1 /*- 2 * Copyright (c) 1992 Diomidis Spinellis. 3 * Copyright (c) 1992 The Regents of the University of California. 4 * All rights reserved. 5 * 6 * This code is derived from software contributed to Berkeley by 7 * Diomidis Spinellis of Imperial College, University of London. 8 * 9 * Redistribution and use in source and binary forms, with or without 10 * modification, are permitted provided that the following conditions 11 * are met: 12 * 1. Redistributions of source code must retain the above copyright 13 * notice, this list of conditions and the following disclaimer. 14 * 2. Redistributions in binary form must reproduce the above copyright 15 * notice, this list of conditions and the following disclaimer in the 16 * documentation and/or other materials provided with the distribution. 17 * 3. All advertising materials mentioning features or use of this software 18 * must display the following acknowledgement: 19 * This product includes software developed by the University of 20 * California, Berkeley and its contributors. 21 * 4. Neither the name of the University nor the names of its contributors 22 * may be used to endorse or promote products derived from this software 23 * without specific prior written permission. 24 * 25 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 26 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 27 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 28 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 29 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 30 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 31 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 32 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 33 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 34 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 35 * SUCH DAMAGE. 36 */ 37 38 #ifndef lint 39 /*static char sccsid[] = "from: @(#)compile.c 5.6 (Berkeley) 11/2/92";*/ 40 static char rcsid[] = "$Id: compile.c,v 1.7 1993/11/20 22:24:17 mycroft Exp $"; 41 #endif /* not lint */ 42 43 #include <sys/types.h> 44 #include <sys/stat.h> 45 46 #include <ctype.h> 47 #include <errno.h> 48 #include <fcntl.h> 49 #include <limits.h> 50 #include <regex.h> 51 #include <stdio.h> 52 #include <stdlib.h> 53 #include <string.h> 54 55 #include "defs.h" 56 #include "extern.h" 57 58 static char *compile_addr __P((char *, struct s_addr *)); 59 static char *compile_ccl __P((char **, char *)); 60 static char *compile_delimited __P((char *, char *)); 61 static char *compile_flags __P((char *, struct s_subst *)); 62 static char *compile_re __P((char *, regex_t **)); 63 static char *compile_subst __P((char *, struct s_subst *)); 64 static char *compile_text __P((void)); 65 static char *compile_tr __P((char *, char **)); 66 static struct s_command 67 **compile_stream __P((char *, struct s_command **, char *)); 68 static char *duptoeol __P((char *)); 69 static struct s_command 70 *findlabel __P((struct s_command *, struct s_command *)); 71 static void fixuplabel __P((struct s_command *, struct s_command *, 72 struct s_command *)); 73 74 /* 75 * Command specification. This is used to drive the command parser. 76 */ 77 struct s_format { 78 char code; /* Command code */ 79 int naddr; /* Number of address args */ 80 enum e_args args; /* Argument type */ 81 }; 82 83 static struct s_format cmd_fmts[] = { 84 {'{', 2, GROUP}, 85 {'a', 1, TEXT}, 86 {'b', 2, BRANCH}, 87 {'c', 2, TEXT}, 88 {'d', 2, EMPTY}, 89 {'D', 2, EMPTY}, 90 {'g', 2, EMPTY}, 91 {'G', 2, EMPTY}, 92 {'h', 2, EMPTY}, 93 {'H', 2, EMPTY}, 94 {'i', 1, TEXT}, 95 {'l', 2, EMPTY}, 96 {'n', 2, EMPTY}, 97 {'N', 2, EMPTY}, 98 {'p', 2, EMPTY}, 99 {'P', 2, EMPTY}, 100 {'q', 1, EMPTY}, 101 {'r', 1, RFILE}, 102 {'s', 2, SUBST}, 103 {'t', 2, BRANCH}, 104 {'w', 2, WFILE}, 105 {'x', 2, EMPTY}, 106 {'y', 2, TR}, 107 {'!', 2, NONSEL}, 108 {':', 0, LABEL}, 109 {'#', 0, COMMENT}, 110 {'=', 1, EMPTY}, 111 {'\0', 0, COMMENT}, 112 }; 113 114 /* The compiled program. */ 115 struct s_command *prog; 116 117 /* 118 * Compile the program into prog. 119 * Initialise appends. 120 */ 121 void 122 compile() 123 { 124 *compile_stream(NULL, &prog, NULL) = NULL; 125 fixuplabel(prog, prog, NULL); 126 appends = xmalloc(sizeof(struct s_appends) * appendnum); 127 match = xmalloc((maxnsub + 1) * sizeof(regmatch_t)); 128 } 129 130 #define EATSPACE() do { \ 131 if (p) \ 132 while (*p && isascii(*p) && isspace(*p)) \ 133 p++; \ 134 } while (0) 135 136 static struct s_command ** 137 compile_stream(terminator, link, p) 138 char *terminator; 139 struct s_command **link; 140 register char *p; 141 { 142 static char lbuf[_POSIX2_LINE_MAX + 1]; /* To save stack */ 143 struct s_command *cmd, *cmd2; 144 struct s_format *fp; 145 int naddr; /* Number of addresses */ 146 147 if (p != NULL) 148 goto semicolon; 149 for (;;) { 150 if ((p = cu_fgets(lbuf, sizeof(lbuf))) == NULL) { 151 if (terminator != NULL) 152 err(COMPILE, "unexpected EOF (pending }'s)"); 153 return (link); 154 } 155 156 semicolon: EATSPACE(); 157 if (p && (*p == '#' || *p == '\0')) 158 continue; 159 if (*p == '}') { 160 if (terminator == NULL) 161 err(COMPILE, "unexpected }"); 162 return (link); 163 } 164 *link = cmd = xmalloc(sizeof(struct s_command)); 165 link = &cmd->next; 166 cmd->nonsel = cmd->inrange = 0; 167 /* First parse the addresses */ 168 naddr = 0; 169 cmd->a1 = cmd->a2 = NULL; 170 171 /* Valid characters to start an address */ 172 #define addrchar(c) (strchr("0123456789/\\$", (c))) 173 if (addrchar(*p)) { 174 naddr++; 175 cmd->a1 = xmalloc(sizeof(struct s_addr)); 176 p = compile_addr(p, cmd->a1); 177 EATSPACE(); /* EXTENSION */ 178 if (*p == ',') { 179 naddr++; 180 p++; 181 EATSPACE(); /* EXTENSION */ 182 cmd->a2 = xmalloc(sizeof(struct s_addr)); 183 p = compile_addr(p, cmd->a2); 184 } 185 } 186 187 nonsel: /* Now parse the command */ 188 EATSPACE(); 189 if (!*p) 190 err(COMPILE, "command expected"); 191 cmd->code = *p; 192 for (fp = cmd_fmts; fp->code; fp++) 193 if (fp->code == *p) 194 break; 195 if (!fp->code) 196 err(COMPILE, "invalid command code %c", *p); 197 if (naddr > fp->naddr) 198 err(COMPILE, 199 "command %c expects up to %d address(es), found %d", *p, fp->naddr, naddr); 200 switch (fp->args) { 201 case NONSEL: /* ! */ 202 cmd->nonsel = ! cmd->nonsel; 203 p++; 204 goto nonsel; 205 case GROUP: /* { */ 206 p++; 207 EATSPACE(); 208 if (!*p) 209 p = NULL; 210 cmd2 = xmalloc(sizeof(struct s_command)); 211 cmd2->code = '}'; 212 cmd2->a1 = cmd2->a2 = NULL; 213 *compile_stream("}", &cmd->u.c, p) = cmd2; 214 cmd->next = cmd2; 215 link = &cmd2->next; 216 break; 217 case EMPTY: /* d D g G h H l n N p P q x = \0 */ 218 p++; 219 EATSPACE(); 220 if (*p == ';') { 221 p++; 222 link = &cmd->next; 223 goto semicolon; 224 } 225 if (*p) 226 err(COMPILE, 227 "extra characters at the end of %c command", cmd->code); 228 break; 229 case TEXT: /* a c i */ 230 p++; 231 EATSPACE(); 232 if (*p != '\\') 233 err(COMPILE, 234 "command %c expects \\ followed by text", cmd->code); 235 p++; 236 EATSPACE(); 237 if (*p) 238 err(COMPILE, 239 "extra characters after \\ at the end of %c command", cmd->code); 240 cmd->t = compile_text(); 241 break; 242 case COMMENT: /* \0 # */ 243 break; 244 case WFILE: /* w */ 245 p++; 246 EATSPACE(); 247 if (*p == '\0') 248 err(COMPILE, "filename expected"); 249 cmd->t = duptoeol(p); 250 if (aflag) 251 cmd->u.fd = -1; 252 else if ((cmd->u.fd = open(p, 253 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, 254 DEFFILEMODE)) == -1) 255 err(FATAL, "%s: %s\n", p, strerror(errno)); 256 break; 257 case RFILE: /* r */ 258 p++; 259 EATSPACE(); 260 if (*p == '\0') 261 err(COMPILE, "filename expected"); 262 else 263 cmd->t = duptoeol(p); 264 break; 265 case BRANCH: /* b t */ 266 p++; 267 EATSPACE(); 268 if (*p == '\0') 269 cmd->t = NULL; 270 else 271 cmd->t = duptoeol(p); 272 break; 273 case LABEL: /* : */ 274 p++; 275 EATSPACE(); 276 cmd->t = duptoeol(p); 277 if (strlen(p) == 0) 278 err(COMPILE, "empty label"); 279 break; 280 case SUBST: /* s */ 281 p++; 282 if (*p == '\0' || *p == '\\') 283 err(COMPILE, 284 "substitute pattern can not be delimited by newline or backslash"); 285 cmd->u.s = xmalloc(sizeof(struct s_subst)); 286 p = compile_re(p, &cmd->u.s->re); 287 if (p == NULL) 288 err(COMPILE, "unterminated substitute pattern"); 289 --p; 290 p = compile_subst(p, cmd->u.s); 291 p = compile_flags(p, cmd->u.s); 292 EATSPACE(); 293 if (*p == ';') { 294 p++; 295 link = &cmd->next; 296 goto semicolon; 297 } 298 break; 299 case TR: /* y */ 300 p++; 301 p = compile_tr(p, (char **)&cmd->u.y); 302 EATSPACE(); 303 if (*p == ';') { 304 p++; 305 link = &cmd->next; 306 goto semicolon; 307 } 308 if (*p) 309 err(COMPILE, 310 "extra text at the end of a transform command"); 311 break; 312 } 313 } 314 } 315 316 /* 317 * Get a delimited string. P points to the delimeter of the string; d points 318 * to a buffer area. Newline and delimiter escapes are processed; other 319 * escapes are ignored. 320 * 321 * Returns a pointer to the first character after the final delimiter or NULL 322 * in the case of a non-terminated string. The character array d is filled 323 * with the processed string. 324 */ 325 static char * 326 compile_delimited(p, d) 327 char *p, *d; 328 { 329 char c; 330 331 c = *p++; 332 if (c == '\0') 333 return (NULL); 334 else if (c == '\\') 335 err(COMPILE, "\\ can not be used as a string delimiter"); 336 else if (c == '\n') 337 err(COMPILE, "newline can not be used as a string delimiter"); 338 while (*p) { 339 if (*p == '[') { 340 if ((d = compile_ccl(&p, d)) == NULL) 341 err(COMPILE, "unbalanced brackets ([])"); 342 continue; 343 } else if (*p == '\\' && p[1] == c) 344 p++; 345 else if (*p == '\\' && p[1] == 'n') { 346 *d++ = '\n'; 347 p += 2; 348 continue; 349 } else if (*p == '\\' && p[1] == '\\') 350 *d++ = *p++; 351 else if (*p == c) { 352 *d = '\0'; 353 return (p + 1); 354 } 355 *d++ = *p++; 356 } 357 return (NULL); 358 } 359 360 361 /* compile_ccl: expand a POSIX character class */ 362 static char * 363 compile_ccl(sp, t) 364 char **sp; 365 char *t; 366 { 367 int c, d; 368 char *s = *sp; 369 370 *t++ = *s++; 371 if (*s == '^') 372 *t++ = *s++; 373 if (*s == ']') 374 *t++ = *s++; 375 for (; *s && (*t = *s) != ']'; s++, t++) 376 if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) { 377 *++t = *++s, t++, s++; 378 for (c = *s; (*t = *s) != ']' || c != d; s++, t++) 379 if ((c = *s) == '\0') 380 return NULL; 381 } else if (*s == '\\' && s[1] == 'n') 382 *t = '\n', s++; 383 return (*s == ']') ? *sp = ++s, ++t : NULL; 384 } 385 386 /* 387 * Get a regular expression. P points to the delimiter of the regular 388 * expression; repp points to the address of a regexp pointer. Newline 389 * and delimiter escapes are processed; other escapes are ignored. 390 * Returns a pointer to the first character after the final delimiter 391 * or NULL in the case of a non terminated regular expression. The regexp 392 * pointer is set to the compiled regular expression. 393 * Cflags are passed to regcomp. 394 */ 395 static char * 396 compile_re(p, repp) 397 char *p; 398 regex_t **repp; 399 { 400 int eval; 401 char re[_POSIX2_LINE_MAX + 1]; 402 403 p = compile_delimited(p, re); 404 if (p && strlen(re) == 0) { 405 *repp = NULL; 406 return (p); 407 } 408 *repp = xmalloc(sizeof(regex_t)); 409 if (p && (eval = regcomp(*repp, re, 0)) != 0) 410 err(COMPILE, "RE error: %s", strregerror(eval, *repp)); 411 if (maxnsub < (*repp)->re_nsub) 412 maxnsub = (*repp)->re_nsub; 413 return (p); 414 } 415 416 /* 417 * Compile the substitution string of a regular expression and set res to 418 * point to a saved copy of it. Nsub is the number of parenthesized regular 419 * expressions. 420 */ 421 static char * 422 compile_subst(p, s) 423 char *p; 424 struct s_subst *s; 425 { 426 static char lbuf[_POSIX2_LINE_MAX + 1]; 427 int asize, ref, size; 428 char c, *text, *op, *sp; 429 430 c = *p++; /* Terminator character */ 431 if (c == '\0') 432 return (NULL); 433 434 s->maxbref = 0; 435 s->linenum = linenum; 436 asize = 2 * _POSIX2_LINE_MAX + 1; 437 text = xmalloc(asize); 438 size = 0; 439 do { 440 op = sp = text + size; 441 for (; *p; p++) { 442 if (*p == '\\') { 443 p++; 444 if (strchr("123456789", *p) != NULL) { 445 *sp++ = '\\'; 446 ref = *p - '0'; 447 if (s->re != NULL && 448 ref > s->re->re_nsub) 449 err(COMPILE, 450 "\\%c not defined in the RE", *p); 451 if (s->maxbref < ref) 452 s->maxbref = ref; 453 } else if (*p == '&' || *p == '\\') 454 *sp++ = '\\'; 455 } else if (*p == c) { 456 p++; 457 *sp++ = '\0'; 458 size += sp - op; 459 s->new = xrealloc(text, size); 460 return (p); 461 } else if (*p == '\n') { 462 err(COMPILE, 463 "unescaped newline inside substitute pattern"); 464 /* NOTREACHED */ 465 } 466 *sp++ = *p; 467 } 468 size += sp - op; 469 if (asize - size < _POSIX2_LINE_MAX + 1) { 470 asize *= 2; 471 text = xmalloc(asize); 472 } 473 } while (cu_fgets(p = lbuf, sizeof(lbuf))); 474 err(COMPILE, "unterminated substitute in regular expression"); 475 /* NOTREACHED */ 476 } 477 478 /* 479 * Compile the flags of the s command 480 */ 481 static char * 482 compile_flags(p, s) 483 char *p; 484 struct s_subst *s; 485 { 486 int gn; /* True if we have seen g or n */ 487 char wfile[_POSIX2_LINE_MAX + 1], *q; 488 489 s->n = 1; /* Default */ 490 s->p = 0; 491 s->wfile = NULL; 492 s->wfd = -1; 493 for (gn = 0;;) { 494 EATSPACE(); /* EXTENSION */ 495 switch (*p) { 496 case 'g': 497 if (gn) 498 err(COMPILE, 499 "more than one number or 'g' in substitute flags"); 500 gn = 1; 501 s->n = 0; 502 break; 503 case '\0': 504 case '\n': 505 case ';': 506 return (p); 507 case 'p': 508 s->p = 1; 509 break; 510 case '1': case '2': case '3': 511 case '4': case '5': case '6': 512 case '7': case '8': case '9': 513 if (gn) 514 err(COMPILE, 515 "more than one number or 'g' in substitute flags"); 516 gn = 1; 517 /* XXX Check for overflow */ 518 s->n = (int)strtol(p, &p, 10); 519 break; 520 case 'w': 521 p++; 522 #ifdef FUSSY_HISTORIC_PRACTICE 523 if (*p != ' ') { 524 err(WARNING, "space missing before w wfile"); 525 return (p); 526 } 527 #endif 528 EATSPACE(); 529 q = wfile; 530 while (*p) { 531 if (*p == '\n') 532 break; 533 *q++ = *p++; 534 } 535 *q = '\0'; 536 if (q == wfile) 537 err(COMPILE, "no wfile specified"); 538 s->wfile = strdup(wfile); 539 if (!aflag && (s->wfd = open(wfile, 540 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, 541 DEFFILEMODE)) == -1) 542 err(FATAL, "%s: %s\n", wfile, strerror(errno)); 543 return (p); 544 default: 545 err(COMPILE, 546 "bad flag in substitute command: '%c'", *p); 547 break; 548 } 549 p++; 550 } 551 } 552 553 /* 554 * Compile a translation set of strings into a lookup table. 555 */ 556 static char * 557 compile_tr(p, transtab) 558 char *p; 559 char **transtab; 560 { 561 int i; 562 char *lt, *op, *np; 563 char old[_POSIX2_LINE_MAX + 1]; 564 char new[_POSIX2_LINE_MAX + 1]; 565 566 if (*p == '\0' || *p == '\\') 567 err(COMPILE, 568 "transform pattern can not be delimited by newline or backslash"); 569 p = compile_delimited(p, old); 570 if (p == NULL) { 571 err(COMPILE, "unterminated transform source string"); 572 return (NULL); 573 } 574 p = compile_delimited(--p, new); 575 if (p == NULL) { 576 err(COMPILE, "unterminated transform target string"); 577 return (NULL); 578 } 579 EATSPACE(); 580 if (strlen(new) != strlen(old)) { 581 err(COMPILE, "transform strings are not the same length"); 582 return (NULL); 583 } 584 /* We assume characters are 8 bits */ 585 lt = xmalloc(UCHAR_MAX); 586 for (i = 0; i <= UCHAR_MAX; i++) 587 lt[i] = (char)i; 588 for (op = old, np = new; *op; op++, np++) 589 lt[(u_char)*op] = *np; 590 *transtab = lt; 591 return (p); 592 } 593 594 /* 595 * Compile the text following an a or i command. 596 */ 597 static char * 598 compile_text() 599 { 600 int asize, size; 601 char *text, *p, *op, *s; 602 char lbuf[_POSIX2_LINE_MAX + 1]; 603 604 asize = 2 * _POSIX2_LINE_MAX + 1; 605 text = xmalloc(asize); 606 size = 0; 607 while (cu_fgets(lbuf, sizeof(lbuf))) { 608 op = s = text + size; 609 p = lbuf; 610 EATSPACE(); 611 for (; *p; p++) { 612 if (*p == '\\') 613 p++; 614 *s++ = *p; 615 } 616 size += s - op; 617 if (p[-2] != '\\') { 618 *s = '\0'; 619 break; 620 } 621 if (asize - size < _POSIX2_LINE_MAX + 1) { 622 asize *= 2; 623 text = xmalloc(asize); 624 } 625 } 626 return (xrealloc(text, size + 1)); 627 } 628 629 /* 630 * Get an address and return a pointer to the first character after 631 * it. Fill the structure pointed to according to the address. 632 */ 633 static char * 634 compile_addr(p, a) 635 char *p; 636 struct s_addr *a; 637 { 638 char *end; 639 640 switch (*p) { 641 case '\\': /* Context address */ 642 ++p; 643 /* FALLTHROUGH */ 644 case '/': /* Context address */ 645 p = compile_re(p, &a->u.r); 646 if (p == NULL) 647 err(COMPILE, "unterminated regular expression"); 648 a->type = AT_RE; 649 return (p); 650 651 case '$': /* Last line */ 652 a->type = AT_LAST; 653 return (p + 1); 654 /* Line number */ 655 case '0': case '1': case '2': case '3': case '4': 656 case '5': case '6': case '7': case '8': case '9': 657 a->type = AT_LINE; 658 a->u.l = strtol(p, &end, 10); 659 return (end); 660 default: 661 err(COMPILE, "expected context address"); 662 return (NULL); 663 } 664 } 665 666 /* 667 * Return a copy of all the characters up to \n or \0 668 */ 669 static char * 670 duptoeol(s) 671 register char *s; 672 { 673 size_t len; 674 char *start; 675 676 for (start = s; *s != '\0' && *s != '\n'; ++s); 677 *s = '\0'; 678 len = s - start + 1; 679 return (memmove(xmalloc(len), start, len)); 680 } 681 682 /* 683 * Find the label contained in the command l in the command linked list cp. 684 * L is excluded from the search. Return NULL if not found. 685 */ 686 static struct s_command * 687 findlabel(l, cp) 688 struct s_command *l, *cp; 689 { 690 struct s_command *r; 691 692 for (; cp; cp = cp->next) 693 if (cp->code == ':' && cp != l && strcmp(l->t, cp->t) == 0) 694 return (cp); 695 else if (cp->code == '{' && (r = findlabel(l, cp->u.c))) 696 return (r); 697 return (NULL); 698 } 699 700 /* 701 * Convert goto label names to addresses. 702 * Detect duplicate labels. 703 * Set appendnum to the number of a and r commands in the script. 704 * Free the memory used by labels in b and t commands (but not by :) 705 * Root is a pointer to the script linked list; cp points to the 706 * search start. 707 * TODO: Remove } nodes 708 */ 709 static void 710 fixuplabel(root, cp, end) 711 struct s_command *root, *cp, *end; 712 { 713 struct s_command *cp2; 714 715 for (; cp != end; cp = cp->next) 716 switch (cp->code) { 717 case ':': 718 if (findlabel(cp, root)) 719 err(COMPILE2, "duplicate label %s", cp->t); 720 break; 721 case 'a': 722 case 'r': 723 appendnum++; 724 break; 725 case 'b': 726 case 't': 727 if (cp->t == NULL) { 728 cp->u.c = NULL; 729 break; 730 } 731 if ((cp2 = findlabel(cp, root)) == NULL) 732 err(COMPILE2, "undefined label '%s'", cp->t); 733 free(cp->t); 734 cp->u.c = cp2; 735 break; 736 case '{': 737 fixuplabel(root, cp->u.c, cp->next); 738 break; 739 } 740 } 741