package/toolbox/src/sed/compile.c

   1 /*      $OpenBSD: compile.c,v 1.42 2017/08/01 18:05:53 martijn Exp $    */
   2
   3 /*-
   4  * Copyright (c) 2016
   5  *      mirabilos <m@mirbsd.org>
   6  * Copyright (c) 1992 Diomidis Spinellis.
   7  * Copyright (c) 1992, 1993
   8  *      The Regents of the University of California.  All rights reserved.
   9  *
  10  * This code is derived from software contributed to Berkeley by
  11  * Diomidis Spinellis of Imperial College, University of London.
  12  *
  13  * Redistribution and use in source and binary forms, with or without
  14  * modification, are permitted provided that the following conditions
  15  * are met:
  16  * 1. Redistributions of source code must retain the above copyright
  17  *    notice, this list of conditions and the following disclaimer.
  18  * 2. Redistributions in binary form must reproduce the above copyright
  19  *    notice, this list of conditions and the following disclaimer in the
  20  *    documentation and/or other materials provided with the distribution.
  21  * 3. 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/types.h>
  39 #include <sys/stat.h>
  40
  41 #include <ctype.h>
  42 #include <errno.h>
  43 #include <fcntl.h>
  44 #include <limits.h>
  45 #include <regex.h>
  46 #include <stdio.h>
  47 #include <stdlib.h>
  48 #include <string.h>
  49
  50 #include "defs.h"
  51 #include "extern.h"
  52
  53 __RCSID("$MirOS: src/usr.bin/sed/compile.c,v 1.3 2017/11/20 01:23:56 tg Exp $");
  54
  55 #define LHSZ    128
  56 #define LHMASK  (LHSZ - 1)
  57 static struct labhash {
  58         struct  labhash *lh_next;
  59         u_int   lh_hash;
  60         struct  s_command *lh_cmd;
  61         int     lh_ref;
  62 } *labels[LHSZ];
  63
  64 static char      *compile_addr(char *, struct s_addr *);
  65 static char      *compile_ccl(char **, char *);
  66 static char      *compile_delimited(char *, char *, int);
  67 static char      *compile_flags(char *, struct s_subst *);
  68 static char      *compile_re(char *, regex_t **);
  69 static char      *compile_subst(char *, struct s_subst *);
  70 static char      *compile_text(void);
  71 static char      *compile_tr(char *, char **);
  72 static struct s_command
  73                 **compile_stream(struct s_command **);
  74 static char      *duptoeol(char *, const char *, char **);
  75 static void       enterlabel(struct s_command *);
  76 static struct s_command
  77                  *findlabel(char *);
  78 static void       fixuplabel(struct s_command *, struct s_command *);
  79 static void       uselabel(void);
  80
  81 /*
  82  * Command specification.  This is used to drive the command parser.
  83  */
  84 struct s_format {
  85         char code;                              /* Command code */
  86         int naddr;                              /* Number of address args */
  87         enum e_args args;                       /* Argument type */
  88 };
  89
  90 static struct s_format cmd_fmts[] = {
  91         {'{', 2, GROUP},
  92         {'}', 0, ENDGROUP},
  93         {'a', 1, TEXT},
  94         {'b', 2, BRANCH},
  95         {'c', 2, TEXT},
  96         {'d', 2, EMPTY},
  97         {'D', 2, EMPTY},
  98         {'g', 2, EMPTY},
  99         {'G', 2, EMPTY},
 100         {'h', 2, EMPTY},
 101         {'H', 2, EMPTY},
 102         {'i', 1, TEXT},
 103         {'l', 2, EMPTY},
 104         {'n', 2, EMPTY},
 105         {'N', 2, EMPTY},
 106         {'p', 2, EMPTY},
 107         {'P', 2, EMPTY},
 108         {'q', 1, EMPTY},
 109         {'r', 1, RFILE},
 110         {'s', 2, SUBST},
 111         {'t', 2, BRANCH},
 112         {'w', 2, WFILE},
 113         {'x', 2, EMPTY},
 114         {'y', 2, TR},
 115         {'!', 2, NONSEL},
 116         {':', 0, LABEL},
 117         {'#', 0, COMMENT},
 118         {'=', 1, EMPTY},
 119         {'\0', 0, COMMENT},
 120 };
 121
 122 /* The compiled program. */
 123 struct s_command *prog;
 124
 125 /*
 126  * Compile the program into prog.
 127  * Initialise appends.
 128  */
 129 void
 130 compile(void)
 131 {
 132         *compile_stream(&prog) = NULL;
 133         fixuplabel(prog, NULL);
 134         uselabel();
 135         appends = xreallocarray(NULL, appendnum, sizeof(struct s_appends));
 136         match = xreallocarray(NULL, maxnsub + 1, sizeof(regmatch_t));
 137 }
 138
 139 #define EATSPACE() do {                                                 \
 140         if (p)                                                          \
 141                 while (isascii((unsigned char)*p) &&                    \
 142                     isspace((unsigned char)*p))                         \
 143                         p++;                                            \
 144         } while (0)
 145
 146 static struct s_command **
 147 compile_stream(struct s_command **link)
 148 {
 149         char *p;
 150         static char *lbuf;      /* To avoid excessive malloc calls */
 151         static size_t bufsize;
 152         struct s_command *cmd, *cmd2, *stack;
 153         struct s_format *fp;
 154         int naddr;                              /* Number of addresses */
 155
 156         stack = 0;
 157         for (;;) {
 158                 if ((p = cu_fgets(&lbuf, &bufsize)) == NULL) {
 159                         if (stack != 0)
 160                                 error(COMPILE, "unexpected EOF (pending }'s)");
 161                         return (link);
 162                 }
 163
 164 semicolon:      EATSPACE();
 165                 if (*p == '#' || *p == '\0')
 166                         continue;
 167                 if (*p == ';') {
 168                         p++;
 169                         goto semicolon;
 170                 }
 171                 *link = cmd = xmalloc(sizeof(struct s_command));
 172                 link = &cmd->next;
 173                 cmd->nonsel = cmd->inrange = 0;
 174                 /* First parse the addresses */
 175                 naddr = 0;
 176
 177 /* Valid characters to start an address */
 178 #define addrchar(c)     (strchr("0123456789/\\$", (c)))
 179                 if (addrchar(*p)) {
 180                         naddr++;
 181                         cmd->a1 = xmalloc(sizeof(struct s_addr));
 182                         p = compile_addr(p, cmd->a1);
 183                         EATSPACE();                             /* EXTENSION */
 184                         if (*p == ',') {
 185                                 p++;
 186                                 EATSPACE();                     /* EXTENSION */
 187                                 naddr++;
 188                                 cmd->a2 = xmalloc(sizeof(struct s_addr));
 189                                 p = compile_addr(p, cmd->a2);
 190                                 EATSPACE();
 191                         } else {
 192                                 cmd->a2 = 0;
 193                         }
 194                 } else {
 195                         cmd->a1 = cmd->a2 = 0;
 196                 }
 197
 198 nonsel:         /* Now parse the command */
 199                 if (!*p)
 200                         error(COMPILE, "command expected");
 201                 cmd->code = *p;
 202                 for (fp = cmd_fmts; fp->code; fp++)
 203                         if (fp->code == *p)
 204                                 break;
 205                 if (!fp->code)
 206                         error(COMPILE, "invalid command code %c", *p);
 207                 if (naddr > fp->naddr)
 208                         error(COMPILE,
 209                             "command %c expects up to %d address(es), found %d",
 210                             *p, fp->naddr, naddr);
 211                 switch (fp->args) {
 212                 case NONSEL:                    /* ! */
 213                         p++;
 214                         EATSPACE();
 215                         cmd->nonsel = 1;
 216                         goto nonsel;
 217                 case GROUP:                     /* { */
 218                         p++;
 219                         EATSPACE();
 220                         cmd->next = stack;
 221                         stack = cmd;
 222                         link = &cmd->u.c;
 223                         if (*p)
 224                                 goto semicolon;
 225                         break;
 226                 case ENDGROUP:
 227                         /*
 228                          * Short-circuit command processing, since end of
 229                          * group is really just a noop.
 230                          */
 231                         cmd->nonsel = 1;
 232                         if (stack == 0)
 233                                 error(COMPILE, "unexpected }");
 234                         cmd2 = stack;
 235                         stack = cmd2->next;
 236                         cmd2->next = cmd;
 237                         /*FALLTHROUGH*/
 238                 case EMPTY:             /* d D g G h H l n N p P q x = \0 */
 239                         p++;
 240                         EATSPACE();
 241                         if (*p == ';') {
 242                                 p++;
 243                                 link = &cmd->next;
 244                                 goto semicolon;
 245                         }
 246                         if (*p)
 247                                 error(COMPILE,
 248 "extra characters at the end of %c command", cmd->code);
 249                         break;
 250                 case TEXT:                      /* a c i */
 251                         p++;
 252                         EATSPACE();
 253                         if (*p != '\\')
 254                                 error(COMPILE, "command %c expects \\ followed by"
 255                                     " text", cmd->code);
 256                         p++;
 257                         EATSPACE();
 258                         if (*p)
 259                                 error(COMPILE, "extra characters after \\ at the"
 260                                     " end of %c command", cmd->code);
 261                         cmd->t = compile_text();
 262                         break;
 263                 case COMMENT:                   /* \0 # */
 264                         break;
 265                 case WFILE:                     /* w */
 266                         p++;
 267                         EATSPACE();
 268                         if (*p == '\0')
 269                                 error(COMPILE, "filename expected");
 270                         cmd->t = duptoeol(p, "w command", NULL);
 271                         if (aflag) {
 272                                 cmd->u.fd = -1;
 273                                 pledge_wpath = 1;
 274                         }
 275                         else if ((cmd->u.fd = open(p,
 276                             O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
 277                             DEFFILEMODE)) == -1)
 278                                 error(FATAL, "%s: %s", p, strerror(errno));
 279                         break;
 280                 case RFILE:                     /* r */
 281                         pledge_rpath = 1;
 282                         p++;
 283                         EATSPACE();
 284                         cmd->t = duptoeol(p, "read command", NULL);
 285                         break;
 286                 case BRANCH:                    /* b t */
 287                         p++;
 288                         EATSPACE();
 289                         if (*p == '\0')
 290                                 cmd->t = NULL;
 291                         else
 292                                 cmd->t = duptoeol(p, "branch", &p);
 293                         if (*p == ';') {
 294                                 p++;
 295                                 goto semicolon;
 296                         }
 297                         break;
 298                 case LABEL:                     /* : */
 299                         p++;
 300                         EATSPACE();
 301                         cmd->t = duptoeol(p, "label", &p);
 302                         if (strlen(cmd->t) == 0)
 303                                 error(COMPILE, "empty label");
 304                         enterlabel(cmd);
 305                         if (*p == ';') {
 306                                 p++;
 307                                 goto semicolon;
 308                         }
 309                         break;
 310                 case SUBST:                     /* s */
 311                         p++;
 312                         if (*p == '\0' || *p == '\\')
 313                                 error(COMPILE, "substitute pattern can not be"
 314                                     " delimited by newline or backslash");
 315                         cmd->u.s = xmalloc(sizeof(struct s_subst));
 316                         p = compile_re(p, &cmd->u.s->re);
 317                         if (p == NULL)
 318                                 error(COMPILE, "unterminated substitute pattern");
 319                         --p;
 320                         p = compile_subst(p, cmd->u.s);
 321                         p = compile_flags(p, cmd->u.s);
 322                         EATSPACE();
 323                         if (*p == ';') {
 324                                 p++;
 325                                 link = &cmd->next;
 326                                 goto semicolon;
 327                         }
 328                         break;
 329                 case TR:                        /* y */
 330                         p++;
 331                         p = compile_tr(p, (char **)&cmd->u.y);
 332                         EATSPACE();
 333                         if (*p == ';') {
 334                                 p++;
 335                                 link = &cmd->next;
 336                                 goto semicolon;
 337                         }
 338                         if (*p)
 339                                 error(COMPILE, "extra text at the end of a"
 340                                     " transform command");
 341                         break;
 342                 }
 343         }
 344 }
 345
 346 /*
 347  * Get a delimited string.  P points to the delimeter of the string; d points
 348  * to a buffer area.  Newline and delimiter escapes are processed; other
 349  * escapes are ignored.
 350  *
 351  * Returns a pointer to the first character after the final delimiter or NULL
 352  * in the case of a non-terminated string.  The character array d is filled
 353  * with the processed string.
 354  */
 355 static char *
 356 compile_delimited(char *p, char *d, int is_tr)
 357 {
 358         char c;
 359
 360         c = *p++;
 361         if (c == '\0')
 362                 return (NULL);
 363         else if (c == '\\')
 364                 error(COMPILE, "\\ can not be used as a string delimiter");
 365         else if (c == '\n')
 366                 error(COMPILE, "newline can not be used as a string delimiter");
 367         while (*p) {
 368                 if (*p == '[' && *p != c) {
 369                         if ((d = compile_ccl(&p, d)) == NULL)
 370                                 error(COMPILE, "unbalanced brackets ([])");
 371                         continue;
 372                 } else if (*p == '\\' && p[1] == '[') {
 373                         *d++ = *p++;
 374                 } else if (*p == '\\' && p[1] == c) {
 375                         p++;
 376                 } else if (*p == '\\' && p[1] == 'n') {
 377                         *d++ = '\n';
 378                         p += 2;
 379                         continue;
 380                 } else if (*p == '\\' && p[1] == '\\') {
 381                         if (is_tr)
 382                                 p++;
 383                         else
 384                                 *d++ = *p++;
 385                 } else if (*p == c) {
 386                         *d = '\0';
 387                         return (p + 1);
 388                 }
 389                 *d++ = *p++;
 390         }
 391         return (NULL);
 392 }
 393
 394
 395 /* compile_ccl: expand a POSIX character class */
 396 static char *
 397 compile_ccl(char **sp, char *t)
 398 {
 399         int c, d;
 400         char *s = *sp;
 401
 402         *t++ = *s++;
 403         if (*s == '^')
 404                 *t++ = *s++;
 405         if (*s == ']')
 406                 *t++ = *s++;
 407         for (; *s && (*t = *s) != ']'; s++, t++)
 408                 if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) {
 409                         *++t = *++s, t++, s++;
 410                         for (c = *s; (*t = *s) != ']' || c != d; s++, t++)
 411                                 if ((c = *s) == '\0')
 412                                         return NULL;
 413                 } else if (*s == '\\' && s[1] == 'n') {
 414                         *t = '\n';
 415                         s++;
 416                 }
 417         if (*s == ']') {
 418                 *sp = ++s;
 419                 return (++t);
 420         } else {
 421                 return (NULL);
 422         }
 423 }
 424
 425 /*
 426  * Get a regular expression.  P points to the delimiter of the regular
 427  * expression; repp points to the address of a regexp pointer.  Newline
 428  * and delimiter escapes are processed; other escapes are ignored.
 429  * Returns a pointer to the first character after the final delimiter
 430  * or NULL in the case of a non terminated regular expression.  The regexp
 431  * pointer is set to the compiled regular expression.
 432  * Cflags are passed to regcomp.
 433  */
 434 static char *
 435 compile_re(char *p, regex_t **repp)
 436 {
 437         int eval;
 438         char *re;
 439
 440         re = xmalloc(strlen(p) + 1); /* strlen(re) <= strlen(p) */
 441         p = compile_delimited(p, re, 0);
 442         if (p && strlen(re) == 0) {
 443                 *repp = NULL;
 444                 free(re);
 445                 return (p);
 446         }
 447         *repp = xmalloc(sizeof(regex_t));
 448         if (p && (eval = regcomp(*repp, re, Eflag ? REG_EXTENDED : 0)) != 0)
 449                 error(COMPILE, "RE error: %s", strregerror(eval, *repp));
 450         if (maxnsub < (*repp)->re_nsub)
 451                 maxnsub = (*repp)->re_nsub;
 452         free(re);
 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;
 465         static size_t bufsize;
 466         size_t asize, ref, size;
 467         char c, *text, *op, *sp;
 468         int 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         text = NULL;
 477         asize = size = 0;
 478         do {
 479                 size_t len = ROUNDLEN(strlen(p) + 1);
 480                 if (asize - size < len) {
 481                         do {
 482                                 asize += len;
 483                         } while (asize - size < len);
 484                         text = xrealloc(text, asize);
 485                 }
 486                 op = sp = text + size;
 487                 for (; *p; p++) {
 488                         if (*p == '\\' || sawesc) {
 489                                 /*
 490                                  * If this is a continuation from the last
 491                                  * buffer, we won't have a character to
 492                                  * skip over.
 493                                  */
 494                                 if (sawesc)
 495                                         sawesc = 0;
 496                                 else
 497                                         p++;
 498
 499                                 if (*p == '\0') {
 500                                         /*
 501                                          * This escaped character is continued
 502                                          * in the next part of the line.  Note
 503                                          * this fact, then cause the loop to
 504                                          * exit w/ normal EOL case and reenter
 505                                          * above with the new buffer.
 506                                          */
 507                                         sawesc = 1;
 508                                         p--;
 509                                         continue;
 510                                 } else if (strchr("123456789", *p) != NULL) {
 511                                         *sp++ = '\\';
 512                                         ref = *p - '0';
 513                                         if (s->re != NULL &&
 514                                             ref > s->re->re_nsub)
 515                                                 error(COMPILE,
 516 "\\%c not defined in the RE", *p);
 517                                         if (s->maxbref < ref)
 518                                                 s->maxbref = ref;
 519                                 } else if (*p == '&' || *p == '\\')
 520                                         *sp++ = '\\';
 521                         } else if (*p == c) {
 522                                 p++;
 523                                 *sp++ = '\0';
 524                                 size += sp - op;
 525                                 s->new = xrealloc(text, size);
 526                                 return (p);
 527                         } else if (*p == '\n') {
 528                                 error(COMPILE,
 529 "unescaped newline inside substitute pattern");
 530                         }
 531                         *sp++ = *p;
 532                 }
 533                 size += sp - op;
 534         } while ((p = cu_fgets(&lbuf, &bufsize)));
 535         error(COMPILE, "unterminated substitute in regular expression");
 536 }
 537
 538 /*
 539  * Compile the flags of the s command
 540  */
 541 static char *
 542 compile_flags(char *p, struct s_subst *s)
 543 {
 544         int gn;                 /* True if we have seen g or n */
 545         long l;
 546         char wfile[PATH_MAX], *q, *eq;
 547
 548         s->n = 1;                               /* Default */
 549         s->p = 0;
 550         s->wfile = NULL;
 551         s->wfd = -1;
 552         for (gn = 0;;) {
 553                 EATSPACE();                     /* EXTENSION */
 554                 switch (*p) {
 555                 case 'g':
 556                         if (gn)
 557                                 error(COMPILE, "more than one number or 'g' in"
 558                                     " substitute flags");
 559                         gn = 1;
 560                         s->n = 0;
 561                         break;
 562                 case '\0':
 563                 case '\n':
 564                 case ';':
 565                         return (p);
 566                 case 'p':
 567                         s->p = 1;
 568                         break;
 569                 case '1': case '2': case '3':
 570                 case '4': case '5': case '6':
 571                 case '7': case '8': case '9':
 572                         if (gn)
 573                                 error(COMPILE, "more than one number or 'g' in"
 574                                     " substitute flags");
 575                         gn = 1;
 576                         l = strtol(p, &p, 10);
 577                         if (l <= 0 || l >= INT_MAX)
 578                                 error(COMPILE,
 579                                     "number in substitute flags out of range");
 580                         s->n = (int)l;
 581                         continue;
 582                 case 'w':
 583                         p++;
 584 #ifdef HISTORIC_PRACTICE
 585                         if (*p != ' ') {
 586                                 warning("space missing before w wfile");
 587                                 return (p);
 588                         }
 589 #endif
 590                         EATSPACE();
 591                         q = wfile;
 592                         eq = wfile + sizeof(wfile) - 1;
 593                         while (*p) {
 594                                 if (*p == '\n')
 595                                         break;
 596                                 if (q >= eq)
 597                                         error(COMPILE, "wfile too long");
 598                                 *q++ = *p++;
 599                         }
 600                         *q = '\0';
 601                         if (q == wfile)
 602                                 error(COMPILE, "no wfile specified");
 603                         s->wfile = strdup(wfile);
 604                         if (aflag)
 605                                 pledge_wpath = 1;
 606                         else if ((s->wfd = open(wfile,
 607                             O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
 608                             DEFFILEMODE)) == -1)
 609                                 error(FATAL, "%s: %s", wfile, strerror(errno));
 610                         return (p);
 611                 default:
 612                         error(COMPILE,
 613                             "bad flag in substitute command: '%c'", *p);
 614                         break;
 615                 }
 616                 p++;
 617         }
 618 }
 619
 620 /*
 621  * Compile a translation set of strings into a lookup table.
 622  */
 623 static char *
 624 compile_tr(char *p, char **transtab)
 625 {
 626         int i;
 627         char *lt, *op, *np;
 628         char *old = NULL, *new = NULL;
 629
 630         if (*p == '\0' || *p == '\\')
 631                 error(COMPILE,
 632 "transform pattern can not be delimited by newline or backslash");
 633         old = xmalloc(strlen(p) + 1);
 634         p = compile_delimited(p, old, 1);
 635         if (p == NULL) {
 636                 error(COMPILE, "unterminated transform source string");
 637                 goto bad;
 638         }
 639         new = xmalloc(strlen(p) + 1);
 640         p = compile_delimited(--p, new, 1);
 641         if (p == NULL) {
 642                 error(COMPILE, "unterminated transform target string");
 643                 goto bad;
 644         }
 645         EATSPACE();
 646         if (strlen(new) != strlen(old)) {
 647                 error(COMPILE, "transform strings are not the same length");
 648                 goto bad;
 649         }
 650         /* We assume characters are 8 bits */
 651         lt = xmalloc(UCHAR_MAX + 1);
 652         for (i = 0; i <= UCHAR_MAX; i++)
 653                 lt[i] = (char)i;
 654         for (op = old, np = new; *op; op++, np++)
 655                 lt[(u_char)*op] = *np;
 656         *transtab = lt;
 657         free(old);
 658         free(new);
 659         return (p);
 660 bad:
 661         free(old);
 662         free(new);
 663         return (NULL);
 664 }
 665
 666 /*
 667  * Compile the text following an a, c, or i command.
 668  */
 669 static char *
 670 compile_text(void)
 671 {
 672         size_t asize, size;
 673         int esc_nl;
 674         char *lbuf, *text, *p, *op, *s;
 675         size_t bufsize;
 676
 677         lbuf = text = NULL;
 678         asize = size = 0;
 679         while ((p = cu_fgets(&lbuf, &bufsize))) {
 680                 size_t len = ROUNDLEN(strlen(p) + 1);
 681                 if (asize - size < len) {
 682                         do {
 683                                 asize += len;
 684                         } while (asize - size < len);
 685                         text = xrealloc(text, asize);
 686                 }
 687                 op = s = text + size;
 688                 for (esc_nl = 0; *p != '\0'; p++) {
 689                         if (*p == '\\' && p[1] != '\0' && *++p == '\n')
 690                                 esc_nl = 1;
 691                         *s++ = *p;
 692                 }
 693                 size += s - op;
 694                 if (!esc_nl) {
 695                         *s = '\0';
 696                         break;
 697                 }
 698         }
 699         free(lbuf);
 700         text = xrealloc(text, size + 1);
 701         text[size] = '\0';
 702         return (text);
 703 }
 704
 705 /*
 706  * Get an address and return a pointer to the first character after
 707  * it.  Fill the structure pointed to according to the address.
 708  */
 709 static char *
 710 compile_addr(char *p, struct s_addr *a)
 711 {
 712         char *end;
 713
 714         switch (*p) {
 715         case '\\':                              /* Context address */
 716                 ++p;
 717                 /* FALLTHROUGH */
 718         case '/':                               /* Context address */
 719                 p = compile_re(p, &a->u.r);
 720                 if (p == NULL)
 721                         error(COMPILE, "unterminated regular expression");
 722                 a->type = AT_RE;
 723                 return (p);
 724
 725         case '$':                               /* Last line */
 726                 a->type = AT_LAST;
 727                 return (p + 1);
 728                                                 /* Line number */
 729         case '0': case '1': case '2': case '3': case '4':
 730         case '5': case '6': case '7': case '8': case '9':
 731                 a->type = AT_LINE;
 732                 a->u.l = strtoul(p, &end, 10);
 733                 return (end);
 734         default:
 735                 error(COMPILE, "expected context address");
 736                 return (NULL);
 737         }
 738 }
 739
 740 /*
 741  * duptoeol --
 742  *      Return a copy of all the characters up to \n or \0.
 743  */
 744 static char *
 745 duptoeol(char *s, const char *ctype, char **semi)
 746 {
 747         size_t len;
 748         int ws;
 749         char *start;
 750
 751         ws = 0;
 752         if (semi) {
 753                 for (start = s; *s != '\0' && *s != '\n' && *s != ';'; ++s)
 754                         ws = isspace((unsigned char)*s);
 755         } else {
 756                 for (start = s; *s != '\0' && *s != '\n'; ++s)
 757                         ws = isspace((unsigned char)*s);
 758                 *s = '\0';
 759         }
 760         if (ws)
 761                 warning("whitespace after %s", ctype);
 762         len = s - start + 1;
 763         if (semi)
 764                 *semi = s;
 765         s = xmalloc(len);
 766         strlcpy(s, start, len);
 767         return (s);
 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(struct s_command *cp, struct s_command *end)
 779 {
 780
 781         for (; cp != end; cp = cp->next)
 782                 switch (cp->code) {
 783                 case 'a':
 784                 case 'r':
 785                         appendnum++;
 786                         break;
 787                 case 'b':
 788                 case 't':
 789                         /* Resolve branch target. */
 790                         if (cp->t == NULL) {
 791                                 cp->u.c = NULL;
 792                                 break;
 793                         }
 794                         if ((cp->u.c = findlabel(cp->t)) == NULL)
 795                                 error(COMPILE, "undefined label '%s'", cp->t);
 796                         free(cp->t);
 797                         break;
 798                 case '{':
 799                         /* Do interior commands. */
 800                         fixuplabel(cp->u.c, cp->next);
 801                         break;
 802                 }
 803 }
 804
 805 /*
 806  * Associate the given command label for later lookup.
 807  */
 808 static void
 809 enterlabel(struct s_command *cp)
 810 {
 811         struct labhash **lhp, *lh;
 812         u_char *p;
 813         u_int h, c;
 814
 815         for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++)
 816                 h = (h << 5) + h + c;
 817         lhp = &labels[h & LHMASK];
 818         for (lh = *lhp; lh != NULL; lh = lh->lh_next)
 819                 if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0)
 820                         error(COMPILE, "duplicate label '%s'", cp->t);
 821         lh = xmalloc(sizeof *lh);
 822         lh->lh_next = *lhp;
 823         lh->lh_hash = h;
 824         lh->lh_cmd = cp;
 825         lh->lh_ref = 0;
 826         *lhp = lh;
 827 }
 828
 829 /*
 830  * Find the label contained in the command l in the command linked
 831  * list cp.  L is excluded from the search.  Return NULL if not found.
 832  */
 833 static struct s_command *
 834 findlabel(char *name)
 835 {
 836         struct labhash *lh;
 837         u_char *p;
 838         u_int h, c;
 839
 840         for (h = 0, p = (u_char *)name; (c = *p) != 0; p++)
 841                 h = (h << 5) + h + c;
 842         for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) {
 843                 if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) {
 844                         lh->lh_ref = 1;
 845                         return (lh->lh_cmd);
 846                 }
 847         }
 848         return (NULL);
 849 }
 850
 851 /*
 852  * Warn about any unused labels.  As a side effect, release the label hash
 853  * table space.
 854  */
 855 static void
 856 uselabel(void)
 857 {
 858         struct labhash *lh, *next;
 859         int i;
 860
 861         for (i = 0; i < LHSZ; i++) {
 862                 for (lh = labels[i]; lh != NULL; lh = next) {
 863                         next = lh->lh_next;
 864                         if (!lh->lh_ref)
 865                                 warning("unused label '%s'",
 866                                     lh->lh_cmd->t);
 867                         free(lh);
 868                 }
 869         }
 870 }