1 /* $NetBSD: compile.c,v 1.39 2013/06/28 15:04:35 joerg Exp $ */ 2 3 /*- 4 * Copyright (c) 1992, 1993 5 * The Regents of the University of California. All rights reserved. 6 * 7 * This code is derived from software contributed to Berkeley by 8 * Diomidis Spinellis of Imperial College, University of London. 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 3. Neither the name of the University nor the names of its contributors 19 * may be used to endorse or promote products derived from this software 20 * without specific prior written permission. 21 * 22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 25 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 32 * SUCH DAMAGE. 33 */ 34 35 /*- 36 * Copyright (c) 1992 Diomidis Spinellis. 37 * 38 * This code is derived from software contributed to Berkeley by 39 * Diomidis Spinellis of Imperial College, University of London. 40 * 41 * Redistribution and use in source and binary forms, with or without 42 * modification, are permitted provided that the following conditions 43 * are met: 44 * 1. Redistributions of source code must retain the above copyright 45 * notice, this list of conditions and the following disclaimer. 46 * 2. Redistributions in binary form must reproduce the above copyright 47 * notice, this list of conditions and the following disclaimer in the 48 * documentation and/or other materials provided with the distribution. 49 * 3. All advertising materials mentioning features or use of this software 50 * must display the following acknowledgement: 51 * This product includes software developed by the University of 52 * California, Berkeley and its contributors. 53 * 4. Neither the name of the University nor the names of its contributors 54 * may be used to endorse or promote products derived from this software 55 * without specific prior written permission. 56 * 57 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 58 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 59 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 60 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 61 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 62 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 63 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 64 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 65 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 66 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 67 * SUCH DAMAGE. 68 */ 69 70 #if HAVE_NBTOOL_CONFIG_H 71 #include "nbtool_config.h" 72 #endif 73 74 #include <sys/cdefs.h> 75 #ifndef lint 76 #if 0 77 static char sccsid[] = "@(#)compile.c 8.2 (Berkeley) 4/28/95"; 78 #else 79 __RCSID("$NetBSD: compile.c,v 1.39 2013/06/28 15:04:35 joerg Exp $"); 80 #endif 81 #endif /* not lint */ 82 83 #include <sys/types.h> 84 #include <sys/stat.h> 85 86 #include <ctype.h> 87 #include <errno.h> 88 #include <fcntl.h> 89 #include <limits.h> 90 #include <regex.h> 91 #include <stdio.h> 92 #include <stdlib.h> 93 #include <string.h> 94 95 #include "defs.h" 96 #include "extern.h" 97 98 #ifndef _POSIX2_LINE_MAX 99 #define _POSIX2_LINE_MAX (2 * BUFSIZ) 100 #endif 101 102 #define LHSZ 128 103 #define LHMASK (LHSZ - 1) 104 static struct labhash { 105 struct labhash *lh_next; 106 u_int lh_hash; 107 struct s_command *lh_cmd; 108 int lh_ref; 109 } *labels[LHSZ]; 110 111 static char *compile_addr(char *, struct s_addr *); 112 static char *compile_ccl(char **, char *); 113 static char *compile_delimited(char *, char *); 114 static char *compile_flags(char *, struct s_subst *); 115 static char *compile_re(char *, regex_t **); 116 static char *compile_subst(char *, struct s_subst *); 117 static char *compile_text(void); 118 static char *compile_tr(char *, char **); 119 static struct s_command 120 **compile_stream(struct s_command **); 121 static char *duptoeol(char *, const char *); 122 static void enterlabel(struct s_command *); 123 static struct s_command 124 *findlabel(char *); 125 static void fixuplabel(struct s_command *, struct s_command *); 126 static void uselabel(void); 127 128 /* 129 * Command specification. This is used to drive the command parser. 130 */ 131 struct s_format { 132 char code; /* Command code */ 133 int naddr; /* Number of address args */ 134 enum e_args args; /* Argument type */ 135 }; 136 137 static struct s_format cmd_fmts[] = { 138 {'{', 2, GROUP}, 139 {'}', 0, ENDGROUP}, 140 {'a', 1, TEXT}, 141 {'b', 2, BRANCH}, 142 {'c', 2, TEXT}, 143 {'d', 2, EMPTY}, 144 {'D', 2, EMPTY}, 145 {'g', 2, EMPTY}, 146 {'G', 2, EMPTY}, 147 {'h', 2, EMPTY}, 148 {'H', 2, EMPTY}, 149 {'i', 1, TEXT}, 150 {'l', 2, EMPTY}, 151 {'n', 2, EMPTY}, 152 {'N', 2, EMPTY}, 153 {'p', 2, EMPTY}, 154 {'P', 2, EMPTY}, 155 {'q', 1, EMPTY}, 156 {'r', 1, RFILE}, 157 {'s', 2, SUBST}, 158 {'t', 2, BRANCH}, 159 {'w', 2, WFILE}, 160 {'x', 2, EMPTY}, 161 {'y', 2, TR}, 162 {'!', 2, NONSEL}, 163 {':', 0, LABEL}, 164 {'#', 0, COMMENT}, 165 {'=', 1, EMPTY}, 166 {'\0', 0, COMMENT}, 167 }; 168 169 /* The compiled program. */ 170 struct s_command *prog; 171 172 /* 173 * Compile the program into prog. 174 * Initialise appends. 175 */ 176 void 177 compile(void) 178 { 179 *compile_stream(&prog) = NULL; 180 fixuplabel(prog, NULL); 181 uselabel(); 182 if (appendnum > 0) 183 appends = xmalloc(sizeof(struct s_appends) * appendnum); 184 match = xmalloc((maxnsub + 1) * sizeof(regmatch_t)); 185 } 186 187 #define EATSPACE() \ 188 while (*p && isascii((unsigned char)*p) && \ 189 isspace((unsigned char)*p)) \ 190 p++ \ 191 192 static struct s_command ** 193 compile_stream(struct s_command **link) 194 { 195 char *p; 196 static char *lbuf; /* To avoid excessive malloc calls */ 197 static size_t bufsize; 198 struct s_command *cmd, *cmd2, *stack; 199 struct s_format *fp; 200 int naddr; /* Number of addresses */ 201 202 stack = 0; 203 for (;;) { 204 if ((p = cu_fgets(&lbuf, &bufsize)) == NULL) { 205 if (stack != 0) 206 err(COMPILE, "unexpected EOF (pending }'s)"); 207 return (link); 208 } 209 210 semicolon: EATSPACE(); 211 if (*p == '#' || *p == '\0') 212 continue; 213 else if (*p == ';') { 214 p++; 215 goto semicolon; 216 } 217 *link = cmd = xmalloc(sizeof(struct s_command)); 218 link = &cmd->next; 219 cmd->nonsel = cmd->inrange = 0; 220 /* First parse the addresses */ 221 naddr = 0; 222 223 /* Valid characters to start an address */ 224 #define addrchar(c) (strchr("0123456789/\\$", (c))) 225 if (addrchar(*p)) { 226 naddr++; 227 cmd->a1 = xmalloc(sizeof(struct s_addr)); 228 p = compile_addr(p, cmd->a1); 229 EATSPACE(); /* EXTENSION */ 230 if (*p == ',') { 231 p++; 232 EATSPACE(); /* EXTENSION */ 233 naddr++; 234 cmd->a2 = xmalloc(sizeof(struct s_addr)); 235 p = compile_addr(p, cmd->a2); 236 EATSPACE(); 237 } else 238 cmd->a2 = 0; 239 } else 240 cmd->a1 = cmd->a2 = 0; 241 242 nonsel: /* Now parse the command */ 243 if (!*p) 244 err(COMPILE, "command expected"); 245 cmd->code = *p; 246 for (fp = cmd_fmts; fp->code; fp++) 247 if (fp->code == *p) 248 break; 249 if (!fp->code) 250 err(COMPILE, "invalid command code %c", *p); 251 if (naddr > fp->naddr) 252 err(COMPILE, 253 "command %c expects up to %d address(es), found %d", *p, fp->naddr, naddr); 254 switch (fp->args) { 255 case NONSEL: /* ! */ 256 p++; 257 EATSPACE(); 258 cmd->nonsel = ! cmd->nonsel; 259 goto nonsel; 260 case GROUP: /* { */ 261 p++; 262 EATSPACE(); 263 cmd->next = stack; 264 stack = cmd; 265 link = &cmd->u.c; 266 if (*p) 267 goto semicolon; 268 break; 269 case ENDGROUP: 270 /* 271 * Short-circuit command processing, since end of 272 * group is really just a noop. 273 */ 274 cmd->nonsel = 1; 275 if (stack == 0) 276 err(COMPILE, "unexpected }"); 277 cmd2 = stack; 278 stack = cmd2->next; 279 cmd2->next = cmd; 280 /*FALLTHROUGH*/ 281 case EMPTY: /* d D g G h H l n N p P q x = \0 */ 282 p++; 283 EATSPACE(); 284 if (*p == ';') { 285 p++; 286 link = &cmd->next; 287 goto semicolon; 288 } 289 if (*p) 290 err(COMPILE, 291 "extra characters at the end of %c command", cmd->code); 292 break; 293 case TEXT: /* a c i */ 294 p++; 295 EATSPACE(); 296 if (*p != '\\') 297 err(COMPILE, 298 "command %c expects \\ followed by text", cmd->code); 299 p++; 300 EATSPACE(); 301 if (*p) 302 err(COMPILE, 303 "extra characters after \\ at the end of %c command", cmd->code); 304 cmd->t = compile_text(); 305 break; 306 case COMMENT: /* \0 # */ 307 break; 308 case WFILE: /* w */ 309 p++; 310 EATSPACE(); 311 if (*p == '\0') 312 err(COMPILE, "filename expected"); 313 cmd->t = duptoeol(p, "w command"); 314 if (aflag) 315 cmd->u.fd = -1; 316 else if ((cmd->u.fd = open(p, 317 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, 318 DEFFILEMODE)) == -1) 319 err(FATAL, "%s: %s", p, strerror(errno)); 320 break; 321 case RFILE: /* r */ 322 p++; 323 EATSPACE(); 324 if (*p == '\0') 325 err(COMPILE, "filename expected"); 326 else 327 cmd->t = duptoeol(p, "read command"); 328 break; 329 case BRANCH: /* b t */ 330 p++; 331 EATSPACE(); 332 if (*p == '\0') 333 cmd->t = NULL; 334 else 335 cmd->t = duptoeol(p, "branch"); 336 break; 337 case LABEL: /* : */ 338 p++; 339 EATSPACE(); 340 cmd->t = duptoeol(p, "label"); 341 if (strlen(p) == 0) 342 err(COMPILE, "empty label"); 343 enterlabel(cmd); 344 break; 345 case SUBST: /* s */ 346 p++; 347 if (*p == '\0' || *p == '\\') 348 err(COMPILE, 349 "substitute pattern can not be delimited by newline or backslash"); 350 cmd->u.s = xmalloc(sizeof(struct s_subst)); 351 p = compile_re(p, &cmd->u.s->re); 352 if (p == NULL) 353 err(COMPILE, "unterminated substitute pattern"); 354 --p; 355 p = compile_subst(p, cmd->u.s); 356 p = compile_flags(p, cmd->u.s); 357 EATSPACE(); 358 if (*p == ';') { 359 p++; 360 link = &cmd->next; 361 goto semicolon; 362 } 363 break; 364 case TR: /* y */ 365 p++; 366 p = compile_tr(p, (char **)(void *)&cmd->u.y); 367 EATSPACE(); 368 if (*p == ';') { 369 p++; 370 link = &cmd->next; 371 goto semicolon; 372 } 373 if (*p) 374 err(COMPILE, 375 "extra text at the end of a transform command"); 376 break; 377 } 378 } 379 } 380 381 /* 382 * Get a delimited string. P points to the delimiter of the string; d points 383 * to a buffer area. Newline and delimiter escapes are processed; other 384 * escapes are ignored. 385 * 386 * Returns a pointer to the first character after the final delimiter or NULL 387 * in the case of a non-terminated string. The character array d is filled 388 * with the processed string. 389 */ 390 static char * 391 compile_delimited(char *p, char *d) 392 { 393 char c; 394 395 c = *p++; 396 if (c == '\0') 397 return (NULL); 398 else if (c == '\\') 399 err(COMPILE, "\\ can not be used as a string delimiter"); 400 else if (c == '\n') 401 err(COMPILE, "newline can not be used as a string delimiter"); 402 while (*p) { 403 if (*p == '[') { 404 if ((d = compile_ccl(&p, d)) == NULL) 405 err(COMPILE, "unbalanced brackets ([])"); 406 continue; 407 } else if (*p == '\\' && p[1] == '[') { 408 *d++ = *p++; 409 } else if (*p == '\\' && p[1] == c) 410 p++; 411 else if (*p == '\\' && p[1] == 'n') { 412 *d++ = '\n'; 413 p += 2; 414 continue; 415 } else if (*p == '\\' && p[1] == '\\') 416 *d++ = *p++; 417 else if (*p == c) { 418 *d = '\0'; 419 return (p + 1); 420 } 421 *d++ = *p++; 422 } 423 return (NULL); 424 } 425 426 427 /* compile_ccl: expand a POSIX character class */ 428 static char * 429 compile_ccl(char **sp, char *t) 430 { 431 int c, d; 432 char *s = *sp; 433 434 *t++ = *s++; 435 if (*s == '^') 436 *t++ = *s++; 437 if (*s == ']') 438 *t++ = *s++; 439 for (; *s && (*t = *s) != ']'; s++, t++) 440 if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) { 441 *++t = *++s, t++, s++; 442 for (c = *s; (*t = *s) != ']' || c != d; s++, t++) 443 if ((c = *s) == '\0') 444 return NULL; 445 } else if (*s == '\\' && s[1] == 'n') 446 *t = '\n', s++; 447 return (*s == ']') ? *sp = ++s, ++t : NULL; 448 } 449 450 /* 451 * Get a regular expression. P points to the delimiter of the regular 452 * expression; repp points to the address of a regexp pointer. Newline 453 * and delimiter escapes are processed; other escapes are ignored. 454 * Returns a pointer to the first character after the final delimiter 455 * or NULL in the case of a non terminated regular expression. The regexp 456 * pointer is set to the compiled regular expression. 457 * Cflags are passed to regcomp. 458 */ 459 static char * 460 compile_re(char *p, regex_t **repp) 461 { 462 int eval; 463 char *re; 464 465 re = xmalloc(strlen(p) + 1); /* strlen(re) <= strlen(p) */ 466 p = compile_delimited(p, re); 467 if (p && strlen(re) == 0) { 468 *repp = NULL; 469 free(re); 470 return (p); 471 } 472 *repp = xmalloc(sizeof(regex_t)); 473 if (p && (eval = regcomp(*repp, re, ere)) != 0) 474 err(COMPILE, "RE error: %s", strregerror(eval, *repp)); 475 if (maxnsub < (*repp)->re_nsub) 476 maxnsub = (*repp)->re_nsub; 477 free(re); 478 return (p); 479 } 480 481 /* 482 * Compile the substitution string of a regular expression and set res to 483 * point to a saved copy of it. Nsub is the number of parenthesized regular 484 * expressions. 485 */ 486 static char * 487 compile_subst(char *p, struct s_subst *s) 488 { 489 static char *lbuf; 490 static size_t bufsize; 491 int asize, ref, size, len; 492 char c, *text, *op, *sp; 493 int sawesc = 0; 494 495 c = *p++; /* Terminator character */ 496 if (c == '\0') 497 return (NULL); 498 499 s->maxbref = 0; 500 s->linenum = linenum; 501 text = NULL; 502 asize = size = 0; 503 do { 504 len = ROUNDLEN(strlen(p) + 1); 505 if (asize - size < len) { 506 do { 507 asize += len; 508 } while (asize - size < len); 509 text = xrealloc(text, asize); 510 } 511 op = sp = text + size; 512 for (; *p; p++) { 513 if (*p == '\\' || sawesc) { 514 /* 515 * If this is a continuation from the last 516 * buffer, we won't have a character to 517 * skip over. 518 */ 519 if (sawesc) 520 sawesc = 0; 521 else 522 p++; 523 524 if (*p == '\0') { 525 /* 526 * This escaped character is continued 527 * in the next part of the line. Note 528 * this fact, then cause the loop to 529 * exit w/ normal EOL case and reenter 530 * above with the new buffer. 531 */ 532 sawesc = 1; 533 p--; 534 continue; 535 } else if (strchr("123456789", *p) != NULL) { 536 *sp++ = '\\'; 537 ref = *p - '0'; 538 if (s->re != NULL && 539 (size_t)ref > s->re->re_nsub) 540 err(COMPILE, 541 "\\%c not defined in the RE", *p); 542 if (s->maxbref < ref) 543 s->maxbref = ref; 544 } else if (*p == '&' || *p == '\\') 545 *sp++ = '\\'; 546 } else if (*p == c) { 547 p++; 548 *sp++ = '\0'; 549 size += sp - op; 550 s->new = xrealloc(text, size); 551 return (p); 552 } else if (*p == '\n') { 553 err(COMPILE, 554 "unescaped newline inside substitute pattern"); 555 /* NOTREACHED */ 556 } 557 *sp++ = *p; 558 } 559 size += sp - op; 560 } while ((p = cu_fgets(&lbuf, &bufsize))); 561 err(COMPILE, "unterminated substitute in regular expression"); 562 /* NOTREACHED */ 563 return (NULL); 564 } 565 566 /* 567 * Compile the flags of the s command 568 */ 569 static char * 570 compile_flags(char *p, struct s_subst *s) 571 { 572 int gn; /* True if we have seen g or n */ 573 char wfile[PATH_MAX], *q; 574 575 s->n = 1; /* Default */ 576 s->p = 0; 577 s->wfile = NULL; 578 s->wfd = -1; 579 for (gn = 0;;) { 580 EATSPACE(); /* EXTENSION */ 581 switch (*p) { 582 case 'g': 583 if (gn) 584 err(COMPILE, 585 "more than one number or 'g' in substitute flags"); 586 gn = 1; 587 s->n = 0; 588 break; 589 case '\0': 590 case '\n': 591 case ';': 592 return (p); 593 case 'p': 594 s->p = 1; 595 break; 596 case '1': case '2': case '3': 597 case '4': case '5': case '6': 598 case '7': case '8': case '9': 599 if (gn) 600 err(COMPILE, 601 "more than one number or 'g' in substitute flags"); 602 gn = 1; 603 /* XXX Check for overflow */ 604 s->n = (int)strtol(p, &p, 10); 605 p--; 606 break; 607 case 'w': 608 p++; 609 #ifdef HISTORIC_PRACTICE 610 if (*p != ' ') { 611 err(WARNING, "space missing before w wfile"); 612 return (p); 613 } 614 #endif 615 EATSPACE(); 616 q = wfile; 617 while (*p) { 618 if (*p == '\n') 619 break; 620 *q++ = *p++; 621 } 622 *q = '\0'; 623 if (q == wfile) 624 err(COMPILE, "no wfile specified"); 625 s->wfile = strdup(wfile); 626 if (!aflag && (s->wfd = open(wfile, 627 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, 628 DEFFILEMODE)) == -1) 629 err(FATAL, "%s: %s", wfile, strerror(errno)); 630 return (p); 631 default: 632 err(COMPILE, 633 "bad flag in substitute command: '%c'", *p); 634 break; 635 } 636 p++; 637 } 638 } 639 640 /* 641 * Compile a translation set of strings into a lookup table. 642 */ 643 static char * 644 compile_tr(char *p, char **transtab) 645 { 646 int i; 647 char *lt, *op, *np; 648 char *old = NULL, *new = NULL; 649 650 if (*p == '\0' || *p == '\\') 651 err(COMPILE, 652 "transform pattern can not be delimited by newline or backslash"); 653 old = xmalloc(strlen(p) + 1); 654 p = compile_delimited(p, old); 655 if (p == NULL) { 656 err(COMPILE, "unterminated transform source string"); 657 goto bad; 658 } 659 new = xmalloc(strlen(p) + 1); 660 p = compile_delimited(p - 1, new); 661 if (p == NULL) { 662 err(COMPILE, "unterminated transform target string"); 663 goto bad; 664 } 665 EATSPACE(); 666 if (strlen(new) != strlen(old)) { 667 err(COMPILE, "transform strings are not the same length"); 668 goto bad; 669 } 670 /* We assume characters are 8 bits */ 671 lt = xmalloc(UCHAR_MAX+1); 672 for (i = 0; i <= UCHAR_MAX; i++) 673 lt[i] = (char)i; 674 for (op = old, np = new; *op; op++, np++) 675 lt[(u_char)*op] = *np; 676 *transtab = lt; 677 free(old); 678 free(new); 679 return (p); 680 bad: 681 free(old); 682 free(new); 683 return (NULL); 684 } 685 686 /* 687 * Compile the text following an a, c, or i command. 688 */ 689 static char * 690 compile_text(void) 691 { 692 int asize, size, len; 693 char *lbuf, *text, *p, *op, *s; 694 size_t bufsize; 695 696 lbuf = text = NULL; 697 asize = size = 0; 698 while ((p = cu_fgets(&lbuf, &bufsize))) { 699 len = ROUNDLEN(strlen(p) + 1); 700 if (asize - size < len) { 701 do { 702 asize += len; 703 } while (asize - size < len); 704 text = xrealloc(text, asize); 705 } 706 op = s = text + size; 707 for (; *p; p++) { 708 if (*p == '\\') 709 p++; 710 *s++ = *p; 711 } 712 size += s - op; 713 if (p[-2] != '\\') { 714 *s = '\0'; 715 break; 716 } 717 } 718 free(lbuf); 719 return (xrealloc(text, size + 1)); 720 } 721 722 /* 723 * Get an address and return a pointer to the first character after 724 * it. Fill the structure pointed to according to the address. 725 */ 726 static char * 727 compile_addr(char *p, struct s_addr *a) 728 { 729 char *end; 730 731 switch (*p) { 732 case '\\': /* Context address */ 733 ++p; 734 /* FALLTHROUGH */ 735 case '/': /* Context address */ 736 p = compile_re(p, &a->u.r); 737 if (p == NULL) 738 err(COMPILE, "unterminated regular expression"); 739 a->type = AT_RE; 740 return (p); 741 742 case '$': /* Last line */ 743 a->type = AT_LAST; 744 return (p + 1); 745 /* Line number */ 746 case '0': case '1': case '2': case '3': case '4': 747 case '5': case '6': case '7': case '8': case '9': 748 a->type = AT_LINE; 749 a->u.l = strtol(p, &end, 10); 750 return (end); 751 default: 752 err(COMPILE, "expected context address"); 753 return (NULL); 754 } 755 } 756 757 /* 758 * duptoeol -- 759 * Return a copy of all the characters up to \n or \0. 760 */ 761 static char * 762 duptoeol(char *s, const char *ctype) 763 { 764 size_t len; 765 int ws; 766 char *start; 767 768 ws = 0; 769 for (start = s; *s != '\0' && *s != '\n'; ++s) 770 ws = isspace((unsigned char)*s); 771 *s = '\0'; 772 if (ws) 773 err(WARNING, "whitespace after %s", ctype); 774 len = s - start + 1; 775 return (memmove(xmalloc(len), start, len)); 776 } 777 778 /* 779 * Convert goto label names to addresses, and count a and r commands, in 780 * the given subset of the script. Free the memory used by labels in b 781 * and t commands (but not by :). 782 * 783 * TODO: Remove } nodes 784 */ 785 static void 786 fixuplabel(struct s_command *cp, struct s_command *end) 787 { 788 789 for (; cp != end; cp = cp->next) 790 switch (cp->code) { 791 case 'a': 792 case 'r': 793 appendnum++; 794 break; 795 case 'b': 796 case 't': 797 /* Resolve branch target. */ 798 if (cp->t == NULL) { 799 cp->u.c = NULL; 800 break; 801 } 802 if ((cp->u.c = findlabel(cp->t)) == NULL) 803 err(COMPILE2, "undefined label '%s'", cp->t); 804 free(cp->t); 805 break; 806 case '{': 807 /* Do interior commands. */ 808 fixuplabel(cp->u.c, cp->next); 809 break; 810 } 811 } 812 813 /* 814 * Associate the given command label for later lookup. 815 */ 816 static void 817 enterlabel(struct s_command *cp) 818 { 819 struct labhash **lhp, *lh; 820 u_char *p; 821 u_int h, c; 822 823 for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++) 824 h = (h << 5) + h + c; 825 lhp = &labels[h & LHMASK]; 826 for (lh = *lhp; lh != NULL; lh = lh->lh_next) 827 if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0) 828 err(COMPILE2, "duplicate label '%s'", cp->t); 829 lh = xmalloc(sizeof *lh); 830 lh->lh_next = *lhp; 831 lh->lh_hash = h; 832 lh->lh_cmd = cp; 833 lh->lh_ref = 0; 834 *lhp = lh; 835 } 836 837 /* 838 * Find the label contained in the command l in the command linked 839 * list cp. L is excluded from the search. Return NULL if not found. 840 */ 841 static struct s_command * 842 findlabel(char *name) 843 { 844 struct labhash *lh; 845 u_char *p; 846 u_int h, c; 847 848 for (h = 0, p = (u_char *)name; (c = *p) != 0; p++) 849 h = (h << 5) + h + c; 850 for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) { 851 if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) { 852 lh->lh_ref = 1; 853 return (lh->lh_cmd); 854 } 855 } 856 return (NULL); 857 } 858 859 /* 860 * Warn about any unused labels. As a side effect, release the label hash 861 * table space. 862 */ 863 static void 864 uselabel(void) 865 { 866 struct labhash *lh, *next; 867 int i; 868 869 for (i = 0; i < LHSZ; i++) { 870 for (lh = labels[i]; lh != NULL; lh = next) { 871 next = lh->lh_next; 872 if (!lh->lh_ref) 873 err(WARNING, "unused label '%s'", 874 lh->lh_cmd->t); 875 free(lh); 876 } 877 } 878 } 879