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[] = "@(#)compile.c 5.6 (Berkeley) 11/2/92"; 40 #endif /* not lint */ 41 42 #include <sys/types.h> 43 #include <sys/stat.h> 44 45 #include <ctype.h> 46 #include <errno.h> 47 #include <fcntl.h> 48 #include <limits.h> 49 #include <regex.h> 50 #include <stdio.h> 51 #include <stdlib.h> 52 #include <string.h> 53 54 #include "defs.h" 55 #include "extern.h" 56 57 static char *compile_addr __P((char *, struct s_addr *)); 58 static char *compile_ccl __P((char **, char *)); 59 static char *compile_delimited __P((char *, char *)); 60 static char *compile_flags __P((char *, struct s_subst *)); 61 static char *compile_re __P((char *, regex_t **)); 62 static char *compile_subst __P((char *, struct s_subst *)); 63 static char *compile_text __P((void)); 64 static char *compile_tr __P((char *, char **)); 65 static struct s_command 66 **compile_stream __P((char *, struct s_command **, char *)); 67 static char *duptoeol __P((char *)); 68 static struct s_command 69 *findlabel __P((struct s_command *, struct s_command *)); 70 static void fixuplabel __P((struct s_command *, struct s_command *, 71 struct s_command *)); 72 73 /* 74 * Command specification. This is used to drive the command parser. 75 */ 76 struct s_format { 77 char code; /* Command code */ 78 int naddr; /* Number of address args */ 79 enum e_args args; /* Argument type */ 80 }; 81 82 static struct s_format cmd_fmts[] = { 83 {'{', 2, GROUP}, 84 {'a', 1, TEXT}, 85 {'b', 2, BRANCH}, 86 {'c', 2, TEXT}, 87 {'d', 2, EMPTY}, 88 {'D', 2, EMPTY}, 89 {'g', 2, EMPTY}, 90 {'G', 2, EMPTY}, 91 {'h', 2, EMPTY}, 92 {'H', 2, EMPTY}, 93 {'i', 1, TEXT}, 94 {'l', 2, EMPTY}, 95 {'n', 2, EMPTY}, 96 {'N', 2, EMPTY}, 97 {'p', 2, EMPTY}, 98 {'P', 2, EMPTY}, 99 {'q', 1, EMPTY}, 100 {'r', 1, RFILE}, 101 {'s', 2, SUBST}, 102 {'t', 2, BRANCH}, 103 {'w', 2, WFILE}, 104 {'x', 2, EMPTY}, 105 {'y', 2, TR}, 106 {'!', 2, NONSEL}, 107 {':', 0, LABEL}, 108 {'#', 0, COMMENT}, 109 {'=', 1, EMPTY}, 110 {'\0', 0, COMMENT}, 111 }; 112 113 /* The compiled program. */ 114 struct s_command *prog; 115 116 /* 117 * Compile the program into prog. 118 * Initialise appends. 119 */ 120 void 121 compile() 122 { 123 *compile_stream(NULL, &prog, NULL) = NULL; 124 fixuplabel(prog, prog, NULL); 125 appends = xmalloc(sizeof(struct s_appends) * appendnum); 126 match = xmalloc((maxnsub + 1) * sizeof(regmatch_t)); 127 } 128 129 #define EATSPACE() do { \ 130 if (p) \ 131 while (*p && isascii(*p) && isspace(*p)) \ 132 p++; \ 133 } while (0) 134 135 static struct s_command ** 136 compile_stream(terminator, link, p) 137 char *terminator; 138 struct s_command **link; 139 register char *p; 140 { 141 static char lbuf[_POSIX2_LINE_MAX + 1]; /* To save stack */ 142 struct s_command *cmd, *cmd2; 143 struct s_format *fp; 144 int naddr; /* Number of addresses */ 145 146 if (p != NULL) 147 goto semicolon; 148 for (;;) { 149 if ((p = cu_fgets(lbuf, sizeof(lbuf))) == NULL) { 150 if (terminator != NULL) 151 err(COMPILE, "unexpected EOF (pending }'s)"); 152 return (link); 153 } 154 155 semicolon: EATSPACE(); 156 if (p && (*p == '#' || *p == '\0')) 157 continue; 158 if (*p == '}') { 159 if (terminator == NULL) 160 err(COMPILE, "unexpected }"); 161 return (link); 162 } 163 *link = cmd = xmalloc(sizeof(struct s_command)); 164 link = &cmd->next; 165 cmd->nonsel = cmd->inrange = 0; 166 /* First parse the addresses */ 167 naddr = 0; 168 cmd->a1 = cmd->a2 = NULL; 169 170 /* Valid characters to start an address */ 171 #define addrchar(c) (strchr("0123456789/\\$", (c))) 172 if (addrchar(*p)) { 173 naddr++; 174 cmd->a1 = xmalloc(sizeof(struct s_addr)); 175 p = compile_addr(p, cmd->a1); 176 EATSPACE(); /* EXTENSION */ 177 if (*p == ',') { 178 naddr++; 179 p++; 180 EATSPACE(); /* EXTENSION */ 181 cmd->a2 = xmalloc(sizeof(struct s_addr)); 182 p = compile_addr(p, cmd->a2); 183 } 184 } 185 186 nonsel: /* Now parse the command */ 187 EATSPACE(); 188 if (!*p) 189 err(COMPILE, "command expected"); 190 cmd->code = *p; 191 for (fp = cmd_fmts; fp->code; fp++) 192 if (fp->code == *p) 193 break; 194 if (!fp->code) 195 err(COMPILE, "invalid command code %c", *p); 196 if (naddr > fp->naddr) 197 err(COMPILE, 198 "command %c expects up to %d address(es), found %d", *p, fp->naddr, naddr); 199 switch (fp->args) { 200 case NONSEL: /* ! */ 201 cmd->nonsel = ! cmd->nonsel; 202 p++; 203 goto nonsel; 204 case GROUP: /* { */ 205 p++; 206 EATSPACE(); 207 if (!*p) 208 p = NULL; 209 cmd2 = xmalloc(sizeof(struct s_command)); 210 cmd2->code = '}'; 211 *compile_stream("}", &cmd->u.c, p) = cmd2; 212 cmd->next = cmd2; 213 link = &cmd2->next; 214 break; 215 case EMPTY: /* d D g G h H l n N p P q x = \0 */ 216 p++; 217 EATSPACE(); 218 if (*p == ';') { 219 p++; 220 link = &cmd->next; 221 goto semicolon; 222 } 223 if (*p) 224 err(COMPILE, 225 "extra characters at the end of %c command", cmd->code); 226 break; 227 case TEXT: /* a c i */ 228 p++; 229 EATSPACE(); 230 if (*p != '\\') 231 err(COMPILE, 232 "command %c expects \\ followed by text", cmd->code); 233 p++; 234 EATSPACE(); 235 if (*p) 236 err(COMPILE, 237 "extra characters after \\ at the end of %c command", cmd->code); 238 cmd->t = compile_text(); 239 break; 240 case COMMENT: /* \0 # */ 241 break; 242 case WFILE: /* w */ 243 p++; 244 EATSPACE(); 245 if (*p == '\0') 246 err(COMPILE, "filename expected"); 247 cmd->t = duptoeol(p); 248 if (aflag) 249 cmd->u.fd = -1; 250 else if ((cmd->u.fd = open(p, 251 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, 252 DEFFILEMODE)) == -1) 253 err(FATAL, "%s: %s\n", p, strerror(errno)); 254 break; 255 case RFILE: /* r */ 256 p++; 257 EATSPACE(); 258 if (*p == '\0') 259 err(COMPILE, "filename expected"); 260 else 261 cmd->t = duptoeol(p); 262 break; 263 case BRANCH: /* b t */ 264 p++; 265 EATSPACE(); 266 if (*p == '\0') 267 cmd->t = NULL; 268 else 269 cmd->t = duptoeol(p); 270 break; 271 case LABEL: /* : */ 272 p++; 273 EATSPACE(); 274 cmd->t = duptoeol(p); 275 if (strlen(p) == 0) 276 err(COMPILE, "empty label"); 277 break; 278 case SUBST: /* s */ 279 p++; 280 if (*p == '\0' || *p == '\\') 281 err(COMPILE, 282 "substitute pattern can not be delimited by newline or backslash"); 283 cmd->u.s = xmalloc(sizeof(struct s_subst)); 284 p = compile_re(p, &cmd->u.s->re); 285 if (p == NULL) 286 err(COMPILE, "unterminated substitute pattern"); 287 --p; 288 p = compile_subst(p, cmd->u.s); 289 p = compile_flags(p, cmd->u.s); 290 EATSPACE(); 291 if (*p == ';') { 292 p++; 293 link = &cmd->next; 294 goto semicolon; 295 } 296 break; 297 case TR: /* y */ 298 p++; 299 p = compile_tr(p, (char **)&cmd->u.y); 300 EATSPACE(); 301 if (*p == ';') { 302 p++; 303 link = &cmd->next; 304 goto semicolon; 305 } 306 if (*p) 307 err(COMPILE, 308 "extra text at the end of a transform command"); 309 break; 310 } 311 } 312 } 313 314 /* 315 * Get a delimited string. P points to the delimeter of the string; d points 316 * to a buffer area. Newline and delimiter escapes are processed; other 317 * escapes are ignored. 318 * 319 * Returns a pointer to the first character after the final delimiter or NULL 320 * in the case of a non-terminated string. The character array d is filled 321 * with the processed string. 322 */ 323 static char * 324 compile_delimited(p, d) 325 char *p, *d; 326 { 327 char c; 328 329 c = *p++; 330 if (c == '\0') 331 return (NULL); 332 else if (c == '\\') 333 err(COMPILE, "\\ can not be used as a string delimiter"); 334 else if (c == '\n') 335 err(COMPILE, "newline can not be used as a string delimiter"); 336 while (*p) { 337 if (*p == '[') { 338 if ((d = compile_ccl(&p, d)) == NULL) 339 err(COMPILE, "unbalanced brackets ([])"); 340 continue; 341 } else if (*p == '\\' && p[1] == c) 342 p++; 343 else if (*p == '\\' && p[1] == 'n') { 344 *d++ = '\n'; 345 p += 2; 346 continue; 347 } else if (*p == '\\' && p[1] == '\\') 348 *d++ = *p++; 349 else if (*p == c) { 350 *d = '\0'; 351 return (p + 1); 352 } 353 *d++ = *p++; 354 } 355 return (NULL); 356 } 357 358 359 /* compile_ccl: expand a POSIX character class */ 360 static char * 361 compile_ccl(sp, t) 362 char **sp; 363 char *t; 364 { 365 int c, d; 366 char *s = *sp; 367 368 *t++ = *s++; 369 if (*s == '^') 370 *t++ = *s++; 371 if (*s == ']') 372 *t++ = *s++; 373 for (; *s && (*t = *s) != ']'; s++, t++) 374 if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) { 375 *++t = *++s, t++, s++; 376 for (c = *s; (*t = *s) != ']' || c != d; s++, t++) 377 if ((c = *s) == '\0') 378 return NULL; 379 } else if (*s == '\\' && s[1] == 'n') 380 *t = '\n', s++; 381 return (*s == ']') ? *sp = ++s, ++t : NULL; 382 } 383 384 /* 385 * Get a regular expression. P points to the delimiter of the regular 386 * expression; repp points to the address of a regexp pointer. Newline 387 * and delimiter escapes are processed; other escapes are ignored. 388 * Returns a pointer to the first character after the final delimiter 389 * or NULL in the case of a non terminated regular expression. The regexp 390 * pointer is set to the compiled regular expression. 391 * Cflags are passed to regcomp. 392 */ 393 static char * 394 compile_re(p, repp) 395 char *p; 396 regex_t **repp; 397 { 398 int eval; 399 char re[_POSIX2_LINE_MAX + 1]; 400 401 p = compile_delimited(p, re); 402 if (p && strlen(re) == 0) { 403 *repp = NULL; 404 return (p); 405 } 406 *repp = xmalloc(sizeof(regex_t)); 407 #ifdef GNU_REGEX 408 /* initialize pattern buffer */ 409 (*repp)->buffer = NULL; 410 (*repp)->allocated = 0L; 411 (*repp)->fastmap = 0; /* fastmap not used by regex > 0.12 */ 412 (*repp)->translate = 0; 413 #endif 414 if (p && (eval = regcomp(*repp, re, 0)) != 0) 415 err(COMPILE, "RE error: %s", strregerror(eval, *repp)); 416 if (maxnsub < (*repp)->re_nsub) 417 maxnsub = (*repp)->re_nsub; 418 return (p); 419 } 420 421 /* 422 * Compile the substitution string of a regular expression and set res to 423 * point to a saved copy of it. Nsub is the number of parenthesized regular 424 * expressions. 425 */ 426 static char * 427 compile_subst(p, s) 428 char *p; 429 struct s_subst *s; 430 { 431 static char lbuf[_POSIX2_LINE_MAX + 1]; 432 int asize, ref, size; 433 char c, *text, *op, *sp; 434 435 c = *p++; /* Terminator character */ 436 if (c == '\0') 437 return (NULL); 438 439 s->maxbref = 0; 440 s->linenum = linenum; 441 asize = 2 * _POSIX2_LINE_MAX + 1; 442 text = xmalloc(asize); 443 size = 0; 444 do { 445 op = sp = text + size; 446 for (; *p; p++) { 447 if (*p == '\\') { 448 p++; 449 if (strchr("123456789", *p) != NULL) { 450 *sp++ = '\\'; 451 ref = *p - '0'; 452 if (s->re != NULL && 453 ref > s->re->re_nsub) 454 err(COMPILE, 455 "\\%c not defined in the RE", *p); 456 if (s->maxbref < ref) 457 s->maxbref = ref; 458 } else if (*p == '&' || *p == '\\') 459 *sp++ = '\\'; 460 } else if (*p == c) { 461 p++; 462 *sp++ = '\0'; 463 size += sp - op; 464 s->new = xrealloc(text, size); 465 return (p); 466 } else if (*p == '\n') { 467 err(COMPILE, 468 "unescaped newline inside substitute pattern"); 469 /* NOTREACHED */ 470 } 471 *sp++ = *p; 472 } 473 size += sp - op; 474 if (asize - size < _POSIX2_LINE_MAX + 1) { 475 asize *= 2; 476 text = xmalloc(asize); 477 } 478 } while (cu_fgets(p = lbuf, sizeof(lbuf))); 479 err(COMPILE, "unterminated substitute in regular expression"); 480 /* NOTREACHED */ 481 } 482 483 /* 484 * Compile the flags of the s command 485 */ 486 static char * 487 compile_flags(p, s) 488 char *p; 489 struct s_subst *s; 490 { 491 int gn; /* True if we have seen g or n */ 492 char wfile[_POSIX2_LINE_MAX + 1], *q; 493 494 s->n = 1; /* Default */ 495 s->p = 0; 496 s->wfile = NULL; 497 s->wfd = -1; 498 for (gn = 0;;) { 499 EATSPACE(); /* EXTENSION */ 500 switch (*p) { 501 case 'g': 502 if (gn) 503 err(COMPILE, 504 "more than one number or 'g' in substitute flags"); 505 gn = 1; 506 s->n = 0; 507 break; 508 case '\0': 509 case '\n': 510 case ';': 511 return (p); 512 case 'p': 513 s->p = 1; 514 break; 515 case '1': case '2': case '3': 516 case '4': case '5': case '6': 517 case '7': case '8': case '9': 518 if (gn) 519 err(COMPILE, 520 "more than one number or 'g' in substitute flags"); 521 gn = 1; 522 /* XXX Check for overflow */ 523 s->n = (int)strtol(p, &p, 10); 524 break; 525 case 'w': 526 p++; 527 #ifdef HISTORIC_PRACTICE 528 if (*p != ' ') { 529 err(WARNING, "space missing before w wfile"); 530 return (p); 531 } 532 #endif 533 EATSPACE(); 534 q = wfile; 535 while (*p) { 536 if (*p == '\n') 537 break; 538 *q++ = *p++; 539 } 540 *q = '\0'; 541 if (q == wfile) 542 err(COMPILE, "no wfile specified"); 543 s->wfile = strdup(wfile); 544 if (!aflag && (s->wfd = open(wfile, 545 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, 546 DEFFILEMODE)) == -1) 547 err(FATAL, "%s: %s\n", wfile, strerror(errno)); 548 return (p); 549 default: 550 err(COMPILE, 551 "bad flag in substitute command: '%c'", *p); 552 break; 553 } 554 p++; 555 } 556 } 557 558 /* 559 * Compile a translation set of strings into a lookup table. 560 */ 561 static char * 562 compile_tr(p, transtab) 563 char *p; 564 char **transtab; 565 { 566 int i; 567 char *lt, *op, *np; 568 char old[_POSIX2_LINE_MAX + 1]; 569 char new[_POSIX2_LINE_MAX + 1]; 570 571 if (*p == '\0' || *p == '\\') 572 err(COMPILE, 573 "transform pattern can not be delimited by newline or backslash"); 574 p = compile_delimited(p, old); 575 if (p == NULL) { 576 err(COMPILE, "unterminated transform source string"); 577 return (NULL); 578 } 579 p = compile_delimited(--p, new); 580 if (p == NULL) { 581 err(COMPILE, "unterminated transform target string"); 582 return (NULL); 583 } 584 EATSPACE(); 585 if (strlen(new) != strlen(old)) { 586 err(COMPILE, "transform strings are not the same length"); 587 return (NULL); 588 } 589 /* We assume characters are 8 bits */ 590 lt = xmalloc(UCHAR_MAX); 591 for (i = 0; i <= UCHAR_MAX; i++) 592 lt[i] = (char)i; 593 for (op = old, np = new; *op; op++, np++) 594 lt[(u_char)*op] = *np; 595 *transtab = lt; 596 return (p); 597 } 598 599 /* 600 * Compile the text following an a or i command. 601 */ 602 static char * 603 compile_text() 604 { 605 int asize, size; 606 char *text, *p, *op, *s; 607 char lbuf[_POSIX2_LINE_MAX + 1]; 608 609 asize = 2 * _POSIX2_LINE_MAX + 1; 610 text = xmalloc(asize); 611 size = 0; 612 while (cu_fgets(lbuf, sizeof(lbuf))) { 613 op = s = text + size; 614 p = lbuf; 615 EATSPACE(); 616 for (; *p; p++) { 617 if (*p == '\\') 618 p++; 619 *s++ = *p; 620 } 621 size += s - op; 622 if (p[-2] != '\\') { 623 *s = '\0'; 624 break; 625 } 626 if (asize - size < _POSIX2_LINE_MAX + 1) { 627 asize *= 2; 628 text = xmalloc(asize); 629 } 630 } 631 return (xrealloc(text, size + 1)); 632 } 633 634 /* 635 * Get an address and return a pointer to the first character after 636 * it. Fill the structure pointed to according to the address. 637 */ 638 static char * 639 compile_addr(p, a) 640 char *p; 641 struct s_addr *a; 642 { 643 char *end; 644 645 switch (*p) { 646 case '\\': /* Context address */ 647 ++p; 648 /* FALLTHROUGH */ 649 case '/': /* Context address */ 650 p = compile_re(p, &a->u.r); 651 if (p == NULL) 652 err(COMPILE, "unterminated regular expression"); 653 a->type = AT_RE; 654 return (p); 655 656 case '$': /* Last line */ 657 a->type = AT_LAST; 658 return (p + 1); 659 /* Line number */ 660 case '0': case '1': case '2': case '3': case '4': 661 case '5': case '6': case '7': case '8': case '9': 662 a->type = AT_LINE; 663 a->u.l = strtol(p, &end, 10); 664 return (end); 665 default: 666 err(COMPILE, "expected context address"); 667 return (NULL); 668 } 669 } 670 671 /* 672 * Return a copy of all the characters up to \n or \0 673 */ 674 static char * 675 duptoeol(s) 676 register char *s; 677 { 678 size_t len; 679 char *start; 680 681 for (start = s; *s != '\0' && *s != '\n'; ++s); 682 *s = '\0'; 683 len = s - start + 1; 684 return (memmove(xmalloc(len), start, len)); 685 } 686 687 /* 688 * Find the label contained in the command l in the command linked list cp. 689 * L is excluded from the search. Return NULL if not found. 690 */ 691 static struct s_command * 692 findlabel(l, cp) 693 struct s_command *l, *cp; 694 { 695 struct s_command *r; 696 697 for (; cp; cp = cp->next) 698 if (cp->code == ':' && cp != l && strcmp(l->t, cp->t) == 0) 699 return (cp); 700 else if (cp->code == '{' && (r = findlabel(l, cp->u.c))) 701 return (r); 702 return (NULL); 703 } 704 705 /* 706 * Convert goto label names to addresses. 707 * Detect duplicate labels. 708 * Set appendnum to the number of a and r commands in the script. 709 * Free the memory used by labels in b and t commands (but not by :) 710 * Root is a pointer to the script linked list; cp points to the 711 * search start. 712 * TODO: Remove } nodes 713 */ 714 static void 715 fixuplabel(root, cp, end) 716 struct s_command *root, *cp, *end; 717 { 718 struct s_command *cp2; 719 720 for (; cp != end; cp = cp->next) 721 switch (cp->code) { 722 case ':': 723 if (findlabel(cp, root)) 724 err(COMPILE2, "duplicate label %s", cp->t); 725 break; 726 case 'a': 727 case 'r': 728 appendnum++; 729 break; 730 case 'b': 731 case 't': 732 if (cp->t == NULL) { 733 cp->u.c = NULL; 734 break; 735 } 736 if ((cp2 = findlabel(cp, root)) == NULL) 737 err(COMPILE2, "undefined label '%s'", cp->t); 738 free(cp->t); 739 cp->u.c = cp2; 740 break; 741 case '{': 742 fixuplabel(root, cp->u.c, cp->next); 743 break; 744 } 745 } 746