text_cmds/sed/compile.c

   1 /*-
   2  * SPDX-License-Identifier: BSD-3-Clause
   3  *
   4  * Copyright (c) 1992 Diomidis Spinellis.
   5  * Copyright (c) 1992, 1993
   6  *      The Regents of the University of California.  All rights reserved.
   7  *
   8  * This code is derived from software contributed to Berkeley by
   9  * Diomidis Spinellis of Imperial College, University of London.
  10  *
  11  * Redistribution and use in source and binary forms, with or without
  12  * modification, are permitted provided that the following conditions
  13  * are met:
  14  * 1. Redistributions of source code must retain the above copyright
  15  *    notice, this list of conditions and the following disclaimer.
  16  * 2. Redistributions in binary form must reproduce the above copyright
  17  *    notice, this list of conditions and the following disclaimer in the
  18  *    documentation and/or other materials provided with the distribution.
  19  * 3. Neither the name of the University nor the names of its contributors
  20  *    may be used to endorse or promote products derived from this software
  21  *    without specific prior written permission.
  22  *
  23  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  24  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  25  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  26  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  27  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  28  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  29  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  30  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  31  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  32  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  33  * SUCH DAMAGE.
  34  */
  35
  36 #include <sys/cdefs.h>
  37 __FBSDID("$FreeBSD$");
  38
  39 #ifndef lint
  40 static const char sccsid[] = "@(#)compile.c     8.1 (Berkeley) 6/6/93";
  41 #endif
  42
  43 #include <sys/types.h>
  44 #include <sys/stat.h>
  45
  46 #include <ctype.h>
  47 #include <err.h>
  48 #include <errno.h>
  49 #include <fcntl.h>
  50 #include <limits.h>
  51 #include <regex.h>
  52 #include <stdio.h>
  53 #include <stdlib.h>
  54 #include <string.h>
  55 #include <wchar.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 *, int);
  72 static char      *compile_flags(char *, struct s_subst *);
  73 static regex_t   *compile_re(char *, int);
  74 static char      *compile_subst(char *, struct s_subst *);
  75 static char      *compile_text(void);
  76 static char      *compile_tr(char *, struct s_tr **);
  77 static struct s_command
  78                 **compile_stream(struct s_command **);
  79 static char      *duptoeol(char *, const char *);
  80 static void       enterlabel(struct s_command *);
  81 static struct s_command
  82                  *findlabel(char *);
  83 static void       fixuplabel(struct s_command *, struct s_command *);
  84 static void       uselabel(void);
  85
  86 /*
  87  * Command specification.  This is used to drive the command parser.
  88  */
  89 struct s_format {
  90         char code;                              /* Command code */
  91         int naddr;                              /* Number of address args */
  92         enum e_args args;                       /* Argument type */
  93 };
  94
  95 static struct s_format cmd_fmts[] = {
  96         {'{', 2, GROUP},
  97         {'}', 0, ENDGROUP},
  98         {'a', 1, TEXT},
  99         {'b', 2, BRANCH},
 100         {'c', 2, TEXT},
 101         {'d', 2, EMPTY},
 102         {'D', 2, EMPTY},
 103         {'g', 2, EMPTY},
 104         {'G', 2, EMPTY},
 105         {'h', 2, EMPTY},
 106         {'H', 2, EMPTY},
 107         {'i', 1, TEXT},
 108         {'l', 2, EMPTY},
 109         {'n', 2, EMPTY},
 110         {'N', 2, EMPTY},
 111         {'p', 2, EMPTY},
 112         {'P', 2, EMPTY},
 113         {'q', 1, EMPTY},
 114         {'r', 1, RFILE},
 115         {'s', 2, SUBST},
 116         {'t', 2, BRANCH},
 117         {'w', 2, WFILE},
 118         {'x', 2, EMPTY},
 119         {'y', 2, TR},
 120         {'!', 2, NONSEL},
 121         {':', 0, LABEL},
 122         {'#', 0, COMMENT},
 123         {'=', 1, EMPTY},
 124         {'\0', 0, COMMENT},
 125 };
 126
 127 /* The compiled program. */
 128 struct s_command *prog;
 129
 130 /*
 131  * Compile the program into prog.
 132  * Initialise appends.
 133  */
 134 void
 135 compile(void)
 136 {
 137         *compile_stream(&prog) = NULL;
 138         fixuplabel(prog, NULL);
 139         uselabel();
 140         if (appendnum == 0)
 141                 appends = NULL;
 142         else if ((appends = malloc(sizeof(struct s_appends) * appendnum)) ==
 143             NULL)
 144                 err(1, "malloc");
 145         if ((match = malloc((maxnsub + 1) * sizeof(regmatch_t))) == NULL)
 146                 err(1, "malloc");
 147 }
 148
 149 #define EATSPACE() do {                                                 \
 150         if (p)                                                          \
 151                 while (*p && isspace((unsigned char)*p))                \
 152                         p++;                                            \
 153         } while (0)
 154
 155 static struct s_command **
 156 compile_stream(struct s_command **link)
 157 {
 158         char *p;
 159         static char lbuf[_POSIX2_LINE_MAX + 1]; /* To save stack */
 160         struct s_command *cmd, *cmd2, *stack;
 161         struct s_format *fp;
 162         char re[_POSIX2_LINE_MAX + 1];
 163         int naddr;                              /* Number of addresses */
 164
 165         stack = NULL;
 166         for (;;) {
 167                 if ((p = cu_fgets(lbuf, sizeof(lbuf), NULL)) == NULL) {
 168                         if (stack != NULL)
 169                                 errx(1, "%lu: %s: unexpected EOF (pending }'s)",
 170                                                         linenum, fname);
 171                         return (link);
 172                 }
 173
 174 semicolon:      EATSPACE();
 175                 if (p) {
 176                         if (*p == '#' || *p == '\0')
 177                                 continue;
 178                         else if (*p == ';') {
 179                                 p++;
 180                                 goto semicolon;
 181                         }
 182                 }
 183                 if ((*link = cmd = malloc(sizeof(struct s_command))) == NULL)
 184                         err(1, "malloc");
 185                 link = &cmd->next;
 186                 cmd->startline = cmd->nonsel = 0;
 187                 /* First parse the addresses */
 188                 naddr = 0;
 189
 190 /* Valid characters to start an address */
 191 #define addrchar(c)     (strchr("0123456789/\\$", (c)))
 192                 if (addrchar(*p)) {
 193                         naddr++;
 194                         if ((cmd->a1 = malloc(sizeof(struct s_addr))) == NULL)
 195                                 err(1, "malloc");
 196                         p = compile_addr(p, cmd->a1);
 197                         EATSPACE();                             /* EXTENSION */
 198                         if (*p == ',') {
 199                                 p++;
 200                                 EATSPACE();                     /* EXTENSION */
 201                                 naddr++;
 202                                 if ((cmd->a2 = malloc(sizeof(struct s_addr)))
 203                                     == NULL)
 204                                         err(1, "malloc");
 205                                 p = compile_addr(p, cmd->a2);
 206                                 EATSPACE();
 207                         } else
 208                                 cmd->a2 = NULL;
 209                 } else
 210                         cmd->a1 = cmd->a2 = NULL;
 211
 212 nonsel:         /* Now parse the command */
 213                 if (!*p)
 214                         errx(1, "%lu: %s: command expected", linenum, fname);
 215                 cmd->code = *p;
 216                 for (fp = cmd_fmts; fp->code; fp++)
 217                         if (fp->code == *p)
 218                                 break;
 219                 if (!fp->code)
 220                         errx(1, "%lu: %s: invalid command code %c", linenum, fname, *p);
 221                 if (naddr > fp->naddr)
 222                         errx(1,
 223                                 "%lu: %s: command %c expects up to %d address(es), found %d",
 224                                 linenum, fname, *p, fp->naddr, naddr);
 225                 switch (fp->args) {
 226                 case NONSEL:                    /* ! */
 227                         p++;
 228                         EATSPACE();
 229                         cmd->nonsel = 1;
 230                         goto nonsel;
 231                 case GROUP:                     /* { */
 232                         p++;
 233                         EATSPACE();
 234                         cmd->next = stack;
 235                         stack = cmd;
 236                         link = &cmd->u.c;
 237                         if (*p)
 238                                 goto semicolon;
 239                         break;
 240                 case ENDGROUP:
 241                         /*
 242                          * Short-circuit command processing, since end of
 243                          * group is really just a noop.
 244                          */
 245                         cmd->nonsel = 1;
 246                         if (stack == NULL)
 247                                 errx(1, "%lu: %s: unexpected }", linenum, fname);
 248                         cmd2 = stack;
 249                         stack = cmd2->next;
 250                         cmd2->next = cmd;
 251                         /*FALLTHROUGH*/
 252                 case EMPTY:             /* d D g G h H l n N p P q x = \0 */
 253                         p++;
 254                         EATSPACE();
 255                         if (*p == ';') {
 256                                 p++;
 257                                 link = &cmd->next;
 258                                 goto semicolon;
 259                         }
 260                         if (*p)
 261                                 errx(1, "%lu: %s: extra characters at the end of %c command",
 262                                                 linenum, fname, cmd->code);
 263                         break;
 264                 case TEXT:                      /* a c i */
 265                         p++;
 266                         EATSPACE();
 267                         if (*p != '\\')
 268                                 errx(1,
 269 "%lu: %s: command %c expects \\ followed by text", linenum, fname, cmd->code);
 270                         p++;
 271                         EATSPACE();
 272                         if (*p)
 273                                 errx(1,
 274                                 "%lu: %s: extra characters after \\ at the end of %c command",
 275                                 linenum, fname, cmd->code);
 276                         cmd->t = compile_text();
 277                         break;
 278                 case COMMENT:                   /* \0 # */
 279                         break;
 280                 case WFILE:                     /* w */
 281                         p++;
 282                         EATSPACE();
 283                         if (*p == '\0')
 284                                 errx(1, "%lu: %s: filename expected", linenum, fname);
 285                         cmd->t = duptoeol(p, "w command");
 286                         if (aflag)
 287                                 cmd->u.fd = -1;
 288                         else if ((cmd->u.fd = open(p,
 289                             O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
 290                             DEFFILEMODE)) == -1)
 291                                 err(1, "%s", p);
 292                         break;
 293                 case RFILE:                     /* r */
 294                         p++;
 295                         EATSPACE();
 296                         if (*p == '\0')
 297                                 errx(1, "%lu: %s: filename expected", linenum, fname);
 298                         else
 299                                 cmd->t = duptoeol(p, "read command");
 300                         break;
 301                 case BRANCH:                    /* b t */
 302                         p++;
 303                         EATSPACE();
 304                         if (*p == '\0')
 305                                 cmd->t = NULL;
 306                         else
 307                                 cmd->t = duptoeol(p, "branch");
 308                         break;
 309                 case LABEL:                     /* : */
 310                         p++;
 311                         EATSPACE();
 312                         cmd->t = duptoeol(p, "label");
 313                         if (strlen(p) == 0)
 314                                 errx(1, "%lu: %s: empty label", linenum, fname);
 315                         enterlabel(cmd);
 316                         break;
 317                 case SUBST:                     /* s */
 318                         p++;
 319                         if (*p == '\0' || *p == '\\')
 320                                 errx(1,
 321 "%lu: %s: substitute pattern can not be delimited by newline or backslash",
 322                                         linenum, fname);
 323                         if ((cmd->u.s = calloc(1, sizeof(struct s_subst))) == NULL)
 324                                 err(1, "malloc");
 325                         p = compile_delimited(p, re, 0);
 326                         if (p == NULL)
 327                                 errx(1,
 328                                 "%lu: %s: unterminated substitute pattern", linenum, fname);
 329
 330                         /* Compile RE with no case sensitivity temporarily */
 331                         if (*re == '\0')
 332                                 cmd->u.s->re = NULL;
 333                         else
 334                                 cmd->u.s->re = compile_re(re, 0);
 335                         --p;
 336                         p = compile_subst(p, cmd->u.s);
 337                         p = compile_flags(p, cmd->u.s);
 338
 339                         /* Recompile RE with case sensitivity from "I" flag if any */
 340                         if (*re == '\0')
 341                                 cmd->u.s->re = NULL;
 342                         else
 343                                 cmd->u.s->re = compile_re(re, cmd->u.s->icase);
 344                         EATSPACE();
 345                         if (*p == ';') {
 346                                 p++;
 347                                 link = &cmd->next;
 348                                 goto semicolon;
 349                         }
 350                         break;
 351                 case TR:                        /* y */
 352                         p++;
 353                         p = compile_tr(p, &cmd->u.y);
 354                         EATSPACE();
 355                         if (*p == ';') {
 356                                 p++;
 357                                 link = &cmd->next;
 358                                 goto semicolon;
 359                         }
 360                         if (*p)
 361                                 errx(1,
 362 "%lu: %s: extra text at the end of a transform command", linenum, fname);
 363                         break;
 364                 }
 365         }
 366 }
 367
 368 /*
 369  * Get a delimited string.  P points to the delimiter of the string; d points
 370  * to a buffer area.  Newline and delimiter escapes are processed; other
 371  * escapes are ignored.
 372  *
 373  * Returns a pointer to the first character after the final delimiter or NULL
 374  * in the case of a non-terminated string.  The character array d is filled
 375  * with the processed string.
 376  */
 377 static char *
 378 compile_delimited(char *p, char *d, int is_tr)
 379 {
 380         char c;
 381
 382         c = *p++;
 383         if (c == '\0')
 384                 return (NULL);
 385         else if (c == '\\')
 386                 errx(1, "%lu: %s: \\ can not be used as a string delimiter",
 387                                 linenum, fname);
 388         else if (c == '\n')
 389                 errx(1, "%lu: %s: newline can not be used as a string delimiter",
 390                                 linenum, fname);
 391         while (*p) {
 392                 if (*p == '[' && *p != c) {
 393                         if ((d = compile_ccl(&p, d)) == NULL)
 394                                 errx(1, "%lu: %s: unbalanced brackets ([])", linenum, fname);
 395                         continue;
 396                 } else if (*p == '\\' && p[1] == '[') {
 397                         *d++ = *p++;
 398                 } else if (*p == '\\' && p[1] == c) {
 399                         p++;
 400                 } else if (*p == '\\' &&
 401                     (p[1] == 'n' || p[1] == 'r' || p[1] == 't')) {
 402                         switch (p[1]) {
 403                         case 'n':
 404                                 *d++ = '\n';
 405                                 break;
 406                         case 'r':
 407                                 *d++ = '\r';
 408                                 break;
 409                         case 't':
 410                                 *d++ = '\t';
 411                                 break;
 412                         }
 413                         p += 2;
 414                         continue;
 415                 } else if (*p == '\\' && p[1] == '\\') {
 416                         if (is_tr)
 417                                 p++;
 418                         else
 419                                 *d++ = *p++;
 420                 } else if (*p == c) {
 421                         *d = '\0';
 422                         return (p + 1);
 423                 }
 424                 *d++ = *p++;
 425         }
 426         return (NULL);
 427 }
 428
 429
 430 /* compile_ccl: expand a POSIX character class */
 431 static char *
 432 compile_ccl(char **sp, char *t)
 433 {
 434         int c, d;
 435         char *s = *sp;
 436
 437         *t++ = *s++;
 438         if (*s == '^')
 439                 *t++ = *s++;
 440         if (*s == ']')
 441                 *t++ = *s++;
 442         for (; *s && (*t = *s) != ']'; s++, t++) {
 443                 if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) {
 444                         *++t = *++s, t++, s++;
 445                         for (c = *s; (*t = *s) != ']' || c != d; s++, t++)
 446                                 if ((c = *s) == '\0')
 447                                         return NULL;
 448                 }
 449                 /*
 450                  * Apple Note: FreeBSD treats the backslash character as the start of
 451                  * several possible escape sequences here, but this is not POSIX-
 452                  * compliant.  See XBD 9.3.3: "The period, left-bracket, and backslash
 453                  * shall be special except when used in a bracket expression".
 454                  */
 455         }
 456         return (*s == ']') ? *sp = ++s, ++t : NULL;
 457 }
 458
 459 /*
 460  * Compiles the regular expression in RE and returns a pointer to the compiled
 461  * regular expression.
 462  * Cflags are passed to regcomp.
 463  */
 464 static regex_t *
 465 compile_re(char *re, int case_insensitive)
 466 {
 467         regex_t *rep;
 468         int eval, flags;
 469
 470
 471         flags = rflags;
 472         if (case_insensitive)
 473                 flags |= REG_ICASE;
 474         if ((rep = malloc(sizeof(regex_t))) == NULL)
 475                 err(1, "malloc");
 476         if ((eval = regcomp(rep, re, flags)) != 0)
 477                 errx(1, "%lu: %s: RE error: %s",
 478                                 linenum, fname, strregerror(eval, rep));
 479         if (maxnsub < rep->re_nsub)
 480                 maxnsub = rep->re_nsub;
 481         return (rep);
 482 }
 483
 484 /*
 485  * Compile the substitution string of a regular expression and set res to
 486  * point to a saved copy of it.  Nsub is the number of parenthesized regular
 487  * expressions.
 488  */
 489 static char *
 490 compile_subst(char *p, struct s_subst *s)
 491 {
 492         static char lbuf[_POSIX2_LINE_MAX + 1];
 493         int asize, size;
 494         u_char ref;
 495         char c, *text, *op, *sp;
 496         int more = 1, sawesc = 0;
 497
 498         c = *p++;                       /* Terminator character */
 499         if (c == '\0')
 500                 return (NULL);
 501
 502         s->maxbref = 0;
 503         s->linenum = linenum;
 504         asize = 2 * _POSIX2_LINE_MAX + 1;
 505         if ((text = malloc(asize)) == NULL)
 506                 err(1, "malloc");
 507         size = 0;
 508         do {
 509                 op = sp = text + size;
 510                 for (; *p; p++) {
 511                         if (*p == '\\' || sawesc) {
 512                                 /*
 513                                  * If this is a continuation from the last
 514                                  * buffer, we won't have a character to
 515                                  * skip over.
 516                                  */
 517                                 if (sawesc)
 518                                         sawesc = 0;
 519                                 else
 520                                         p++;
 521
 522                                 if (*p == '\0') {
 523                                         /*
 524                                          * This escaped character is continued
 525                                          * in the next part of the line.  Note
 526                                          * this fact, then cause the loop to
 527                                          * exit w/ normal EOL case and reenter
 528                                          * above with the new buffer.
 529                                          */
 530                                         sawesc = 1;
 531                                         p--;
 532                                         continue;
 533                                 } else if (strchr("123456789", *p) != NULL) {
 534                                         *sp++ = '\\';
 535                                         ref = *p - '0';
 536                                         if (s->re != NULL &&
 537                                             ref > s->re->re_nsub)
 538                                                 errx(1, "%lu: %s: \\%c not defined in the RE",
 539                                                                 linenum, fname, *p);
 540                                         if (s->maxbref < ref)
 541                                                 s->maxbref = ref;
 542                                 } else {
 543                                         switch (*p) {
 544                                         case '&':
 545                                         case '\\':
 546                                                 *sp++ = '\\';
 547                                                 break;
 548                                         case 'n':
 549                                                 *p = '\n';
 550                                                 break;
 551                                         case 'r':
 552                                                 *p = '\r';
 553                                                 break;
 554                                         case 't':
 555                                                 *p = '\t';
 556                                                 break;
 557                                         }
 558                                 }
 559                         } else if (*p == c) {
 560                                 if (*++p == '\0' && more) {
 561                                         if (cu_fgets(lbuf, sizeof(lbuf), &more))
 562                                                 p = lbuf;
 563                                 }
 564                                 *sp++ = '\0';
 565                                 size += sp - op;
 566                                 if ((s->new = realloc(text, size)) == NULL)
 567                                         err(1, "realloc");
 568                                 return (p);
 569                         } else if (*p == '\n') {
 570                                 errx(1,
 571 "%lu: %s: unescaped newline inside substitute pattern", linenum, fname);
 572                                 /* NOTREACHED */
 573                         }
 574                         *sp++ = *p;
 575                 }
 576                 size += sp - op;
 577                 if (asize - size < _POSIX2_LINE_MAX + 1) {
 578                         asize *= 2;
 579                         if ((text = realloc(text, asize)) == NULL)
 580                                 err(1, "realloc");
 581                 }
 582         } while (cu_fgets(p = lbuf, sizeof(lbuf), &more) != NULL);
 583         errx(1, "%lu: %s: unterminated substitute in regular expression",
 584                         linenum, fname);
 585         /* NOTREACHED */
 586 }
 587
 588 /*
 589  * Compile the flags of the s command
 590  */
 591 static char *
 592 compile_flags(char *p, struct s_subst *s)
 593 {
 594         int gn;                 /* True if we have seen g or n */
 595         unsigned long nval;
 596         char wfile[_POSIX2_LINE_MAX + 1], *q, *eq;
 597
 598         s->n = 1;                               /* Default */
 599         s->p = 0;
 600         s->wfile = NULL;
 601         s->wfd = -1;
 602         s->icase = 0;
 603         for (gn = 0;;) {
 604                 EATSPACE();                     /* EXTENSION */
 605                 switch (*p) {
 606                 case 'g':
 607                         if (gn)
 608                                 errx(1,
 609 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname);
 610                         gn = 1;
 611                         s->n = 0;
 612                         break;
 613                 case '\0':
 614                 case '\n':
 615                 case ';':
 616                         return (p);
 617                 case 'p':
 618                         s->p = 1;
 619                         break;
 620                 case 'i':
 621                 case 'I':
 622                         s->icase = 1;
 623                         break;
 624                 case '1': case '2': case '3':
 625                 case '4': case '5': case '6':
 626                 case '7': case '8': case '9':
 627                         if (gn)
 628                                 errx(1,
 629 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname);
 630                         gn = 1;
 631                         errno = 0;
 632                         nval = strtol(p, &p, 10);
 633                         if (errno == ERANGE || nval > INT_MAX)
 634                                 errx(1,
 635 "%lu: %s: overflow in the 'N' substitute flag", linenum, fname);
 636                         s->n = nval;
 637                         p--;
 638                         break;
 639                 case 'w':
 640                         p++;
 641 #ifdef HISTORIC_PRACTICE
 642                         if (*p != ' ') {
 643                                 warnx("%lu: %s: space missing before w wfile", linenum, fname);
 644                                 return (p);
 645                         }
 646 #endif
 647                         EATSPACE();
 648                         q = wfile;
 649                         eq = wfile + sizeof(wfile) - 1;
 650                         while (*p) {
 651                                 if (*p == '\n')
 652                                         break;
 653                                 if (q >= eq)
 654                                         err(1, "wfile too long");
 655                                 *q++ = *p++;
 656                         }
 657                         *q = '\0';
 658                         if (q == wfile)
 659                                 errx(1, "%lu: %s: no wfile specified", linenum, fname);
 660                         s->wfile = strdup(wfile);
 661                         if (!aflag && (s->wfd = open(wfile,
 662                             O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
 663                             DEFFILEMODE)) == -1)
 664                                 err(1, "%s", wfile);
 665                         return (p);
 666                 default:
 667                         errx(1, "%lu: %s: bad flag in substitute command: '%c'",
 668                                         linenum, fname, *p);
 669                         break;
 670                 }
 671                 p++;
 672         }
 673 }
 674
 675 /*
 676  * Compile a translation set of strings into a lookup table.
 677  */
 678 static char *
 679 compile_tr(char *p, struct s_tr **py)
 680 {
 681         struct s_tr *y;
 682         int i;
 683         const char *op, *np;
 684         char old[_POSIX2_LINE_MAX + 1];
 685         char new[_POSIX2_LINE_MAX + 1];
 686         size_t oclen, oldlen, nclen, newlen;
 687         mbstate_t mbs1, mbs2;
 688
 689         if ((*py = y = malloc(sizeof(*y))) == NULL)
 690                 err(1, NULL);
 691         y->multis = NULL;
 692         y->nmultis = 0;
 693
 694         if (*p == '\0' || *p == '\\')
 695                 errx(1,
 696         "%lu: %s: transform pattern can not be delimited by newline or backslash",
 697                         linenum, fname);
 698         p = compile_delimited(p, old, 1);
 699         if (p == NULL)
 700                 errx(1, "%lu: %s: unterminated transform source string",
 701                                 linenum, fname);
 702         p = compile_delimited(p - 1, new, 1);
 703         if (p == NULL)
 704                 errx(1, "%lu: %s: unterminated transform target string",
 705                                 linenum, fname);
 706         EATSPACE();
 707         op = old;
 708         oldlen = mbsrtowcs(NULL, &op, 0, NULL);
 709         if (oldlen == (size_t)-1)
 710                 err(1, NULL);
 711         np = new;
 712         newlen = mbsrtowcs(NULL, &np, 0, NULL);
 713         if (newlen == (size_t)-1)
 714                 err(1, NULL);
 715         if (newlen != oldlen)
 716                 errx(1, "%lu: %s: transform strings are not the same length",
 717                                 linenum, fname);
 718         if (MB_CUR_MAX == 1) {
 719                 /*
 720                  * The single-byte encoding case is easy: generate a
 721                  * lookup table.
 722                  */
 723                 for (i = 0; i <= UCHAR_MAX; i++)
 724                         y->bytetab[i] = (char)i;
 725                 for (; *op; op++, np++)
 726                         y->bytetab[(u_char)*op] = *np;
 727         } else {
 728                 /*
 729                  * Multi-byte encoding case: generate a lookup table as
 730                  * above, but only for single-byte characters. The first
 731                  * bytes of multi-byte characters have their lookup table
 732                  * entries set to 0, which causes do_tr() to search through
 733                  * an auxiliary vector of multi-byte mappings.
 734                  */
 735                 memset(&mbs1, 0, sizeof(mbs1));
 736                 memset(&mbs2, 0, sizeof(mbs2));
 737                 for (i = 0; i <= UCHAR_MAX; i++)
 738                         y->bytetab[i] = (btowc(i) != WEOF) ? i : 0;
 739                 while (*op != '\0') {
 740                         oclen = mbrlen(op, MB_LEN_MAX, &mbs1);
 741                         if (oclen == (size_t)-1 || oclen == (size_t)-2)
 742                                 errc(1, EILSEQ, NULL);
 743                         nclen = mbrlen(np, MB_LEN_MAX, &mbs2);
 744                         if (nclen == (size_t)-1 || nclen == (size_t)-2)
 745                                 errc(1, EILSEQ, NULL);
 746                         if (oclen == 1 && nclen == 1)
 747                                 y->bytetab[(u_char)*op] = *np;
 748                         else {
 749                                 y->bytetab[(u_char)*op] = 0;
 750                                 y->multis = realloc(y->multis,
 751                                     (y->nmultis + 1) * sizeof(*y->multis));
 752                                 if (y->multis == NULL)
 753                                         err(1, NULL);
 754                                 i = y->nmultis++;
 755                                 y->multis[i].fromlen = oclen;
 756                                 memcpy(y->multis[i].from, op, oclen);
 757                                 y->multis[i].tolen = nclen;
 758                                 memcpy(y->multis[i].to, np, nclen);
 759                         }
 760                         op += oclen;
 761                         np += nclen;
 762                 }
 763         }
 764         return (p);
 765 }
 766
 767 /*
 768  * Compile the text following an a, c, or i command.
 769  */
 770 static char *
 771 compile_text(void)
 772 {
 773         int asize, esc_nl, size;
 774         char *text, *p, *op, *s;
 775         char lbuf[_POSIX2_LINE_MAX + 1];
 776
 777         asize = 2 * _POSIX2_LINE_MAX + 1;
 778         if ((text = malloc(asize)) == NULL)
 779                 err(1, "malloc");
 780         size = 0;
 781         while (cu_fgets(lbuf, sizeof(lbuf), NULL) != NULL) {
 782                 op = s = text + size;
 783                 p = lbuf;
 784 #ifdef LEGACY_BSDSED_COMPAT
 785                 EATSPACE();
 786 #endif
 787                 for (esc_nl = 0; *p != '\0'; p++) {
 788                         if (*p == '\\' && p[1] != '\0' && *++p == '\n')
 789                                 esc_nl = 1;
 790                         *s++ = *p;
 791                 }
 792                 size += s - op;
 793                 if (!esc_nl) {
 794                         *s = '\0';
 795                         break;
 796                 }
 797                 if (asize - size < _POSIX2_LINE_MAX + 1) {
 798                         asize *= 2;
 799                         if ((text = realloc(text, asize)) == NULL)
 800                                 err(1, "realloc");
 801                 }
 802         }
 803         text[size] = '\0';
 804         if ((p = realloc(text, size + 1)) == NULL)
 805                 err(1, "realloc");
 806         return (p);
 807 }
 808
 809 /*
 810  * Get an address and return a pointer to the first character after
 811  * it.  Fill the structure pointed to according to the address.
 812  */
 813 static char *
 814 compile_addr(char *p, struct s_addr *a)
 815 {
 816         char *end, re[_POSIX2_LINE_MAX + 1];
 817         int icase;
 818
 819         icase = 0;
 820
 821         a->type = 0;
 822         switch (*p) {
 823         case '\\':                              /* Context address */
 824                 ++p;
 825                 /* FALLTHROUGH */
 826         case '/':                               /* Context address */
 827                 p = compile_delimited(p, re, 0);
 828                 if (p == NULL)
 829                         errx(1, "%lu: %s: unterminated regular expression", linenum, fname);
 830                 /* Check for case insensitive regexp flag */
 831                 if (*p == 'I') {
 832                         icase = 1;
 833                         p++;
 834                 }
 835                 if (*re == '\0')
 836                         a->u.r = NULL;
 837                 else
 838                         a->u.r = compile_re(re, icase);
 839                 a->type = AT_RE;
 840                 return (p);
 841
 842         case '$':                               /* Last line */
 843                 a->type = AT_LAST;
 844                 return (p + 1);
 845
 846         case '+':                               /* Relative line number */
 847                 a->type = AT_RELLINE;
 848                 p++;
 849                 /* FALLTHROUGH */
 850                                                 /* Line number */
 851         case '0': case '1': case '2': case '3': case '4':
 852         case '5': case '6': case '7': case '8': case '9':
 853                 if (a->type == 0)
 854                         a->type = AT_LINE;
 855                 a->u.l = strtol(p, &end, 10);
 856                 return (end);
 857         default:
 858                 errx(1, "%lu: %s: expected context address", linenum, fname);
 859                 return (NULL);
 860         }
 861 }
 862
 863 /*
 864  * duptoeol --
 865  *      Return a copy of all the characters up to \n or \0.
 866  */
 867 static char *
 868 duptoeol(char *s, const char *ctype)
 869 {
 870         size_t len;
 871         int ws;
 872         char *p, *start;
 873
 874         ws = 0;
 875         for (start = s; *s != '\0' && *s != '\n'; ++s)
 876                 ws = isspace((unsigned char)*s);
 877         *s = '\0';
 878         if (ws)
 879                 warnx("%lu: %s: whitespace after %s", linenum, fname, ctype);
 880         len = s - start + 1;
 881         if ((p = malloc(len)) == NULL)
 882                 err(1, "malloc");
 883         return (memmove(p, start, len));
 884 }
 885
 886 /*
 887  * Convert goto label names to addresses, and count a and r commands, in
 888  * the given subset of the script.  Free the memory used by labels in b
 889  * and t commands (but not by :).
 890  *
 891  * TODO: Remove } nodes
 892  */
 893 static void
 894 fixuplabel(struct s_command *cp, struct s_command *end)
 895 {
 896
 897         for (; cp != end; cp = cp->next)
 898                 switch (cp->code) {
 899                 case 'a':
 900                 case 'r':
 901                         appendnum++;
 902                         break;
 903                 case 'b':
 904                 case 't':
 905                         /* Resolve branch target. */
 906                         if (cp->t == NULL) {
 907                                 cp->u.c = NULL;
 908                                 break;
 909                         }
 910                         if ((cp->u.c = findlabel(cp->t)) == NULL)
 911                                 errx(1, "%lu: %s: undefined label '%s'", linenum, fname, cp->t);
 912                         free(cp->t);
 913                         break;
 914                 case '{':
 915                         /* Do interior commands. */
 916                         fixuplabel(cp->u.c, cp->next);
 917                         break;
 918                 }
 919 }
 920
 921 /*
 922  * Associate the given command label for later lookup.
 923  */
 924 static void
 925 enterlabel(struct s_command *cp)
 926 {
 927         struct labhash **lhp, *lh;
 928         u_char *p;
 929         u_int h, c;
 930
 931         for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++)
 932                 h = (h << 5) + h + c;
 933         lhp = &labels[h & LHMASK];
 934         for (lh = *lhp; lh != NULL; lh = lh->lh_next)
 935                 if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0)
 936                         errx(1, "%lu: %s: duplicate label '%s'", linenum, fname, cp->t);
 937         if ((lh = malloc(sizeof *lh)) == NULL)
 938                 err(1, "malloc");
 939         lh->lh_next = *lhp;
 940         lh->lh_hash = h;
 941         lh->lh_cmd = cp;
 942         lh->lh_ref = 0;
 943         *lhp = lh;
 944 }
 945
 946 /*
 947  * Find the label contained in the command l in the command linked
 948  * list cp.  L is excluded from the search.  Return NULL if not found.
 949  */
 950 static struct s_command *
 951 findlabel(char *name)
 952 {
 953         struct labhash *lh;
 954         u_char *p;
 955         u_int h, c;
 956
 957         for (h = 0, p = (u_char *)name; (c = *p) != 0; p++)
 958                 h = (h << 5) + h + c;
 959         for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) {
 960                 if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) {
 961                         lh->lh_ref = 1;
 962                         return (lh->lh_cmd);
 963                 }
 964         }
 965         return (NULL);
 966 }
 967
 968 /*
 969  * Warn about any unused labels.  As a side effect, release the label hash
 970  * table space.
 971  */
 972 static void
 973 uselabel(void)
 974 {
 975         struct labhash *lh, *next;
 976         int i;
 977
 978         for (i = 0; i < LHSZ; i++) {
 979                 for (lh = labels[i]; lh != NULL; lh = next) {
 980                         next = lh->lh_next;
 981                         if (!lh->lh_ref)
 982                                 warnx("%lu: %s: unused label '%s'",
 983                                     linenum, fname, lh->lh_cmd->t);
 984                         free(lh);
 985                 }
 986         }
 987 }