Annotation of src/usr.bin/m4/main.c, Revision 1.26
1.26 ! espie 1: /* $OpenBSD: main.c,v 1.25 2000/01/11 14:10:01 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
1.26 ! espie 50: static char rcsid[] = "$OpenBSD: main.c,v 1.25 2000/01/11 14:10:01 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 */
1.26 ! espie 79: struct input_file infile[MAXINP];/* input file stack (0=stdin) */
1.1 deraadt 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 */
1.26 ! espie 195: set_input(infile+0, stdin, "stdin");
! 196: /* default input (naturally) */
1.1 deraadt 197: macro();
198: } else
199: for (; argc--; ++argv) {
200: p = *argv;
1.13 espie 201: if (p[0] == '-' && p[1] == EOS)
1.26 ! espie 202: set_input(infile, stdin, "stdin");
! 203: else if (fopen_trypath(infile, p) == NULL)
1.11 espie 204: err(1, "%s", p);
1.1 deraadt 205: sp = -1;
206: fp = 0;
207: macro();
1.26 ! espie 208: release_input(infile);
1.1 deraadt 209: }
210:
211: if (*m4wraps) { /* anything for rundown ?? */
212: ilevel = 0; /* in case m4wrap includes.. */
213: bufbase = bp = buf; /* use the entire buffer */
214: putback(EOF); /* eof is a must !! */
215: pbstr(m4wraps); /* user-defined wrapup act */
216: macro(); /* last will and testament */
217: }
218:
219: if (active != stdout)
220: active = stdout; /* reset output just in case */
221: for (n = 1; n < MAXOUT; n++) /* default wrap-up: undivert */
222: if (outfile[n] != NULL)
223: getdiv(n);
224: /* remove bitbucket if used */
225: if (outfile[0] != NULL) {
226: (void) fclose(outfile[0]);
227: }
228:
229: return 0;
230: }
231:
232: /*
1.21 espie 233: * Look ahead for `token'.
1.2 deraadt 234: * (on input `t == token[0]')
235: * Used for comment and quoting delimiters.
236: * Returns 1 if `token' present; copied to output.
237: * 0 if `token' not found; all characters pushed back
238: */
1.18 espie 239: static int
1.2 deraadt 240: do_look_ahead(t, token)
241: int t;
1.18 espie 242: const char *token;
1.2 deraadt 243: {
244: int i;
245:
1.21 espie 246: assert(t == token[0]);
1.2 deraadt 247:
248: for (i = 1; *++token; i++) {
249: t = gpbc();
250: if (t == EOF || t != *token) {
251: if (t != EOF)
252: putback(t);
253: while (--i)
254: putback(*--token);
255: return 0;
256: }
257: }
258: return 1;
259: }
260:
261: #define LOOK_AHEAD(t, token) ((t)==(token)[0] && do_look_ahead(t,token))
262:
263: /*
1.1 deraadt 264: * macro - the work horse..
265: */
1.18 espie 266: static void
1.17 espie 267: macro()
268: {
1.7 deraadt 269: char token[MAXTOK], chars[2];
1.17 espie 270: char *s;
271: int t, l;
272: ndptr p;
273: int nlpar;
1.1 deraadt 274:
275: cycle {
1.2 deraadt 276: t = gpbc();
277: if (t == '_' || isalpha(t)) {
1.10 deraadt 278: s = token;
1.25 espie 279: p = inspect(t, s);
1.24 espie 280: if (p != nil)
281: putback(l = gpbc());
282: if (p == nil || (l != LPAREN &&
283: (p->type & NEEDARGS) != 0))
284: outputstr(s);
1.1 deraadt 285: else {
286: /*
287: * real thing.. First build a call frame:
288: */
289: pushf(fp); /* previous call frm */
290: pushf(p->type); /* type of the call */
291: pushf(0); /* parenthesis level */
292: fp = sp; /* new frame pointer */
293: /*
294: * now push the string arguments:
295: */
296: pushs(p->defn); /* defn string */
297: pushs(p->name); /* macro name */
298: pushs(ep); /* start next..*/
299:
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.. */
1.26 ! espie 311: release_input(infile+ilevel--);
1.1 deraadt 312: bufbase = bbase[ilevel];
313: continue;
314: }
315: /*
1.7 deraadt 316: * non-alpha token possibly seen..
1.1 deraadt 317: * [the order of else if .. stmts is important.]
318: */
1.2 deraadt 319: else if (LOOK_AHEAD(t,lquote)) { /* strip quotes */
1.1 deraadt 320: nlpar = 1;
321: do {
1.7 deraadt 322:
1.2 deraadt 323: l = gpbc();
1.7 deraadt 324: if (LOOK_AHEAD(l,rquote)) {
1.1 deraadt 325: nlpar--;
1.7 deraadt 326: s = rquote;
327: } else if (LOOK_AHEAD(l,lquote)) {
1.1 deraadt 328: nlpar++;
1.7 deraadt 329: s = lquote;
1.17 espie 330: } else if (l == EOF) {
331: if (nlpar == 1)
332: errx(1, "missing right quote.");
333: else
334: errx(1, "missing %d right quotes.", nlpar);
335: } else {
1.7 deraadt 336: chars[0] = l;
1.13 espie 337: chars[1] = EOS;
1.7 deraadt 338: s = chars;
339: }
1.24 espie 340: if (nlpar > 0)
341: outputstr(s);
1.1 deraadt 342: }
343: while (nlpar != 0);
344: }
345:
1.2 deraadt 346: else if (sp < 0 && LOOK_AHEAD(t, scommt)) {
1.21 espie 347: fputs(scommt, active);
1.2 deraadt 348:
349: for(;;) {
350: t = gpbc();
351: if (LOOK_AHEAD(t, ecommt)) {
1.21 espie 352: fputs(ecommt, active);
1.2 deraadt 353: break;
354: }
355: if (t == EOF)
356: break;
1.1 deraadt 357: putc(t, active);
358: }
1.2 deraadt 359: }
360:
361: else if (sp < 0) { /* not in a macro at all */
1.1 deraadt 362: putc(t, active); /* output directly.. */
363: }
364:
365: else switch(t) {
366:
367: case LPAREN:
368: if (PARLEV > 0)
369: chrsave(t);
370: while (isspace(l = gpbc()))
371: ; /* skip blank, tab, nl.. */
372: putback(l);
373: PARLEV++;
374: break;
375:
376: case RPAREN:
377: if (--PARLEV > 0)
378: chrsave(t);
379: else { /* end of argument list */
380: chrsave(EOS);
381:
382: if (sp == STACKMAX)
1.11 espie 383: errx(1, "internal stack overflow");
1.1 deraadt 384:
385: if (CALTYP == MACRTYPE)
1.18 espie 386: expand((const char **) mstack+fp+1, sp-fp);
1.1 deraadt 387: else
1.18 espie 388: eval((const char **) mstack+fp+1, sp-fp, CALTYP);
1.1 deraadt 389:
390: ep = PREVEP; /* flush strspace */
391: sp = PREVSP; /* previous sp.. */
392: fp = PREVFP; /* rewind stack...*/
393: }
394: break;
395:
396: case COMMA:
397: if (PARLEV == 1) {
398: chrsave(EOS); /* new argument */
399: while (isspace(l = gpbc()))
400: ;
401: putback(l);
402: pushs(ep);
403: } else
404: chrsave(t);
405: break;
406:
407: default:
1.22 espie 408: if (LOOK_AHEAD(t, scommt)) {
409: char *p;
410: for (p = scommt; *p; p++)
411: chrsave(*p);
412: for(;;) {
413: t = gpbc();
414: if (LOOK_AHEAD(t, ecommt)) {
415: for (p = ecommt; *p; p++)
416: chrsave(*p);
417: break;
418: }
419: if (t == EOF)
420: break;
421: chrsave(t);
422: }
423: } else
424: chrsave(t); /* stack the char */
1.1 deraadt 425: break;
426: }
427: }
428: }
429:
1.24 espie 430: /*
431: * output string directly, without pushing it for reparses.
432: */
433: void
434: outputstr(s)
435: const char *s;
436: {
437: if (sp < 0)
438: while (*s)
439: putc(*s++, active);
440: else
441: while (*s)
442: chrsave(*s++);
443: }
444:
1.1 deraadt 445: /*
446: * build an input token..
447: * consider only those starting with _ or A-Za-z. This is a
448: * combo with lookup to speed things up.
449: */
1.18 espie 450: static ndptr
1.25 espie 451: inspect(c, tp)
452: char c;
1.17 espie 453: char *tp;
1.1 deraadt 454: {
1.17 espie 455: char *name = tp;
456: char *etp = tp+MAXTOK;
457: ndptr p;
1.25 espie 458: unsigned int h;
459:
460: h = *tp++ = c;
1.1 deraadt 461:
462: while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
463: h = (h << 5) + h + (*tp++ = c);
464: putback(c);
465: if (tp == etp)
1.11 espie 466: errx(1, "token too long");
1.1 deraadt 467:
468: *tp = EOS;
469:
1.19 espie 470: for (p = hashtab[h % HASHSIZE]; p != nil; p = p->nxtptr)
471: if (h == p->hv && STREQ(name, p->name))
1.1 deraadt 472: break;
473: return p;
474: }
475:
476: /*
477: * initkwds - initialise m4 keywords as fast as possible.
478: * This very similar to install, but without certain overheads,
479: * such as calling lookup. Malloc is not used for storing the
1.17 espie 480: * keyword strings, since we simply use the static pointers
1.1 deraadt 481: * within keywrds block.
482: */
1.18 espie 483: static void
1.17 espie 484: initkwds()
485: {
486: size_t i;
1.20 millert 487: unsigned int h;
1.17 espie 488: ndptr p;
1.1 deraadt 489:
490: for (i = 0; i < MAXKEYS; i++) {
491: h = hash(keywrds[i].knam);
492: p = (ndptr) xalloc(sizeof(struct ndblock));
1.19 espie 493: p->nxtptr = hashtab[h % HASHSIZE];
494: hashtab[h % HASHSIZE] = p;
1.1 deraadt 495: p->name = keywrds[i].knam;
496: p->defn = null;
1.19 espie 497: p->hv = h;
1.24 espie 498: p->type = (keywrds[i].ktyp & TYPEMASK) | STATIC;
499: if ((keywrds[i].ktyp & NOARGS) == 0)
500: p->type |= NEEDARGS;
1.1 deraadt 501: }
502: }
1.17 espie 503: