Annotation of src/usr.bin/m4/main.c, Revision 1.7
1.7 ! deraadt 1: /* $OpenBSD: main.c,v 1.6 1997/12/10 20:24:15 deraadt Exp $ */
! 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
1.7 ! deraadt 50: static char rcsid[] = "$OpenBSD: main.c,v 1.6 1997/12/10 20:24:15 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>
67: #include "mdef.h"
68: #include "stdd.h"
69: #include "extern.h"
70: #include "pathnames.h"
71:
72: ndptr hashtab[HASHSIZE]; /* hash table for macros etc. */
1.6 deraadt 73: pbent buf[BUFSIZE]; /* push-back buffer */
74: pbent *bufbase = buf; /* the base for current ilevel */
75: pbent *bbase[MAXINP]; /* the base for each ilevel */
76: pbent *bp = buf; /* first available character */
77: pbent *endpbb = buf+BUFSIZE; /* end of push-back buffer */
1.1 deraadt 78: stae mstack[STACKMAX+1]; /* stack of m4 machine */
79: char strspace[STRSPMAX+1]; /* string space for evaluation */
80: char *ep = strspace; /* first free char in strspace */
81: char *endest= strspace+STRSPMAX;/* end of string space */
82: int sp; /* current m4 stack pointer */
83: int fp; /* m4 call frame pointer */
84: FILE *infile[MAXINP]; /* input file stack (0=stdin) */
85: FILE *outfile[MAXOUT]; /* diversion array(0=bitbucket)*/
86: FILE *active; /* active output file pointer */
87: char *m4temp; /* filename for diversions */
88: int ilevel = 0; /* input file stack pointer */
89: int oindex = 0; /* diversion index.. */
90: char *null = ""; /* as it says.. just a null.. */
91: char *m4wraps = ""; /* m4wrap string default.. */
92: char *progname; /* name of this program */
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 */
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,
116: #ifdef EXTENDED
117: "paste", PASTTYPE,
118: "spaste", SPASTYPE,
119: #endif
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,
134:
1.4 deraadt 135: #if defined(unix) || defined(__NetBSD__) || defined(__OpenBSD__)
1.1 deraadt 136: "unix", MACRTYPE,
137: #else
138: #ifdef vms
139: "vms", MACRTYPE,
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: progname = basename(argv[0]);
164:
165: if (signal(SIGINT, SIG_IGN) != SIG_IGN)
166: signal(SIGINT, onintr);
167:
168: initkwds();
169:
1.5 millert 170: while ((c = getopt(argc, argv, "tD:U:o:")) != -1)
1.1 deraadt 171: switch(c) {
172:
173: case 'D': /* define something..*/
174: for (p = optarg; *p; p++)
175: if (*p == '=')
176: break;
177: if (*p)
178: *p++ = EOS;
179: dodefine(optarg, p);
180: break;
181: case 'U': /* undefine... */
182: remhash(optarg, TOP);
183: break;
184: case 'o': /* specific output */
185: case '?':
186: usage();
187: }
188:
189: argc -= optind;
190: argv += optind;
191:
192: active = stdout; /* default active output */
193: /* filename for diversions */
194: m4temp = mktemp(xstrdup(_PATH_DIVNAME));
195:
196: bbase[0] = bufbase;
197: if (!argc) {
198: sp = -1; /* stack pointer initialized */
199: fp = 0; /* frame pointer initialized */
200: infile[0] = stdin; /* default input (naturally) */
201: macro();
202: } else
203: for (; argc--; ++argv) {
204: p = *argv;
205: if (p[0] == '-' && p[1] == '\0')
206: ifp = stdin;
207: else if ((ifp = fopen(p, "r")) == NULL)
208: oops("%s: %s", p, strerror(errno));
209: sp = -1;
210: fp = 0;
211: infile[0] = ifp;
212: macro();
213: if (ifp != stdin)
214: (void)fclose(ifp);
215: }
216:
217: if (*m4wraps) { /* anything for rundown ?? */
218: ilevel = 0; /* in case m4wrap includes.. */
219: bufbase = bp = buf; /* use the entire buffer */
220: putback(EOF); /* eof is a must !! */
221: pbstr(m4wraps); /* user-defined wrapup act */
222: macro(); /* last will and testament */
223: }
224:
225: if (active != stdout)
226: active = stdout; /* reset output just in case */
227: for (n = 1; n < MAXOUT; n++) /* default wrap-up: undivert */
228: if (outfile[n] != NULL)
229: getdiv(n);
230: /* remove bitbucket if used */
231: if (outfile[0] != NULL) {
232: (void) fclose(outfile[0]);
233: m4temp[UNIQUE] = '0';
234: #ifdef vms
235: (void) remove(m4temp);
236: #else
237: (void) unlink(m4temp);
238: #endif
239: }
240:
241: return 0;
242: }
243:
244: ndptr inspect();
245:
246: /*
1.2 deraadt 247: * Look ahead (at most MAXCCHARS characters) for `token'.
248: * (on input `t == token[0]')
249: * Used for comment and quoting delimiters.
250: * Returns 1 if `token' present; copied to output.
251: * 0 if `token' not found; all characters pushed back
252: */
253: int
254: do_look_ahead(t, token)
255: int t;
256: char *token;
257: {
258: int i;
259:
260: if (t != token[0])
261: oops("internal error", "");
262:
263: for (i = 1; *++token; i++) {
264: t = gpbc();
265: if (t == EOF || t != *token) {
266: if (t != EOF)
267: putback(t);
268: while (--i)
269: putback(*--token);
270: return 0;
271: }
272: }
273: return 1;
274: }
275:
276: #define LOOK_AHEAD(t, token) ((t)==(token)[0] && do_look_ahead(t,token))
277:
278: /*
1.1 deraadt 279: * macro - the work horse..
280: */
281: void
282: macro() {
1.7 ! deraadt 283: char token[MAXTOK], chars[2];
1.1 deraadt 284: register char *s;
285: register int t, l;
286: register ndptr p;
287: register int nlpar;
288:
289: cycle {
1.2 deraadt 290: t = gpbc();
291: if (t == '_' || isalpha(t)) {
1.1 deraadt 292: putback(t);
293: if ((p = inspect(s = token)) == nil) {
294: if (sp < 0)
295: while (*s)
296: putc(*s++, active);
297: else
298: while (*s)
299: chrsave(*s++);
300: }
301: else {
302: /*
303: * real thing.. First build a call frame:
304: */
305: pushf(fp); /* previous call frm */
306: pushf(p->type); /* type of the call */
307: pushf(0); /* parenthesis level */
308: fp = sp; /* new frame pointer */
309: /*
310: * now push the string arguments:
311: */
312: pushs(p->defn); /* defn string */
313: pushs(p->name); /* macro name */
314: pushs(ep); /* start next..*/
315:
316: putback(l = gpbc());
317: if (l != LPAREN) { /* add bracks */
318: putback(RPAREN);
319: putback(LPAREN);
320: }
321: }
322: }
323: else if (t == EOF) {
324: if (sp > -1)
325: oops("unexpected end of input", "");
326: if (ilevel <= 0)
327: break; /* all done thanks.. */
328: --ilevel;
329: (void) fclose(infile[ilevel+1]);
330: bufbase = bbase[ilevel];
331: continue;
332: }
333: /*
1.7 ! deraadt 334: * non-alpha token possibly seen..
1.1 deraadt 335: * [the order of else if .. stmts is important.]
336: */
1.2 deraadt 337: else if (LOOK_AHEAD(t,lquote)) { /* strip quotes */
1.1 deraadt 338: nlpar = 1;
339: do {
1.7 ! deraadt 340:
1.2 deraadt 341: l = gpbc();
1.7 ! deraadt 342: if (LOOK_AHEAD(l,rquote)) {
1.1 deraadt 343: nlpar--;
1.7 ! deraadt 344: s = rquote;
! 345: } else if (LOOK_AHEAD(l,lquote)) {
1.1 deraadt 346: nlpar++;
1.7 ! deraadt 347: s = lquote;
! 348: } else if (l == EOF)
1.1 deraadt 349: oops("missing right quote", "");
1.7 ! deraadt 350: else {
! 351: chars[0] = l;
! 352: chars[1] = '\0';
! 353: s = chars;
! 354: }
1.1 deraadt 355: if (nlpar > 0) {
356: if (sp < 0)
1.7 ! deraadt 357: while (*s)
! 358: putc(*s++, active);
1.1 deraadt 359: else
1.7 ! deraadt 360: while (*s)
! 361: chrsave(*s++);
1.1 deraadt 362: }
363: }
364: while (nlpar != 0);
365: }
366:
1.2 deraadt 367: else if (sp < 0 && LOOK_AHEAD(t, scommt)) {
368: int i;
369: for (i = 0; i < MAXCCHARS && scommt[i]; i++)
370: putc(scommt[i], active);
371:
372: for(;;) {
373: t = gpbc();
374: if (LOOK_AHEAD(t, ecommt)) {
375: for (i = 0; i < MAXCCHARS && ecommt[i];
376: i++)
377: putc(ecommt[i], active);
378: break;
379: }
380: if (t == EOF)
381: break;
1.1 deraadt 382: putc(t, active);
383: }
1.2 deraadt 384: }
385:
386: else if (sp < 0) { /* not in a macro at all */
1.1 deraadt 387: putc(t, active); /* output directly.. */
388: }
389:
390: else switch(t) {
391:
392: case LPAREN:
393: if (PARLEV > 0)
394: chrsave(t);
395: while (isspace(l = gpbc()))
396: ; /* skip blank, tab, nl.. */
397: putback(l);
398: PARLEV++;
399: break;
400:
401: case RPAREN:
402: if (--PARLEV > 0)
403: chrsave(t);
404: else { /* end of argument list */
405: chrsave(EOS);
406:
407: if (sp == STACKMAX)
408: oops("internal stack overflow", "");
409:
410: if (CALTYP == MACRTYPE)
411: expand((char **) mstack+fp+1, sp-fp);
412: else
413: eval((char **) mstack+fp+1, sp-fp, CALTYP);
414:
415: ep = PREVEP; /* flush strspace */
416: sp = PREVSP; /* previous sp.. */
417: fp = PREVFP; /* rewind stack...*/
418: }
419: break;
420:
421: case COMMA:
422: if (PARLEV == 1) {
423: chrsave(EOS); /* new argument */
424: while (isspace(l = gpbc()))
425: ;
426: putback(l);
427: pushs(ep);
428: } else
429: chrsave(t);
430: break;
431:
432: default:
433: chrsave(t); /* stack the char */
434: break;
435: }
436: }
437: }
438:
439: /*
440: * build an input token..
441: * consider only those starting with _ or A-Za-z. This is a
442: * combo with lookup to speed things up.
443: */
444: ndptr
445: inspect(tp)
446: register char *tp;
447: {
448: register char c;
449: register char *name = tp;
450: register char *etp = tp+MAXTOK;
451: register ndptr p;
452: register unsigned long h = 0;
453:
454: while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
455: h = (h << 5) + h + (*tp++ = c);
456: putback(c);
457: if (tp == etp)
458: oops("token too long", "");
459:
460: *tp = EOS;
461:
462: for (p = hashtab[h%HASHSIZE]; p != nil; p = p->nxtptr)
463: if (STREQ(name, p->name))
464: break;
465: return p;
466: }
467:
468: /*
469: * initkwds - initialise m4 keywords as fast as possible.
470: * This very similar to install, but without certain overheads,
471: * such as calling lookup. Malloc is not used for storing the
472: * keyword strings, since we simply use the static pointers
473: * within keywrds block.
474: */
475: void
476: initkwds() {
477: register int i;
478: register int h;
479: register ndptr p;
480:
481: for (i = 0; i < MAXKEYS; i++) {
482: h = hash(keywrds[i].knam);
483: p = (ndptr) xalloc(sizeof(struct ndblock));
484: p->nxtptr = hashtab[h];
485: hashtab[h] = p;
486: p->name = keywrds[i].knam;
487: p->defn = null;
488: p->type = keywrds[i].ktyp | STATIC;
489: }
490: }