usr.bin/sed/compile.c

   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  * @(#)compile.c        8.1 (Berkeley) 6/6/93
  38  * $FreeBSD: src/usr.bin/sed/compile.c,v 1.13.2.8 2002/08/17 05:47:06 tjr Exp $
  39  * $DragonFly: src/usr.bin/sed/compile.c,v 1.3 2003/10/04 20:36:50 hmp Exp $
  40  */
  41
  42 #include <sys/types.h>
  43 #include <sys/stat.h>
  44
  45 #include <ctype.h>
  46 #include <err.h>
  47 #include <fcntl.h>
  48 #include <limits.h>
  49 #include <regex.h>
  50 #include <stdio.h>
  51 #include <stdlib.h>
  52 #include <string.h>
  53
  54 #include "defs.h"
  55 #include "extern.h"
  56
  57 #define LHSZ    128
  58 #define LHMASK  (LHSZ - 1)
  59 static struct labhash {
  60         struct  labhash *lh_next;
  61         u_int   lh_hash;
  62         struct  s_command *lh_cmd;
  63         int     lh_ref;
  64 } *labels[LHSZ];
  65
  66 static char      *compile_addr(char *, struct s_addr *);
  67 static char      *compile_ccl(char **, char *);
  68 static char      *compile_delimited(char *, char *);
  69 static char      *compile_flags(char *, struct s_subst *);
  70 static char      *compile_re(char *, regex_t **);
  71 static char      *compile_subst(char *, struct s_subst *);
  72 static char      *compile_text(void);
  73 static char      *compile_tr(char *, char **);
  74 static struct s_command
  75                 **compile_stream(struct s_command **);
  76 static char      *duptoeol(char *, const char *);
  77 static void       enterlabel(struct s_command *);
  78 static struct s_command
  79                  *findlabel(char *);
  80 static void       fixuplabel(struct s_command *, struct s_command *);
  81 static void       uselabel(void);
  82
  83 /*
  84  * Command specification.  This is used to drive the command parser.
  85  */
  86 struct s_format {
  87         char code;                              /* Command code */
  88         int naddr;                              /* Number of address args */
  89         enum e_args args;                       /* Argument type */
  90 };
  91
  92 static struct s_format cmd_fmts[] = {
  93         {'{', 2, GROUP},
  94         {'}', 0, ENDGROUP},
  95         {'a', 1, TEXT},
  96         {'b', 2, BRANCH},
  97         {'c', 2, TEXT},
  98         {'d', 2, EMPTY},
  99         {'D', 2, EMPTY},
 100         {'g', 2, EMPTY},
 101         {'G', 2, EMPTY},
 102         {'h', 2, EMPTY},
 103         {'H', 2, EMPTY},
 104         {'i', 1, TEXT},
 105         {'l', 2, EMPTY},
 106         {'n', 2, EMPTY},
 107         {'N', 2, EMPTY},
 108         {'p', 2, EMPTY},
 109         {'P', 2, EMPTY},
 110         {'q', 1, EMPTY},
 111         {'r', 1, RFILE},
 112         {'s', 2, SUBST},
 113         {'t', 2, BRANCH},
 114         {'w', 2, WFILE},
 115         {'x', 2, EMPTY},
 116         {'y', 2, TR},
 117         {'!', 2, NONSEL},
 118         {':', 0, LABEL},
 119         {'#', 0, COMMENT},
 120         {'=', 1, EMPTY},
 121         {'\0', 0, COMMENT},
 122 };
 123
 124 /* The compiled program. */
 125 struct s_command *prog;
 126
 127 /*
 128  * Compile the program into prog.
 129  * Initialise appends.
 130  */
 131 void
 132 compile(void)
 133 {
 134         *compile_stream(&prog) = NULL;
 135         fixuplabel(prog, NULL);
 136         uselabel();
 137         if (appendnum == 0)
 138                 appends = NULL;
 139         else if ((appends = malloc(sizeof(struct s_appends) * appendnum)) ==
 140             NULL)
 141                 err(1, "malloc");
 142         if ((match = malloc((maxnsub + 1) * sizeof(regmatch_t))) == NULL)
 143                 err(1, "malloc");
 144 }
 145
 146 #define EATSPACE() do {                                                 \
 147         if (p)                                                          \
 148                 while (*p && isspace((unsigned char)*p))                \
 149                         p++;                                            \
 150         } while (0)
 151
 152 static struct s_command **
 153 compile_stream(struct s_command **link)
 154 {
 155         char *p;
 156         static char lbuf[_POSIX2_LINE_MAX + 1]; /* To save stack */
 157         struct s_command *cmd, *cmd2, *stack;
 158         struct s_format *fp;
 159         int naddr;                              /* Number of addresses */
 160
 161         stack = 0;
 162         for (;;) {
 163                 if ((p = cu_fgets(lbuf, sizeof(lbuf), NULL)) == NULL) {
 164                         if (stack != 0)
 165                                 errx(1, "%lu: %s: unexpected EOF (pending }'s)",
 166                                                         linenum, fname);
 167                         return (link);
 168                 }
 169
 170 semicolon:      EATSPACE();
 171                 if (p) {
 172                         if (*p == '#' || *p == '\0')
 173                                 continue;
 174                         else if (*p == ';') {
 175                                 p++;
 176                                 goto semicolon;
 177                         }
 178                 }
 179                 if ((*link = cmd = malloc(sizeof(struct s_command))) == NULL)
 180                         err(1, "malloc");
 181                 link = &cmd->next;
 182                 cmd->nonsel = cmd->inrange = 0;
 183                 /* First parse the addresses */
 184                 naddr = 0;
 185
 186 /* Valid characters to start an address */
 187 #define addrchar(c)     (strchr("0123456789/\\$", (c)))
 188                 if (addrchar(*p)) {
 189                         naddr++;
 190                         if ((cmd->a1 = malloc(sizeof(struct s_addr))) == NULL)
 191                                 err(1, "malloc");
 192                         p = compile_addr(p, cmd->a1);
 193                         EATSPACE();                             /* EXTENSION */
 194                         if (*p == ',') {
 195                                 p++;
 196                                 EATSPACE();                     /* EXTENSION */
 197                                 naddr++;
 198                                 if ((cmd->a2 = malloc(sizeof(struct s_addr)))
 199                                     == NULL)
 200                                         err(1, "malloc");
 201                                 p = compile_addr(p, cmd->a2);
 202                                 EATSPACE();
 203                         } else
 204                                 cmd->a2 = 0;
 205                 } else
 206                         cmd->a1 = cmd->a2 = 0;
 207
 208 nonsel:         /* Now parse the command */
 209                 if (!*p)
 210                         errx(1, "%lu: %s: command expected", linenum, fname);
 211                 cmd->code = *p;
 212                 for (fp = cmd_fmts; fp->code; fp++)
 213                         if (fp->code == *p)
 214                                 break;
 215                 if (!fp->code)
 216                         errx(1, "%lu: %s: invalid command code %c", linenum, fname, *p);
 217                 if (naddr > fp->naddr)
 218                         errx(1,
 219                                 "%lu: %s: command %c expects up to %d address(es), found %d",
 220                                 linenum, fname, *p, fp->naddr, naddr);
 221                 switch (fp->args) {
 222                 case NONSEL:                    /* ! */
 223                         p++;
 224                         EATSPACE();
 225                         cmd->nonsel = ! cmd->nonsel;
 226                         goto nonsel;
 227                 case GROUP:                     /* { */
 228                         p++;
 229                         EATSPACE();
 230                         cmd->next = stack;
 231                         stack = cmd;
 232                         link = &cmd->u.c;
 233                         if (*p)
 234                                 goto semicolon;
 235                         break;
 236                 case ENDGROUP:
 237                         /*
 238                          * Short-circuit command processing, since end of
 239                          * group is really just a noop.
 240                          */
 241                         cmd->nonsel = 1;
 242                         if (stack == 0)
 243                                 errx(1, "%lu: %s: unexpected }", linenum, fname);
 244                         cmd2 = stack;
 245                         stack = cmd2->next;
 246                         cmd2->next = cmd;
 247                         /*FALLTHROUGH*/
 248                 case EMPTY:             /* d D g G h H l n N p P q x = \0 */
 249                         p++;
 250                         EATSPACE();
 251                         if (*p == ';') {
 252                                 p++;
 253                                 link = &cmd->next;
 254                                 goto semicolon;
 255                         }
 256                         if (*p)
 257                                 errx(1, "%lu: %s: extra characters at the end of %c command",
 258                                                 linenum, fname, cmd->code);
 259                         break;
 260                 case TEXT:                      /* a c i */
 261                         p++;
 262                         EATSPACE();
 263                         if (*p != '\\')
 264                                 errx(1,
 265 "%lu: %s: command %c expects \\ followed by text", linenum, fname, cmd->code);
 266                         p++;
 267                         EATSPACE();
 268                         if (*p)
 269                                 errx(1,
 270                                 "%lu: %s: extra characters after \\ at the end of %c command",
 271                                 linenum, fname, cmd->code);
 272                         cmd->t = compile_text();
 273                         break;
 274                 case COMMENT:                   /* \0 # */
 275                         break;
 276                 case WFILE:                     /* w */
 277                         p++;
 278                         EATSPACE();
 279                         if (*p == '\0')
 280                                 errx(1, "%lu: %s: filename expected", linenum, fname);
 281                         cmd->t = duptoeol(p, "w command");
 282                         if (aflag)
 283                                 cmd->u.fd = -1;
 284                         else if ((cmd->u.fd = open(p,
 285                             O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
 286                             DEFFILEMODE)) == -1)
 287                                 err(1, "%s", p);
 288                         break;
 289                 case RFILE:                     /* r */
 290                         p++;
 291                         EATSPACE();
 292                         if (*p == '\0')
 293                                 errx(1, "%lu: %s: filename expected", linenum, fname);
 294                         else
 295                                 cmd->t = duptoeol(p, "read command");
 296                         break;
 297                 case BRANCH:                    /* b t */
 298                         p++;
 299                         EATSPACE();
 300                         if (*p == '\0')
 301                                 cmd->t = NULL;
 302                         else
 303                                 cmd->t = duptoeol(p, "branch");
 304                         break;
 305                 case LABEL:                     /* : */
 306                         p++;
 307                         EATSPACE();
 308                         cmd->t = duptoeol(p, "label");
 309                         if (strlen(p) == 0)
 310                                 errx(1, "%lu: %s: empty label", linenum, fname);
 311                         enterlabel(cmd);
 312                         break;
 313                 case SUBST:                     /* s */
 314                         p++;
 315                         if (*p == '\0' || *p == '\\')
 316                                 errx(1,
 317 "%lu: %s: substitute pattern can not be delimited by newline or backslash",
 318                                         linenum, fname);
 319                         if ((cmd->u.s = malloc(sizeof(struct s_subst))) == NULL)
 320                                 err(1, "malloc");
 321                         p = compile_re(p, &cmd->u.s->re);
 322                         if (p == NULL)
 323                                 errx(1,
 324                                 "%lu: %s: unterminated substitute pattern", linenum, fname);
 325                         --p;
 326                         p = compile_subst(p, cmd->u.s);
 327                         p = compile_flags(p, cmd->u.s);
 328                         EATSPACE();
 329                         if (*p == ';') {
 330                                 p++;
 331                                 link = &cmd->next;
 332                                 goto semicolon;
 333                         }
 334                         break;
 335                 case TR:                        /* y */
 336                         p++;
 337                         p = compile_tr(p, (char **)&cmd->u.y);
 338                         EATSPACE();
 339                         if (*p == ';') {
 340                                 p++;
 341                                 link = &cmd->next;
 342                                 goto semicolon;
 343                         }
 344                         if (*p)
 345                                 errx(1,
 346 "%lu: %s: extra text at the end of a transform command", linenum, fname);
 347                         break;
 348                 }
 349         }
 350 }
 351
 352 /*
 353  * Get a delimited string.  P points to the delimeter of the string; d points
 354  * to a buffer area.  Newline and delimiter escapes are processed; other
 355  * escapes are ignored.
 356  *
 357  * Returns a pointer to the first character after the final delimiter or NULL
 358  * in the case of a non-terminated string.  The character array d is filled
 359  * with the processed string.
 360  */
 361 static char *
 362 compile_delimited(char *p, char *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(char **sp, char *t)
 403 {
 404         int c, d;
 405         char *s = *sp;
 406
 407         *t++ = *s++;
 408         if (*s == '^')
 409                 *t++ = *s++;
 410         if (*s == ']')
 411                 *t++ = *s++;
 412         for (; *s && (*t = *s) != ']'; s++, t++)
 413                 if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) {
 414                         *++t = *++s, t++, s++;
 415                         for (c = *s; (*t = *s) != ']' || c != d; s++, t++)
 416                                 if ((c = *s) == '\0')
 417                                         return NULL;
 418                 } else if (*s == '\\' && s[1] == 'n')
 419                             *t = '\n', s++;
 420         return (*s == ']') ? *sp = ++s, ++t : NULL;
 421 }
 422
 423 /*
 424  * Get a regular expression.  P points to the delimiter of the regular
 425  * expression; repp points to the address of a regexp pointer.  Newline
 426  * and delimiter escapes are processed; other escapes are ignored.
 427  * Returns a pointer to the first character after the final delimiter
 428  * or NULL in the case of a non terminated regular expression.  The regexp
 429  * pointer is set to the compiled regular expression.
 430  * Cflags are passed to regcomp.
 431  */
 432 static char *
 433 compile_re(char *p, regex_t **repp)
 434 {
 435         int eval;
 436         char re[_POSIX2_LINE_MAX + 1];
 437
 438         p = compile_delimited(p, re);
 439         if (p && strlen(re) == 0) {
 440                 *repp = NULL;
 441                 return (p);
 442         }
 443         if ((*repp = malloc(sizeof(regex_t))) == NULL)
 444                 err(1, "malloc");
 445         if (p && (eval = regcomp(*repp, re, rflags)) != 0)
 446                 errx(1, "%lu: %s: RE error: %s",
 447                                 linenum, fname, strregerror(eval, *repp));
 448         if (maxnsub < (*repp)->re_nsub)
 449                 maxnsub = (*repp)->re_nsub;
 450         return (p);
 451 }
 452
 453 /*
 454  * Compile the substitution string of a regular expression and set res to
 455  * point to a saved copy of it.  Nsub is the number of parenthesized regular
 456  * expressions.
 457  */
 458 static char *
 459 compile_subst(char *p, struct s_subst *s)
 460 {
 461         static char lbuf[_POSIX2_LINE_MAX + 1];
 462         int asize, size;
 463         u_char ref;
 464         char c, *text, *op, *sp;
 465         int more = 1, sawesc = 0;
 466
 467         c = *p++;                       /* Terminator character */
 468         if (c == '\0')
 469                 return (NULL);
 470
 471         s->maxbref = 0;
 472         s->linenum = linenum;
 473         asize = 2 * _POSIX2_LINE_MAX + 1;
 474         if ((text = malloc(asize)) == NULL)
 475                 err(1, "malloc");
 476         size = 0;
 477         do {
 478                 op = sp = text + size;
 479                 for (; *p; p++) {
 480                         if (*p == '\\' || sawesc) {
 481                                 /*
 482                                  * If this is a continuation from the last
 483                                  * buffer, we won't have a character to
 484                                  * skip over.
 485                                  */
 486                                 if (sawesc)
 487                                         sawesc = 0;
 488                                 else
 489                                         p++;
 490
 491                                 if (*p == '\0') {
 492                                         /*
 493                                          * This escaped character is continued
 494                                          * in the next part of the line.  Note
 495                                          * this fact, then cause the loop to
 496                                          * exit w/ normal EOL case and reenter
 497                                          * above with the new buffer.
 498                                          */
 499                                         sawesc = 1;
 500                                         p--;
 501                                         continue;
 502                                 } else if (strchr("123456789", *p) != NULL) {
 503                                         *sp++ = '\\';
 504                                         ref = *p - '0';
 505                                         if (s->re != NULL &&
 506                                             ref > s->re->re_nsub)
 507                                                 errx(1, "%lu: %s: \\%c not defined in the RE",
 508                                                                 linenum, fname, *p);
 509                                         if (s->maxbref < ref)
 510                                                 s->maxbref = ref;
 511                                 } else if (*p == '&' || *p == '\\')
 512                                         *sp++ = '\\';
 513                         } else if (*p == c) {
 514                                 if (*++p == '\0' && more) {
 515                                         if (cu_fgets(lbuf, sizeof(lbuf), &more))
 516                                                 p = lbuf;
 517                                 }
 518                                 *sp++ = '\0';
 519                                 size += sp - op;
 520                                 if ((s->new = realloc(text, size)) == NULL)
 521                                         err(1, "realloc");
 522                                 return (p);
 523                         } else if (*p == '\n') {
 524                                 errx(1,
 525 "%lu: %s: unescaped newline inside substitute pattern", linenum, fname);
 526                                 /* NOTREACHED */
 527                         }
 528                         *sp++ = *p;
 529                 }
 530                 size += sp - op;
 531                 if (asize - size < _POSIX2_LINE_MAX + 1) {
 532                         asize *= 2;
 533                         if ((text = realloc(text, asize)) == NULL)
 534                                 err(1, "realloc");
 535                 }
 536         } while (cu_fgets(p = lbuf, sizeof(lbuf), &more));
 537         errx(1, "%lu: %s: unterminated substitute in regular expression",
 538                         linenum, fname);
 539         /* NOTREACHED */
 540 }
 541
 542 /*
 543  * Compile the flags of the s command
 544  */
 545 static char *
 546 compile_flags(char *p, struct s_subst *s)
 547 {
 548         int gn;                 /* True if we have seen g or n */
 549         char wfile[_POSIX2_LINE_MAX + 1], *q;
 550
 551         s->n = 1;                               /* Default */
 552         s->p = 0;
 553         s->wfile = NULL;
 554         s->wfd = -1;
 555         for (gn = 0;;) {
 556                 EATSPACE();                     /* EXTENSION */
 557                 switch (*p) {
 558                 case 'g':
 559                         if (gn)
 560                                 errx(1,
 561 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname);
 562                         gn = 1;
 563                         s->n = 0;
 564                         break;
 565                 case '\0':
 566                 case '\n':
 567                 case ';':
 568                         return (p);
 569                 case 'p':
 570                         s->p = 1;
 571                         break;
 572                 case '1': case '2': case '3':
 573                 case '4': case '5': case '6':
 574                 case '7': case '8': case '9':
 575                         if (gn)
 576                                 errx(1,
 577 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname);
 578                         gn = 1;
 579                         /* XXX Check for overflow */
 580                         s->n = (int)strtol(p, &p, 10);
 581                         break;
 582                 case 'w':
 583                         p++;
 584 #ifdef HISTORIC_PRACTICE
 585                         if (*p != ' ') {
 586                                 warnx("%lu: %s: space missing before w wfile", linenum, fname);
 587                                 return (p);
 588                         }
 589 #endif
 590                         EATSPACE();
 591                         q = wfile;
 592                         while (*p) {
 593                                 if (*p == '\n')
 594                                         break;
 595                                 *q++ = *p++;
 596                         }
 597                         *q = '\0';
 598                         if (q == wfile)
 599                                 errx(1, "%lu: %s: no wfile specified", linenum, fname);
 600                         s->wfile = strdup(wfile);
 601                         if (!aflag && (s->wfd = open(wfile,
 602                             O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
 603                             DEFFILEMODE)) == -1)
 604                                 err(1, "%s", wfile);
 605                         return (p);
 606                 default:
 607                         errx(1, "%lu: %s: bad flag in substitute command: '%c'",
 608                                         linenum, fname, *p);
 609                         break;
 610                 }
 611                 p++;
 612         }
 613 }
 614
 615 /*
 616  * Compile a translation set of strings into a lookup table.
 617  */
 618 static char *
 619 compile_tr(char *p, char **transtab)
 620 {
 621         int i;
 622         char *lt, *op, *np;
 623         char old[_POSIX2_LINE_MAX + 1];
 624         char new[_POSIX2_LINE_MAX + 1];
 625
 626         if (*p == '\0' || *p == '\\')
 627                 errx(1,
 628         "%lu: %s: transform pattern can not be delimited by newline or backslash",
 629                         linenum, fname);
 630         p = compile_delimited(p, old);
 631         if (p == NULL)
 632                 errx(1, "%lu: %s: unterminated transform source string",
 633                                 linenum, fname);
 634         p = compile_delimited(--p, new);
 635         if (p == NULL)
 636                 errx(1, "%lu: %s: unterminated transform target string",
 637                                 linenum, fname);
 638         EATSPACE();
 639         if (strlen(new) != strlen(old))
 640                 errx(1, "%lu: %s: transform strings are not the same length",
 641                                 linenum, fname);
 642         /* We assume characters are 8 bits */
 643         if ((lt = malloc(UCHAR_MAX)) == NULL)
 644                 err(1, "malloc");
 645         for (i = 0; i <= UCHAR_MAX; i++)
 646                 lt[i] = (char)i;
 647         for (op = old, np = new; *op; op++, np++)
 648                 lt[(u_char)*op] = *np;
 649         *transtab = lt;
 650         return (p);
 651 }
 652
 653 /*
 654  * Compile the text following an a or i command.
 655  */
 656 static char *
 657 compile_text(void)
 658 {
 659         int asize, esc_nl, size;
 660         char *text, *p, *op, *s;
 661         char lbuf[_POSIX2_LINE_MAX + 1];
 662
 663         asize = 2 * _POSIX2_LINE_MAX + 1;
 664         if ((text = malloc(asize)) == NULL)
 665                 err(1, "malloc");
 666         size = 0;
 667         while (cu_fgets(lbuf, sizeof(lbuf), NULL)) {
 668                 op = s = text + size;
 669                 p = lbuf;
 670                 EATSPACE();
 671                 for (esc_nl = 0; *p != '\0'; p++) {
 672                         if (*p == '\\' && p[1] != '\0' && *++p == '\n')
 673                                 esc_nl = 1;
 674                         *s++ = *p;
 675                 }
 676                 size += s - op;
 677                 if (!esc_nl) {
 678                         *s = '\0';
 679                         break;
 680                 }
 681                 if (asize - size < _POSIX2_LINE_MAX + 1) {
 682                         asize *= 2;
 683                         if ((text = realloc(text, asize)) == NULL)
 684                                 err(1, "realloc");
 685                 }
 686         }
 687         text[size] = '\0';
 688         if ((p = realloc(text, size + 1)) == NULL)
 689                 err(1, "realloc");
 690         return (p);
 691 }
 692
 693 /*
 694  * Get an address and return a pointer to the first character after
 695  * it.  Fill the structure pointed to according to the address.
 696  */
 697 static char *
 698 compile_addr(char *p, struct s_addr *a)
 699 {
 700         char *end;
 701
 702         switch (*p) {
 703         case '\\':                              /* Context address */
 704                 ++p;
 705                 /* FALLTHROUGH */
 706         case '/':                               /* Context address */
 707                 p = compile_re(p, &a->u.r);
 708                 if (p == NULL)
 709                         errx(1, "%lu: %s: unterminated regular expression", linenum, fname);
 710                 a->type = AT_RE;
 711                 return (p);
 712
 713         case '$':                               /* Last line */
 714                 a->type = AT_LAST;
 715                 return (p + 1);
 716                                                 /* Line number */
 717         case '0': case '1': case '2': case '3': case '4':
 718         case '5': case '6': case '7': case '8': case '9':
 719                 a->type = AT_LINE;
 720                 a->u.l = strtol(p, &end, 10);
 721                 return (end);
 722         default:
 723                 errx(1, "%lu: %s: expected context address", linenum, fname);
 724                 return (NULL);
 725         }
 726 }
 727
 728 /*
 729  * duptoeol --
 730  *      Return a copy of all the characters up to \n or \0.
 731  */
 732 static char *
 733 duptoeol(char *s, const char *ctype)
 734 {
 735         size_t len;
 736         int ws;
 737         char *p, *start;
 738
 739         ws = 0;
 740         for (start = s; *s != '\0' && *s != '\n'; ++s)
 741                 ws = isspace((unsigned char)*s);
 742         *s = '\0';
 743         if (ws)
 744                 warnx("%lu: %s: whitespace after %s", linenum, fname, ctype);
 745         len = s - start + 1;
 746         if ((p = malloc(len)) == NULL)
 747                 err(1, "malloc");
 748         return (memmove(p, start, len));
 749 }
 750
 751 /*
 752  * Convert goto label names to addresses, and count a and r commands, in
 753  * the given subset of the script.  Free the memory used by labels in b
 754  * and t commands (but not by :).
 755  *
 756  * TODO: Remove } nodes
 757  */
 758 static void
 759 fixuplabel(struct s_command *cp, struct s_command *end)
 760 {
 761
 762         for (; cp != end; cp = cp->next)
 763                 switch (cp->code) {
 764                 case 'a':
 765                 case 'r':
 766                         appendnum++;
 767                         break;
 768                 case 'b':
 769                 case 't':
 770                         /* Resolve branch target. */
 771                         if (cp->t == NULL) {
 772                                 cp->u.c = NULL;
 773                                 break;
 774                         }
 775                         if ((cp->u.c = findlabel(cp->t)) == NULL)
 776                                 errx(1, "%lu: %s: undefined label '%s'", linenum, fname, cp->t);
 777                         free(cp->t);
 778                         break;
 779                 case '{':
 780                         /* Do interior commands. */
 781                         fixuplabel(cp->u.c, cp->next);
 782                         break;
 783                 }
 784 }
 785
 786 /*
 787  * Associate the given command label for later lookup.
 788  */
 789 static void
 790 enterlabel(struct s_command *cp)
 791 {
 792         struct labhash **lhp, *lh;
 793         u_char *p;
 794         u_int h, c;
 795
 796         for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++)
 797                 h = (h << 5) + h + c;
 798         lhp = &labels[h & LHMASK];
 799         for (lh = *lhp; lh != NULL; lh = lh->lh_next)
 800                 if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0)
 801                         errx(1, "%lu: %s: duplicate label '%s'", linenum, fname, cp->t);
 802         if ((lh = malloc(sizeof *lh)) == NULL)
 803                 err(1, "malloc");
 804         lh->lh_next = *lhp;
 805         lh->lh_hash = h;
 806         lh->lh_cmd = cp;
 807         lh->lh_ref = 0;
 808         *lhp = lh;
 809 }
 810
 811 /*
 812  * Find the label contained in the command l in the command linked
 813  * list cp.  L is excluded from the search.  Return NULL if not found.
 814  */
 815 static struct s_command *
 816 findlabel(char *name)
 817 {
 818         struct labhash *lh;
 819         u_char *p;
 820         u_int h, c;
 821
 822         for (h = 0, p = (u_char *)name; (c = *p) != 0; p++)
 823                 h = (h << 5) + h + c;
 824         for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) {
 825                 if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) {
 826                         lh->lh_ref = 1;
 827                         return (lh->lh_cmd);
 828                 }
 829         }
 830         return (NULL);
 831 }
 832
 833 /*
 834  * Warn about any unused labels.  As a side effect, release the label hash
 835  * table space.
 836  */
 837 static void
 838 uselabel(void)
 839 {
 840         struct labhash *lh, *next;
 841         int i;
 842
 843         for (i = 0; i < LHSZ; i++) {
 844                 for (lh = labels[i]; lh != NULL; lh = next) {
 845                         next = lh->lh_next;
 846                         if (!lh->lh_ref)
 847                                 warnx("%lu: %s: unused label '%s'",
 848                                     linenum, fname, lh->lh_cmd->t);
 849                         free(lh);
 850                 }
 851         }
 852 }