Annotation of src/usr.bin/m4/main.c, Revision 1.23
1.23 ! espie 1: /* $OpenBSD: main.c,v 1.22 1999/11/30 22:24:20 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
1.23 ! espie 50: static char rcsid[] = "$OpenBSD: main.c,v 1.22 1999/11/30 22:24:20 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 */
79: FILE *infile[MAXINP]; /* input file stack (0=stdin) */
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 },
96: { "divert", DIVRTYPE },
97: { "expr", EXPRTYPE },
98: { "eval", EXPRTYPE },
99: { "substr", SUBSTYPE },
100: { "ifelse", IFELTYPE },
101: { "ifdef", IFDFTYPE },
102: { "len", LENGTYPE },
103: { "incr", INCRTYPE },
104: { "decr", DECRTYPE },
105: { "dnl", DNLNTYPE },
106: { "changequote", CHNQTYPE },
107: { "changecom", CHNCTYPE },
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 },
115: { "dumpdef", DUMPTYPE },
116: { "shift", SHIFTYPE },
117: { "translit", TRNLTYPE },
118: { "undefine", UNDFTYPE },
119: { "undivert", UNDVTYPE },
120: { "divnum", DIVNTYPE },
121: { "maketemp", MKTMTYPE },
122: { "errprint", ERRPTYPE },
123: { "m4wrap", M4WRTYPE },
124: { "m4exit", EXITTYPE },
125: { "syscmd", SYSCTYPE },
126: { "sysval", SYSVTYPE },
1.1 deraadt 127:
1.23 ! espie 128: #if defined(unix) || defined(__unix__)
! 129: { "unix", SELFTYPE },
1.1 deraadt 130: #else
131: #ifdef vms
1.23 ! espie 132: { "vms", SELFTYPE },
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:
1.18 espie 142: static void macro __P((void));
143: static void initkwds __P((void));
144: static ndptr inspect __P((char *));
145: static int do_look_ahead __P((int, const char *));
146:
147: int main __P((int, char *[]));
1.1 deraadt 148:
149: int
150: main(argc,argv)
151: int argc;
152: char *argv[];
153: {
1.17 espie 154: int c;
155: int n;
1.1 deraadt 156: char *p;
1.17 espie 157: FILE *ifp;
1.1 deraadt 158:
159: if (signal(SIGINT, SIG_IGN) != SIG_IGN)
160: signal(SIGINT, onintr);
161:
162: initkwds();
1.14 espie 163: initspaces();
1.1 deraadt 164:
1.16 espie 165: while ((c = getopt(argc, argv, "tD:U:o:I:")) != -1)
1.1 deraadt 166: switch(c) {
167:
168: case 'D': /* define something..*/
169: for (p = optarg; *p; p++)
170: if (*p == '=')
171: break;
172: if (*p)
173: *p++ = EOS;
174: dodefine(optarg, p);
1.16 espie 175: break;
176: case 'I':
177: addtoincludepath(optarg);
1.1 deraadt 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;
1.15 espie 202: else if ((ifp = fopen_trypath(p)) == 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: /*
1.21 espie 234: * Look ahead for `token'.
1.2 deraadt 235: * (on input `t == token[0]')
236: * Used for comment and quoting delimiters.
237: * Returns 1 if `token' present; copied to output.
238: * 0 if `token' not found; all characters pushed back
239: */
1.18 espie 240: static int
1.2 deraadt 241: do_look_ahead(t, token)
242: int t;
1.18 espie 243: const char *token;
1.2 deraadt 244: {
245: int i;
246:
1.21 espie 247: assert(t == token[0]);
1.2 deraadt 248:
249: for (i = 1; *++token; i++) {
250: t = gpbc();
251: if (t == EOF || t != *token) {
252: if (t != EOF)
253: putback(t);
254: while (--i)
255: putback(*--token);
256: return 0;
257: }
258: }
259: return 1;
260: }
261:
262: #define LOOK_AHEAD(t, token) ((t)==(token)[0] && do_look_ahead(t,token))
263:
264: /*
1.1 deraadt 265: * macro - the work horse..
266: */
1.18 espie 267: static void
1.17 espie 268: macro()
269: {
1.7 deraadt 270: char token[MAXTOK], chars[2];
1.17 espie 271: char *s;
272: int t, l;
273: ndptr p;
274: int nlpar;
1.1 deraadt 275:
276: cycle {
1.2 deraadt 277: t = gpbc();
278: if (t == '_' || isalpha(t)) {
1.1 deraadt 279: putback(t);
1.10 deraadt 280: s = token;
1.9 mickey 281: if ((p = inspect(s)) == nil) {
1.1 deraadt 282: if (sp < 0)
283: while (*s)
284: putc(*s++, active);
285: else
286: while (*s)
287: chrsave(*s++);
288: }
289: else {
290: /*
291: * real thing.. First build a call frame:
292: */
293: pushf(fp); /* previous call frm */
294: pushf(p->type); /* type of the call */
295: pushf(0); /* parenthesis level */
296: fp = sp; /* new frame pointer */
297: /*
298: * now push the string arguments:
299: */
300: pushs(p->defn); /* defn string */
301: pushs(p->name); /* macro name */
302: pushs(ep); /* start next..*/
303:
304: putback(l = gpbc());
305: if (l != LPAREN) { /* add bracks */
306: putback(RPAREN);
307: putback(LPAREN);
308: }
309: }
310: }
311: else if (t == EOF) {
312: if (sp > -1)
1.11 espie 313: errx(1, "unexpected end of input");
1.1 deraadt 314: if (ilevel <= 0)
315: break; /* all done thanks.. */
316: --ilevel;
317: (void) fclose(infile[ilevel+1]);
318: bufbase = bbase[ilevel];
319: continue;
320: }
321: /*
1.7 deraadt 322: * non-alpha token possibly seen..
1.1 deraadt 323: * [the order of else if .. stmts is important.]
324: */
1.2 deraadt 325: else if (LOOK_AHEAD(t,lquote)) { /* strip quotes */
1.1 deraadt 326: nlpar = 1;
327: do {
1.7 deraadt 328:
1.2 deraadt 329: l = gpbc();
1.7 deraadt 330: if (LOOK_AHEAD(l,rquote)) {
1.1 deraadt 331: nlpar--;
1.7 deraadt 332: s = rquote;
333: } else if (LOOK_AHEAD(l,lquote)) {
1.1 deraadt 334: nlpar++;
1.7 deraadt 335: s = lquote;
1.17 espie 336: } else if (l == EOF) {
337: if (nlpar == 1)
338: errx(1, "missing right quote.");
339: else
340: errx(1, "missing %d right quotes.", nlpar);
341: } else {
1.7 deraadt 342: chars[0] = l;
1.13 espie 343: chars[1] = EOS;
1.7 deraadt 344: s = chars;
345: }
1.1 deraadt 346: if (nlpar > 0) {
347: if (sp < 0)
1.7 deraadt 348: while (*s)
349: putc(*s++, active);
1.1 deraadt 350: else
1.7 deraadt 351: while (*s)
352: chrsave(*s++);
1.1 deraadt 353: }
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);
385: PARLEV++;
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:
442: /*
443: * build an input token..
444: * consider only those starting with _ or A-Za-z. This is a
445: * combo with lookup to speed things up.
446: */
1.18 espie 447: static ndptr
1.1 deraadt 448: inspect(tp)
1.17 espie 449: char *tp;
1.1 deraadt 450: {
1.17 espie 451: char c;
452: char *name = tp;
453: char *etp = tp+MAXTOK;
454: ndptr p;
1.20 millert 455: unsigned int h = 0;
1.1 deraadt 456:
457: while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
458: h = (h << 5) + h + (*tp++ = c);
459: putback(c);
460: if (tp == etp)
1.11 espie 461: errx(1, "token too long");
1.1 deraadt 462:
463: *tp = EOS;
464:
1.19 espie 465: for (p = hashtab[h % HASHSIZE]; p != nil; p = p->nxtptr)
466: if (h == p->hv && STREQ(name, p->name))
1.1 deraadt 467: break;
468: return p;
469: }
470:
471: /*
472: * initkwds - initialise m4 keywords as fast as possible.
473: * This very similar to install, but without certain overheads,
474: * such as calling lookup. Malloc is not used for storing the
1.17 espie 475: * keyword strings, since we simply use the static pointers
1.1 deraadt 476: * within keywrds block.
477: */
1.18 espie 478: static void
1.17 espie 479: initkwds()
480: {
481: size_t i;
1.20 millert 482: unsigned int h;
1.17 espie 483: ndptr p;
1.1 deraadt 484:
485: for (i = 0; i < MAXKEYS; i++) {
486: h = hash(keywrds[i].knam);
487: p = (ndptr) xalloc(sizeof(struct ndblock));
1.19 espie 488: p->nxtptr = hashtab[h % HASHSIZE];
489: hashtab[h % HASHSIZE] = p;
1.1 deraadt 490: p->name = keywrds[i].knam;
491: p->defn = null;
1.19 espie 492: p->hv = h;
1.1 deraadt 493: p->type = keywrds[i].ktyp | STATIC;
494: }
495: }
1.17 espie 496: