAX);
1.34 espie 626: mstack = realloc(mstack, sizeof(stae) * STACKMAX);
627: sstack = realloc(sstack, STACKMAX);
628: if (mstack == NULL || sstack == NULL)
629: errx(1, "Evaluation stack overflow (%lu)",
630: (unsigned long)STACKMAX);
1.27 espie 631: }