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 #ifndef lint 39 #if 0 40 static char sccsid[] = "@(#)compile.c 8.1 (Berkeley) 6/6/93"; 41 #endif 42 static const char rcsid[] = 43 "$FreeBSD$"; 44 #endif /* not lint */ 45 46 #include <sys/types.h> 47 #include <sys/stat.h> 48 49 #include <ctype.h> 50 #include <err.h> 51 #include <fcntl.h> 52 #include <limits.h> 53 #include <regex.h> 54 #include <stdio.h> 55 #include <stdlib.h> 56 #include <string.h> 57 58 #include "defs.h" 59 #include "extern.h" 60 61 #define LHSZ 128 62 #define LHMASK (LHSZ - 1) 63 static struct labhash { 64 struct labhash *lh_next; 65 u_int lh_hash; 66 struct s_command *lh_cmd; 67 int lh_ref; 68 } *labels[LHSZ]; 69 70 static char *compile_addr __P((char *, struct s_addr *)); 71 static char *compile_ccl __P((char **, char *)); 72 static char *compile_delimited __P((char *, char *)); 73 static char *compile_flags __P((char *, struct s_subst *)); 74 static char *compile_re __P((char *, regex_t **)); 75 static char *compile_subst __P((char *, struct s_subst *)); 76 static char *compile_text __P((void)); 77 static char *compile_tr __P((char *, char **)); 78 static struct s_command 79 **compile_stream __P((struct s_command **)); 80 static char *duptoeol __P((char *, char *)); 81 static void enterlabel __P((struct s_command *)); 82 static struct s_command 83 *findlabel __P((char *)); 84 static void fixuplabel __P((struct s_command *, struct s_command *)); 85 static void uselabel __P((void)); 86 87 /* 88 * Command specification. This is used to drive the command parser. 89 */ 90 struct s_format { 91 char code; /* Command code */ 92 int naddr; /* Number of address args */ 93 enum e_args args; /* Argument type */ 94 }; 95 96 static struct s_format cmd_fmts[] = { 97 {'{', 2, GROUP}, 98 {'}', 0, ENDGROUP}, 99 {'a', 1, TEXT}, 100 {'b', 2, BRANCH}, 101 {'c', 2, TEXT}, 102 {'d', 2, EMPTY}, 103 {'D', 2, EMPTY}, 104 {'g', 2, EMPTY}, 105 {'G', 2, EMPTY}, 106 {'h', 2, EMPTY}, 107 {'H', 2, EMPTY}, 108 {'i', 1, TEXT}, 109 {'l', 2, EMPTY}, 110 {'n', 2, EMPTY}, 111 {'N', 2, EMPTY}, 112 {'p', 2, EMPTY}, 113 {'P', 2, EMPTY}, 114 {'q', 1, EMPTY}, 115 {'r', 1, RFILE}, 116 {'s', 2, SUBST}, 117 {'t', 2, BRANCH}, 118 {'w', 2, WFILE}, 119 {'x', 2, EMPTY}, 120 {'y', 2, TR}, 121 {'!', 2, NONSEL}, 122 {':', 0, LABEL}, 123 {'#', 0, COMMENT}, 124 {'=', 1, EMPTY}, 125 {'\0', 0, COMMENT}, 126 }; 127 128 /* The compiled program. */ 129 struct s_command *prog; 130 131 /* 132 * Compile the program into prog. 133 * Initialise appends. 134 */ 135 void 136 compile() 137 { 138 *compile_stream(&prog) = NULL; 139 fixuplabel(prog, NULL); 140 uselabel(); 141 if (appendnum == 0) 142 appends = NULL; 143 else if ((appends = malloc(sizeof(struct s_appends) * appendnum)) == 144 NULL) 145 err(1, "malloc"); 146 if ((match = malloc((maxnsub + 1) * sizeof(regmatch_t))) == NULL) 147 err(1, "malloc"); 148 } 149 150 #define EATSPACE() do { \ 151 if (p) \ 152 while (*p && isspace((unsigned char)*p)) \ 153 p++; \ 154 } while (0) 155 156 static struct s_command ** 157 compile_stream(link) 158 struct s_command **link; 159 { 160 register char *p; 161 static char lbuf[_POSIX2_LINE_MAX + 1]; /* To save stack */ 162 struct s_command *cmd, *cmd2, *stack; 163 struct s_format *fp; 164 int naddr; /* Number of addresses */ 165 166 stack = 0; 167 for (;;) { 168 if ((p = cu_fgets(lbuf, sizeof(lbuf), NULL)) == NULL) { 169 if (stack != 0) 170 errx(1, "%lu: %s: unexpected EOF (pending }'s)", 171 linenum, fname); 172 return (link); 173 } 174 175 semicolon: EATSPACE(); 176 if (p && (*p == '#' || *p == '\0')) 177 continue; 178 if ((*link = cmd = malloc(sizeof(struct s_command))) == NULL) 179 err(1, "malloc"); 180 link = &cmd->next; 181 cmd->nonsel = cmd->inrange = 0; 182 /* First parse the addresses */ 183 naddr = 0; 184 185 /* Valid characters to start an address */ 186 #define addrchar(c) (strchr("0123456789/\\$", (c))) 187 if (addrchar(*p)) { 188 naddr++; 189 if ((cmd->a1 = malloc(sizeof(struct s_addr))) == NULL) 190 err(1, "malloc"); 191 p = compile_addr(p, cmd->a1); 192 EATSPACE(); /* EXTENSION */ 193 if (*p == ',') { 194 p++; 195 EATSPACE(); /* EXTENSION */ 196 naddr++; 197 if ((cmd->a2 = malloc(sizeof(struct s_addr))) 198 == NULL) 199 err(1, "malloc"); 200 p = compile_addr(p, cmd->a2); 201 EATSPACE(); 202 } else 203 cmd->a2 = 0; 204 } else 205 cmd->a1 = cmd->a2 = 0; 206 207 nonsel: /* Now parse the command */ 208 if (!*p) 209 errx(1, "%lu: %s: command expected", linenum, fname); 210 cmd->code = *p; 211 for (fp = cmd_fmts; fp->code; fp++) 212 if (fp->code == *p) 213 break; 214 if (!fp->code) 215 errx(1, "%lu: %s: invalid command code %c", linenum, fname, *p); 216 if (naddr > fp->naddr) 217 errx(1, 218 "%lu: %s: command %c expects up to %d address(es), found %d", 219 linenum, fname, *p, fp->naddr, naddr); 220 switch (fp->args) { 221 case NONSEL: /* ! */ 222 p++; 223 EATSPACE(); 224 cmd->nonsel = ! cmd->nonsel; 225 goto nonsel; 226 case GROUP: /* { */ 227 p++; 228 EATSPACE(); 229 cmd->next = stack; 230 stack = cmd; 231 link = &cmd->u.c; 232 if (*p) 233 goto semicolon; 234 break; 235 case ENDGROUP: 236 /* 237 * Short-circuit command processing, since end of 238 * group is really just a noop. 239 */ 240 cmd->nonsel = 1; 241 if (stack == 0) 242 errx(1, "%lu: %s: unexpected }", linenum, fname); 243 cmd2 = stack; 244 stack = cmd2->next; 245 cmd2->next = cmd; 246 /*FALLTHROUGH*/ 247 case EMPTY: /* d D g G h H l n N p P q x = \0 */ 248 p++; 249 EATSPACE(); 250 if (*p == ';') { 251 p++; 252 link = &cmd->next; 253 goto semicolon; 254 } 255 if (*p) 256 errx(1, "%lu: %s: extra characters at the end of %c command", 257 linenum, fname, cmd->code); 258 break; 259 case TEXT: /* a c i */ 260 p++; 261 EATSPACE(); 262 if (*p != '\\') 263 errx(1, 264 "%lu: %s: command %c expects \\ followed by text", linenum, fname, cmd->code); 265 p++; 266 EATSPACE(); 267 if (*p) 268 errx(1, 269 "%lu: %s: extra characters after \\ at the end of %c command", 270 linenum, fname, cmd->code); 271 cmd->t = compile_text(); 272 break; 273 case COMMENT: /* \0 # */ 274 break; 275 case WFILE: /* w */ 276 p++; 277 EATSPACE(); 278 if (*p == '\0') 279 errx(1, "%lu: %s: filename expected", linenum, fname); 280 cmd->t = duptoeol(p, "w command"); 281 if (aflag) 282 cmd->u.fd = -1; 283 else if ((cmd->u.fd = open(p, 284 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, 285 DEFFILEMODE)) == -1) 286 err(1, "%s", p); 287 break; 288 case RFILE: /* r */ 289 p++; 290 EATSPACE(); 291 if (*p == '\0') 292 errx(1, "%lu: %s: filename expected", linenum, fname); 293 else 294 cmd->t = duptoeol(p, "read command"); 295 break; 296 case BRANCH: /* b t */ 297 p++; 298 EATSPACE(); 299 if (*p == '\0') 300 cmd->t = NULL; 301 else 302 cmd->t = duptoeol(p, "branch"); 303 break; 304 case LABEL: /* : */ 305 p++; 306 EATSPACE(); 307 cmd->t = duptoeol(p, "label"); 308 if (strlen(p) == 0) 309 errx(1, "%lu: %s: empty label", linenum, fname); 310 enterlabel(cmd); 311 break; 312 case SUBST: /* s */ 313 p++; 314 if (*p == '\0' || *p == '\\') 315 errx(1, 316 "%lu: %s: substitute pattern can not be delimited by newline or backslash", 317 linenum, fname); 318 if ((cmd->u.s = malloc(sizeof(struct s_subst))) == NULL) 319 err(1, "malloc"); 320 p = compile_re(p, &cmd->u.s->re); 321 if (p == NULL) 322 errx(1, 323 "%lu: %s: unterminated substitute pattern", linenum, fname); 324 --p; 325 p = compile_subst(p, cmd->u.s); 326 p = compile_flags(p, cmd->u.s); 327 EATSPACE(); 328 if (*p == ';') { 329 p++; 330 link = &cmd->next; 331 goto semicolon; 332 } 333 break; 334 case TR: /* y */ 335 p++; 336 p = compile_tr(p, (char **)&cmd->u.y); 337 EATSPACE(); 338 if (*p == ';') { 339 p++; 340 link = &cmd->next; 341 goto semicolon; 342 } 343 if (*p) 344 errx(1, 345 "%lu: %s: extra text at the end of a transform command", linenum, fname); 346 break; 347 } 348 } 349 } 350 351 /* 352 * Get a delimited string. P points to the delimeter of the string; d points 353 * to a buffer area. Newline and delimiter escapes are processed; other 354 * escapes are ignored. 355 * 356 * Returns a pointer to the first character after the final delimiter or NULL 357 * in the case of a non-terminated string. The character array d is filled 358 * with the processed string. 359 */ 360 static char * 361 compile_delimited(p, d) 362 char *p, *d; 363 { 364 char c; 365 366 c = *p++; 367 if (c == '\0') 368 return (NULL); 369 else if (c == '\\') 370 errx(1, "%lu: %s: \\ can not be used as a string delimiter", 371 linenum, fname); 372 else if (c == '\n') 373 errx(1, "%lu: %s: newline can not be used as a string delimiter", 374 linenum, fname); 375 while (*p) { 376 if (*p == '[') { 377 if ((d = compile_ccl(&p, d)) == NULL) 378 errx(1, "%lu: %s: unbalanced brackets ([])", linenum, fname); 379 continue; 380 } else if (*p == '\\' && p[1] == '[') { 381 *d++ = *p++; 382 } else if (*p == '\\' && p[1] == c) 383 p++; 384 else if (*p == '\\' && p[1] == 'n') { 385 *d++ = '\n'; 386 p += 2; 387 continue; 388 } else if (*p == '\\' && p[1] == '\\') 389 *d++ = *p++; 390 else if (*p == c) { 391 *d = '\0'; 392 return (p + 1); 393 } 394 *d++ = *p++; 395 } 396 return (NULL); 397 } 398 399 400 /* compile_ccl: expand a POSIX character class */ 401 static char * 402 compile_ccl(sp, t) 403 char **sp; 404 char *t; 405 { 406 int c, d; 407 char *s = *sp; 408 409 *t++ = *s++; 410 if (*s == '^') 411 *t++ = *s++; 412 if (*s == ']') 413 *t++ = *s++; 414 for (; *s && (*t = *s) != ']'; s++, t++) 415 if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) { 416 *++t = *++s, t++, s++; 417 for (c = *s; (*t = *s) != ']' || c != d; s++, t++) 418 if ((c = *s) == '\0') 419 return NULL; 420 } else if (*s == '\\' && s[1] == 'n') 421 *t = '\n', s++; 422 return (*s == ']') ? *sp = ++s, ++t : NULL; 423 } 424 425 /* 426 * Get a regular expression. P points to the delimiter of the regular 427 * expression; repp points to the address of a regexp pointer. Newline 428 * and delimiter escapes are processed; other escapes are ignored. 429 * Returns a pointer to the first character after the final delimiter 430 * or NULL in the case of a non terminated regular expression. The regexp 431 * pointer is set to the compiled regular expression. 432 * Cflags are passed to regcomp. 433 */ 434 static char * 435 compile_re(p, repp) 436 char *p; 437 regex_t **repp; 438 { 439 int eval; 440 char re[_POSIX2_LINE_MAX + 1]; 441 442 p = compile_delimited(p, re); 443 if (p && strlen(re) == 0) { 444 *repp = NULL; 445 return (p); 446 } 447 if ((*repp = malloc(sizeof(regex_t))) == NULL) 448 err(1, "malloc"); 449 if (p && (eval = regcomp(*repp, re, rflags)) != 0) 450 errx(1, "%lu: %s: RE error: %s", 451 linenum, fname, strregerror(eval, *repp)); 452 if (maxnsub < (*repp)->re_nsub) 453 maxnsub = (*repp)->re_nsub; 454 return (p); 455 } 456 457 /* 458 * Compile the substitution string of a regular expression and set res to 459 * point to a saved copy of it. Nsub is the number of parenthesized regular 460 * expressions. 461 */ 462 static char * 463 compile_subst(p, s) 464 char *p; 465 struct s_subst *s; 466 { 467 static char lbuf[_POSIX2_LINE_MAX + 1]; 468 int asize, ref, size; 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 register char *s; 726 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 register struct labhash **lhp, *lh; 788 register u_char *p; 789 register 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 register struct labhash *lh; 815 register u_char *p; 816 register 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 register struct labhash *lh, *next; 837 register 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