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 && (*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 = 1, sawesc = 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 == '\\' || sawesc) { 486 /* 487 * If this is a continuation from the last 488 * buffer, we won't have a character to 489 * skip over. 490 */ 491 if (sawesc) 492 sawesc = 0; 493 else 494 p++; 495 496 if (*p == '\0') { 497 /* 498 * This escaped character is continued 499 * in the next part of the line. Note 500 * this fact, then cause the loop to 501 * exit w/ normal EOL case and reenter 502 * above with the new buffer. 503 */ 504 sawesc = 1; 505 p--; 506 continue; 507 } else if (strchr("123456789", *p) != NULL) { 508 *sp++ = '\\'; 509 ref = *p - '0'; 510 if (s->re != NULL && 511 ref > s->re->re_nsub) 512 errx(1, "%lu: %s: \\%c not defined in the RE", 513 linenum, fname, *p); 514 if (s->maxbref < ref) 515 s->maxbref = ref; 516 } else if (*p == '&' || *p == '\\') 517 *sp++ = '\\'; 518 } else if (*p == c) { 519 if (*++p == '\0' && more) { 520 if (cu_fgets(lbuf, sizeof(lbuf), &more)) 521 p = lbuf; 522 } 523 *sp++ = '\0'; 524 size += sp - op; 525 if ((s->new = realloc(text, size)) == NULL) 526 err(1, "realloc"); 527 return (p); 528 } else if (*p == '\n') { 529 errx(1, 530 "%lu: %s: unescaped newline inside substitute pattern", linenum, fname); 531 /* NOTREACHED */ 532 } 533 *sp++ = *p; 534 } 535 size += sp - op; 536 if (asize - size < _POSIX2_LINE_MAX + 1) { 537 asize *= 2; 538 if ((text = realloc(text, asize)) == NULL) 539 err(1, "realloc"); 540 } 541 } while (cu_fgets(p = lbuf, sizeof(lbuf), &more)); 542 errx(1, "%lu: %s: unterminated substitute in regular expression", 543 linenum, fname); 544 /* NOTREACHED */ 545 } 546 547 /* 548 * Compile the flags of the s command 549 */ 550 static char * 551 compile_flags(p, s) 552 char *p; 553 struct s_subst *s; 554 { 555 int gn; /* True if we have seen g or n */ 556 char wfile[_POSIX2_LINE_MAX + 1], *q; 557 558 s->n = 1; /* Default */ 559 s->p = 0; 560 s->wfile = NULL; 561 s->wfd = -1; 562 for (gn = 0;;) { 563 EATSPACE(); /* EXTENSION */ 564 switch (*p) { 565 case 'g': 566 if (gn) 567 errx(1, 568 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname); 569 gn = 1; 570 s->n = 0; 571 break; 572 case '\0': 573 case '\n': 574 case ';': 575 return (p); 576 case 'p': 577 s->p = 1; 578 break; 579 case '1': case '2': case '3': 580 case '4': case '5': case '6': 581 case '7': case '8': case '9': 582 if (gn) 583 errx(1, 584 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname); 585 gn = 1; 586 /* XXX Check for overflow */ 587 s->n = (int)strtol(p, &p, 10); 588 break; 589 case 'w': 590 p++; 591 #ifdef HISTORIC_PRACTICE 592 if (*p != ' ') { 593 warnx("%lu: %s: space missing before w wfile", linenum, fname); 594 return (p); 595 } 596 #endif 597 EATSPACE(); 598 q = wfile; 599 while (*p) { 600 if (*p == '\n') 601 break; 602 *q++ = *p++; 603 } 604 *q = '\0'; 605 if (q == wfile) 606 errx(1, "%lu: %s: no wfile specified", linenum, fname); 607 s->wfile = strdup(wfile); 608 if (!aflag && (s->wfd = open(wfile, 609 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, 610 DEFFILEMODE)) == -1) 611 err(1, "%s", wfile); 612 return (p); 613 default: 614 errx(1, "%lu: %s: bad flag in substitute command: '%c'", 615 linenum, fname, *p); 616 break; 617 } 618 p++; 619 } 620 } 621 622 /* 623 * Compile a translation set of strings into a lookup table. 624 */ 625 static char * 626 compile_tr(p, transtab) 627 char *p; 628 char **transtab; 629 { 630 int i; 631 char *lt, *op, *np; 632 char old[_POSIX2_LINE_MAX + 1]; 633 char new[_POSIX2_LINE_MAX + 1]; 634 635 if (*p == '\0' || *p == '\\') 636 errx(1, 637 "%lu: %s: transform pattern can not be delimited by newline or backslash", 638 linenum, fname); 639 p = compile_delimited(p, old); 640 if (p == NULL) 641 errx(1, "%lu: %s: unterminated transform source string", 642 linenum, fname); 643 p = compile_delimited(--p, new); 644 if (p == NULL) 645 errx(1, "%lu: %s: unterminated transform target string", 646 linenum, fname); 647 EATSPACE(); 648 if (strlen(new) != strlen(old)) 649 errx(1, "%lu: %s: transform strings are not the same length", 650 linenum, fname); 651 /* We assume characters are 8 bits */ 652 if ((lt = malloc(UCHAR_MAX)) == NULL) 653 err(1, "malloc"); 654 for (i = 0; i <= UCHAR_MAX; i++) 655 lt[i] = (char)i; 656 for (op = old, np = new; *op; op++, np++) 657 lt[(u_char)*op] = *np; 658 *transtab = lt; 659 return (p); 660 } 661 662 /* 663 * Compile the text following an a or i command. 664 */ 665 static char * 666 compile_text() 667 { 668 int asize, esc_nl, size; 669 char *text, *p, *op, *s; 670 char lbuf[_POSIX2_LINE_MAX + 1]; 671 672 asize = 2 * _POSIX2_LINE_MAX + 1; 673 if ((text = malloc(asize)) == NULL) 674 err(1, "malloc"); 675 size = 0; 676 while (cu_fgets(lbuf, sizeof(lbuf), NULL)) { 677 op = s = text + size; 678 p = lbuf; 679 EATSPACE(); 680 for (esc_nl = 0; *p != '\0'; p++) { 681 if (*p == '\\' && p[1] != '\0' && *++p == '\n') 682 esc_nl = 1; 683 *s++ = *p; 684 } 685 size += s - op; 686 if (!esc_nl) { 687 *s = '\0'; 688 break; 689 } 690 if (asize - size < _POSIX2_LINE_MAX + 1) { 691 asize *= 2; 692 if ((text = realloc(text, asize)) == NULL) 693 err(1, "realloc"); 694 } 695 } 696 text[size] = '\0'; 697 if ((p = realloc(text, size + 1)) == NULL) 698 err(1, "realloc"); 699 return (p); 700 } 701 702 /* 703 * Get an address and return a pointer to the first character after 704 * it. Fill the structure pointed to according to the address. 705 */ 706 static char * 707 compile_addr(p, a) 708 char *p; 709 struct s_addr *a; 710 { 711 char *end; 712 713 switch (*p) { 714 case '\\': /* Context address */ 715 ++p; 716 /* FALLTHROUGH */ 717 case '/': /* Context address */ 718 p = compile_re(p, &a->u.r); 719 if (p == NULL) 720 errx(1, "%lu: %s: unterminated regular expression", linenum, fname); 721 a->type = AT_RE; 722 return (p); 723 724 case '$': /* Last line */ 725 a->type = AT_LAST; 726 return (p + 1); 727 /* Line number */ 728 case '0': case '1': case '2': case '3': case '4': 729 case '5': case '6': case '7': case '8': case '9': 730 a->type = AT_LINE; 731 a->u.l = strtol(p, &end, 10); 732 return (end); 733 default: 734 errx(1, "%lu: %s: expected context address", linenum, fname); 735 return (NULL); 736 } 737 } 738 739 /* 740 * duptoeol -- 741 * Return a copy of all the characters up to \n or \0. 742 */ 743 static char * 744 duptoeol(s, ctype) 745 char *s; 746 const char *ctype; 747 { 748 size_t len; 749 int ws; 750 char *p, *start; 751 752 ws = 0; 753 for (start = s; *s != '\0' && *s != '\n'; ++s) 754 ws = isspace((unsigned char)*s); 755 *s = '\0'; 756 if (ws) 757 warnx("%lu: %s: whitespace after %s", linenum, fname, ctype); 758 len = s - start + 1; 759 if ((p = malloc(len)) == NULL) 760 err(1, "malloc"); 761 return (memmove(p, start, len)); 762 } 763 764 /* 765 * Convert goto label names to addresses, and count a and r commands, in 766 * the given subset of the script. Free the memory used by labels in b 767 * and t commands (but not by :). 768 * 769 * TODO: Remove } nodes 770 */ 771 static void 772 fixuplabel(cp, end) 773 struct s_command *cp, *end; 774 { 775 776 for (; cp != end; cp = cp->next) 777 switch (cp->code) { 778 case 'a': 779 case 'r': 780 appendnum++; 781 break; 782 case 'b': 783 case 't': 784 /* Resolve branch target. */ 785 if (cp->t == NULL) { 786 cp->u.c = NULL; 787 break; 788 } 789 if ((cp->u.c = findlabel(cp->t)) == NULL) 790 errx(1, "%lu: %s: undefined label '%s'", linenum, fname, cp->t); 791 free(cp->t); 792 break; 793 case '{': 794 /* Do interior commands. */ 795 fixuplabel(cp->u.c, cp->next); 796 break; 797 } 798 } 799 800 /* 801 * Associate the given command label for later lookup. 802 */ 803 static void 804 enterlabel(cp) 805 struct s_command *cp; 806 { 807 struct labhash **lhp, *lh; 808 u_char *p; 809 u_int h, c; 810 811 for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++) 812 h = (h << 5) + h + c; 813 lhp = &labels[h & LHMASK]; 814 for (lh = *lhp; lh != NULL; lh = lh->lh_next) 815 if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0) 816 errx(1, "%lu: %s: duplicate label '%s'", linenum, fname, cp->t); 817 if ((lh = malloc(sizeof *lh)) == NULL) 818 err(1, "malloc"); 819 lh->lh_next = *lhp; 820 lh->lh_hash = h; 821 lh->lh_cmd = cp; 822 lh->lh_ref = 0; 823 *lhp = lh; 824 } 825 826 /* 827 * Find the label contained in the command l in the command linked 828 * list cp. L is excluded from the search. Return NULL if not found. 829 */ 830 static struct s_command * 831 findlabel(name) 832 char *name; 833 { 834 struct labhash *lh; 835 u_char *p; 836 u_int h, c; 837 838 for (h = 0, p = (u_char *)name; (c = *p) != 0; p++) 839 h = (h << 5) + h + c; 840 for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) { 841 if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) { 842 lh->lh_ref = 1; 843 return (lh->lh_cmd); 844 } 845 } 846 return (NULL); 847 } 848 849 /* 850 * Warn about any unused labels. As a side effect, release the label hash 851 * table space. 852 */ 853 static void 854 uselabel() 855 { 856 struct labhash *lh, *next; 857 int i; 858 859 for (i = 0; i < LHSZ; i++) { 860 for (lh = labels[i]; lh != NULL; lh = next) { 861 next = lh->lh_next; 862 if (!lh->lh_ref) 863 warnx("%lu: %s: unused label '%s'", 864 linenum, fname, lh->lh_cmd->t); 865 free(lh); 866 } 867 } 868 } 869