Annotation of src/usr.bin/m4/main.c, Revision 1.29
1.29 ! espie 1: /* $OpenBSD: main.c,v 1.28 2000/01/15 14:26:00 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
1.29 ! espie 50: static char rcsid[] = "$OpenBSD: main.c,v 1.28 2000/01/15 14:26:00 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>
1.11 espie 69: #include <err.h>
1.1 deraadt 70: #include "mdef.h"
71: #include "stdd.h"
72: #include "extern.h"
73: #include "pathnames.h"
74:
75: ndptr hashtab[HASHSIZE]; /* hash table for macros etc. */
76: stae mstack[STACKMAX+1]; /* stack of m4 machine */
77: int sp; /* current m4 stack pointer */
78: int fp; /* m4 call frame pointer */
1.26 espie 79: struct input_file infile[MAXINP];/* input file stack (0=stdin) */
1.1 deraadt 80: FILE *outfile[MAXOUT]; /* diversion array(0=bitbucket)*/
81: FILE *active; /* active output file pointer */
82: int ilevel = 0; /* input file stack pointer */
83: int oindex = 0; /* diversion index.. */
84: char *null = ""; /* as it says.. just a null.. */
85: char *m4wraps = ""; /* m4wrap string default.. */
1.2 deraadt 86: char lquote[MAXCCHARS+1] = {LQUOTE}; /* left quote character (`) */
87: char rquote[MAXCCHARS+1] = {RQUOTE}; /* right quote character (') */
88: char scommt[MAXCCHARS+1] = {SCOMMT}; /* start character for comment */
89: char ecommt[MAXCCHARS+1] = {ECOMMT}; /* end character for comment */
1.1 deraadt 90:
91: struct keyblk keywrds[] = { /* m4 keywords to be installed */
1.8 millert 92: { "include", INCLTYPE },
93: { "sinclude", SINCTYPE },
94: { "define", DEFITYPE },
95: { "defn", DEFNTYPE },
1.24 espie 96: { "divert", DIVRTYPE | NOARGS },
1.8 millert 97: { "expr", EXPRTYPE },
98: { "eval", EXPRTYPE },
99: { "substr", SUBSTYPE },
100: { "ifelse", IFELTYPE },
101: { "ifdef", IFDFTYPE },
102: { "len", LENGTYPE },
103: { "incr", INCRTYPE },
104: { "decr", DECRTYPE },
1.24 espie 105: { "dnl", DNLNTYPE | NOARGS },
106: { "changequote", CHNQTYPE | NOARGS },
107: { "changecom", CHNCTYPE | NOARGS },
1.8 millert 108: { "index", INDXTYPE },
1.1 deraadt 109: #ifdef EXTENDED
1.8 millert 110: { "paste", PASTTYPE },
111: { "spaste", SPASTYPE },
1.1 deraadt 112: #endif
1.8 millert 113: { "popdef", POPDTYPE },
114: { "pushdef", PUSDTYPE },
1.24 espie 115: { "dumpdef", DUMPTYPE | NOARGS },
116: { "shift", SHIFTYPE | NOARGS },
1.8 millert 117: { "translit", TRNLTYPE },
118: { "undefine", UNDFTYPE },
1.24 espie 119: { "undivert", UNDVTYPE | NOARGS },
120: { "divnum", DIVNTYPE | NOARGS },
1.8 millert 121: { "maketemp", MKTMTYPE },
1.24 espie 122: { "errprint", ERRPTYPE | NOARGS },
123: { "m4wrap", M4WRTYPE | NOARGS },
124: { "m4exit", EXITTYPE | NOARGS },
1.8 millert 125: { "syscmd", SYSCTYPE },
1.24 espie 126: { "sysval", SYSVTYPE | NOARGS },
1.1 deraadt 127:
1.24 espie 128: #if defined(unix) || defined(__unix__)
129: { "unix", SELFTYPE | NOARGS },
1.1 deraadt 130: #else
131: #ifdef vms
1.24 espie 132: { "vms", SELFTYPE | NOARGS },
1.1 deraadt 133: #endif
134: #endif
135: };
136:
137: #define MAXKEYS (sizeof(keywrds)/sizeof(struct keyblk))
138:
139: extern int optind;
140: extern char *optarg;
141:
1.27 espie 142: #define MAXRECORD 50
143: static struct position {
144: char *name;
145: unsigned long line;
146: } quotes[MAXRECORD], paren[MAXRECORD];
147:
148: static void record __P((struct position *, int));
149: static void dump_stack __P((struct position *, int));
150:
1.18 espie 151: static void macro __P((void));
152: static void initkwds __P((void));
1.25 espie 153: static ndptr inspect __P((char, char *));
1.18 espie 154: static int do_look_ahead __P((int, const char *));
155:
156: int main __P((int, char *[]));
1.1 deraadt 157:
158: int
159: main(argc,argv)
160: int argc;
161: char *argv[];
162: {
1.17 espie 163: int c;
164: int n;
1.1 deraadt 165: char *p;
1.17 espie 166: FILE *ifp;
1.1 deraadt 167:
168: if (signal(SIGINT, SIG_IGN) != SIG_IGN)
169: signal(SIGINT, onintr);
170:
171: initkwds();
1.14 espie 172: initspaces();
1.1 deraadt 173:
1.16 espie 174: while ((c = getopt(argc, argv, "tD:U:o:I:")) != -1)
1.1 deraadt 175: switch(c) {
176:
177: case 'D': /* define something..*/
178: for (p = optarg; *p; p++)
179: if (*p == '=')
180: break;
181: if (*p)
182: *p++ = EOS;
183: dodefine(optarg, p);
1.16 espie 184: break;
185: case 'I':
186: addtoincludepath(optarg);
1.1 deraadt 187: break;
188: case 'U': /* undefine... */
189: remhash(optarg, TOP);
190: break;
191: case 'o': /* specific output */
192: case '?':
193: usage();
194: }
195:
196: argc -= optind;
197: argv += optind;
198:
199: active = stdout; /* default active output */
200: bbase[0] = bufbase;
201: if (!argc) {
202: sp = -1; /* stack pointer initialized */
203: fp = 0; /* frame pointer initialized */
1.26 espie 204: set_input(infile+0, stdin, "stdin");
205: /* default input (naturally) */
1.1 deraadt 206: macro();
207: } else
208: for (; argc--; ++argv) {
209: p = *argv;
1.13 espie 210: if (p[0] == '-' && p[1] == EOS)
1.26 espie 211: set_input(infile, stdin, "stdin");
212: else if (fopen_trypath(infile, p) == NULL)
1.11 espie 213: err(1, "%s", p);
1.1 deraadt 214: sp = -1;
215: fp = 0;
216: macro();
1.26 espie 217: release_input(infile);
1.1 deraadt 218: }
219:
220: if (*m4wraps) { /* anything for rundown ?? */
221: ilevel = 0; /* in case m4wrap includes.. */
222: bufbase = bp = buf; /* use the entire buffer */
223: pbstr(m4wraps); /* user-defined wrapup act */
224: macro(); /* last will and testament */
225: }
226:
227: if (active != stdout)
228: active = stdout; /* reset output just in case */
229: for (n = 1; n < MAXOUT; n++) /* default wrap-up: undivert */
230: if (outfile[n] != NULL)
231: getdiv(n);
232: /* remove bitbucket if used */
233: if (outfile[0] != NULL) {
234: (void) fclose(outfile[0]);
235: }
236:
237: return 0;
238: }
239:
240: /*
1.21 espie 241: * Look ahead for `token'.
1.2 deraadt 242: * (on input `t == token[0]')
243: * Used for comment and quoting delimiters.
244: * Returns 1 if `token' present; copied to output.
245: * 0 if `token' not found; all characters pushed back
246: */
1.18 espie 247: static int
1.2 deraadt 248: do_look_ahead(t, token)
249: int t;
1.18 espie 250: const char *token;
1.2 deraadt 251: {
252: int i;
253:
1.21 espie 254: assert(t == token[0]);
1.2 deraadt 255:
256: for (i = 1; *++token; i++) {
257: t = gpbc();
258: if (t == EOF || t != *token) {
1.28 espie 259: putback(t);
1.2 deraadt 260: while (--i)
261: putback(*--token);
262: return 0;
263: }
264: }
265: return 1;
266: }
267:
268: #define LOOK_AHEAD(t, token) ((t)==(token)[0] && do_look_ahead(t,token))
269:
270: /*
1.1 deraadt 271: * macro - the work horse..
272: */
1.18 espie 273: static void
1.17 espie 274: macro()
275: {
1.29 ! espie 276: char token[MAXTOK];
1.17 espie 277: int t, l;
278: ndptr p;
279: int nlpar;
1.1 deraadt 280:
281: cycle {
1.2 deraadt 282: t = gpbc();
283: if (t == '_' || isalpha(t)) {
1.29 ! espie 284: p = inspect(t, token);
1.24 espie 285: if (p != nil)
286: putback(l = gpbc());
287: if (p == nil || (l != LPAREN &&
288: (p->type & NEEDARGS) != 0))
1.29 ! espie 289: outputstr(token);
1.1 deraadt 290: else {
291: /*
292: * real thing.. First build a call frame:
293: */
294: pushf(fp); /* previous call frm */
295: pushf(p->type); /* type of the call */
296: pushf(0); /* parenthesis level */
297: fp = sp; /* new frame pointer */
298: /*
299: * now push the string arguments:
300: */
301: pushs(p->defn); /* defn string */
302: pushs(p->name); /* macro name */
303: pushs(ep); /* start next..*/
304:
305: if (l != LPAREN) { /* add bracks */
306: putback(RPAREN);
307: putback(LPAREN);
308: }
309: }
310: }
311: else if (t == EOF) {
1.27 espie 312: if (sp > -1) {
313: warnx( "unexpected end of input, unclosed parenthesis:");
314: dump_stack(paren, PARLEV);
315: exit(1);
316: }
1.1 deraadt 317: if (ilevel <= 0)
318: break; /* all done thanks.. */
1.26 espie 319: release_input(infile+ilevel--);
1.1 deraadt 320: bufbase = bbase[ilevel];
321: continue;
322: }
323: /*
1.7 deraadt 324: * non-alpha token possibly seen..
1.1 deraadt 325: * [the order of else if .. stmts is important.]
326: */
1.2 deraadt 327: else if (LOOK_AHEAD(t,lquote)) { /* strip quotes */
1.27 espie 328: nlpar = 0;
329: record(quotes, nlpar++);
1.1 deraadt 330: do {
1.7 deraadt 331:
1.2 deraadt 332: l = gpbc();
1.7 deraadt 333: if (LOOK_AHEAD(l,rquote)) {
1.29 ! espie 334: if (--nlpar > 0)
! 335: outputstr(rquote);
1.7 deraadt 336: } else if (LOOK_AHEAD(l,lquote)) {
1.27 espie 337: record(quotes, nlpar++);
1.29 ! espie 338: outputstr(lquote);
1.17 espie 339: } else if (l == EOF) {
340: if (nlpar == 1)
1.27 espie 341: warnx("unclosed quote:");
1.17 espie 342: else
1.27 espie 343: warnx("%d unclosed quotes:", nlpar);
344: dump_stack(quotes, nlpar);
345: exit(1);
1.17 espie 346: } else {
1.29 ! espie 347: if (nlpar > 0) {
! 348: if (sp < 0)
! 349: putc(l, active);
! 350: else
! 351: chrsave(l);
! 352: }
1.7 deraadt 353: }
1.1 deraadt 354: }
355: while (nlpar != 0);
356: }
357:
1.2 deraadt 358: else if (sp < 0 && LOOK_AHEAD(t, scommt)) {
1.21 espie 359: fputs(scommt, active);
1.2 deraadt 360:
361: for(;;) {
362: t = gpbc();
363: if (LOOK_AHEAD(t, ecommt)) {
1.21 espie 364: fputs(ecommt, active);
1.2 deraadt 365: break;
366: }
367: if (t == EOF)
368: break;
1.1 deraadt 369: putc(t, active);
370: }
1.2 deraadt 371: }
372:
373: else if (sp < 0) { /* not in a macro at all */
1.1 deraadt 374: putc(t, active); /* output directly.. */
375: }
376:
377: else switch(t) {
378:
379: case LPAREN:
380: if (PARLEV > 0)
381: chrsave(t);
382: while (isspace(l = gpbc()))
383: ; /* skip blank, tab, nl.. */
384: putback(l);
1.27 espie 385: record(paren, PARLEV++);
1.1 deraadt 386: break;
387:
388: case RPAREN:
389: if (--PARLEV > 0)
390: chrsave(t);
391: else { /* end of argument list */
392: chrsave(EOS);
393:
394: if (sp == STACKMAX)
1.11 espie 395: errx(1, "internal stack overflow");
1.1 deraadt 396:
397: if (CALTYP == MACRTYPE)
1.18 espie 398: expand((const char **) mstack+fp+1, sp-fp);
1.1 deraadt 399: else
1.18 espie 400: eval((const char **) mstack+fp+1, sp-fp, CALTYP);
1.1 deraadt 401:
402: ep = PREVEP; /* flush strspace */
403: sp = PREVSP; /* previous sp.. */
404: fp = PREVFP; /* rewind stack...*/
405: }
406: break;
407:
408: case COMMA:
409: if (PARLEV == 1) {
410: chrsave(EOS); /* new argument */
411: while (isspace(l = gpbc()))
412: ;
413: putback(l);
414: pushs(ep);
415: } else
416: chrsave(t);
417: break;
418:
419: default:
1.22 espie 420: if (LOOK_AHEAD(t, scommt)) {
421: char *p;
422: for (p = scommt; *p; p++)
423: chrsave(*p);
424: for(;;) {
425: t = gpbc();
426: if (LOOK_AHEAD(t, ecommt)) {
427: for (p = ecommt; *p; p++)
428: chrsave(*p);
429: break;
430: }
431: if (t == EOF)
432: break;
433: chrsave(t);
434: }
435: } else
436: chrsave(t); /* stack the char */
1.1 deraadt 437: break;
438: }
439: }
440: }
441:
1.24 espie 442: /*
443: * output string directly, without pushing it for reparses.
444: */
445: void
446: outputstr(s)
447: const char *s;
448: {
449: if (sp < 0)
450: while (*s)
451: putc(*s++, active);
452: else
453: while (*s)
454: chrsave(*s++);
455: }
456:
1.1 deraadt 457: /*
458: * build an input token..
459: * consider only those starting with _ or A-Za-z. This is a
460: * combo with lookup to speed things up.
461: */
1.18 espie 462: static ndptr
1.25 espie 463: inspect(c, tp)
464: char c;
1.17 espie 465: char *tp;
1.1 deraadt 466: {
1.17 espie 467: char *name = tp;
468: char *etp = tp+MAXTOK;
469: ndptr p;
1.25 espie 470: unsigned int h;
471:
472: h = *tp++ = c;
1.1 deraadt 473:
474: while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
475: h = (h << 5) + h + (*tp++ = c);
476: putback(c);
477: if (tp == etp)
1.11 espie 478: errx(1, "token too long");
1.1 deraadt 479:
480: *tp = EOS;
481:
1.19 espie 482: for (p = hashtab[h % HASHSIZE]; p != nil; p = p->nxtptr)
483: if (h == p->hv && STREQ(name, p->name))
1.1 deraadt 484: break;
485: return p;
486: }
487:
488: /*
489: * initkwds - initialise m4 keywords as fast as possible.
490: * This very similar to install, but without certain overheads,
491: * such as calling lookup. Malloc is not used for storing the
1.17 espie 492: * keyword strings, since we simply use the static pointers
1.1 deraadt 493: * within keywrds block.
494: */
1.18 espie 495: static void
1.17 espie 496: initkwds()
497: {
498: size_t i;
1.20 millert 499: unsigned int h;
1.17 espie 500: ndptr p;
1.1 deraadt 501:
502: for (i = 0; i < MAXKEYS; i++) {
503: h = hash(keywrds[i].knam);
504: p = (ndptr) xalloc(sizeof(struct ndblock));
1.19 espie 505: p->nxtptr = hashtab[h % HASHSIZE];
506: hashtab[h % HASHSIZE] = p;
1.1 deraadt 507: p->name = keywrds[i].knam;
508: p->defn = null;
1.19 espie 509: p->hv = h;
1.24 espie 510: p->type = (keywrds[i].ktyp & TYPEMASK) | STATIC;
511: if ((keywrds[i].ktyp & NOARGS) == 0)
512: p->type |= NEEDARGS;
1.1 deraadt 513: }
514: }
1.17 espie 515:
1.27 espie 516: static void
517: record(t, lev)
518: struct position *t;
519: int lev;
520: {
521: if (lev < MAXRECORD) {
522: t[lev].name = CURRENT_NAME;
523: t[lev].line = CURRENT_LINE;
524: }
525: }
526:
527: static void
528: dump_stack(t, lev)
529: struct position *t;
530: int lev;
531: {
532: int i;
533:
534: for (i = 0; i < lev; i++) {
535: if (i == MAXRECORD) {
536: fprintf(stderr, " ...\n");
537: break;
538: }
539: fprintf(stderr, " %s at line %lu\n",
540: t[i].name, t[i].line);
541: }
542: }