Annotation of src/usr.bin/m4/main.c, Revision 1.13
1.11 espie 1: /* $OpenBSD: main.c,v 1.10 1998/06/02 20:46:40 deraadt 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
1.11 espie 50: static char rcsid[] = "$OpenBSD: main.c,v 1.10 1998/06/02 20:46:40 deraadt 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>
61: #include <signal.h>
62: #include <errno.h>
63: #include <unistd.h>
64: #include <stdio.h>
65: #include <ctype.h>
66: #include <string.h>
1.13 ! espie 67: #include <stddef.h>
1.11 espie 68: #include <err.h>
1.1 deraadt 69: #include "mdef.h"
70: #include "stdd.h"
71: #include "extern.h"
72: #include "pathnames.h"
73:
74: ndptr hashtab[HASHSIZE]; /* hash table for macros etc. */
1.6 deraadt 75: pbent buf[BUFSIZE]; /* push-back buffer */
76: pbent *bufbase = buf; /* the base for current ilevel */
77: pbent *bbase[MAXINP]; /* the base for each ilevel */
78: pbent *bp = buf; /* first available character */
79: pbent *endpbb = buf+BUFSIZE; /* end of push-back buffer */
1.1 deraadt 80: stae mstack[STACKMAX+1]; /* stack of m4 machine */
81: char strspace[STRSPMAX+1]; /* string space for evaluation */
82: char *ep = strspace; /* first free char in strspace */
83: char *endest= strspace+STRSPMAX;/* end of string space */
84: int sp; /* current m4 stack pointer */
85: int fp; /* m4 call frame pointer */
86: FILE *infile[MAXINP]; /* input file stack (0=stdin) */
87: FILE *outfile[MAXOUT]; /* diversion array(0=bitbucket)*/
88: FILE *active; /* active output file pointer */
89: int ilevel = 0; /* input file stack pointer */
90: int oindex = 0; /* diversion index.. */
91: char *null = ""; /* as it says.. just a null.. */
92: char *m4wraps = ""; /* m4wrap string default.. */
1.2 deraadt 93: char lquote[MAXCCHARS+1] = {LQUOTE}; /* left quote character (`) */
94: char rquote[MAXCCHARS+1] = {RQUOTE}; /* right quote character (') */
95: char scommt[MAXCCHARS+1] = {SCOMMT}; /* start character for comment */
96: char ecommt[MAXCCHARS+1] = {ECOMMT}; /* end character for comment */
1.1 deraadt 97:
98: struct keyblk keywrds[] = { /* m4 keywords to be installed */
1.8 millert 99: { "include", INCLTYPE },
100: { "sinclude", SINCTYPE },
101: { "define", DEFITYPE },
102: { "defn", DEFNTYPE },
103: { "divert", DIVRTYPE },
104: { "expr", EXPRTYPE },
105: { "eval", EXPRTYPE },
106: { "substr", SUBSTYPE },
107: { "ifelse", IFELTYPE },
108: { "ifdef", IFDFTYPE },
109: { "len", LENGTYPE },
110: { "incr", INCRTYPE },
111: { "decr", DECRTYPE },
112: { "dnl", DNLNTYPE },
113: { "changequote", CHNQTYPE },
114: { "changecom", CHNCTYPE },
115: { "index", INDXTYPE },
1.1 deraadt 116: #ifdef EXTENDED
1.8 millert 117: { "paste", PASTTYPE },
118: { "spaste", SPASTYPE },
1.1 deraadt 119: #endif
1.8 millert 120: { "popdef", POPDTYPE },
121: { "pushdef", PUSDTYPE },
122: { "dumpdef", DUMPTYPE },
123: { "shift", SHIFTYPE },
124: { "translit", TRNLTYPE },
125: { "undefine", UNDFTYPE },
126: { "undivert", UNDVTYPE },
127: { "divnum", DIVNTYPE },
128: { "maketemp", MKTMTYPE },
129: { "errprint", ERRPTYPE },
130: { "m4wrap", M4WRTYPE },
131: { "m4exit", EXITTYPE },
132: { "syscmd", SYSCTYPE },
133: { "sysval", SYSVTYPE },
1.1 deraadt 134:
1.4 deraadt 135: #if defined(unix) || defined(__NetBSD__) || defined(__OpenBSD__)
1.8 millert 136: { "unix", MACRTYPE },
1.1 deraadt 137: #else
138: #ifdef vms
1.8 millert 139: { "vms", MACRTYPE },
1.1 deraadt 140: #endif
141: #endif
142: };
143:
144: #define MAXKEYS (sizeof(keywrds)/sizeof(struct keyblk))
145:
146: extern int optind;
147: extern char *optarg;
148:
149: void macro();
150: void initkwds();
151: extern int getopt();
152:
153: int
154: main(argc,argv)
155: int argc;
156: char *argv[];
157: {
158: register int c;
159: register int n;
160: char *p;
161: register FILE *ifp;
162:
163: if (signal(SIGINT, SIG_IGN) != SIG_IGN)
164: signal(SIGINT, onintr);
165:
166: initkwds();
167:
1.5 millert 168: while ((c = getopt(argc, argv, "tD:U:o:")) != -1)
1.1 deraadt 169: switch(c) {
170:
171: case 'D': /* define something..*/
172: for (p = optarg; *p; p++)
173: if (*p == '=')
174: break;
175: if (*p)
176: *p++ = EOS;
177: dodefine(optarg, p);
178: break;
179: case 'U': /* undefine... */
180: remhash(optarg, TOP);
181: break;
182: case 'o': /* specific output */
183: case '?':
184: usage();
185: }
186:
187: argc -= optind;
188: argv += optind;
189:
190: active = stdout; /* default active output */
191: bbase[0] = bufbase;
192: if (!argc) {
193: sp = -1; /* stack pointer initialized */
194: fp = 0; /* frame pointer initialized */
195: infile[0] = stdin; /* default input (naturally) */
196: macro();
197: } else
198: for (; argc--; ++argv) {
199: p = *argv;
1.13 ! espie 200: if (p[0] == '-' && p[1] == EOS)
1.1 deraadt 201: ifp = stdin;
202: else if ((ifp = fopen(p, "r")) == NULL)
1.11 espie 203: err(1, "%s", p);
1.1 deraadt 204: sp = -1;
205: fp = 0;
206: infile[0] = ifp;
207: macro();
208: if (ifp != stdin)
209: (void)fclose(ifp);
210: }
211:
212: if (*m4wraps) { /* anything for rundown ?? */
213: ilevel = 0; /* in case m4wrap includes.. */
214: bufbase = bp = buf; /* use the entire buffer */
215: putback(EOF); /* eof is a must !! */
216: pbstr(m4wraps); /* user-defined wrapup act */
217: macro(); /* last will and testament */
218: }
219:
220: if (active != stdout)
221: active = stdout; /* reset output just in case */
222: for (n = 1; n < MAXOUT; n++) /* default wrap-up: undivert */
223: if (outfile[n] != NULL)
224: getdiv(n);
225: /* remove bitbucket if used */
226: if (outfile[0] != NULL) {
227: (void) fclose(outfile[0]);
228: }
229:
230: return 0;
231: }
232:
233: ndptr inspect();
234:
235: /*
1.2 deraadt 236: * Look ahead (at most MAXCCHARS characters) for `token'.
237: * (on input `t == token[0]')
238: * Used for comment and quoting delimiters.
239: * Returns 1 if `token' present; copied to output.
240: * 0 if `token' not found; all characters pushed back
241: */
242: int
243: do_look_ahead(t, token)
244: int t;
245: char *token;
246: {
247: int i;
248:
249: if (t != token[0])
1.11 espie 250: errx(1, "internal error");
1.2 deraadt 251:
252: for (i = 1; *++token; i++) {
253: t = gpbc();
254: if (t == EOF || t != *token) {
255: if (t != EOF)
256: putback(t);
257: while (--i)
258: putback(*--token);
259: return 0;
260: }
261: }
262: return 1;
263: }
264:
265: #define LOOK_AHEAD(t, token) ((t)==(token)[0] && do_look_ahead(t,token))
266:
267: /*
1.1 deraadt 268: * macro - the work horse..
269: */
270: void
271: macro() {
1.7 deraadt 272: char token[MAXTOK], chars[2];
1.1 deraadt 273: register char *s;
274: register int t, l;
275: register ndptr p;
276: register int nlpar;
277:
278: cycle {
1.2 deraadt 279: t = gpbc();
280: if (t == '_' || isalpha(t)) {
1.1 deraadt 281: putback(t);
1.10 deraadt 282: s = token;
1.9 mickey 283: if ((p = inspect(s)) == nil) {
1.1 deraadt 284: if (sp < 0)
285: while (*s)
286: putc(*s++, active);
287: else
288: while (*s)
289: chrsave(*s++);
290: }
291: else {
292: /*
293: * real thing.. First build a call frame:
294: */
295: pushf(fp); /* previous call frm */
296: pushf(p->type); /* type of the call */
297: pushf(0); /* parenthesis level */
298: fp = sp; /* new frame pointer */
299: /*
300: * now push the string arguments:
301: */
302: pushs(p->defn); /* defn string */
303: pushs(p->name); /* macro name */
304: pushs(ep); /* start next..*/
305:
306: putback(l = gpbc());
307: if (l != LPAREN) { /* add bracks */
308: putback(RPAREN);
309: putback(LPAREN);
310: }
311: }
312: }
313: else if (t == EOF) {
314: if (sp > -1)
1.11 espie 315: errx(1, "unexpected end of input");
1.1 deraadt 316: if (ilevel <= 0)
317: break; /* all done thanks.. */
318: --ilevel;
319: (void) fclose(infile[ilevel+1]);
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 */
1.1 deraadt 328: nlpar = 1;
329: do {
1.7 deraadt 330:
1.2 deraadt 331: l = gpbc();
1.7 deraadt 332: if (LOOK_AHEAD(l,rquote)) {
1.1 deraadt 333: nlpar--;
1.7 deraadt 334: s = rquote;
335: } else if (LOOK_AHEAD(l,lquote)) {
1.1 deraadt 336: nlpar++;
1.7 deraadt 337: s = lquote;
338: } else if (l == EOF)
1.11 espie 339: errx(1, "missing right quote");
1.7 deraadt 340: else {
341: chars[0] = l;
1.13 ! espie 342: chars[1] = EOS;
1.7 deraadt 343: s = chars;
344: }
1.1 deraadt 345: if (nlpar > 0) {
346: if (sp < 0)
1.7 deraadt 347: while (*s)
348: putc(*s++, active);
1.1 deraadt 349: else
1.7 deraadt 350: while (*s)
351: chrsave(*s++);
1.1 deraadt 352: }
353: }
354: while (nlpar != 0);
355: }
356:
1.2 deraadt 357: else if (sp < 0 && LOOK_AHEAD(t, scommt)) {
358: int i;
359: for (i = 0; i < MAXCCHARS && scommt[i]; i++)
360: putc(scommt[i], active);
361:
362: for(;;) {
363: t = gpbc();
364: if (LOOK_AHEAD(t, ecommt)) {
365: for (i = 0; i < MAXCCHARS && ecommt[i];
366: i++)
367: putc(ecommt[i], active);
368: break;
369: }
370: if (t == EOF)
371: break;
1.1 deraadt 372: putc(t, active);
373: }
1.2 deraadt 374: }
375:
376: else if (sp < 0) { /* not in a macro at all */
1.1 deraadt 377: putc(t, active); /* output directly.. */
378: }
379:
380: else switch(t) {
381:
382: case LPAREN:
383: if (PARLEV > 0)
384: chrsave(t);
385: while (isspace(l = gpbc()))
386: ; /* skip blank, tab, nl.. */
387: putback(l);
388: PARLEV++;
389: break;
390:
391: case RPAREN:
392: if (--PARLEV > 0)
393: chrsave(t);
394: else { /* end of argument list */
395: chrsave(EOS);
396:
397: if (sp == STACKMAX)
1.11 espie 398: errx(1, "internal stack overflow");
1.1 deraadt 399:
400: if (CALTYP == MACRTYPE)
401: expand((char **) mstack+fp+1, sp-fp);
402: else
403: eval((char **) mstack+fp+1, sp-fp, CALTYP);
404:
405: ep = PREVEP; /* flush strspace */
406: sp = PREVSP; /* previous sp.. */
407: fp = PREVFP; /* rewind stack...*/
408: }
409: break;
410:
411: case COMMA:
412: if (PARLEV == 1) {
413: chrsave(EOS); /* new argument */
414: while (isspace(l = gpbc()))
415: ;
416: putback(l);
417: pushs(ep);
418: } else
419: chrsave(t);
420: break;
421:
422: default:
423: chrsave(t); /* stack the char */
424: break;
425: }
426: }
427: }
428:
429: /*
430: * build an input token..
431: * consider only those starting with _ or A-Za-z. This is a
432: * combo with lookup to speed things up.
433: */
434: ndptr
435: inspect(tp)
436: register char *tp;
437: {
438: register char c;
439: register char *name = tp;
440: register char *etp = tp+MAXTOK;
441: register ndptr p;
442: register unsigned long h = 0;
443:
444: while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
445: h = (h << 5) + h + (*tp++ = c);
446: putback(c);
447: if (tp == etp)
1.11 espie 448: errx(1, "token too long");
1.1 deraadt 449:
450: *tp = EOS;
451:
452: for (p = hashtab[h%HASHSIZE]; p != nil; p = p->nxtptr)
453: if (STREQ(name, p->name))
454: break;
455: return p;
456: }
457:
458: /*
459: * initkwds - initialise m4 keywords as fast as possible.
460: * This very similar to install, but without certain overheads,
461: * such as calling lookup. Malloc is not used for storing the
462: * keyword strings, since we simply use the static pointers
463: * within keywrds block.
464: */
465: void
466: initkwds() {
467: register int i;
468: register int h;
469: register ndptr p;
470:
471: for (i = 0; i < MAXKEYS; i++) {
472: h = hash(keywrds[i].knam);
473: p = (ndptr) xalloc(sizeof(struct ndblock));
474: p->nxtptr = hashtab[h];
475: hashtab[h] = p;
476: p->name = keywrds[i].knam;
477: p->defn = null;
478: p->type = keywrds[i].ktyp | STATIC;
479: }
480: }