Annotation of src/usr.bin/m4/main.c, Revision 1.25
1.25 ! espie 1: /* $OpenBSD: main.c,v 1.24 2000/01/11 14:06:11 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
1.25 ! espie 50: static char rcsid[] = "$OpenBSD: main.c,v 1.24 2000/01/11 14:06:11 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 },
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:
1.18 espie 142: static void macro __P((void));
143: static void initkwds __P((void));
1.25 ! espie 144: static ndptr inspect __P((char, char *));
1.18 espie 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)) {
1.10 deraadt 279: s = token;
1.25 ! espie 280: p = inspect(t, s);
1.24 espie 281: if (p != nil)
282: putback(l = gpbc());
283: if (p == nil || (l != LPAREN &&
284: (p->type & NEEDARGS) != 0))
285: outputstr(s);
1.1 deraadt 286: else {
287: /*
288: * real thing.. First build a call frame:
289: */
290: pushf(fp); /* previous call frm */
291: pushf(p->type); /* type of the call */
292: pushf(0); /* parenthesis level */
293: fp = sp; /* new frame pointer */
294: /*
295: * now push the string arguments:
296: */
297: pushs(p->defn); /* defn string */
298: pushs(p->name); /* macro name */
299: pushs(ep); /* start next..*/
300:
301: if (l != LPAREN) { /* add bracks */
302: putback(RPAREN);
303: putback(LPAREN);
304: }
305: }
306: }
307: else if (t == EOF) {
308: if (sp > -1)
1.11 espie 309: errx(1, "unexpected end of input");
1.1 deraadt 310: if (ilevel <= 0)
311: break; /* all done thanks.. */
312: --ilevel;
313: (void) fclose(infile[ilevel+1]);
314: bufbase = bbase[ilevel];
315: continue;
316: }
317: /*
1.7 deraadt 318: * non-alpha token possibly seen..
1.1 deraadt 319: * [the order of else if .. stmts is important.]
320: */
1.2 deraadt 321: else if (LOOK_AHEAD(t,lquote)) { /* strip quotes */
1.1 deraadt 322: nlpar = 1;
323: do {
1.7 deraadt 324:
1.2 deraadt 325: l = gpbc();
1.7 deraadt 326: if (LOOK_AHEAD(l,rquote)) {
1.1 deraadt 327: nlpar--;
1.7 deraadt 328: s = rquote;
329: } else if (LOOK_AHEAD(l,lquote)) {
1.1 deraadt 330: nlpar++;
1.7 deraadt 331: s = lquote;
1.17 espie 332: } else if (l == EOF) {
333: if (nlpar == 1)
334: errx(1, "missing right quote.");
335: else
336: errx(1, "missing %d right quotes.", nlpar);
337: } else {
1.7 deraadt 338: chars[0] = l;
1.13 espie 339: chars[1] = EOS;
1.7 deraadt 340: s = chars;
341: }
1.24 espie 342: if (nlpar > 0)
343: outputstr(s);
1.1 deraadt 344: }
345: while (nlpar != 0);
346: }
347:
1.2 deraadt 348: else if (sp < 0 && LOOK_AHEAD(t, scommt)) {
1.21 espie 349: fputs(scommt, active);
1.2 deraadt 350:
351: for(;;) {
352: t = gpbc();
353: if (LOOK_AHEAD(t, ecommt)) {
1.21 espie 354: fputs(ecommt, active);
1.2 deraadt 355: break;
356: }
357: if (t == EOF)
358: break;
1.1 deraadt 359: putc(t, active);
360: }
1.2 deraadt 361: }
362:
363: else if (sp < 0) { /* not in a macro at all */
1.1 deraadt 364: putc(t, active); /* output directly.. */
365: }
366:
367: else switch(t) {
368:
369: case LPAREN:
370: if (PARLEV > 0)
371: chrsave(t);
372: while (isspace(l = gpbc()))
373: ; /* skip blank, tab, nl.. */
374: putback(l);
375: PARLEV++;
376: break;
377:
378: case RPAREN:
379: if (--PARLEV > 0)
380: chrsave(t);
381: else { /* end of argument list */
382: chrsave(EOS);
383:
384: if (sp == STACKMAX)
1.11 espie 385: errx(1, "internal stack overflow");
1.1 deraadt 386:
387: if (CALTYP == MACRTYPE)
1.18 espie 388: expand((const char **) mstack+fp+1, sp-fp);
1.1 deraadt 389: else
1.18 espie 390: eval((const char **) mstack+fp+1, sp-fp, CALTYP);
1.1 deraadt 391:
392: ep = PREVEP; /* flush strspace */
393: sp = PREVSP; /* previous sp.. */
394: fp = PREVFP; /* rewind stack...*/
395: }
396: break;
397:
398: case COMMA:
399: if (PARLEV == 1) {
400: chrsave(EOS); /* new argument */
401: while (isspace(l = gpbc()))
402: ;
403: putback(l);
404: pushs(ep);
405: } else
406: chrsave(t);
407: break;
408:
409: default:
1.22 espie 410: if (LOOK_AHEAD(t, scommt)) {
411: char *p;
412: for (p = scommt; *p; p++)
413: chrsave(*p);
414: for(;;) {
415: t = gpbc();
416: if (LOOK_AHEAD(t, ecommt)) {
417: for (p = ecommt; *p; p++)
418: chrsave(*p);
419: break;
420: }
421: if (t == EOF)
422: break;
423: chrsave(t);
424: }
425: } else
426: chrsave(t); /* stack the char */
1.1 deraadt 427: break;
428: }
429: }
430: }
431:
1.24 espie 432: /*
433: * output string directly, without pushing it for reparses.
434: */
435: void
436: outputstr(s)
437: const char *s;
438: {
439: if (sp < 0)
440: while (*s)
441: putc(*s++, active);
442: else
443: while (*s)
444: chrsave(*s++);
445: }
446:
1.1 deraadt 447: /*
448: * build an input token..
449: * consider only those starting with _ or A-Za-z. This is a
450: * combo with lookup to speed things up.
451: */
1.18 espie 452: static ndptr
1.25 ! espie 453: inspect(c, tp)
! 454: char c;
1.17 espie 455: char *tp;
1.1 deraadt 456: {
1.17 espie 457: char *name = tp;
458: char *etp = tp+MAXTOK;
459: ndptr p;
1.25 ! espie 460: unsigned int h;
! 461:
! 462: h = *tp++ = c;
1.1 deraadt 463:
464: while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
465: h = (h << 5) + h + (*tp++ = c);
466: putback(c);
467: if (tp == etp)
1.11 espie 468: errx(1, "token too long");
1.1 deraadt 469:
470: *tp = EOS;
471:
1.19 espie 472: for (p = hashtab[h % HASHSIZE]; p != nil; p = p->nxtptr)
473: if (h == p->hv && STREQ(name, p->name))
1.1 deraadt 474: break;
475: return p;
476: }
477:
478: /*
479: * initkwds - initialise m4 keywords as fast as possible.
480: * This very similar to install, but without certain overheads,
481: * such as calling lookup. Malloc is not used for storing the
1.17 espie 482: * keyword strings, since we simply use the static pointers
1.1 deraadt 483: * within keywrds block.
484: */
1.18 espie 485: static void
1.17 espie 486: initkwds()
487: {
488: size_t i;
1.20 millert 489: unsigned int h;
1.17 espie 490: ndptr p;
1.1 deraadt 491:
492: for (i = 0; i < MAXKEYS; i++) {
493: h = hash(keywrds[i].knam);
494: p = (ndptr) xalloc(sizeof(struct ndblock));
1.19 espie 495: p->nxtptr = hashtab[h % HASHSIZE];
496: hashtab[h % HASHSIZE] = p;
1.1 deraadt 497: p->name = keywrds[i].knam;
498: p->defn = null;
1.19 espie 499: p->hv = h;
1.24 espie 500: p->type = (keywrds[i].ktyp & TYPEMASK) | STATIC;
501: if ((keywrds[i].ktyp & NOARGS) == 0)
502: p->type |= NEEDARGS;
1.1 deraadt 503: }
504: }
1.17 espie 505: