1 /*- 2 * Copyright (c) 1992 Diomidis Spinellis. 3 * Copyright (c) 1992, 1993 4 * The Regents of the University of California. 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 #include <sys/cdefs.h> 39 __FBSDID("$FreeBSD$"); 40 41 #ifndef lint 42 static const char sccsid[] = "@(#)compile.c 8.1 (Berkeley) 6/6/93"; 43 #endif 44 45 #include <sys/types.h> 46 #include <sys/stat.h> 47 48 #include <ctype.h> 49 #include <err.h> 50 #include <fcntl.h> 51 #include <limits.h> 52 #include <regex.h> 53 #include <stdio.h> 54 #include <stdlib.h> 55 #include <string.h> 56 57 #include "defs.h" 58 #include "extern.h" 59 60 #define LHSZ 128 61 #define LHMASK (LHSZ - 1) 62 static struct labhash { 63 struct labhash *lh_next; 64 u_int lh_hash; 65 struct s_command *lh_cmd; 66 int lh_ref; 67 } *labels[LHSZ]; 68 69 static char *compile_addr __P((char *, struct s_addr *)); 70 static char *compile_ccl __P((char **, char *)); 71 static char *compile_delimited __P((char *, char *)); 72 static char *compile_flags __P((char *, struct s_subst *)); 73 static char *compile_re __P((char *, regex_t **)); 74 static char *compile_subst __P((char *, struct s_subst *)); 75 static char *compile_text __P((void)); 76 static char *compile_tr __P((char *, char **)); 77 static struct s_command 78 **compile_stream __P((struct s_command **)); 79 static char *duptoeol __P((char *, const char *)); 80 static void enterlabel __P((struct s_command *)); 81 static struct s_command 82 *findlabel __P((char *)); 83 static void fixuplabel __P((struct s_command *, struct s_command *)); 84 static void uselabel __P((void)); 85 86 /* 87 * Command specification. This is used to drive the command parser. 88 */ 89 struct s_format { 90 char code; /* Command code */ 91 int naddr; /* Number of address args */ 92 enum e_args args; /* Argument type */ 93 }; 94 95 static struct s_format cmd_fmts[] = { 96 {'{', 2, GROUP}, 97 {'}', 0, ENDGROUP}, 98 {'a', 1, TEXT}, 99 {'b', 2, BRANCH}, 100 {'c', 2, TEXT}, 101 {'d', 2, EMPTY}, 102 {'D', 2, EMPTY}, 103 {'g', 2, EMPTY}, 104 {'G', 2, EMPTY}, 105 {'h', 2, EMPTY}, 106 {'H', 2, EMPTY}, 107 {'i', 1, TEXT}, 108 {'l', 2, EMPTY}, 109 {'n', 2, EMPTY}, 110 {'N', 2, EMPTY}, 111 {'p', 2, EMPTY}, 112 {'P', 2, EMPTY}, 113 {'q', 1, EMPTY}, 114 {'r', 1, RFILE}, 115 {'s', 2, SUBST}, 116 {'t', 2, BRANCH}, 117 {'w', 2, WFILE}, 118 {'x', 2, EMPTY}, 119 {'y', 2, TR}, 120 {'!', 2, NONSEL}, 121 {':', 0, LABEL}, 122 {'#', 0, COMMENT}, 123 {'=', 1, EMPTY}, 124 {'\0', 0, COMMENT}, 125 }; 126 127 /* The compiled program. */ 128 struct s_command *prog; 129 130 /* 131 * Compile the program into prog. 132 * Initialise appends. 133 */ 134 void 135 compile() 136 { 137 *compile_stream(&prog) = NULL; 138 fixuplabel(prog, NULL); 139 uselabel(); 140 if (appendnum == 0) 141 appends = NULL; 142 else if ((appends = malloc(sizeof(struct s_appends) * appendnum)) == 143 NULL) 144 err(1, "malloc"); 145 if ((match = malloc((maxnsub + 1) * sizeof(regmatch_t))) == NULL) 146 err(1, "malloc"); 147 } 148 149 #define EATSPACE() do { \ 150 if (p) \ 151 while (*p && isspace((unsigned char)*p)) \ 152 p++; \ 153 } while (0) 154 155 static struct s_command ** 156 compile_stream(link) 157 struct s_command **link; 158 { 159 char *p; 160 static char lbuf[_POSIX2_LINE_MAX + 1]; /* To save stack */ 161 struct s_command *cmd, *cmd2, *stack; 162 struct s_format *fp; 163 int naddr; /* Number of addresses */ 164 165 stack = 0; 166 for (;;) { 167 if ((p = cu_fgets(lbuf, sizeof(lbuf), NULL)) == NULL) { 168 if (stack != 0) 169 errx(1, "%lu: %s: unexpected EOF (pending }'s)", 170 linenum, fname); 171 return (link); 172 } 173 174 semicolon: EATSPACE(); 175 if (p && (*p == '#' || *p == '\0')) 176 continue; 177 if ((*link = cmd = malloc(sizeof(struct s_command))) == NULL) 178 err(1, "malloc"); 179 link = &cmd->next; 180 cmd->nonsel = cmd->inrange = 0; 181 /* First parse the addresses */ 182 naddr = 0; 183 184 /* Valid characters to start an address */ 185 #define addrchar(c) (strchr("0123456789/\\$", (c))) 186 if (addrchar(*p)) { 187 naddr++; 188 if ((cmd->a1 = malloc(sizeof(struct s_addr))) == NULL) 189 err(1, "malloc"); 190 p = compile_addr(p, cmd->a1); 191 EATSPACE(); /* EXTENSION */ 192 if (*p == ',') { 193 p++; 194 EATSPACE(); /* EXTENSION */ 195 naddr++; 196 if ((cmd->a2 = malloc(sizeof(struct s_addr))) 197 == NULL) 198 err(1, "malloc"); 199 p = compile_addr(p, cmd->a2); 200 EATSPACE(); 201 } else 202 cmd->a2 = 0; 203 } else 204 cmd->a1 = cmd->a2 = 0; 205 206 nonsel: /* Now parse the command */ 207 if (!*p) 208 errx(1, "%lu: %s: command expected", linenum, fname); 209 cmd->code = *p; 210 for (fp = cmd_fmts; fp->code; fp++) 211 if (fp->code == *p) 212 break; 213 if (!fp->code) 214 errx(1, "%lu: %s: invalid command code %c", linenum, fname, *p); 215 if (naddr > fp->naddr) 216 errx(1, 217 "%lu: %s: command %c expects up to %d address(es), found %d", 218 linenum, fname, *p, fp->naddr, naddr); 219 switch (fp->args) { 220 case NONSEL: /* ! */ 221 p++; 222 EATSPACE(); 223 cmd->nonsel = ! cmd->nonsel; 224 goto nonsel; 225 case GROUP: /* { */ 226 p++; 227 EATSPACE(); 228 cmd->next = stack; 229 stack = cmd; 230 link = &cmd->u.c; 231 if (*p) 232 goto semicolon; 233 break; 234 case ENDGROUP: 235 /* 236 * Short-circuit command processing, since end of 237 * group is really just a noop. 238 */ 239 cmd->nonsel = 1; 240 if (stack == 0) 241 errx(1, "%lu: %s: unexpected }", linenum, fname); 242 cmd2 = stack; 243 stack = cmd2->next; 244 cmd2->next = cmd; 245 /*FALLTHROUGH*/ 246 case EMPTY: /* d D g G h H l n N p P q x = \0 */ 247 p++; 248 EATSPACE(); 249 if (*p == ';') { 250 p++; 251 link = &cmd->next; 252 goto semicolon; 253 } 254 if (*p) 255 errx(1, "%lu: %s: extra characters at the end of %c command", 256 linenum, fname, cmd->code); 257 break; 258 case TEXT: /* a c i */ 259 p++; 260 EATSPACE(); 261 if (*p != '\\') 262 errx(1, 263 "%lu: %s: command %c expects \\ followed by text", linenum, fname, cmd->code); 264 p++; 265 EATSPACE(); 266 if (*p) 267 errx(1, 268 "%lu: %s: extra characters after \\ at the end of %c command", 269 linenum, fname, cmd->code); 270 cmd->t = compile_text(); 271 break; 272 case COMMENT: /* \0 # */ 273 break; 274 case WFILE: /* w */ 275 p++; 276 EATSPACE(); 277 if (*p == '\0') 278 errx(1, "%lu: %s: filename expected", linenum, fname); 279 cmd->t = duptoeol(p, "w command"); 280 if (aflag) 281 cmd->u.fd = -1; 282 else if ((cmd->u.fd = open(p, 283 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, 284 DEFFILEMODE)) == -1) 285 err(1, "%s", p); 286 break; 287 case RFILE: /* r */ 288 p++; 289 EATSPACE(); 290 if (*p == '\0') 291 errx(1, "%lu: %s: filename expected", linenum, fname); 292 else 293 cmd->t = duptoeol(p, "read command"); 294 break; 295 case BRANCH: /* b t */ 296 p++; 297 EATSPACE(); 298 if (*p == '\0') 299 cmd->t = NULL; 300 else 301 cmd->t = duptoeol(p, "branch"); 302 break; 303 case LABEL: /* : */ 304 p++; 305 EATSPACE(); 306 cmd->t = duptoeol(p, "label"); 307 if (strlen(p) == 0) 308 errx(1, "%lu: %s: empty label", linenum, fname); 309 enterlabel(cmd); 310 break; 311 case SUBST: /* s */ 312 p++; 313 if (*p == '\0' || *p == '\\') 314 errx(1, 315 "%lu: %s: substitute pattern can not be delimited by newline or backslash", 316 linenum, fname); 317 if ((cmd->u.s = malloc(sizeof(struct s_subst))) == NULL) 318 err(1, "malloc"); 319 p = compile_re(p, &cmd->u.s->re); 320 if (p == NULL) 321 errx(1, 322 "%lu: %s: unterminated substitute pattern", linenum, fname); 323 --p; 324 p = compile_subst(p, cmd->u.s); 325 p = compile_flags(p, cmd->u.s); 326 EATSPACE(); 327 if (*p == ';') { 328 p++; 329 link = &cmd->next; 330 goto semicolon; 331 } 332 break; 333 case TR: /* y */ 334 p++; 335 p = compile_tr(p, (char **)&cmd->u.y); 336 EATSPACE(); 337 if (*p == ';') { 338 p++; 339 link = &cmd->next; 340 goto semicolon; 341 } 342 if (*p) 343 errx(1, 344 "%lu: %s: extra text at the end of a transform command", linenum, fname); 345 break; 346 } 347 } 348 } 349 350 /* 351 * Get a delimited string. P points to the delimeter of the string; d points 352 * to a buffer area. Newline and delimiter escapes are processed; other 353 * escapes are ignored. 354 * 355 * Returns a pointer to the first character after the final delimiter or NULL 356 * in the case of a non-terminated string. The character array d is filled 357 * with the processed string. 358 */ 359 static char * 360 compile_delimited(p, d) 361 char *p, *d; 362 { 363 char c; 364 365 c = *p++; 366 if (c == '\0') 367 return (NULL); 368 else if (c == '\\') 369 errx(1, "%lu: %s: \\ can not be used as a string delimiter", 370 linenum, fname); 371 else if (c == '\n') 372 errx(1, "%lu: %s: newline can not be used as a string delimiter", 373 linenum, fname); 374 while (*p) { 375 if (*p == '[') { 376 if ((d = compile_ccl(&p, d)) == NULL) 377 errx(1, "%lu: %s: unbalanced brackets ([])", linenum, fname); 378 continue; 379 } else if (*p == '\\' && p[1] == '[') { 380 *d++ = *p++; 381 } else if (*p == '\\' && p[1] == c) 382 p++; 383 else if (*p == '\\' && p[1] == 'n') { 384 *d++ = '\n'; 385 p += 2; 386 continue; 387 } else if (*p == '\\' && p[1] == '\\') 388 *d++ = *p++; 389 else if (*p == c) { 390 *d = '\0'; 391 return (p + 1); 392 } 393 *d++ = *p++; 394 } 395 return (NULL); 396 } 397 398 399 /* compile_ccl: expand a POSIX character class */ 400 static char * 401 compile_ccl(sp, t) 402 char **sp; 403 char *t; 404 { 405 int c, d; 406 char *s = *sp; 407 408 *t++ = *s++; 409 if (*s == '^') 410 *t++ = *s++; 411 if (*s == ']') 412 *t++ = *s++; 413 for (; *s && (*t = *s) != ']'; s++, t++) 414 if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) { 415 *++t = *++s, t++, s++; 416 for (c = *s; (*t = *s) != ']' || c != d; s++, t++) 417 if ((c = *s) == '\0') 418 return NULL; 419 } else if (*s == '\\' && s[1] == 'n') 420 *t = '\n', s++; 421 return (*s == ']') ? *sp = ++s, ++t : NULL; 422 } 423 424 /* 425 * Get a regular expression. P points to the delimiter of the regular 426 * expression; repp points to the address of a regexp pointer. Newline 427 * and delimiter escapes are processed; other escapes are ignored. 428 * Returns a pointer to the first character after the final delimiter 429 * or NULL in the case of a non terminated regular expression. The regexp 430 * pointer is set to the compiled regular expression. 431 * Cflags are passed to regcomp. 432 */ 433 static char * 434 compile_re(p, repp) 435 char *p; 436 regex_t **repp; 437 { 438 int eval; 439 char re[_POSIX2_LINE_MAX + 1]; 440 441 p = compile_delimited(p, re); 442 if (p && strlen(re) == 0) { 443 *repp = NULL; 444 return (p); 445 } 446 if ((*repp = malloc(sizeof(regex_t))) == NULL) 447 err(1, "malloc"); 448 if (p && (eval = regcomp(*repp, re, rflags)) != 0) 449 errx(1, "%lu: %s: RE error: %s", 450 linenum, fname, strregerror(eval, *repp)); 451 if (maxnsub < (*repp)->re_nsub) 452 maxnsub = (*repp)->re_nsub; 453 return (p); 454 } 455 456 /* 457 * Compile the substitution string of a regular expression and set res to 458 * point to a saved copy of it. Nsub is the number of parenthesized regular 459 * expressions. 460 */ 461 static char * 462 compile_subst(p, s) 463 char *p; 464 struct s_subst *s; 465 { 466 static char lbuf[_POSIX2_LINE_MAX + 1]; 467 int asize, size; 468 u_char ref; 469 char c, *text, *op, *sp; 470 int more = 0; 471 472 c = *p++; /* Terminator character */ 473 if (c == '\0') 474 return (NULL); 475 476 s->maxbref = 0; 477 s->linenum = linenum; 478 asize = 2 * _POSIX2_LINE_MAX + 1; 479 if ((text = malloc(asize)) == NULL) 480 err(1, "malloc"); 481 size = 0; 482 do { 483 op = sp = text + size; 484 for (; *p; p++) { 485 if (*p == '\\') { 486 p++; 487 if (strchr("123456789", *p) != NULL) { 488 *sp++ = '\\'; 489 ref = *p - '0'; 490 if (s->re != NULL && 491 ref > s->re->re_nsub) 492 errx(1, "%lu: %s: \\%c not defined in the RE", 493 linenum, fname, *p); 494 if (s->maxbref < ref) 495 s->maxbref = ref; 496 } else if (*p == '&' || *p == '\\') 497 *sp++ = '\\'; 498 } else if (*p == c) { 499 if (*++p == '\0' && more) { 500 if (cu_fgets(lbuf, sizeof(lbuf), &more)) 501 p = lbuf; 502 } 503 *sp++ = '\0'; 504 size += sp - op; 505 if ((s->new = realloc(text, size)) == NULL) 506 err(1, "realloc"); 507 return (p); 508 } else if (*p == '\n') { 509 errx(1, 510 "%lu: %s: unescaped newline inside substitute pattern", linenum, fname); 511 /* NOTREACHED */ 512 } 513 *sp++ = *p; 514 } 515 size += sp - op; 516 if (asize - size < _POSIX2_LINE_MAX + 1) { 517 asize *= 2; 518 if ((text = realloc(text, asize)) == NULL) 519 err(1, "realloc"); 520 } 521 } while (cu_fgets(p = lbuf, sizeof(lbuf), &more)); 522 errx(1, "%lu: %s: unterminated substitute in regular expression", 523 linenum, fname); 524 /* NOTREACHED */ 525 } 526 527 /* 528 * Compile the flags of the s command 529 */ 530 static char * 531 compile_flags(p, s) 532 char *p; 533 struct s_subst *s; 534 { 535 int gn; /* True if we have seen g or n */ 536 char wfile[_POSIX2_LINE_MAX + 1], *q; 537 538 s->n = 1; /* Default */ 539 s->p = 0; 540 s->wfile = NULL; 541 s->wfd = -1; 542 for (gn = 0;;) { 543 EATSPACE(); /* EXTENSION */ 544 switch (*p) { 545 case 'g': 546 if (gn) 547 errx(1, 548 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname); 549 gn = 1; 550 s->n = 0; 551 break; 552 case '\0': 553 case '\n': 554 case ';': 555 return (p); 556 case 'p': 557 s->p = 1; 558 break; 559 case '1': case '2': case '3': 560 case '4': case '5': case '6': 561 case '7': case '8': case '9': 562 if (gn) 563 errx(1, 564 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname); 565 gn = 1; 566 /* XXX Check for overflow */ 567 s->n = (int)strtol(p, &p, 10); 568 break; 569 case 'w': 570 p++; 571 #ifdef HISTORIC_PRACTICE 572 if (*p != ' ') { 573 warnx("%lu: %s: space missing before w wfile", linenum, fname); 574 return (p); 575 } 576 #endif 577 EATSPACE(); 578 q = wfile; 579 while (*p) { 580 if (*p == '\n') 581 break; 582 *q++ = *p++; 583 } 584 *q = '\0'; 585 if (q == wfile) 586 errx(1, "%lu: %s: no wfile specified", linenum, fname); 587 s->wfile = strdup(wfile); 588 if (!aflag && (s->wfd = open(wfile, 589 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, 590 DEFFILEMODE)) == -1) 591 err(1, "%s", wfile); 592 return (p); 593 default: 594 errx(1, "%lu: %s: bad flag in substitute command: '%c'", 595 linenum, fname, *p); 596 break; 597 } 598 p++; 599 } 600 } 601 602 /* 603 * Compile a translation set of strings into a lookup table. 604 */ 605 static char * 606 compile_tr(p, transtab) 607 char *p; 608 char **transtab; 609 { 610 int i; 611 char *lt, *op, *np; 612 char old[_POSIX2_LINE_MAX + 1]; 613 char new[_POSIX2_LINE_MAX + 1]; 614 615 if (*p == '\0' || *p == '\\') 616 errx(1, 617 "%lu: %s: transform pattern can not be delimited by newline or backslash", 618 linenum, fname); 619 p = compile_delimited(p, old); 620 if (p == NULL) 621 errx(1, "%lu: %s: unterminated transform source string", 622 linenum, fname); 623 p = compile_delimited(--p, new); 624 if (p == NULL) 625 errx(1, "%lu: %s: unterminated transform target string", 626 linenum, fname); 627 EATSPACE(); 628 if (strlen(new) != strlen(old)) 629 errx(1, "%lu: %s: transform strings are not the same length", 630 linenum, fname); 631 /* We assume characters are 8 bits */ 632 if ((lt = malloc(UCHAR_MAX)) == NULL) 633 err(1, "malloc"); 634 for (i = 0; i <= UCHAR_MAX; i++) 635 lt[i] = (char)i; 636 for (op = old, np = new; *op; op++, np++) 637 lt[(u_char)*op] = *np; 638 *transtab = lt; 639 return (p); 640 } 641 642 /* 643 * Compile the text following an a or i command. 644 */ 645 static char * 646 compile_text() 647 { 648 int asize, esc_nl, size; 649 char *text, *p, *op, *s; 650 char lbuf[_POSIX2_LINE_MAX + 1]; 651 652 asize = 2 * _POSIX2_LINE_MAX + 1; 653 if ((text = malloc(asize)) == NULL) 654 err(1, "malloc"); 655 size = 0; 656 while (cu_fgets(lbuf, sizeof(lbuf), NULL)) { 657 op = s = text + size; 658 p = lbuf; 659 EATSPACE(); 660 for (esc_nl = 0; *p != '\0'; p++) { 661 if (*p == '\\' && p[1] != '\0' && *++p == '\n') 662 esc_nl = 1; 663 *s++ = *p; 664 } 665 size += s - op; 666 if (!esc_nl) { 667 *s = '\0'; 668 break; 669 } 670 if (asize - size < _POSIX2_LINE_MAX + 1) { 671 asize *= 2; 672 if ((text = realloc(text, asize)) == NULL) 673 err(1, "realloc"); 674 } 675 } 676 text[size] = '\0'; 677 if ((p = realloc(text, size + 1)) == NULL) 678 err(1, "realloc"); 679 return (p); 680 } 681 682 /* 683 * Get an address and return a pointer to the first character after 684 * it. Fill the structure pointed to according to the address. 685 */ 686 static char * 687 compile_addr(p, a) 688 char *p; 689 struct s_addr *a; 690 { 691 char *end; 692 693 switch (*p) { 694 case '\\': /* Context address */ 695 ++p; 696 /* FALLTHROUGH */ 697 case '/': /* Context address */ 698 p = compile_re(p, &a->u.r); 699 if (p == NULL) 700 errx(1, "%lu: %s: unterminated regular expression", linenum, fname); 701 a->type = AT_RE; 702 return (p); 703 704 case '$': /* Last line */ 705 a->type = AT_LAST; 706 return (p + 1); 707 /* Line number */ 708 case '0': case '1': case '2': case '3': case '4': 709 case '5': case '6': case '7': case '8': case '9': 710 a->type = AT_LINE; 711 a->u.l = strtol(p, &end, 10); 712 return (end); 713 default: 714 errx(1, "%lu: %s: expected context address", linenum, fname); 715 return (NULL); 716 } 717 } 718 719 /* 720 * duptoeol -- 721 * Return a copy of all the characters up to \n or \0. 722 */ 723 static char * 724 duptoeol(s, ctype) 725 char *s; 726 const char *ctype; 727 { 728 size_t len; 729 int ws; 730 char *p, *start; 731 732 ws = 0; 733 for (start = s; *s != '\0' && *s != '\n'; ++s) 734 ws = isspace((unsigned char)*s); 735 *s = '\0'; 736 if (ws) 737 warnx("%lu: %s: whitespace after %s", linenum, fname, ctype); 738 len = s - start + 1; 739 if ((p = malloc(len)) == NULL) 740 err(1, "malloc"); 741 return (memmove(p, start, len)); 742 } 743 744 /* 745 * Convert goto label names to addresses, and count a and r commands, in 746 * the given subset of the script. Free the memory used by labels in b 747 * and t commands (but not by :). 748 * 749 * TODO: Remove } nodes 750 */ 751 static void 752 fixuplabel(cp, end) 753 struct s_command *cp, *end; 754 { 755 756 for (; cp != end; cp = cp->next) 757 switch (cp->code) { 758 case 'a': 759 case 'r': 760 appendnum++; 761 break; 762 case 'b': 763 case 't': 764 /* Resolve branch target. */ 765 if (cp->t == NULL) { 766 cp->u.c = NULL; 767 break; 768 } 769 if ((cp->u.c = findlabel(cp->t)) == NULL) 770 errx(1, "%lu: %s: undefined label '%s'", linenum, fname, cp->t); 771 free(cp->t); 772 break; 773 case '{': 774 /* Do interior commands. */ 775 fixuplabel(cp->u.c, cp->next); 776 break; 777 } 778 } 779 780 /* 781 * Associate the given command label for later lookup. 782 */ 783 static void 784 enterlabel(cp) 785 struct s_command *cp; 786 { 787 struct labhash **lhp, *lh; 788 u_char *p; 789 u_int h, c; 790 791 for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++) 792 h = (h << 5) + h + c; 793 lhp = &labels[h & LHMASK]; 794 for (lh = *lhp; lh != NULL; lh = lh->lh_next) 795 if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0) 796 errx(1, "%lu: %s: duplicate label '%s'", linenum, fname, cp->t); 797 if ((lh = malloc(sizeof *lh)) == NULL) 798 err(1, "malloc"); 799 lh->lh_next = *lhp; 800 lh->lh_hash = h; 801 lh->lh_cmd = cp; 802 lh->lh_ref = 0; 803 *lhp = lh; 804 } 805 806 /* 807 * Find the label contained in the command l in the command linked 808 * list cp. L is excluded from the search. Return NULL if not found. 809 */ 810 static struct s_command * 811 findlabel(name) 812 char *name; 813 { 814 struct labhash *lh; 815 u_char *p; 816 u_int h, c; 817 818 for (h = 0, p = (u_char *)name; (c = *p) != 0; p++) 819 h = (h << 5) + h + c; 820 for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) { 821 if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) { 822 lh->lh_ref = 1; 823 return (lh->lh_cmd); 824 } 825 } 826 return (NULL); 827 } 828 829 /* 830 * Warn about any unused labels. As a side effect, release the label hash 831 * table space. 832 */ 833 static void 834 uselabel() 835 { 836 struct labhash *lh, *next; 837 int i; 838 839 for (i = 0; i < LHSZ; i++) { 840 for (lh = labels[i]; lh != NULL; lh = next) { 841 next = lh->lh_next; 842 if (!lh->lh_ref) 843 warnx("%lu: %s: unused label '%s'", 844 linenum, fname, lh->lh_cmd->t); 845 free(lh); 846 } 847 } 848 } 849