Annotation of src/usr.bin/m4/main.c, Revision 1.27
1.27 ! espie 1: /* $OpenBSD: main.c,v 1.26 2000/01/12 17:49:53 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
1.27 ! espie 50: static char rcsid[] = "$OpenBSD: main.c,v 1.26 2000/01/12 17:49:53 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:
1.27 ! espie 142: #define MAXRECORD 50
! 143: static struct position {
! 144: char *name;
! 145: unsigned long line;
! 146: } quotes[MAXRECORD], paren[MAXRECORD];
! 147:
! 148: static void record __P((struct position *, int));
! 149: static void dump_stack __P((struct position *, int));
! 150:
1.18 espie 151: static void macro __P((void));
152: static void initkwds __P((void));
1.25 espie 153: static ndptr inspect __P((char, char *));
1.18 espie 154: static int do_look_ahead __P((int, const char *));
155:
156: int main __P((int, char *[]));
1.1 deraadt 157:
158: int
159: main(argc,argv)
160: int argc;
161: char *argv[];
162: {
1.17 espie 163: int c;
164: int n;
1.1 deraadt 165: char *p;
1.17 espie 166: FILE *ifp;
1.1 deraadt 167:
168: if (signal(SIGINT, SIG_IGN) != SIG_IGN)
169: signal(SIGINT, onintr);
170:
171: initkwds();
1.14 espie 172: initspaces();
1.1 deraadt 173:
1.16 espie 174: while ((c = getopt(argc, argv, "tD:U:o:I:")) != -1)
1.1 deraadt 175: switch(c) {
176:
177: case 'D': /* define something..*/
178: for (p = optarg; *p; p++)
179: if (*p == '=')
180: break;
181: if (*p)
182: *p++ = EOS;
183: dodefine(optarg, p);
1.16 espie 184: break;
185: case 'I':
186: addtoincludepath(optarg);
1.1 deraadt 187: break;
188: case 'U': /* undefine... */
189: remhash(optarg, TOP);
190: break;
191: case 'o': /* specific output */
192: case '?':
193: usage();
194: }
195:
196: argc -= optind;
197: argv += optind;
198:
199: active = stdout; /* default active output */
200: bbase[0] = bufbase;
201: if (!argc) {
202: sp = -1; /* stack pointer initialized */
203: fp = 0; /* frame pointer initialized */
1.26 espie 204: set_input(infile+0, stdin, "stdin");
205: /* default input (naturally) */
1.1 deraadt 206: macro();
207: } else
208: for (; argc--; ++argv) {
209: p = *argv;
1.13 espie 210: if (p[0] == '-' && p[1] == EOS)
1.26 espie 211: set_input(infile, stdin, "stdin");
212: else if (fopen_trypath(infile, p) == NULL)
1.11 espie 213: err(1, "%s", p);
1.1 deraadt 214: sp = -1;
215: fp = 0;
216: macro();
1.26 espie 217: release_input(infile);
1.1 deraadt 218: }
219:
220: if (*m4wraps) { /* anything for rundown ?? */
221: ilevel = 0; /* in case m4wrap includes.. */
222: bufbase = bp = buf; /* use the entire buffer */
223: putback(EOF); /* eof is a must !! */
224: pbstr(m4wraps); /* user-defined wrapup act */
225: macro(); /* last will and testament */
226: }
227:
228: if (active != stdout)
229: active = stdout; /* reset output just in case */
230: for (n = 1; n < MAXOUT; n++) /* default wrap-up: undivert */
231: if (outfile[n] != NULL)
232: getdiv(n);
233: /* remove bitbucket if used */
234: if (outfile[0] != NULL) {
235: (void) fclose(outfile[0]);
236: }
237:
238: return 0;
239: }
240:
241: /*
1.21 espie 242: * Look ahead for `token'.
1.2 deraadt 243: * (on input `t == token[0]')
244: * Used for comment and quoting delimiters.
245: * Returns 1 if `token' present; copied to output.
246: * 0 if `token' not found; all characters pushed back
247: */
1.18 espie 248: static int
1.2 deraadt 249: do_look_ahead(t, token)
250: int t;
1.18 espie 251: const char *token;
1.2 deraadt 252: {
253: int i;
254:
1.21 espie 255: assert(t == token[0]);
1.2 deraadt 256:
257: for (i = 1; *++token; i++) {
258: t = gpbc();
259: if (t == EOF || t != *token) {
260: if (t != EOF)
261: putback(t);
262: while (--i)
263: putback(*--token);
264: return 0;
265: }
266: }
267: return 1;
268: }
269:
270: #define LOOK_AHEAD(t, token) ((t)==(token)[0] && do_look_ahead(t,token))
271:
272: /*
1.1 deraadt 273: * macro - the work horse..
274: */
1.18 espie 275: static void
1.17 espie 276: macro()
277: {
1.7 deraadt 278: char token[MAXTOK], chars[2];
1.17 espie 279: char *s;
280: int t, l;
281: ndptr p;
282: int nlpar;
1.1 deraadt 283:
284: cycle {
1.2 deraadt 285: t = gpbc();
286: if (t == '_' || isalpha(t)) {
1.10 deraadt 287: s = token;
1.25 espie 288: p = inspect(t, s);
1.24 espie 289: if (p != nil)
290: putback(l = gpbc());
291: if (p == nil || (l != LPAREN &&
292: (p->type & NEEDARGS) != 0))
293: outputstr(s);
1.1 deraadt 294: else {
295: /*
296: * real thing.. First build a call frame:
297: */
298: pushf(fp); /* previous call frm */
299: pushf(p->type); /* type of the call */
300: pushf(0); /* parenthesis level */
301: fp = sp; /* new frame pointer */
302: /*
303: * now push the string arguments:
304: */
305: pushs(p->defn); /* defn string */
306: pushs(p->name); /* macro name */
307: pushs(ep); /* start next..*/
308:
309: if (l != LPAREN) { /* add bracks */
310: putback(RPAREN);
311: putback(LPAREN);
312: }
313: }
314: }
315: else if (t == EOF) {
1.27 ! espie 316: if (sp > -1) {
! 317: warnx( "unexpected end of input, unclosed parenthesis:");
! 318: dump_stack(paren, PARLEV);
! 319: exit(1);
! 320: }
1.1 deraadt 321: if (ilevel <= 0)
322: break; /* all done thanks.. */
1.26 espie 323: release_input(infile+ilevel--);
1.1 deraadt 324: bufbase = bbase[ilevel];
325: continue;
326: }
327: /*
1.7 deraadt 328: * non-alpha token possibly seen..
1.1 deraadt 329: * [the order of else if .. stmts is important.]
330: */
1.2 deraadt 331: else if (LOOK_AHEAD(t,lquote)) { /* strip quotes */
1.27 ! espie 332: nlpar = 0;
! 333: record(quotes, nlpar++);
1.1 deraadt 334: do {
1.7 deraadt 335:
1.2 deraadt 336: l = gpbc();
1.7 deraadt 337: if (LOOK_AHEAD(l,rquote)) {
1.1 deraadt 338: nlpar--;
1.7 deraadt 339: s = rquote;
340: } else if (LOOK_AHEAD(l,lquote)) {
1.27 ! espie 341: record(quotes, nlpar++);
1.7 deraadt 342: s = lquote;
1.17 espie 343: } else if (l == EOF) {
344: if (nlpar == 1)
1.27 ! espie 345: warnx("unclosed quote:");
1.17 espie 346: else
1.27 ! espie 347: warnx("%d unclosed quotes:", nlpar);
! 348: dump_stack(quotes, nlpar);
! 349: exit(1);
1.17 espie 350: } else {
1.7 deraadt 351: chars[0] = l;
1.13 espie 352: chars[1] = EOS;
1.7 deraadt 353: s = chars;
354: }
1.24 espie 355: if (nlpar > 0)
356: outputstr(s);
1.1 deraadt 357: }
358: while (nlpar != 0);
359: }
360:
1.2 deraadt 361: else if (sp < 0 && LOOK_AHEAD(t, scommt)) {
1.21 espie 362: fputs(scommt, active);
1.2 deraadt 363:
364: for(;;) {
365: t = gpbc();
366: if (LOOK_AHEAD(t, ecommt)) {
1.21 espie 367: fputs(ecommt, active);
1.2 deraadt 368: break;
369: }
370: if (t == EOF)
371: break;
1.1 deraadt 372: putc(t, active);
373: }
1.2 deraadt 374: }
375:
376: else if (sp < 0) { /* not in a macro at all */
1.1 deraadt 377: putc(t, active); /* output directly.. */
378: }
379:
380: else switch(t) {
381:
382: case LPAREN:
383: if (PARLEV > 0)
384: chrsave(t);
385: while (isspace(l = gpbc()))
386: ; /* skip blank, tab, nl.. */
387: putback(l);
1.27 ! espie 388: record(paren, PARLEV++);
1.1 deraadt 389: break;
390:
391: case RPAREN:
392: if (--PARLEV > 0)
393: chrsave(t);
394: else { /* end of argument list */
395: chrsave(EOS);
396:
397: if (sp == STACKMAX)
1.11 espie 398: errx(1, "internal stack overflow");
1.1 deraadt 399:
400: if (CALTYP == MACRTYPE)
1.18 espie 401: expand((const char **) mstack+fp+1, sp-fp);
1.1 deraadt 402: else
1.18 espie 403: eval((const char **) mstack+fp+1, sp-fp, CALTYP);
1.1 deraadt 404:
405: ep = PREVEP; /* flush strspace */
406: sp = PREVSP; /* previous sp.. */
407: fp = PREVFP; /* rewind stack...*/
408: }
409: break;
410:
411: case COMMA:
412: if (PARLEV == 1) {
413: chrsave(EOS); /* new argument */
414: while (isspace(l = gpbc()))
415: ;
416: putback(l);
417: pushs(ep);
418: } else
419: chrsave(t);
420: break;
421:
422: default:
1.22 espie 423: if (LOOK_AHEAD(t, scommt)) {
424: char *p;
425: for (p = scommt; *p; p++)
426: chrsave(*p);
427: for(;;) {
428: t = gpbc();
429: if (LOOK_AHEAD(t, ecommt)) {
430: for (p = ecommt; *p; p++)
431: chrsave(*p);
432: break;
433: }
434: if (t == EOF)
435: break;
436: chrsave(t);
437: }
438: } else
439: chrsave(t); /* stack the char */
1.1 deraadt 440: break;
441: }
442: }
443: }
444:
1.24 espie 445: /*
446: * output string directly, without pushing it for reparses.
447: */
448: void
449: outputstr(s)
450: const char *s;
451: {
452: if (sp < 0)
453: while (*s)
454: putc(*s++, active);
455: else
456: while (*s)
457: chrsave(*s++);
458: }
459:
1.1 deraadt 460: /*
461: * build an input token..
462: * consider only those starting with _ or A-Za-z. This is a
463: * combo with lookup to speed things up.
464: */
1.18 espie 465: static ndptr
1.25 espie 466: inspect(c, tp)
467: char c;
1.17 espie 468: char *tp;
1.1 deraadt 469: {
1.17 espie 470: char *name = tp;
471: char *etp = tp+MAXTOK;
472: ndptr p;
1.25 espie 473: unsigned int h;
474:
475: h = *tp++ = c;
1.1 deraadt 476:
477: while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
478: h = (h << 5) + h + (*tp++ = c);
479: putback(c);
480: if (tp == etp)
1.11 espie 481: errx(1, "token too long");
1.1 deraadt 482:
483: *tp = EOS;
484:
1.19 espie 485: for (p = hashtab[h % HASHSIZE]; p != nil; p = p->nxtptr)
486: if (h == p->hv && STREQ(name, p->name))
1.1 deraadt 487: break;
488: return p;
489: }
490:
491: /*
492: * initkwds - initialise m4 keywords as fast as possible.
493: * This very similar to install, but without certain overheads,
494: * such as calling lookup. Malloc is not used for storing the
1.17 espie 495: * keyword strings, since we simply use the static pointers
1.1 deraadt 496: * within keywrds block.
497: */
1.18 espie 498: static void
1.17 espie 499: initkwds()
500: {
501: size_t i;
1.20 millert 502: unsigned int h;
1.17 espie 503: ndptr p;
1.1 deraadt 504:
505: for (i = 0; i < MAXKEYS; i++) {
506: h = hash(keywrds[i].knam);
507: p = (ndptr) xalloc(sizeof(struct ndblock));
1.19 espie 508: p->nxtptr = hashtab[h % HASHSIZE];
509: hashtab[h % HASHSIZE] = p;
1.1 deraadt 510: p->name = keywrds[i].knam;
511: p->defn = null;
1.19 espie 512: p->hv = h;
1.24 espie 513: p->type = (keywrds[i].ktyp & TYPEMASK) | STATIC;
514: if ((keywrds[i].ktyp & NOARGS) == 0)
515: p->type |= NEEDARGS;
1.1 deraadt 516: }
517: }
1.17 espie 518:
1.27 ! espie 519: static void
! 520: record(t, lev)
! 521: struct position *t;
! 522: int lev;
! 523: {
! 524: if (lev < MAXRECORD) {
! 525: t[lev].name = CURRENT_NAME;
! 526: t[lev].line = CURRENT_LINE;
! 527: }
! 528: }
! 529:
! 530: static void
! 531: dump_stack(t, lev)
! 532: struct position *t;
! 533: int lev;
! 534: {
! 535: int i;
! 536:
! 537: for (i = 0; i < lev; i++) {
! 538: if (i == MAXRECORD) {
! 539: fprintf(stderr, " ...\n");
! 540: break;
! 541: }
! 542: fprintf(stderr, " %s at line %lu\n",
! 543: t[i].name, t[i].line);
! 544: }
! 545: }