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() 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) { 176 if (*p == '#' || *p == '\0') 177 continue; 178 else if (*p == ';') { 179 p++; 180 goto semicolon; 181 } 182 } 183 if ((*link = cmd = malloc(sizeof(struct s_command))) == NULL) 184 err(1, "malloc"); 185 link = &cmd->next; 186 cmd->nonsel = cmd->inrange = 0; 187 /* First parse the addresses */ 188 naddr = 0; 189 190 /* Valid characters to start an address */ 191 #define addrchar(c) (strchr("0123456789/\\$", (c))) 192 if (addrchar(*p)) { 193 naddr++; 194 if ((cmd->a1 = malloc(sizeof(struct s_addr))) == NULL) 195 err(1, "malloc"); 196 p = compile_addr(p, cmd->a1); 197 EATSPACE(); /* EXTENSION */ 198 if (*p == ',') { 199 p++; 200 EATSPACE(); /* EXTENSION */ 201 naddr++; 202 if ((cmd->a2 = malloc(sizeof(struct s_addr))) 203 == NULL) 204 err(1, "malloc"); 205 p = compile_addr(p, cmd->a2); 206 EATSPACE(); 207 } else 208 cmd->a2 = 0; 209 } else 210 cmd->a1 = cmd->a2 = 0; 211 212 nonsel: /* Now parse the command */ 213 if (!*p) 214 errx(1, "%lu: %s: command expected", linenum, fname); 215 cmd->code = *p; 216 for (fp = cmd_fmts; fp->code; fp++) 217 if (fp->code == *p) 218 break; 219 if (!fp->code) 220 errx(1, "%lu: %s: invalid command code %c", linenum, fname, *p); 221 if (naddr > fp->naddr) 222 errx(1, 223 "%lu: %s: command %c expects up to %d address(es), found %d", 224 linenum, fname, *p, fp->naddr, naddr); 225 switch (fp->args) { 226 case NONSEL: /* ! */ 227 p++; 228 EATSPACE(); 229 cmd->nonsel = ! cmd->nonsel; 230 goto nonsel; 231 case GROUP: /* { */ 232 p++; 233 EATSPACE(); 234 cmd->next = stack; 235 stack = cmd; 236 link = &cmd->u.c; 237 if (*p) 238 goto semicolon; 239 break; 240 case ENDGROUP: 241 /* 242 * Short-circuit command processing, since end of 243 * group is really just a noop. 244 */ 245 cmd->nonsel = 1; 246 if (stack == 0) 247 errx(1, "%lu: %s: unexpected }", linenum, fname); 248 cmd2 = stack; 249 stack = cmd2->next; 250 cmd2->next = cmd; 251 /*FALLTHROUGH*/ 252 case EMPTY: /* d D g G h H l n N p P q x = \0 */ 253 p++; 254 EATSPACE(); 255 if (*p == ';') { 256 p++; 257 link = &cmd->next; 258 goto semicolon; 259 } 260 if (*p) 261 errx(1, "%lu: %s: extra characters at the end of %c command", 262 linenum, fname, cmd->code); 263 break; 264 case TEXT: /* a c i */ 265 p++; 266 EATSPACE(); 267 if (*p != '\\') 268 errx(1, 269 "%lu: %s: command %c expects \\ followed by text", linenum, fname, cmd->code); 270 p++; 271 EATSPACE(); 272 if (*p) 273 errx(1, 274 "%lu: %s: extra characters after \\ at the end of %c command", 275 linenum, fname, cmd->code); 276 cmd->t = compile_text(); 277 break; 278 case COMMENT: /* \0 # */ 279 break; 280 case WFILE: /* w */ 281 p++; 282 EATSPACE(); 283 if (*p == '\0') 284 errx(1, "%lu: %s: filename expected", linenum, fname); 285 cmd->t = duptoeol(p, "w command"); 286 if (aflag) 287 cmd->u.fd = -1; 288 else if ((cmd->u.fd = open(p, 289 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, 290 DEFFILEMODE)) == -1) 291 err(1, "%s", p); 292 break; 293 case RFILE: /* r */ 294 p++; 295 EATSPACE(); 296 if (*p == '\0') 297 errx(1, "%lu: %s: filename expected", linenum, fname); 298 else 299 cmd->t = duptoeol(p, "read command"); 300 break; 301 case BRANCH: /* b t */ 302 p++; 303 EATSPACE(); 304 if (*p == '\0') 305 cmd->t = NULL; 306 else 307 cmd->t = duptoeol(p, "branch"); 308 break; 309 case LABEL: /* : */ 310 p++; 311 EATSPACE(); 312 cmd->t = duptoeol(p, "label"); 313 if (strlen(p) == 0) 314 errx(1, "%lu: %s: empty label", linenum, fname); 315 enterlabel(cmd); 316 break; 317 case SUBST: /* s */ 318 p++; 319 if (*p == '\0' || *p == '\\') 320 errx(1, 321 "%lu: %s: substitute pattern can not be delimited by newline or backslash", 322 linenum, fname); 323 if ((cmd->u.s = malloc(sizeof(struct s_subst))) == NULL) 324 err(1, "malloc"); 325 p = compile_re(p, &cmd->u.s->re); 326 if (p == NULL) 327 errx(1, 328 "%lu: %s: unterminated substitute pattern", linenum, fname); 329 --p; 330 p = compile_subst(p, cmd->u.s); 331 p = compile_flags(p, cmd->u.s); 332 EATSPACE(); 333 if (*p == ';') { 334 p++; 335 link = &cmd->next; 336 goto semicolon; 337 } 338 break; 339 case TR: /* y */ 340 p++; 341 p = compile_tr(p, (char **)&cmd->u.y); 342 EATSPACE(); 343 if (*p == ';') { 344 p++; 345 link = &cmd->next; 346 goto semicolon; 347 } 348 if (*p) 349 errx(1, 350 "%lu: %s: extra text at the end of a transform command", linenum, fname); 351 break; 352 } 353 } 354 } 355 356 /* 357 * Get a delimited string. P points to the delimeter of the string; d points 358 * to a buffer area. Newline and delimiter escapes are processed; other 359 * escapes are ignored. 360 * 361 * Returns a pointer to the first character after the final delimiter or NULL 362 * in the case of a non-terminated string. The character array d is filled 363 * with the processed string. 364 */ 365 static char * 366 compile_delimited(p, d) 367 char *p, *d; 368 { 369 char c; 370 371 c = *p++; 372 if (c == '\0') 373 return (NULL); 374 else if (c == '\\') 375 errx(1, "%lu: %s: \\ can not be used as a string delimiter", 376 linenum, fname); 377 else if (c == '\n') 378 errx(1, "%lu: %s: newline can not be used as a string delimiter", 379 linenum, fname); 380 while (*p) { 381 if (*p == '[') { 382 if ((d = compile_ccl(&p, d)) == NULL) 383 errx(1, "%lu: %s: unbalanced brackets ([])", linenum, fname); 384 continue; 385 } else if (*p == '\\' && p[1] == '[') { 386 *d++ = *p++; 387 } else if (*p == '\\' && p[1] == c) 388 p++; 389 else if (*p == '\\' && p[1] == 'n') { 390 *d++ = '\n'; 391 p += 2; 392 continue; 393 } else if (*p == '\\' && p[1] == '\\') 394 *d++ = *p++; 395 else if (*p == c) { 396 *d = '\0'; 397 return (p + 1); 398 } 399 *d++ = *p++; 400 } 401 return (NULL); 402 } 403 404 405 /* compile_ccl: expand a POSIX character class */ 406 static char * 407 compile_ccl(sp, t) 408 char **sp; 409 char *t; 410 { 411 int c, d; 412 char *s = *sp; 413 414 *t++ = *s++; 415 if (*s == '^') 416 *t++ = *s++; 417 if (*s == ']') 418 *t++ = *s++; 419 for (; *s && (*t = *s) != ']'; s++, t++) 420 if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) { 421 *++t = *++s, t++, s++; 422 for (c = *s; (*t = *s) != ']' || c != d; s++, t++) 423 if ((c = *s) == '\0') 424 return NULL; 425 } else if (*s == '\\' && s[1] == 'n') 426 *t = '\n', s++; 427 return (*s == ']') ? *sp = ++s, ++t : NULL; 428 } 429 430 /* 431 * Get a regular expression. P points to the delimiter of the regular 432 * expression; repp points to the address of a regexp pointer. Newline 433 * and delimiter escapes are processed; other escapes are ignored. 434 * Returns a pointer to the first character after the final delimiter 435 * or NULL in the case of a non terminated regular expression. The regexp 436 * pointer is set to the compiled regular expression. 437 * Cflags are passed to regcomp. 438 */ 439 static char * 440 compile_re(p, repp) 441 char *p; 442 regex_t **repp; 443 { 444 int eval; 445 char re[_POSIX2_LINE_MAX + 1]; 446 447 p = compile_delimited(p, re); 448 if (p && strlen(re) == 0) { 449 *repp = NULL; 450 return (p); 451 } 452 if ((*repp = malloc(sizeof(regex_t))) == NULL) 453 err(1, "malloc"); 454 if (p && (eval = regcomp(*repp, re, rflags)) != 0) 455 errx(1, "%lu: %s: RE error: %s", 456 linenum, fname, strregerror(eval, *repp)); 457 if (maxnsub < (*repp)->re_nsub) 458 maxnsub = (*repp)->re_nsub; 459 return (p); 460 } 461 462 /* 463 * Compile the substitution string of a regular expression and set res to 464 * point to a saved copy of it. Nsub is the number of parenthesized regular 465 * expressions. 466 */ 467 static char * 468 compile_subst(p, s) 469 char *p; 470 struct s_subst *s; 471 { 472 static char lbuf[_POSIX2_LINE_MAX + 1]; 473 int asize, size; 474 u_char ref; 475 char c, *text, *op, *sp; 476 int more = 1, sawesc = 0; 477 478 c = *p++; /* Terminator character */ 479 if (c == '\0') 480 return (NULL); 481 482 s->maxbref = 0; 483 s->linenum = linenum; 484 asize = 2 * _POSIX2_LINE_MAX + 1; 485 if ((text = malloc(asize)) == NULL) 486 err(1, "malloc"); 487 size = 0; 488 do { 489 op = sp = text + size; 490 for (; *p; p++) { 491 if (*p == '\\' || sawesc) { 492 /* 493 * If this is a continuation from the last 494 * buffer, we won't have a character to 495 * skip over. 496 */ 497 if (sawesc) 498 sawesc = 0; 499 else 500 p++; 501 502 if (*p == '\0') { 503 /* 504 * This escaped character is continued 505 * in the next part of the line. Note 506 * this fact, then cause the loop to 507 * exit w/ normal EOL case and reenter 508 * above with the new buffer. 509 */ 510 sawesc = 1; 511 p--; 512 continue; 513 } else if (strchr("123456789", *p) != NULL) { 514 *sp++ = '\\'; 515 ref = *p - '0'; 516 if (s->re != NULL && 517 ref > s->re->re_nsub) 518 errx(1, "%lu: %s: \\%c not defined in the RE", 519 linenum, fname, *p); 520 if (s->maxbref < ref) 521 s->maxbref = ref; 522 } else if (*p == '&' || *p == '\\') 523 *sp++ = '\\'; 524 } else if (*p == c) { 525 if (*++p == '\0' && more) { 526 if (cu_fgets(lbuf, sizeof(lbuf), &more)) 527 p = lbuf; 528 } 529 *sp++ = '\0'; 530 size += sp - op; 531 if ((s->new = realloc(text, size)) == NULL) 532 err(1, "realloc"); 533 return (p); 534 } else if (*p == '\n') { 535 errx(1, 536 "%lu: %s: unescaped newline inside substitute pattern", linenum, fname); 537 /* NOTREACHED */ 538 } 539 *sp++ = *p; 540 } 541 size += sp - op; 542 if (asize - size < _POSIX2_LINE_MAX + 1) { 543 asize *= 2; 544 if ((text = realloc(text, asize)) == NULL) 545 err(1, "realloc"); 546 } 547 } while (cu_fgets(p = lbuf, sizeof(lbuf), &more)); 548 errx(1, "%lu: %s: unterminated substitute in regular expression", 549 linenum, fname); 550 /* NOTREACHED */ 551 } 552 553 /* 554 * Compile the flags of the s command 555 */ 556 static char * 557 compile_flags(p, s) 558 char *p; 559 struct s_subst *s; 560 { 561 int gn; /* True if we have seen g or n */ 562 char wfile[_POSIX2_LINE_MAX + 1], *q; 563 564 s->n = 1; /* Default */ 565 s->p = 0; 566 s->wfile = NULL; 567 s->wfd = -1; 568 for (gn = 0;;) { 569 EATSPACE(); /* EXTENSION */ 570 switch (*p) { 571 case 'g': 572 if (gn) 573 errx(1, 574 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname); 575 gn = 1; 576 s->n = 0; 577 break; 578 case '\0': 579 case '\n': 580 case ';': 581 return (p); 582 case 'p': 583 s->p = 1; 584 break; 585 case '1': case '2': case '3': 586 case '4': case '5': case '6': 587 case '7': case '8': case '9': 588 if (gn) 589 errx(1, 590 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname); 591 gn = 1; 592 /* XXX Check for overflow */ 593 s->n = (int)strtol(p, &p, 10); 594 break; 595 case 'w': 596 p++; 597 #ifdef HISTORIC_PRACTICE 598 if (*p != ' ') { 599 warnx("%lu: %s: space missing before w wfile", linenum, fname); 600 return (p); 601 } 602 #endif 603 EATSPACE(); 604 q = wfile; 605 while (*p) { 606 if (*p == '\n') 607 break; 608 *q++ = *p++; 609 } 610 *q = '\0'; 611 if (q == wfile) 612 errx(1, "%lu: %s: no wfile specified", linenum, fname); 613 s->wfile = strdup(wfile); 614 if (!aflag && (s->wfd = open(wfile, 615 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, 616 DEFFILEMODE)) == -1) 617 err(1, "%s", wfile); 618 return (p); 619 default: 620 errx(1, "%lu: %s: bad flag in substitute command: '%c'", 621 linenum, fname, *p); 622 break; 623 } 624 p++; 625 } 626 } 627 628 /* 629 * Compile a translation set of strings into a lookup table. 630 */ 631 static char * 632 compile_tr(p, transtab) 633 char *p; 634 char **transtab; 635 { 636 int i; 637 char *lt, *op, *np; 638 char old[_POSIX2_LINE_MAX + 1]; 639 char new[_POSIX2_LINE_MAX + 1]; 640 641 if (*p == '\0' || *p == '\\') 642 errx(1, 643 "%lu: %s: transform pattern can not be delimited by newline or backslash", 644 linenum, fname); 645 p = compile_delimited(p, old); 646 if (p == NULL) 647 errx(1, "%lu: %s: unterminated transform source string", 648 linenum, fname); 649 p = compile_delimited(--p, new); 650 if (p == NULL) 651 errx(1, "%lu: %s: unterminated transform target string", 652 linenum, fname); 653 EATSPACE(); 654 if (strlen(new) != strlen(old)) 655 errx(1, "%lu: %s: transform strings are not the same length", 656 linenum, fname); 657 /* We assume characters are 8 bits */ 658 if ((lt = malloc(UCHAR_MAX)) == NULL) 659 err(1, "malloc"); 660 for (i = 0; i <= UCHAR_MAX; i++) 661 lt[i] = (char)i; 662 for (op = old, np = new; *op; op++, np++) 663 lt[(u_char)*op] = *np; 664 *transtab = lt; 665 return (p); 666 } 667 668 /* 669 * Compile the text following an a or i command. 670 */ 671 static char * 672 compile_text() 673 { 674 int asize, esc_nl, size; 675 char *text, *p, *op, *s; 676 char lbuf[_POSIX2_LINE_MAX + 1]; 677 678 asize = 2 * _POSIX2_LINE_MAX + 1; 679 if ((text = malloc(asize)) == NULL) 680 err(1, "malloc"); 681 size = 0; 682 while (cu_fgets(lbuf, sizeof(lbuf), NULL)) { 683 op = s = text + size; 684 p = lbuf; 685 EATSPACE(); 686 for (esc_nl = 0; *p != '\0'; p++) { 687 if (*p == '\\' && p[1] != '\0' && *++p == '\n') 688 esc_nl = 1; 689 *s++ = *p; 690 } 691 size += s - op; 692 if (!esc_nl) { 693 *s = '\0'; 694 break; 695 } 696 if (asize - size < _POSIX2_LINE_MAX + 1) { 697 asize *= 2; 698 if ((text = realloc(text, asize)) == NULL) 699 err(1, "realloc"); 700 } 701 } 702 text[size] = '\0'; 703 if ((p = realloc(text, size + 1)) == NULL) 704 err(1, "realloc"); 705 return (p); 706 } 707 708 /* 709 * Get an address and return a pointer to the first character after 710 * it. Fill the structure pointed to according to the address. 711 */ 712 static char * 713 compile_addr(p, a) 714 char *p; 715 struct s_addr *a; 716 { 717 char *end; 718 719 switch (*p) { 720 case '\\': /* Context address */ 721 ++p; 722 /* FALLTHROUGH */ 723 case '/': /* Context address */ 724 p = compile_re(p, &a->u.r); 725 if (p == NULL) 726 errx(1, "%lu: %s: unterminated regular expression", linenum, fname); 727 a->type = AT_RE; 728 return (p); 729 730 case '$': /* Last line */ 731 a->type = AT_LAST; 732 return (p + 1); 733 /* Line number */ 734 case '0': case '1': case '2': case '3': case '4': 735 case '5': case '6': case '7': case '8': case '9': 736 a->type = AT_LINE; 737 a->u.l = strtol(p, &end, 10); 738 return (end); 739 default: 740 errx(1, "%lu: %s: expected context address", linenum, fname); 741 return (NULL); 742 } 743 } 744 745 /* 746 * duptoeol -- 747 * Return a copy of all the characters up to \n or \0. 748 */ 749 static char * 750 duptoeol(s, ctype) 751 char *s; 752 const char *ctype; 753 { 754 size_t len; 755 int ws; 756 char *p, *start; 757 758 ws = 0; 759 for (start = s; *s != '\0' && *s != '\n'; ++s) 760 ws = isspace((unsigned char)*s); 761 *s = '\0'; 762 if (ws) 763 warnx("%lu: %s: whitespace after %s", linenum, fname, ctype); 764 len = s - start + 1; 765 if ((p = malloc(len)) == NULL) 766 err(1, "malloc"); 767 return (memmove(p, start, len)); 768 } 769 770 /* 771 * Convert goto label names to addresses, and count a and r commands, in 772 * the given subset of the script. Free the memory used by labels in b 773 * and t commands (but not by :). 774 * 775 * TODO: Remove } nodes 776 */ 777 static void 778 fixuplabel(cp, end) 779 struct s_command *cp, *end; 780 { 781 782 for (; cp != end; cp = cp->next) 783 switch (cp->code) { 784 case 'a': 785 case 'r': 786 appendnum++; 787 break; 788 case 'b': 789 case 't': 790 /* Resolve branch target. */ 791 if (cp->t == NULL) { 792 cp->u.c = NULL; 793 break; 794 } 795 if ((cp->u.c = findlabel(cp->t)) == NULL) 796 errx(1, "%lu: %s: undefined label '%s'", linenum, fname, cp->t); 797 free(cp->t); 798 break; 799 case '{': 800 /* Do interior commands. */ 801 fixuplabel(cp->u.c, cp->next); 802 break; 803 } 804 } 805 806 /* 807 * Associate the given command label for later lookup. 808 */ 809 static void 810 enterlabel(cp) 811 struct s_command *cp; 812 { 813 struct labhash **lhp, *lh; 814 u_char *p; 815 u_int h, c; 816 817 for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++) 818 h = (h << 5) + h + c; 819 lhp = &labels[h & LHMASK]; 820 for (lh = *lhp; lh != NULL; lh = lh->lh_next) 821 if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0) 822 errx(1, "%lu: %s: duplicate label '%s'", linenum, fname, cp->t); 823 if ((lh = malloc(sizeof *lh)) == NULL) 824 err(1, "malloc"); 825 lh->lh_next = *lhp; 826 lh->lh_hash = h; 827 lh->lh_cmd = cp; 828 lh->lh_ref = 0; 829 *lhp = lh; 830 } 831 832 /* 833 * Find the label contained in the command l in the command linked 834 * list cp. L is excluded from the search. Return NULL if not found. 835 */ 836 static struct s_command * 837 findlabel(name) 838 char *name; 839 { 840 struct labhash *lh; 841 u_char *p; 842 u_int h, c; 843 844 for (h = 0, p = (u_char *)name; (c = *p) != 0; p++) 845 h = (h << 5) + h + c; 846 for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) { 847 if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) { 848 lh->lh_ref = 1; 849 return (lh->lh_cmd); 850 } 851 } 852 return (NULL); 853 } 854 855 /* 856 * Warn about any unused labels. As a side effect, release the label hash 857 * table space. 858 */ 859 static void 860 uselabel() 861 { 862 struct labhash *lh, *next; 863 int i; 864 865 for (i = 0; i < LHSZ; i++) { 866 for (lh = labels[i]; lh != NULL; lh = next) { 867 next = lh->lh_next; 868 if (!lh->lh_ref) 869 warnx("%lu: %s: unused label '%s'", 870 linenum, fname, lh->lh_cmd->t); 871 free(lh); 872 } 873 } 874 } 875