Annotation of src/usr.bin/m4/main.c, Revision 1.14
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. */
75: stae mstack[STACKMAX+1]; /* stack of m4 machine */
76: int sp; /* current m4 stack pointer */
77: int fp; /* m4 call frame pointer */
78: FILE *infile[MAXINP]; /* input file stack (0=stdin) */
79: FILE *outfile[MAXOUT]; /* diversion array(0=bitbucket)*/
80: FILE *active; /* active output file pointer */
81: int ilevel = 0; /* input file stack pointer */
82: int oindex = 0; /* diversion index.. */
83: char *null = ""; /* as it says.. just a null.. */
84: char *m4wraps = ""; /* m4wrap string default.. */
1.2 deraadt 85: char lquote[MAXCCHARS+1] = {LQUOTE}; /* left quote character (`) */
86: char rquote[MAXCCHARS+1] = {RQUOTE}; /* right quote character (') */
87: char scommt[MAXCCHARS+1] = {SCOMMT}; /* start character for comment */
88: char ecommt[MAXCCHARS+1] = {ECOMMT}; /* end character for comment */
1.1 deraadt 89:
90: struct keyblk keywrds[] = { /* m4 keywords to be installed */
1.8 millert 91: { "include", INCLTYPE },
92: { "sinclude", SINCTYPE },
93: { "define", DEFITYPE },
94: { "defn", DEFNTYPE },
95: { "divert", DIVRTYPE },
96: { "expr", EXPRTYPE },
97: { "eval", EXPRTYPE },
98: { "substr", SUBSTYPE },
99: { "ifelse", IFELTYPE },
100: { "ifdef", IFDFTYPE },
101: { "len", LENGTYPE },
102: { "incr", INCRTYPE },
103: { "decr", DECRTYPE },
104: { "dnl", DNLNTYPE },
105: { "changequote", CHNQTYPE },
106: { "changecom", CHNCTYPE },
107: { "index", INDXTYPE },
1.1 deraadt 108: #ifdef EXTENDED
1.8 millert 109: { "paste", PASTTYPE },
110: { "spaste", SPASTYPE },
1.1 deraadt 111: #endif
1.8 millert 112: { "popdef", POPDTYPE },
113: { "pushdef", PUSDTYPE },
114: { "dumpdef", DUMPTYPE },
115: { "shift", SHIFTYPE },
116: { "translit", TRNLTYPE },
117: { "undefine", UNDFTYPE },
118: { "undivert", UNDVTYPE },
119: { "divnum", DIVNTYPE },
120: { "maketemp", MKTMTYPE },
121: { "errprint", ERRPTYPE },
122: { "m4wrap", M4WRTYPE },
123: { "m4exit", EXITTYPE },
124: { "syscmd", SYSCTYPE },
125: { "sysval", SYSVTYPE },
1.1 deraadt 126:
1.4 deraadt 127: #if defined(unix) || defined(__NetBSD__) || defined(__OpenBSD__)
1.8 millert 128: { "unix", MACRTYPE },
1.1 deraadt 129: #else
130: #ifdef vms
1.8 millert 131: { "vms", MACRTYPE },
1.1 deraadt 132: #endif
133: #endif
134: };
135:
136: #define MAXKEYS (sizeof(keywrds)/sizeof(struct keyblk))
137:
138: extern int optind;
139: extern char *optarg;
140:
141: void macro();
142: void initkwds();
143: extern int getopt();
144:
145: int
146: main(argc,argv)
147: int argc;
148: char *argv[];
149: {
150: register int c;
151: register int n;
152: char *p;
153: register FILE *ifp;
154:
155: if (signal(SIGINT, SIG_IGN) != SIG_IGN)
156: signal(SIGINT, onintr);
157:
158: initkwds();
1.14 ! espie 159: initspaces();
1.1 deraadt 160:
1.5 millert 161: while ((c = getopt(argc, argv, "tD:U:o:")) != -1)
1.1 deraadt 162: switch(c) {
163:
164: case 'D': /* define something..*/
165: for (p = optarg; *p; p++)
166: if (*p == '=')
167: break;
168: if (*p)
169: *p++ = EOS;
170: dodefine(optarg, p);
171: break;
172: case 'U': /* undefine... */
173: remhash(optarg, TOP);
174: break;
175: case 'o': /* specific output */
176: case '?':
177: usage();
178: }
179:
180: argc -= optind;
181: argv += optind;
182:
183: active = stdout; /* default active output */
184: bbase[0] = bufbase;
185: if (!argc) {
186: sp = -1; /* stack pointer initialized */
187: fp = 0; /* frame pointer initialized */
188: infile[0] = stdin; /* default input (naturally) */
189: macro();
190: } else
191: for (; argc--; ++argv) {
192: p = *argv;
1.13 espie 193: if (p[0] == '-' && p[1] == EOS)
1.1 deraadt 194: ifp = stdin;
195: else if ((ifp = fopen(p, "r")) == NULL)
1.11 espie 196: err(1, "%s", p);
1.1 deraadt 197: sp = -1;
198: fp = 0;
199: infile[0] = ifp;
200: macro();
201: if (ifp != stdin)
202: (void)fclose(ifp);
203: }
204:
205: if (*m4wraps) { /* anything for rundown ?? */
206: ilevel = 0; /* in case m4wrap includes.. */
207: bufbase = bp = buf; /* use the entire buffer */
208: putback(EOF); /* eof is a must !! */
209: pbstr(m4wraps); /* user-defined wrapup act */
210: macro(); /* last will and testament */
211: }
212:
213: if (active != stdout)
214: active = stdout; /* reset output just in case */
215: for (n = 1; n < MAXOUT; n++) /* default wrap-up: undivert */
216: if (outfile[n] != NULL)
217: getdiv(n);
218: /* remove bitbucket if used */
219: if (outfile[0] != NULL) {
220: (void) fclose(outfile[0]);
221: }
222:
223: return 0;
224: }
225:
226: ndptr inspect();
227:
228: /*
1.2 deraadt 229: * Look ahead (at most MAXCCHARS characters) for `token'.
230: * (on input `t == token[0]')
231: * Used for comment and quoting delimiters.
232: * Returns 1 if `token' present; copied to output.
233: * 0 if `token' not found; all characters pushed back
234: */
235: int
236: do_look_ahead(t, token)
237: int t;
238: char *token;
239: {
240: int i;
241:
242: if (t != token[0])
1.11 espie 243: errx(1, "internal error");
1.2 deraadt 244:
245: for (i = 1; *++token; i++) {
246: t = gpbc();
247: if (t == EOF || t != *token) {
248: if (t != EOF)
249: putback(t);
250: while (--i)
251: putback(*--token);
252: return 0;
253: }
254: }
255: return 1;
256: }
257:
258: #define LOOK_AHEAD(t, token) ((t)==(token)[0] && do_look_ahead(t,token))
259:
260: /*
1.1 deraadt 261: * macro - the work horse..
262: */
263: void
264: macro() {
1.7 deraadt 265: char token[MAXTOK], chars[2];
1.1 deraadt 266: register char *s;
267: register int t, l;
268: register ndptr p;
269: register int nlpar;
270:
271: cycle {
1.2 deraadt 272: t = gpbc();
273: if (t == '_' || isalpha(t)) {
1.1 deraadt 274: putback(t);
1.10 deraadt 275: s = token;
1.9 mickey 276: if ((p = inspect(s)) == nil) {
1.1 deraadt 277: if (sp < 0)
278: while (*s)
279: putc(*s++, active);
280: else
281: while (*s)
282: chrsave(*s++);
283: }
284: else {
285: /*
286: * real thing.. First build a call frame:
287: */
288: pushf(fp); /* previous call frm */
289: pushf(p->type); /* type of the call */
290: pushf(0); /* parenthesis level */
291: fp = sp; /* new frame pointer */
292: /*
293: * now push the string arguments:
294: */
295: pushs(p->defn); /* defn string */
296: pushs(p->name); /* macro name */
297: pushs(ep); /* start next..*/
298:
299: putback(l = gpbc());
300: if (l != LPAREN) { /* add bracks */
301: putback(RPAREN);
302: putback(LPAREN);
303: }
304: }
305: }
306: else if (t == EOF) {
307: if (sp > -1)
1.11 espie 308: errx(1, "unexpected end of input");
1.1 deraadt 309: if (ilevel <= 0)
310: break; /* all done thanks.. */
311: --ilevel;
312: (void) fclose(infile[ilevel+1]);
313: bufbase = bbase[ilevel];
314: continue;
315: }
316: /*
1.7 deraadt 317: * non-alpha token possibly seen..
1.1 deraadt 318: * [the order of else if .. stmts is important.]
319: */
1.2 deraadt 320: else if (LOOK_AHEAD(t,lquote)) { /* strip quotes */
1.1 deraadt 321: nlpar = 1;
322: do {
1.7 deraadt 323:
1.2 deraadt 324: l = gpbc();
1.7 deraadt 325: if (LOOK_AHEAD(l,rquote)) {
1.1 deraadt 326: nlpar--;
1.7 deraadt 327: s = rquote;
328: } else if (LOOK_AHEAD(l,lquote)) {
1.1 deraadt 329: nlpar++;
1.7 deraadt 330: s = lquote;
331: } else if (l == EOF)
1.11 espie 332: errx(1, "missing right quote");
1.7 deraadt 333: else {
334: chars[0] = l;
1.13 espie 335: chars[1] = EOS;
1.7 deraadt 336: s = chars;
337: }
1.1 deraadt 338: if (nlpar > 0) {
339: if (sp < 0)
1.7 deraadt 340: while (*s)
341: putc(*s++, active);
1.1 deraadt 342: else
1.7 deraadt 343: while (*s)
344: chrsave(*s++);
1.1 deraadt 345: }
346: }
347: while (nlpar != 0);
348: }
349:
1.2 deraadt 350: else if (sp < 0 && LOOK_AHEAD(t, scommt)) {
351: int i;
352: for (i = 0; i < MAXCCHARS && scommt[i]; i++)
353: putc(scommt[i], active);
354:
355: for(;;) {
356: t = gpbc();
357: if (LOOK_AHEAD(t, ecommt)) {
358: for (i = 0; i < MAXCCHARS && ecommt[i];
359: i++)
360: putc(ecommt[i], active);
361: break;
362: }
363: if (t == EOF)
364: break;
1.1 deraadt 365: putc(t, active);
366: }
1.2 deraadt 367: }
368:
369: else if (sp < 0) { /* not in a macro at all */
1.1 deraadt 370: putc(t, active); /* output directly.. */
371: }
372:
373: else switch(t) {
374:
375: case LPAREN:
376: if (PARLEV > 0)
377: chrsave(t);
378: while (isspace(l = gpbc()))
379: ; /* skip blank, tab, nl.. */
380: putback(l);
381: PARLEV++;
382: break;
383:
384: case RPAREN:
385: if (--PARLEV > 0)
386: chrsave(t);
387: else { /* end of argument list */
388: chrsave(EOS);
389:
390: if (sp == STACKMAX)
1.11 espie 391: errx(1, "internal stack overflow");
1.1 deraadt 392:
393: if (CALTYP == MACRTYPE)
394: expand((char **) mstack+fp+1, sp-fp);
395: else
396: eval((char **) mstack+fp+1, sp-fp, CALTYP);
397:
398: ep = PREVEP; /* flush strspace */
399: sp = PREVSP; /* previous sp.. */
400: fp = PREVFP; /* rewind stack...*/
401: }
402: break;
403:
404: case COMMA:
405: if (PARLEV == 1) {
406: chrsave(EOS); /* new argument */
407: while (isspace(l = gpbc()))
408: ;
409: putback(l);
410: pushs(ep);
411: } else
412: chrsave(t);
413: break;
414:
415: default:
416: chrsave(t); /* stack the char */
417: break;
418: }
419: }
420: }
421:
422: /*
423: * build an input token..
424: * consider only those starting with _ or A-Za-z. This is a
425: * combo with lookup to speed things up.
426: */
427: ndptr
428: inspect(tp)
429: register char *tp;
430: {
431: register char c;
432: register char *name = tp;
433: register char *etp = tp+MAXTOK;
434: register ndptr p;
435: register unsigned long h = 0;
436:
437: while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
438: h = (h << 5) + h + (*tp++ = c);
439: putback(c);
440: if (tp == etp)
1.11 espie 441: errx(1, "token too long");
1.1 deraadt 442:
443: *tp = EOS;
444:
445: for (p = hashtab[h%HASHSIZE]; p != nil; p = p->nxtptr)
446: if (STREQ(name, p->name))
447: break;
448: return p;
449: }
450:
451: /*
452: * initkwds - initialise m4 keywords as fast as possible.
453: * This very similar to install, but without certain overheads,
454: * such as calling lookup. Malloc is not used for storing the
455: * keyword strings, since we simply use the static pointers
456: * within keywrds block.
457: */
458: void
459: initkwds() {
460: register int i;
461: register int h;
462: register ndptr p;
463:
464: for (i = 0; i < MAXKEYS; i++) {
465: h = hash(keywrds[i].knam);
466: p = (ndptr) xalloc(sizeof(struct ndblock));
467: p->nxtptr = hashtab[h];
468: hashtab[h] = p;
469: p->name = keywrds[i].knam;
470: p->defn = null;
471: p->type = keywrds[i].ktyp | STATIC;
472: }
473: }