1 /* $OpenBSD: main.c,v 1.84 2014/12/21 09:33:12 espie Exp $ */ 2 /* $NetBSD: main.c,v 1.12 1997/02/08 23:54:49 cgd Exp $ */ 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. Neither the name of the University nor the names of its contributors 20 * may be used to endorse or promote products derived from this software 21 * without specific prior written permission. 22 * 23 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 24 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 25 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 26 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 27 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 28 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 29 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 30 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 31 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 32 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 33 * SUCH DAMAGE. 34 */ 35 36 /* 37 * main.c 38 * Facility: m4 macro processor 39 * by: oz 40 */ 41 #include <sys/cdefs.h> 42 __FBSDID("$FreeBSD$"); 43 44 #include <assert.h> 45 #include <signal.h> 46 #include <err.h> 47 #include <errno.h> 48 #include <unistd.h> 49 #include <stdio.h> 50 #include <ctype.h> 51 #include <string.h> 52 #include <stddef.h> 53 #include <stdint.h> 54 #include <stdlib.h> 55 #include <ohash.h> 56 #include "mdef.h" 57 #include "stdd.h" 58 #include "extern.h" 59 #include "pathnames.h" 60 61 stae *mstack; /* stack of m4 machine */ 62 char *sstack; /* shadow stack, for string space extension */ 63 static size_t STACKMAX; /* current maximum size of stack */ 64 int sp; /* current m4 stack pointer */ 65 int fp; /* m4 call frame pointer */ 66 struct input_file infile[MAXINP];/* input file stack (0=stdin) */ 67 FILE **outfile; /* diversion array(0=bitbucket)*/ 68 int maxout; 69 FILE *active; /* active output file pointer */ 70 int ilevel = 0; /* input file stack pointer */ 71 int oindex = 0; /* diversion index.. */ 72 const char *null = ""; /* as it says.. just a null.. */ 73 char **m4wraps = NULL; /* m4wraps array. */ 74 int maxwraps = 0; /* size of m4wraps array */ 75 int wrapindex = 0; /* current offset in m4wraps */ 76 char lquote[MAXCCHARS+1] = {LQUOTE}; /* left quote character (`) */ 77 char rquote[MAXCCHARS+1] = {RQUOTE}; /* right quote character (') */ 78 char scommt[MAXCCHARS+1] = {SCOMMT}; /* start character for comment */ 79 char ecommt[MAXCCHARS+1] = {ECOMMT}; /* end character for comment */ 80 int synch_lines = 0; /* line synchronisation for C preprocessor */ 81 int prefix_builtins = 0; /* -P option to prefix builtin keywords */ 82 83 struct keyblk { 84 const char *knam; /* keyword name */ 85 int ktyp; /* keyword type */ 86 }; 87 88 static struct keyblk keywrds[] = { /* m4 keywords to be installed */ 89 { "include", INCLTYPE }, 90 { "sinclude", SINCTYPE }, 91 { "define", DEFITYPE }, 92 { "defn", DEFNTYPE }, 93 { "divert", DIVRTYPE | NOARGS }, 94 { "expr", EXPRTYPE }, 95 { "eval", EXPRTYPE }, 96 { "substr", SUBSTYPE }, 97 { "ifelse", IFELTYPE }, 98 { "ifdef", IFDFTYPE }, 99 { "len", LENGTYPE }, 100 { "incr", INCRTYPE }, 101 { "decr", DECRTYPE }, 102 { "dnl", DNLNTYPE | NOARGS }, 103 { "changequote", CHNQTYPE | NOARGS }, 104 { "changecom", CHNCTYPE | NOARGS }, 105 { "index", INDXTYPE }, 106 #ifdef EXTENDED 107 { "paste", PASTTYPE }, 108 { "spaste", SPASTYPE }, 109 /* Newer extensions, needed to handle gnu-m4 scripts */ 110 { "indir", INDIRTYPE}, 111 { "builtin", BUILTINTYPE}, 112 { "patsubst", PATSTYPE}, 113 { "regexp", REGEXPTYPE}, 114 { "esyscmd", ESYSCMDTYPE}, 115 { "__file__", FILENAMETYPE | NOARGS}, 116 { "__line__", LINETYPE | NOARGS}, 117 #endif 118 { "popdef", POPDTYPE }, 119 { "pushdef", PUSDTYPE }, 120 { "dumpdef", DUMPTYPE | NOARGS }, 121 { "shift", SHIFTYPE | NOARGS }, 122 { "translit", TRNLTYPE }, 123 { "undefine", UNDFTYPE }, 124 { "undivert", UNDVTYPE | NOARGS }, 125 { "divnum", DIVNTYPE | NOARGS }, 126 { "maketemp", MKTMTYPE }, 127 { "mkstemp", MKTMTYPE }, 128 { "errprint", ERRPTYPE | NOARGS }, 129 { "m4wrap", M4WRTYPE | NOARGS }, 130 { "m4exit", EXITTYPE | NOARGS }, 131 { "syscmd", SYSCTYPE }, 132 { "sysval", SYSVTYPE | NOARGS }, 133 { "traceon", TRACEONTYPE | NOARGS }, 134 { "traceoff", TRACEOFFTYPE | NOARGS }, 135 136 #if defined(unix) || defined(__unix__) 137 { "unix", SELFTYPE | NOARGS }, 138 #else 139 #ifdef vms 140 { "vms", SELFTYPE | NOARGS }, 141 #endif 142 #endif 143 }; 144 145 #define MAXKEYS (sizeof(keywrds)/sizeof(struct keyblk)) 146 147 extern int optind; 148 extern char *optarg; 149 150 #define MAXRECORD 50 151 static struct position { 152 char *name; 153 unsigned long line; 154 } quotes[MAXRECORD], paren[MAXRECORD]; 155 156 static void record(struct position *, int); 157 static void dump_stack(struct position *, int); 158 159 static void macro(void); 160 static void initkwds(void); 161 static ndptr inspect(int, char *); 162 static int do_look_ahead(int, const char *); 163 static void reallyoutputstr(const char *); 164 static void reallyputchar(int); 165 166 static void enlarge_stack(void); 167 168 int main(int, char *[]); 169 170 int exit_code = 0; 171 172 int 173 main(int argc, char *argv[]) 174 { 175 int c; 176 int n; 177 char *p; 178 179 if (signal(SIGINT, SIG_IGN) != SIG_IGN) 180 signal(SIGINT, onintr); 181 182 init_macros(); 183 initspaces(); 184 STACKMAX = INITSTACKMAX; 185 186 mstack = xreallocarray(NULL, STACKMAX, sizeof(stae), NULL); 187 sstack = xalloc(STACKMAX, NULL); 188 189 maxout = 0; 190 outfile = NULL; 191 resizedivs(MAXOUT); 192 193 while ((c = getopt(argc, argv, "gst:d:D:U:o:I:P")) != -1) 194 switch(c) { 195 196 case 'D': /* define something..*/ 197 for (p = optarg; *p; p++) 198 if (*p == '=') 199 break; 200 if (*p) 201 *p++ = EOS; 202 dodefine(optarg, p); 203 break; 204 case 'I': 205 addtoincludepath(optarg); 206 break; 207 case 'P': 208 prefix_builtins = 1; 209 break; 210 case 'U': /* undefine... */ 211 macro_popdef(optarg); 212 break; 213 case 'g': 214 mimic_gnu = 1; 215 break; 216 case 'd': 217 set_trace_flags(optarg); 218 break; 219 case 's': 220 synch_lines = 1; 221 break; 222 case 't': 223 mark_traced(optarg, 1); 224 break; 225 case 'o': 226 trace_file(optarg); 227 break; 228 case '?': 229 usage(); 230 } 231 232 argc -= optind; 233 argv += optind; 234 235 initkwds(); 236 if (mimic_gnu) 237 setup_builtin("format", FORMATTYPE); 238 239 active = stdout; /* default active output */ 240 bbase[0] = bufbase; 241 if (!argc) { 242 sp = -1; /* stack pointer initialized */ 243 fp = 0; /* frame pointer initialized */ 244 set_input(infile+0, stdin, "stdin"); 245 /* default input (naturally) */ 246 macro(); 247 } else 248 for (; argc--; ++argv) { 249 p = *argv; 250 if (p[0] == '-' && p[1] == EOS) 251 set_input(infile, stdin, "stdin"); 252 else if (fopen_trypath(infile, p) == NULL) 253 err(1, "%s", p); 254 sp = -1; 255 fp = 0; 256 macro(); 257 release_input(infile); 258 } 259 260 if (wrapindex) { 261 int i; 262 263 ilevel = 0; /* in case m4wrap includes.. */ 264 bufbase = bp = buf; /* use the entire buffer */ 265 if (mimic_gnu) { 266 while (wrapindex != 0) { 267 for (i = 0; i < wrapindex; i++) 268 pbstr(m4wraps[i]); 269 wrapindex =0; 270 macro(); 271 } 272 } else { 273 for (i = 0; i < wrapindex; i++) { 274 pbstr(m4wraps[i]); 275 macro(); 276 } 277 } 278 } 279 280 if (active != stdout) 281 active = stdout; /* reset output just in case */ 282 for (n = 1; n < maxout; n++) /* default wrap-up: undivert */ 283 if (outfile[n] != NULL) 284 getdiv(n); 285 /* remove bitbucket if used */ 286 if (outfile[0] != NULL) { 287 (void) fclose(outfile[0]); 288 } 289 290 return exit_code; 291 } 292 293 /* 294 * Look ahead for `token'. 295 * (on input `t == token[0]') 296 * Used for comment and quoting delimiters. 297 * Returns 1 if `token' present; copied to output. 298 * 0 if `token' not found; all characters pushed back 299 */ 300 static int 301 do_look_ahead(int t, const char *token) 302 { 303 int i; 304 305 assert((unsigned char)t == (unsigned char)token[0]); 306 307 for (i = 1; *++token; i++) { 308 t = gpbc(); 309 if (t == EOF || (unsigned char)t != (unsigned char)*token) { 310 pushback(t); 311 while (--i) 312 pushback(*--token); 313 return 0; 314 } 315 } 316 return 1; 317 } 318 319 #define LOOK_AHEAD(t, token) (t != EOF && \ 320 (unsigned char)(t)==(unsigned char)(token)[0] && \ 321 do_look_ahead(t,token)) 322 323 /* 324 * macro - the work horse.. 325 */ 326 static void 327 macro(void) 328 { 329 char token[MAXTOK+1]; 330 int t, l; 331 ndptr p; 332 int nlpar; 333 334 cycle { 335 t = gpbc(); 336 337 if (LOOK_AHEAD(t,lquote)) { /* strip quotes */ 338 nlpar = 0; 339 record(quotes, nlpar++); 340 /* 341 * Opening quote: scan forward until matching 342 * closing quote has been found. 343 */ 344 do { 345 346 l = gpbc(); 347 if (LOOK_AHEAD(l,rquote)) { 348 if (--nlpar > 0) 349 outputstr(rquote); 350 } else if (LOOK_AHEAD(l,lquote)) { 351 record(quotes, nlpar++); 352 outputstr(lquote); 353 } else if (l == EOF) { 354 if (nlpar == 1) 355 warnx("unclosed quote:"); 356 else 357 warnx("%d unclosed quotes:", nlpar); 358 dump_stack(quotes, nlpar); 359 exit(1); 360 } else { 361 if (nlpar > 0) { 362 if (sp < 0) 363 reallyputchar(l); 364 else 365 CHRSAVE(l); 366 } 367 } 368 } 369 while (nlpar != 0); 370 } else if (sp < 0 && LOOK_AHEAD(t, scommt)) { 371 reallyoutputstr(scommt); 372 373 for(;;) { 374 t = gpbc(); 375 if (LOOK_AHEAD(t, ecommt)) { 376 reallyoutputstr(ecommt); 377 break; 378 } 379 if (t == EOF) 380 break; 381 reallyputchar(t); 382 } 383 } else if (t == '_' || isalpha(t)) { 384 p = inspect(t, token); 385 if (p != NULL) 386 pushback(l = gpbc()); 387 if (p == NULL || (l != LPAREN && 388 (macro_getdef(p)->type & NEEDARGS) != 0)) 389 outputstr(token); 390 else { 391 /* 392 * real thing.. First build a call frame: 393 */ 394 pushf(fp); /* previous call frm */ 395 pushf(macro_getdef(p)->type); /* type of the call */ 396 pushf(is_traced(p)); 397 pushf(0); /* parenthesis level */ 398 fp = sp; /* new frame pointer */ 399 /* 400 * now push the string arguments: 401 */ 402 pushdef(p); /* defn string */ 403 pushs1((char *)macro_name(p)); /* macro name */ 404 pushs(ep); /* start next..*/ 405 406 if (l != LPAREN && PARLEV == 0) { 407 /* no bracks */ 408 chrsave(EOS); 409 410 if (sp == (int)STACKMAX) 411 errx(1, "internal stack overflow"); 412 eval((const char **) mstack+fp+1, 2, 413 CALTYP, TRACESTATUS); 414 415 ep = PREVEP; /* flush strspace */ 416 sp = PREVSP; /* previous sp.. */ 417 fp = PREVFP; /* rewind stack...*/ 418 } 419 } 420 } else if (t == EOF) { 421 if (!mimic_gnu /* you can puke right there */ 422 && sp > -1 && ilevel <= 0) { 423 warnx( "unexpected end of input, unclosed parenthesis:"); 424 dump_stack(paren, PARLEV); 425 exit(1); 426 } 427 if (ilevel <= 0) 428 break; /* all done thanks.. */ 429 release_input(infile+ilevel--); 430 emit_synchline(); 431 bufbase = bbase[ilevel]; 432 continue; 433 } else if (sp < 0) { /* not in a macro at all */ 434 reallyputchar(t); /* output directly.. */ 435 } 436 437 else switch(t) { 438 439 case LPAREN: 440 if (PARLEV > 0) 441 chrsave(t); 442 while (isspace(l = gpbc())) /* skip blank, tab, nl.. */ 443 if (PARLEV > 0) 444 chrsave(l); 445 pushback(l); 446 record(paren, PARLEV++); 447 break; 448 449 case RPAREN: 450 if (--PARLEV > 0) 451 chrsave(t); 452 else { /* end of argument list */ 453 chrsave(EOS); 454 455 if (sp == (int)STACKMAX) 456 errx(1, "internal stack overflow"); 457 458 eval((const char **) mstack+fp+1, sp-fp, 459 CALTYP, TRACESTATUS); 460 461 ep = PREVEP; /* flush strspace */ 462 sp = PREVSP; /* previous sp.. */ 463 fp = PREVFP; /* rewind stack...*/ 464 } 465 break; 466 467 case COMMA: 468 if (PARLEV == 1) { 469 chrsave(EOS); /* new argument */ 470 while (isspace(l = gpbc())) 471 ; 472 pushback(l); 473 pushs(ep); 474 } else 475 chrsave(t); 476 break; 477 478 default: 479 if (LOOK_AHEAD(t, scommt)) { 480 char *cp; 481 for (cp = scommt; *cp; cp++) 482 chrsave(*cp); 483 for(;;) { 484 t = gpbc(); 485 if (LOOK_AHEAD(t, ecommt)) { 486 for (cp = ecommt; *cp; cp++) 487 chrsave(*cp); 488 break; 489 } 490 if (t == EOF) 491 break; 492 CHRSAVE(t); 493 } 494 } else 495 CHRSAVE(t); /* stack the char */ 496 break; 497 } 498 } 499 } 500 501 /* 502 * output string directly, without pushing it for reparses. 503 */ 504 void 505 outputstr(const char *s) 506 { 507 if (sp < 0) 508 reallyoutputstr(s); 509 else 510 while (*s) 511 CHRSAVE(*s++); 512 } 513 514 void 515 reallyoutputstr(const char *s) 516 { 517 if (synch_lines) { 518 while (*s) { 519 fputc(*s, active); 520 if (*s++ == '\n') { 521 infile[ilevel].synch_lineno++; 522 if (infile[ilevel].synch_lineno != 523 infile[ilevel].lineno) 524 do_emit_synchline(); 525 } 526 } 527 } else 528 fputs(s, active); 529 } 530 531 void 532 reallyputchar(int c) 533 { 534 putc(c, active); 535 if (synch_lines && c == '\n') { 536 infile[ilevel].synch_lineno++; 537 if (infile[ilevel].synch_lineno != infile[ilevel].lineno) 538 do_emit_synchline(); 539 } 540 } 541 542 /* 543 * build an input token.. 544 * consider only those starting with _ or A-Za-z. 545 */ 546 static ndptr 547 inspect(int c, char *tp) 548 { 549 char *name = tp; 550 char *etp = tp+MAXTOK; 551 ndptr p; 552 553 *tp++ = c; 554 555 while ((isalnum(c = gpbc()) || c == '_') && tp < etp) 556 *tp++ = c; 557 if (c != EOF) 558 PUSHBACK(c); 559 *tp = EOS; 560 /* token is too long, it won't match anything, but it can still 561 * be output. */ 562 if (tp == ep) { 563 outputstr(name); 564 while (isalnum(c = gpbc()) || c == '_') { 565 if (sp < 0) 566 reallyputchar(c); 567 else 568 CHRSAVE(c); 569 } 570 *name = EOS; 571 return NULL; 572 } 573 574 p = ohash_find(¯os, ohash_qlookupi(¯os, name, (const char **)&tp)); 575 if (p == NULL) 576 return NULL; 577 if (macro_getdef(p) == NULL) 578 return NULL; 579 return p; 580 } 581 582 /* 583 * initkwds - initialise m4 keywords as fast as possible. 584 * This very similar to install, but without certain overheads, 585 * such as calling lookup. Malloc is not used for storing the 586 * keyword strings, since we simply use the static pointers 587 * within keywrds block. 588 */ 589 static void 590 initkwds(void) 591 { 592 unsigned int type; 593 int i; 594 595 for (i = 0; i < (int)MAXKEYS; i++) { 596 type = keywrds[i].ktyp & TYPEMASK; 597 if ((keywrds[i].ktyp & NOARGS) == 0) 598 type |= NEEDARGS; 599 setup_builtin(keywrds[i].knam, type); 600 } 601 } 602 603 static void 604 record(struct position *t, int lev) 605 { 606 if (lev < MAXRECORD) { 607 t[lev].name = CURRENT_NAME; 608 t[lev].line = CURRENT_LINE; 609 } 610 } 611 612 static void 613 dump_stack(struct position *t, int lev) 614 { 615 int i; 616 617 for (i = 0; i < lev; i++) { 618 if (i == MAXRECORD) { 619 fprintf(stderr, " ...\n"); 620 break; 621 } 622 fprintf(stderr, " %s at line %lu\n", 623 t[i].name, t[i].line); 624 } 625 } 626 627 628 static void 629 enlarge_stack(void) 630 { 631 STACKMAX += STACKMAX/2; 632 mstack = xreallocarray(mstack, STACKMAX, sizeof(stae), 633 "Evaluation stack overflow (%lu)", 634 (unsigned long)STACKMAX); 635 sstack = xrealloc(sstack, STACKMAX, 636 "Evaluation stack overflow (%lu)", 637 (unsigned long)STACKMAX); 638 } 639