Annotation of src/usr.bin/m4/main.c, Revision 1.37
1.37 ! deraadt 1: /* $OpenBSD: main.c,v 1.36 2000/07/27 17:44:33 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
1.37 ! deraadt 50: static char rcsid[] = "$OpenBSD: main.c,v 1.36 2000/07/27 17:44:33 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));
1.25 espie 165: static ndptr inspect __P((char, 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;
1.17 espie 180: FILE *ifp;
1.1 deraadt 181:
182: if (signal(SIGINT, SIG_IGN) != SIG_IGN)
183: signal(SIGINT, onintr);
184:
185: initkwds();
1.14 espie 186: initspaces();
1.34 espie 187: STACKMAX = INITSTACKMAX;
188:
189: mstack = (stae *)xalloc(sizeof(stae) * STACKMAX);
190: sstack = (char *)xalloc(STACKMAX);
1.1 deraadt 191:
1.36 espie 192: maxout = 0;
193: outfile = NULL;
194: resizedivs(MAXOUT);
195:
1.32 espie 196: while ((c = getopt(argc, argv, "gtD:U:o:I:")) != -1)
1.1 deraadt 197: switch(c) {
198:
199: case 'D': /* define something..*/
200: for (p = optarg; *p; p++)
201: if (*p == '=')
202: break;
203: if (*p)
204: *p++ = EOS;
205: dodefine(optarg, p);
1.16 espie 206: break;
207: case 'I':
208: addtoincludepath(optarg);
1.1 deraadt 209: break;
210: case 'U': /* undefine... */
211: remhash(optarg, TOP);
1.32 espie 212: break;
213: case 'g':
214: mimic_gnu = 1;
1.1 deraadt 215: break;
216: case 'o': /* specific output */
217: case '?':
218: usage();
219: }
220:
221: argc -= optind;
222: argv += optind;
223:
224: active = stdout; /* default active output */
225: bbase[0] = bufbase;
226: if (!argc) {
227: sp = -1; /* stack pointer initialized */
228: fp = 0; /* frame pointer initialized */
1.26 espie 229: set_input(infile+0, stdin, "stdin");
230: /* default input (naturally) */
1.1 deraadt 231: macro();
232: } else
233: for (; argc--; ++argv) {
234: p = *argv;
1.13 espie 235: if (p[0] == '-' && p[1] == EOS)
1.26 espie 236: set_input(infile, stdin, "stdin");
237: else if (fopen_trypath(infile, p) == NULL)
1.11 espie 238: err(1, "%s", p);
1.1 deraadt 239: sp = -1;
240: fp = 0;
241: macro();
1.26 espie 242: release_input(infile);
1.1 deraadt 243: }
244:
245: if (*m4wraps) { /* anything for rundown ?? */
246: ilevel = 0; /* in case m4wrap includes.. */
247: bufbase = bp = buf; /* use the entire buffer */
248: pbstr(m4wraps); /* user-defined wrapup act */
249: macro(); /* last will and testament */
250: }
251:
252: if (active != stdout)
253: active = stdout; /* reset output just in case */
1.36 espie 254: for (n = 1; n < maxout; n++) /* default wrap-up: undivert */
1.1 deraadt 255: if (outfile[n] != NULL)
256: getdiv(n);
257: /* remove bitbucket if used */
258: if (outfile[0] != NULL) {
259: (void) fclose(outfile[0]);
260: }
261:
262: return 0;
263: }
264:
265: /*
1.21 espie 266: * Look ahead for `token'.
1.2 deraadt 267: * (on input `t == token[0]')
268: * Used for comment and quoting delimiters.
269: * Returns 1 if `token' present; copied to output.
270: * 0 if `token' not found; all characters pushed back
271: */
1.18 espie 272: static int
1.2 deraadt 273: do_look_ahead(t, token)
274: int t;
1.18 espie 275: const char *token;
1.2 deraadt 276: {
277: int i;
278:
1.21 espie 279: assert(t == token[0]);
1.2 deraadt 280:
281: for (i = 1; *++token; i++) {
282: t = gpbc();
283: if (t == EOF || t != *token) {
1.28 espie 284: putback(t);
1.2 deraadt 285: while (--i)
286: putback(*--token);
287: return 0;
288: }
289: }
290: return 1;
291: }
292:
293: #define LOOK_AHEAD(t, token) ((t)==(token)[0] && do_look_ahead(t,token))
294:
295: /*
1.1 deraadt 296: * macro - the work horse..
297: */
1.18 espie 298: static void
1.17 espie 299: macro()
300: {
1.34 espie 301: char token[MAXTOK+1];
1.17 espie 302: int t, l;
303: ndptr p;
304: int nlpar;
1.1 deraadt 305:
306: cycle {
1.2 deraadt 307: t = gpbc();
308: if (t == '_' || isalpha(t)) {
1.29 espie 309: p = inspect(t, token);
1.24 espie 310: if (p != nil)
311: putback(l = gpbc());
312: if (p == nil || (l != LPAREN &&
313: (p->type & NEEDARGS) != 0))
1.29 espie 314: outputstr(token);
1.1 deraadt 315: else {
316: /*
317: * real thing.. First build a call frame:
318: */
319: pushf(fp); /* previous call frm */
320: pushf(p->type); /* type of the call */
321: pushf(0); /* parenthesis level */
322: fp = sp; /* new frame pointer */
323: /*
324: * now push the string arguments:
325: */
1.34 espie 326: pushs1(p->defn); /* defn string */
327: pushs1(p->name); /* macro name */
328: pushs(ep); /* start next..*/
1.1 deraadt 329:
330: if (l != LPAREN) { /* add bracks */
331: putback(RPAREN);
332: putback(LPAREN);
333: }
334: }
335: }
336: else if (t == EOF) {
1.27 espie 337: if (sp > -1) {
338: warnx( "unexpected end of input, unclosed parenthesis:");
339: dump_stack(paren, PARLEV);
340: exit(1);
341: }
1.1 deraadt 342: if (ilevel <= 0)
343: break; /* all done thanks.. */
1.26 espie 344: release_input(infile+ilevel--);
1.1 deraadt 345: bufbase = bbase[ilevel];
346: continue;
347: }
348: /*
1.7 deraadt 349: * non-alpha token possibly seen..
1.1 deraadt 350: * [the order of else if .. stmts is important.]
351: */
1.2 deraadt 352: else if (LOOK_AHEAD(t,lquote)) { /* strip quotes */
1.27 espie 353: nlpar = 0;
354: record(quotes, nlpar++);
1.30 espie 355: /*
356: * Opening quote: scan forward until matching
357: * closing quote has been found.
358: */
1.1 deraadt 359: do {
1.7 deraadt 360:
1.2 deraadt 361: l = gpbc();
1.7 deraadt 362: if (LOOK_AHEAD(l,rquote)) {
1.29 espie 363: if (--nlpar > 0)
364: outputstr(rquote);
1.7 deraadt 365: } else if (LOOK_AHEAD(l,lquote)) {
1.27 espie 366: record(quotes, nlpar++);
1.29 espie 367: outputstr(lquote);
1.17 espie 368: } else if (l == EOF) {
369: if (nlpar == 1)
1.27 espie 370: warnx("unclosed quote:");
1.17 espie 371: else
1.27 espie 372: warnx("%d unclosed quotes:", nlpar);
373: dump_stack(quotes, nlpar);
374: exit(1);
1.17 espie 375: } else {
1.29 espie 376: if (nlpar > 0) {
377: if (sp < 0)
378: putc(l, active);
379: else
380: chrsave(l);
381: }
1.7 deraadt 382: }
1.1 deraadt 383: }
384: while (nlpar != 0);
385: }
386:
1.2 deraadt 387: else if (sp < 0 && LOOK_AHEAD(t, scommt)) {
1.21 espie 388: fputs(scommt, active);
1.2 deraadt 389:
390: for(;;) {
391: t = gpbc();
392: if (LOOK_AHEAD(t, ecommt)) {
1.21 espie 393: fputs(ecommt, active);
1.2 deraadt 394: break;
395: }
396: if (t == EOF)
397: break;
1.1 deraadt 398: putc(t, active);
399: }
1.2 deraadt 400: }
401:
402: else if (sp < 0) { /* not in a macro at all */
1.1 deraadt 403: putc(t, active); /* output directly.. */
404: }
405:
406: else switch(t) {
407:
408: case LPAREN:
409: if (PARLEV > 0)
410: chrsave(t);
411: while (isspace(l = gpbc()))
412: ; /* skip blank, tab, nl.. */
413: putback(l);
1.27 espie 414: record(paren, PARLEV++);
1.1 deraadt 415: break;
416:
417: case RPAREN:
418: if (--PARLEV > 0)
419: chrsave(t);
420: else { /* end of argument list */
421: chrsave(EOS);
422:
423: if (sp == STACKMAX)
1.11 espie 424: errx(1, "internal stack overflow");
1.1 deraadt 425:
426: if (CALTYP == MACRTYPE)
1.18 espie 427: expand((const char **) mstack+fp+1, sp-fp);
1.1 deraadt 428: else
1.18 espie 429: eval((const char **) mstack+fp+1, sp-fp, CALTYP);
1.1 deraadt 430:
431: ep = PREVEP; /* flush strspace */
432: sp = PREVSP; /* previous sp.. */
433: fp = PREVFP; /* rewind stack...*/
434: }
435: break;
436:
437: case COMMA:
438: if (PARLEV == 1) {
439: chrsave(EOS); /* new argument */
440: while (isspace(l = gpbc()))
441: ;
442: putback(l);
443: pushs(ep);
444: } else
445: chrsave(t);
446: break;
447:
448: default:
1.22 espie 449: if (LOOK_AHEAD(t, scommt)) {
450: char *p;
451: for (p = scommt; *p; p++)
452: chrsave(*p);
453: for(;;) {
454: t = gpbc();
455: if (LOOK_AHEAD(t, ecommt)) {
456: for (p = ecommt; *p; p++)
457: chrsave(*p);
458: break;
459: }
460: if (t == EOF)
461: break;
462: chrsave(t);
463: }
464: } else
465: chrsave(t); /* stack the char */
1.1 deraadt 466: break;
467: }
468: }
469: }
470:
1.24 espie 471: /*
472: * output string directly, without pushing it for reparses.
473: */
474: void
475: outputstr(s)
476: const char *s;
477: {
478: if (sp < 0)
479: while (*s)
480: putc(*s++, active);
481: else
482: while (*s)
483: chrsave(*s++);
484: }
485:
1.1 deraadt 486: /*
487: * build an input token..
488: * consider only those starting with _ or A-Za-z. This is a
489: * combo with lookup to speed things up.
490: */
1.18 espie 491: static ndptr
1.25 espie 492: inspect(c, tp)
493: char c;
1.17 espie 494: char *tp;
1.1 deraadt 495: {
1.17 espie 496: char *name = tp;
497: char *etp = tp+MAXTOK;
498: ndptr p;
1.25 espie 499: unsigned int h;
500:
501: h = *tp++ = c;
1.1 deraadt 502:
503: while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
504: h = (h << 5) + h + (*tp++ = c);
505: putback(c);
506: *tp = EOS;
1.33 espie 507: /* token is too long, it won't match anything, but it can still
508: * be output. */
509: if (tp == ep) {
510: outputstr(name);
511: while (isalnum(c = gpbc()) || c == '_') {
512: if (sp < 0)
513: putc(c, active);
514: else
515: chrsave(c);
516: }
517: *name = EOS;
518: return nil;
519: }
1.1 deraadt 520:
1.19 espie 521: for (p = hashtab[h % HASHSIZE]; p != nil; p = p->nxtptr)
522: if (h == p->hv && STREQ(name, p->name))
1.1 deraadt 523: break;
524: return p;
525: }
526:
527: /*
528: * initkwds - initialise m4 keywords as fast as possible.
529: * This very similar to install, but without certain overheads,
530: * such as calling lookup. Malloc is not used for storing the
1.17 espie 531: * keyword strings, since we simply use the static pointers
1.1 deraadt 532: * within keywrds block.
533: */
1.18 espie 534: static void
1.17 espie 535: initkwds()
536: {
537: size_t i;
1.20 millert 538: unsigned int h;
1.17 espie 539: ndptr p;
1.1 deraadt 540:
541: for (i = 0; i < MAXKEYS; i++) {
542: h = hash(keywrds[i].knam);
543: p = (ndptr) xalloc(sizeof(struct ndblock));
1.19 espie 544: p->nxtptr = hashtab[h % HASHSIZE];
545: hashtab[h % HASHSIZE] = p;
1.1 deraadt 546: p->name = keywrds[i].knam;
547: p->defn = null;
1.19 espie 548: p->hv = h;
1.24 espie 549: p->type = (keywrds[i].ktyp & TYPEMASK) | STATIC;
550: if ((keywrds[i].ktyp & NOARGS) == 0)
551: p->type |= NEEDARGS;
1.1 deraadt 552: }
553: }
1.31 espie 554:
555: /* Look up a builtin type, even if overridden by the user */
556: int
557: builtin_type(key)
558: const char *key;
559: {
560: int i;
561:
562: for (i = 0; i != MAXKEYS; i++)
563: if (STREQ(keywrds[i].knam, key))
564: return keywrds[i].ktyp;
565: return -1;
566: }
567:
1.17 espie 568:
1.27 espie 569: static void
570: record(t, lev)
571: struct position *t;
572: int lev;
573: {
574: if (lev < MAXRECORD) {
575: t[lev].name = CURRENT_NAME;
576: t[lev].line = CURRENT_LINE;
577: }
578: }
579:
580: static void
581: dump_stack(t, lev)
582: struct position *t;
583: int lev;
584: {
585: int i;
586:
587: for (i = 0; i < lev; i++) {
588: if (i == MAXRECORD) {
589: fprintf(stderr, " ...\n");
590: break;
591: }
592: fprintf(stderr, " %s at line %lu\n",
593: t[i].name, t[i].line);
594: }
1.34 espie 595: }
596:
597:
598: static void
599: enlarge_stack()
600: {
601: STACKMAX *= 2;
1.37 ! deraadt 602: fprintf(stderr, "%lu\n", (unsigned long)STACKMAX);
1.34 espie 603: mstack = realloc(mstack, sizeof(stae) * STACKMAX);
604: sstack = realloc(sstack, STACKMAX);
605: if (mstack == NULL || sstack == NULL)
606: errx(1, "Evaluation stack overflow (%lu)",
607: (unsigned long)STACKMAX);
1.27 espie 608: }