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(char *, struct s_addr *); 70 static char *compile_ccl(char **, char *); 71 static char *compile_delimited(char *, char *); 72 static char *compile_flags(char *, struct s_subst *); 73 static char *compile_re(char *, regex_t **); 74 static char *compile_subst(char *, struct s_subst *); 75 static char *compile_text(void); 76 static char *compile_tr(char *, char **); 77 static struct s_command 78 **compile_stream(struct s_command **); 79 static char *duptoeol(char *, const char *); 80 static void enterlabel(struct s_command *); 81 static struct s_command 82 *findlabel(char *); 83 static void fixuplabel(struct s_command *, struct s_command *); 84 static void uselabel(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(void) 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(struct s_command **link) 157 { 158 char *p; 159 static char lbuf[_POSIX2_LINE_MAX + 1]; /* To save stack */ 160 struct s_command *cmd, *cmd2, *stack; 161 struct s_format *fp; 162 int naddr; /* Number of addresses */ 163 164 stack = 0; 165 for (;;) { 166 if ((p = cu_fgets(lbuf, sizeof(lbuf), NULL)) == NULL) { 167 if (stack != 0) 168 errx(1, "%lu: %s: unexpected EOF (pending }'s)", 169 linenum, fname); 170 return (link); 171 } 172 173 semicolon: EATSPACE(); 174 if (p) { 175 if (*p == '#' || *p == '\0') 176 continue; 177 else if (*p == ';') { 178 p++; 179 goto semicolon; 180 } 181 } 182 if ((*link = cmd = malloc(sizeof(struct s_command))) == NULL) 183 err(1, "malloc"); 184 link = &cmd->next; 185 cmd->nonsel = cmd->inrange = 0; 186 /* First parse the addresses */ 187 naddr = 0; 188 189 /* Valid characters to start an address */ 190 #define addrchar(c) (strchr("0123456789/\\$", (c))) 191 if (addrchar(*p)) { 192 naddr++; 193 if ((cmd->a1 = malloc(sizeof(struct s_addr))) == NULL) 194 err(1, "malloc"); 195 p = compile_addr(p, cmd->a1); 196 EATSPACE(); /* EXTENSION */ 197 if (*p == ',') { 198 p++; 199 EATSPACE(); /* EXTENSION */ 200 naddr++; 201 if ((cmd->a2 = malloc(sizeof(struct s_addr))) 202 == NULL) 203 err(1, "malloc"); 204 p = compile_addr(p, cmd->a2); 205 EATSPACE(); 206 } else 207 cmd->a2 = 0; 208 } else 209 cmd->a1 = cmd->a2 = 0; 210 211 nonsel: /* Now parse the command */ 212 if (!*p) 213 errx(1, "%lu: %s: command expected", linenum, fname); 214 cmd->code = *p; 215 for (fp = cmd_fmts; fp->code; fp++) 216 if (fp->code == *p) 217 break; 218 if (!fp->code) 219 errx(1, "%lu: %s: invalid command code %c", linenum, fname, *p); 220 if (naddr > fp->naddr) 221 errx(1, 222 "%lu: %s: command %c expects up to %d address(es), found %d", 223 linenum, fname, *p, fp->naddr, naddr); 224 switch (fp->args) { 225 case NONSEL: /* ! */ 226 p++; 227 EATSPACE(); 228 cmd->nonsel = ! cmd->nonsel; 229 goto nonsel; 230 case GROUP: /* { */ 231 p++; 232 EATSPACE(); 233 cmd->next = stack; 234 stack = cmd; 235 link = &cmd->u.c; 236 if (*p) 237 goto semicolon; 238 break; 239 case ENDGROUP: 240 /* 241 * Short-circuit command processing, since end of 242 * group is really just a noop. 243 */ 244 cmd->nonsel = 1; 245 if (stack == 0) 246 errx(1, "%lu: %s: unexpected }", linenum, fname); 247 cmd2 = stack; 248 stack = cmd2->next; 249 cmd2->next = cmd; 250 /*FALLTHROUGH*/ 251 case EMPTY: /* d D g G h H l n N p P q x = \0 */ 252 p++; 253 EATSPACE(); 254 if (*p == ';') { 255 p++; 256 link = &cmd->next; 257 goto semicolon; 258 } 259 if (*p) 260 errx(1, "%lu: %s: extra characters at the end of %c command", 261 linenum, fname, cmd->code); 262 break; 263 case TEXT: /* a c i */ 264 p++; 265 EATSPACE(); 266 if (*p != '\\') 267 errx(1, 268 "%lu: %s: command %c expects \\ followed by text", linenum, fname, cmd->code); 269 p++; 270 EATSPACE(); 271 if (*p) 272 errx(1, 273 "%lu: %s: extra characters after \\ at the end of %c command", 274 linenum, fname, cmd->code); 275 cmd->t = compile_text(); 276 break; 277 case COMMENT: /* \0 # */ 278 break; 279 case WFILE: /* w */ 280 p++; 281 EATSPACE(); 282 if (*p == '\0') 283 errx(1, "%lu: %s: filename expected", linenum, fname); 284 cmd->t = duptoeol(p, "w command"); 285 if (aflag) 286 cmd->u.fd = -1; 287 else if ((cmd->u.fd = open(p, 288 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, 289 DEFFILEMODE)) == -1) 290 err(1, "%s", p); 291 break; 292 case RFILE: /* r */ 293 p++; 294 EATSPACE(); 295 if (*p == '\0') 296 errx(1, "%lu: %s: filename expected", linenum, fname); 297 else 298 cmd->t = duptoeol(p, "read command"); 299 break; 300 case BRANCH: /* b t */ 301 p++; 302 EATSPACE(); 303 if (*p == '\0') 304 cmd->t = NULL; 305 else 306 cmd->t = duptoeol(p, "branch"); 307 break; 308 case LABEL: /* : */ 309 p++; 310 EATSPACE(); 311 cmd->t = duptoeol(p, "label"); 312 if (strlen(p) == 0) 313 errx(1, "%lu: %s: empty label", linenum, fname); 314 enterlabel(cmd); 315 break; 316 case SUBST: /* s */ 317 p++; 318 if (*p == '\0' || *p == '\\') 319 errx(1, 320 "%lu: %s: substitute pattern can not be delimited by newline or backslash", 321 linenum, fname); 322 if ((cmd->u.s = malloc(sizeof(struct s_subst))) == NULL) 323 err(1, "malloc"); 324 p = compile_re(p, &cmd->u.s->re); 325 if (p == NULL) 326 errx(1, 327 "%lu: %s: unterminated substitute pattern", linenum, fname); 328 --p; 329 p = compile_subst(p, cmd->u.s); 330 p = compile_flags(p, cmd->u.s); 331 EATSPACE(); 332 if (*p == ';') { 333 p++; 334 link = &cmd->next; 335 goto semicolon; 336 } 337 break; 338 case TR: /* y */ 339 p++; 340 p = compile_tr(p, (char **)&cmd->u.y); 341 EATSPACE(); 342 if (*p == ';') { 343 p++; 344 link = &cmd->next; 345 goto semicolon; 346 } 347 if (*p) 348 errx(1, 349 "%lu: %s: extra text at the end of a transform command", linenum, fname); 350 break; 351 } 352 } 353 } 354 355 /* 356 * Get a delimited string. P points to the delimeter of the string; d points 357 * to a buffer area. Newline and delimiter escapes are processed; other 358 * escapes are ignored. 359 * 360 * Returns a pointer to the first character after the final delimiter or NULL 361 * in the case of a non-terminated string. The character array d is filled 362 * with the processed string. 363 */ 364 static char * 365 compile_delimited(char *p, char *d) 366 { 367 char c; 368 369 c = *p++; 370 if (c == '\0') 371 return (NULL); 372 else if (c == '\\') 373 errx(1, "%lu: %s: \\ can not be used as a string delimiter", 374 linenum, fname); 375 else if (c == '\n') 376 errx(1, "%lu: %s: newline can not be used as a string delimiter", 377 linenum, fname); 378 while (*p) { 379 if (*p == '[') { 380 if ((d = compile_ccl(&p, d)) == NULL) 381 errx(1, "%lu: %s: unbalanced brackets ([])", linenum, fname); 382 continue; 383 } else if (*p == '\\' && p[1] == '[') { 384 *d++ = *p++; 385 } else if (*p == '\\' && p[1] == c) 386 p++; 387 else if (*p == '\\' && p[1] == 'n') { 388 *d++ = '\n'; 389 p += 2; 390 continue; 391 } else if (*p == '\\' && p[1] == '\\') 392 *d++ = *p++; 393 else if (*p == c) { 394 *d = '\0'; 395 return (p + 1); 396 } 397 *d++ = *p++; 398 } 399 return (NULL); 400 } 401 402 403 /* compile_ccl: expand a POSIX character class */ 404 static char * 405 compile_ccl(char **sp, char *t) 406 { 407 int c, d; 408 char *s = *sp; 409 410 *t++ = *s++; 411 if (*s == '^') 412 *t++ = *s++; 413 if (*s == ']') 414 *t++ = *s++; 415 for (; *s && (*t = *s) != ']'; s++, t++) 416 if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) { 417 *++t = *++s, t++, s++; 418 for (c = *s; (*t = *s) != ']' || c != d; s++, t++) 419 if ((c = *s) == '\0') 420 return NULL; 421 } else if (*s == '\\' && s[1] == 'n') 422 *t = '\n', s++; 423 return (*s == ']') ? *sp = ++s, ++t : NULL; 424 } 425 426 /* 427 * Get a regular expression. P points to the delimiter of the regular 428 * expression; repp points to the address of a regexp pointer. Newline 429 * and delimiter escapes are processed; other escapes are ignored. 430 * Returns a pointer to the first character after the final delimiter 431 * or NULL in the case of a non terminated regular expression. The regexp 432 * pointer is set to the compiled regular expression. 433 * Cflags are passed to regcomp. 434 */ 435 static char * 436 compile_re(char *p, 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(char *p, struct s_subst *s) 463 { 464 static char lbuf[_POSIX2_LINE_MAX + 1]; 465 int asize, size; 466 u_char ref; 467 char c, *text, *op, *sp; 468 int more = 1, sawesc = 0; 469 470 c = *p++; /* Terminator character */ 471 if (c == '\0') 472 return (NULL); 473 474 s->maxbref = 0; 475 s->linenum = linenum; 476 asize = 2 * _POSIX2_LINE_MAX + 1; 477 if ((text = malloc(asize)) == NULL) 478 err(1, "malloc"); 479 size = 0; 480 do { 481 op = sp = text + size; 482 for (; *p; p++) { 483 if (*p == '\\' || sawesc) { 484 /* 485 * If this is a continuation from the last 486 * buffer, we won't have a character to 487 * skip over. 488 */ 489 if (sawesc) 490 sawesc = 0; 491 else 492 p++; 493 494 if (*p == '\0') { 495 /* 496 * This escaped character is continued 497 * in the next part of the line. Note 498 * this fact, then cause the loop to 499 * exit w/ normal EOL case and reenter 500 * above with the new buffer. 501 */ 502 sawesc = 1; 503 p--; 504 continue; 505 } else if (strchr("123456789", *p) != NULL) { 506 *sp++ = '\\'; 507 ref = *p - '0'; 508 if (s->re != NULL && 509 ref > s->re->re_nsub) 510 errx(1, "%lu: %s: \\%c not defined in the RE", 511 linenum, fname, *p); 512 if (s->maxbref < ref) 513 s->maxbref = ref; 514 } else if (*p == '&' || *p == '\\') 515 *sp++ = '\\'; 516 } else if (*p == c) { 517 if (*++p == '\0' && more) { 518 if (cu_fgets(lbuf, sizeof(lbuf), &more)) 519 p = lbuf; 520 } 521 *sp++ = '\0'; 522 size += sp - op; 523 if ((s->new = realloc(text, size)) == NULL) 524 err(1, "realloc"); 525 return (p); 526 } else if (*p == '\n') { 527 errx(1, 528 "%lu: %s: unescaped newline inside substitute pattern", linenum, fname); 529 /* NOTREACHED */ 530 } 531 *sp++ = *p; 532 } 533 size += sp - op; 534 if (asize - size < _POSIX2_LINE_MAX + 1) { 535 asize *= 2; 536 if ((text = realloc(text, asize)) == NULL) 537 err(1, "realloc"); 538 } 539 } while (cu_fgets(p = lbuf, sizeof(lbuf), &more)); 540 errx(1, "%lu: %s: unterminated substitute in regular expression", 541 linenum, fname); 542 /* NOTREACHED */ 543 } 544 545 /* 546 * Compile the flags of the s command 547 */ 548 static char * 549 compile_flags(char *p, struct s_subst *s) 550 { 551 int gn; /* True if we have seen g or n */ 552 char wfile[_POSIX2_LINE_MAX + 1], *q; 553 554 s->n = 1; /* Default */ 555 s->p = 0; 556 s->wfile = NULL; 557 s->wfd = -1; 558 for (gn = 0;;) { 559 EATSPACE(); /* EXTENSION */ 560 switch (*p) { 561 case 'g': 562 if (gn) 563 errx(1, 564 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname); 565 gn = 1; 566 s->n = 0; 567 break; 568 case '\0': 569 case '\n': 570 case ';': 571 return (p); 572 case 'p': 573 s->p = 1; 574 break; 575 case '1': case '2': case '3': 576 case '4': case '5': case '6': 577 case '7': case '8': case '9': 578 if (gn) 579 errx(1, 580 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname); 581 gn = 1; 582 /* XXX Check for overflow */ 583 s->n = (int)strtol(p, &p, 10); 584 break; 585 case 'w': 586 p++; 587 #ifdef HISTORIC_PRACTICE 588 if (*p != ' ') { 589 warnx("%lu: %s: space missing before w wfile", linenum, fname); 590 return (p); 591 } 592 #endif 593 EATSPACE(); 594 q = wfile; 595 while (*p) { 596 if (*p == '\n') 597 break; 598 *q++ = *p++; 599 } 600 *q = '\0'; 601 if (q == wfile) 602 errx(1, "%lu: %s: no wfile specified", linenum, fname); 603 s->wfile = strdup(wfile); 604 if (!aflag && (s->wfd = open(wfile, 605 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, 606 DEFFILEMODE)) == -1) 607 err(1, "%s", wfile); 608 return (p); 609 default: 610 errx(1, "%lu: %s: bad flag in substitute command: '%c'", 611 linenum, fname, *p); 612 break; 613 } 614 p++; 615 } 616 } 617 618 /* 619 * Compile a translation set of strings into a lookup table. 620 */ 621 static char * 622 compile_tr(char *p, char **transtab) 623 { 624 int i; 625 char *lt, *op, *np; 626 char old[_POSIX2_LINE_MAX + 1]; 627 char new[_POSIX2_LINE_MAX + 1]; 628 629 if (*p == '\0' || *p == '\\') 630 errx(1, 631 "%lu: %s: transform pattern can not be delimited by newline or backslash", 632 linenum, fname); 633 p = compile_delimited(p, old); 634 if (p == NULL) 635 errx(1, "%lu: %s: unterminated transform source string", 636 linenum, fname); 637 p = compile_delimited(--p, new); 638 if (p == NULL) 639 errx(1, "%lu: %s: unterminated transform target string", 640 linenum, fname); 641 EATSPACE(); 642 if (strlen(new) != strlen(old)) 643 errx(1, "%lu: %s: transform strings are not the same length", 644 linenum, fname); 645 /* We assume characters are 8 bits */ 646 if ((lt = malloc(UCHAR_MAX)) == NULL) 647 err(1, "malloc"); 648 for (i = 0; i <= UCHAR_MAX; i++) 649 lt[i] = (char)i; 650 for (op = old, np = new; *op; op++, np++) 651 lt[(u_char)*op] = *np; 652 *transtab = lt; 653 return (p); 654 } 655 656 /* 657 * Compile the text following an a or i command. 658 */ 659 static char * 660 compile_text(void) 661 { 662 int asize, esc_nl, size; 663 char *text, *p, *op, *s; 664 char lbuf[_POSIX2_LINE_MAX + 1]; 665 666 asize = 2 * _POSIX2_LINE_MAX + 1; 667 if ((text = malloc(asize)) == NULL) 668 err(1, "malloc"); 669 size = 0; 670 while (cu_fgets(lbuf, sizeof(lbuf), NULL)) { 671 op = s = text + size; 672 p = lbuf; 673 EATSPACE(); 674 for (esc_nl = 0; *p != '\0'; p++) { 675 if (*p == '\\' && p[1] != '\0' && *++p == '\n') 676 esc_nl = 1; 677 *s++ = *p; 678 } 679 size += s - op; 680 if (!esc_nl) { 681 *s = '\0'; 682 break; 683 } 684 if (asize - size < _POSIX2_LINE_MAX + 1) { 685 asize *= 2; 686 if ((text = realloc(text, asize)) == NULL) 687 err(1, "realloc"); 688 } 689 } 690 text[size] = '\0'; 691 if ((p = realloc(text, size + 1)) == NULL) 692 err(1, "realloc"); 693 return (p); 694 } 695 696 /* 697 * Get an address and return a pointer to the first character after 698 * it. Fill the structure pointed to according to the address. 699 */ 700 static char * 701 compile_addr(char *p, struct s_addr *a) 702 { 703 char *end; 704 705 switch (*p) { 706 case '\\': /* Context address */ 707 ++p; 708 /* FALLTHROUGH */ 709 case '/': /* Context address */ 710 p = compile_re(p, &a->u.r); 711 if (p == NULL) 712 errx(1, "%lu: %s: unterminated regular expression", linenum, fname); 713 a->type = AT_RE; 714 return (p); 715 716 case '$': /* Last line */ 717 a->type = AT_LAST; 718 return (p + 1); 719 /* Line number */ 720 case '0': case '1': case '2': case '3': case '4': 721 case '5': case '6': case '7': case '8': case '9': 722 a->type = AT_LINE; 723 a->u.l = strtol(p, &end, 10); 724 return (end); 725 default: 726 errx(1, "%lu: %s: expected context address", linenum, fname); 727 return (NULL); 728 } 729 } 730 731 /* 732 * duptoeol -- 733 * Return a copy of all the characters up to \n or \0. 734 */ 735 static char * 736 duptoeol(char *s, const char *ctype) 737 { 738 size_t len; 739 int ws; 740 char *p, *start; 741 742 ws = 0; 743 for (start = s; *s != '\0' && *s != '\n'; ++s) 744 ws = isspace((unsigned char)*s); 745 *s = '\0'; 746 if (ws) 747 warnx("%lu: %s: whitespace after %s", linenum, fname, ctype); 748 len = s - start + 1; 749 if ((p = malloc(len)) == NULL) 750 err(1, "malloc"); 751 return (memmove(p, start, len)); 752 } 753 754 /* 755 * Convert goto label names to addresses, and count a and r commands, in 756 * the given subset of the script. Free the memory used by labels in b 757 * and t commands (but not by :). 758 * 759 * TODO: Remove } nodes 760 */ 761 static void 762 fixuplabel(struct s_command *cp, struct s_command *end) 763 { 764 765 for (; cp != end; cp = cp->next) 766 switch (cp->code) { 767 case 'a': 768 case 'r': 769 appendnum++; 770 break; 771 case 'b': 772 case 't': 773 /* Resolve branch target. */ 774 if (cp->t == NULL) { 775 cp->u.c = NULL; 776 break; 777 } 778 if ((cp->u.c = findlabel(cp->t)) == NULL) 779 errx(1, "%lu: %s: undefined label '%s'", linenum, fname, cp->t); 780 free(cp->t); 781 break; 782 case '{': 783 /* Do interior commands. */ 784 fixuplabel(cp->u.c, cp->next); 785 break; 786 } 787 } 788 789 /* 790 * Associate the given command label for later lookup. 791 */ 792 static void 793 enterlabel(struct s_command *cp) 794 { 795 struct labhash **lhp, *lh; 796 u_char *p; 797 u_int h, c; 798 799 for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++) 800 h = (h << 5) + h + c; 801 lhp = &labels[h & LHMASK]; 802 for (lh = *lhp; lh != NULL; lh = lh->lh_next) 803 if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0) 804 errx(1, "%lu: %s: duplicate label '%s'", linenum, fname, cp->t); 805 if ((lh = malloc(sizeof *lh)) == NULL) 806 err(1, "malloc"); 807 lh->lh_next = *lhp; 808 lh->lh_hash = h; 809 lh->lh_cmd = cp; 810 lh->lh_ref = 0; 811 *lhp = lh; 812 } 813 814 /* 815 * Find the label contained in the command l in the command linked 816 * list cp. L is excluded from the search. Return NULL if not found. 817 */ 818 static struct s_command * 819 findlabel(char *name) 820 { 821 struct labhash *lh; 822 u_char *p; 823 u_int h, c; 824 825 for (h = 0, p = (u_char *)name; (c = *p) != 0; p++) 826 h = (h << 5) + h + c; 827 for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) { 828 if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) { 829 lh->lh_ref = 1; 830 return (lh->lh_cmd); 831 } 832 } 833 return (NULL); 834 } 835 836 /* 837 * Warn about any unused labels. As a side effect, release the label hash 838 * table space. 839 */ 840 static void 841 uselabel(void) 842 { 843 struct labhash *lh, *next; 844 int i; 845 846 for (i = 0; i < LHSZ; i++) { 847 for (lh = labels[i]; lh != NULL; lh = next) { 848 next = lh->lh_next; 849 if (!lh->lh_ref) 850 warnx("%lu: %s: unused label '%s'", 851 linenum, fname, lh->lh_cmd->t); 852 free(lh); 853 } 854 } 855 } 856