package/toolbox/src/sed/process.c

   1 /*      $OpenBSD: process.c,v 1.32 2017/02/22 14:09:09 tom Exp $        */
   2
   3 /*-
   4  * Copyright (c) 2015
   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 #include <sys/uio.h>
  41
  42 __SCCSID("@(#)process.c 8.1 (Berkeley) 6/6/93");
  43 __RCSID("$MirOS: src/usr.bin/sed/process.c,v 1.5 2017/11/20 01:23:57 tg Exp $");
  44
  45 #include <ctype.h>
  46 #include <errno.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 #include <unistd.h>
  54
  55 #include "defs.h"
  56 #include "extern.h"
  57
  58 static SPACE HS, PS, SS;
  59 #define pd              PS.deleted
  60 #define ps              PS.space
  61 #define psl             PS.len
  62 #define psanl           PS.append_newline
  63 #define hs              HS.space
  64 #define hsl             HS.len
  65
  66 static inline int        applies(struct s_command *);
  67 static void              flush_appends(void);
  68 static void              lputs(char *);
  69 static inline int        regexec_e(regex_t *, const char *, int, int, size_t,
  70                              size_t);
  71 static void              regsub(SPACE *, char *, char *);
  72 static int               substitute(struct s_command *);
  73
  74 struct s_appends *appends;      /* Array of pointers to strings to append. */
  75 static size_t appendx;          /* Index into appends array. */
  76 size_t appendnum;               /* Size of appends array. */
  77
  78 static int lastaddr;            /* Set by applies if last address of a range. */
  79 static int sdone;               /* If any substitutes since last line input. */
  80                                 /* Iov structure for 'w' commands. */
  81 static regex_t *defpreg;
  82 size_t maxnsub;
  83 regmatch_t *match;
  84
  85 #define OUT() do {\
  86         fwrite(ps, 1, psl, outfile);\
  87         if (psanl) fputc('\n', outfile);\
  88 } while (0)
  89
  90 void
  91 process(void)
  92 {
  93         struct s_command *cp;
  94         SPACE tspace;
  95         size_t len, oldpsl;
  96         char *p;
  97
  98         for (linenum = 0; mf_fgets(&PS, REPLACE);) {
  99                 pd = 0;
 100 top:
 101                 cp = prog;
 102 redirect:
 103                 while (cp != NULL) {
 104                         if (!applies(cp)) {
 105                                 cp = cp->next;
 106                                 continue;
 107                         }
 108                         switch (cp->code) {
 109                         case '{':
 110                                 cp = cp->u.c;
 111                                 goto redirect;
 112                         case 'a':
 113                                 if (appendx >= appendnum) {
 114                                         appends = xreallocarray(appends,
 115                                             appendnum,
 116                                             2 * sizeof(struct s_appends));
 117                                         appendnum *= 2;
 118                                 }
 119                                 appends[appendx].type = AP_STRING;
 120                                 appends[appendx].s = cp->t;
 121                                 appends[appendx].len = strlen(cp->t);
 122                                 appendx++;
 123                                 break;
 124                         case 'b':
 125                                 cp = cp->u.c;
 126                                 goto redirect;
 127                         case 'c':
 128                                 pd = 1;
 129                                 psl = 0;
 130                                 if (cp->a2 == NULL || lastaddr || lastline())
 131                                         (void)fprintf(outfile, "%s", cp->t);
 132                                 break;
 133                         case 'd':
 134                                 pd = 1;
 135                                 goto new;
 136                         case 'D':
 137                                 if (pd)
 138                                         goto new;
 139                                 if (psl == 0 ||
 140                                     (p = memchr(ps, '\n', psl)) == NULL) {
 141                                         pd = 1;
 142                                         goto new;
 143                                 } else {
 144                                         psl -= (p + 1) - ps;
 145                                         memmove(ps, p + 1, psl);
 146                                         goto top;
 147                                 }
 148                         case 'g':
 149                                 cspace(&PS, hs, hsl, REPLACE);
 150                                 break;
 151                         case 'G':
 152                                 cspace(&PS, "\n", 1, 0);
 153                                 cspace(&PS, hs, hsl, 0);
 154                                 break;
 155                         case 'h':
 156                                 cspace(&HS, ps, psl, REPLACE);
 157                                 break;
 158                         case 'H':
 159                                 cspace(&HS, "\n", 1, 0);
 160                                 cspace(&HS, ps, psl, 0);
 161                                 break;
 162                         case 'i':
 163                                 (void)fprintf(outfile, "%s", cp->t);
 164                                 break;
 165                         case 'l':
 166                                 lputs(ps);
 167                                 break;
 168                         case 'n':
 169                                 if (!nflag && !pd)
 170                                         OUT();
 171                                 flush_appends();
 172                                 if (!mf_fgets(&PS, REPLACE))
 173                                         exit(0);
 174                                 pd = 0;
 175                                 break;
 176                         case 'N':
 177                                 flush_appends();
 178                                 cspace(&PS, "\n", 1, 0);
 179                                 if (!mf_fgets(&PS, 0))
 180                                         exit(0);
 181                                 break;
 182                         case 'p':
 183                                 if (pd)
 184                                         break;
 185                                 OUT();
 186                                 break;
 187                         case 'P':
 188                                 if (pd)
 189                                         break;
 190                                 if ((p = memchr(ps, '\n', psl)) != NULL) {
 191                                         oldpsl = psl;
 192                                         psl = p - ps;
 193                                         psanl = 1;
 194                                         OUT();
 195                                         psl = oldpsl;
 196                                 } else {
 197                                         OUT();
 198                                 }
 199                                 break;
 200                         case 'q':
 201                                 if (!nflag && !pd)
 202                                         OUT();
 203                                 flush_appends();
 204                                 exit(0);
 205                         case 'r':
 206                                 if (appendx >= appendnum) {
 207                                         appends = xreallocarray(appends,
 208                                             appendnum,
 209                                             2 * sizeof(struct s_appends));
 210                                         appendnum *= 2;
 211                                 }
 212                                 appends[appendx].type = AP_FILE;
 213                                 appends[appendx].s = cp->t;
 214                                 appends[appendx].len = strlen(cp->t);
 215                                 appendx++;
 216                                 break;
 217                         case 's':
 218                                 sdone |= substitute(cp);
 219                                 break;
 220                         case 't':
 221                                 if (sdone) {
 222                                         sdone = 0;
 223                                         cp = cp->u.c;
 224                                         goto redirect;
 225                                 }
 226                                 break;
 227                         case 'w':
 228                                 if (pd)
 229                                         break;
 230                                 if (cp->u.fd == -1 && (cp->u.fd = open(cp->t,
 231                                     O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
 232                                     DEFFILEMODE)) == -1)
 233                                         error(FATAL, "%s: %s",
 234                                             cp->t, strerror(errno));
 235                                 if ((size_t)write(cp->u.fd, ps, psl) != psl ||
 236                                     write(cp->u.fd, "\n", 1) != 1)
 237                                         error(FATAL, "%s: %s",
 238                                             cp->t, strerror(errno));
 239                                 break;
 240                         case 'x':
 241                                 if (hs == NULL)
 242                                         cspace(&HS, "", 0, REPLACE);
 243                                 tspace = PS;
 244                                 PS = HS;
 245                                 psanl = tspace.append_newline;
 246                                 HS = tspace;
 247                                 break;
 248                         case 'y':
 249                                 if (pd || psl == 0)
 250                                         break;
 251                                 for (p = ps, len = psl; len--; ++p)
 252                                         *p = cp->u.y[(unsigned char)*p];
 253                                 break;
 254                         case ':':
 255                         case '}':
 256                                 break;
 257                         case '=':
 258                                 (void)fprintf(outfile, "%lu\n", linenum);
 259                         }
 260                         cp = cp->next;
 261                 } /* for all cp */
 262
 263 new:            if (!nflag && !pd)
 264                         OUT();
 265                 flush_appends();
 266         } /* for all lines */
 267 }
 268
 269 /*
 270  * TRUE if the address passed matches the current program state
 271  * (lastline, linenumber, ps).
 272  */
 273 #define MATCH(a)                                                \
 274         (a)->type == AT_RE ? regexec_e((a)->u.r, ps, 0, 1, 0, psl) :    \
 275             (a)->type == AT_LINE ? linenum == (a)->u.l : lastline()
 276
 277 /*
 278  * Return TRUE if the command applies to the current line.  Sets the inrange
 279  * flag to process ranges.  Interprets the non-select (``!'') flag.
 280  */
 281 static inline int
 282 applies(struct s_command *cp)
 283 {
 284         int r;
 285
 286         lastaddr = 0;
 287         if (cp->a1 == NULL && cp->a2 == NULL)
 288                 r = 1;
 289         else if (cp->a2)
 290                 if (cp->inrange) {
 291                         if (MATCH(cp->a2)) {
 292                                 cp->inrange = 0;
 293                                 lastaddr = 1;
 294                         }
 295                         r = 1;
 296                 } else if (MATCH(cp->a1)) {
 297                         /*
 298                          * If the second address is a number less than or
 299                          * equal to the line number first selected, only
 300                          * one line shall be selected.
 301                          *      -- POSIX 1003.2
 302                          */
 303                         if (cp->a2->type == AT_LINE &&
 304                             linenum >= cp->a2->u.l)
 305                                 lastaddr = 1;
 306                         else
 307                                 cp->inrange = 1;
 308                         r = 1;
 309                 } else
 310                         r = 0;
 311         else
 312                 r = MATCH(cp->a1);
 313         return (cp->nonsel ? !r : r);
 314 }
 315
 316 /*
 317  * Reset all inrange markers.
 318  */
 319 void
 320 resetranges(void)
 321 {
 322         struct s_command *cp;
 323
 324         for (cp = prog; cp; cp = cp->code == '{' ? cp->u.c : cp->next)
 325                 if (cp->a2)
 326                         cp->inrange = 0;
 327 }
 328
 329 /*
 330  * substitute --
 331  *      Do substitutions in the pattern space.  Currently, we build a
 332  *      copy of the new pattern space in the substitute space structure
 333  *      and then swap them.
 334  */
 335 static int
 336 substitute(struct s_command *cp)
 337 {
 338         SPACE tspace;
 339         regex_t *re;
 340         regoff_t slen;
 341         int n, lastempty;
 342         size_t le = 0;
 343         char *s;
 344
 345         s = ps;
 346         re = cp->u.s->re;
 347         if (re == NULL) {
 348                 if (defpreg != NULL && cp->u.s->maxbref > defpreg->re_nsub) {
 349                         linenum = cp->u.s->linenum;
 350                         error(COMPILE, "\\%zu not defined in the RE",
 351                             cp->u.s->maxbref);
 352                 }
 353         }
 354         if (!regexec_e(re, ps, 0, 0, 0, psl))
 355                 return (0);
 356
 357         SS.len = 0;                             /* Clean substitute space. */
 358         slen = psl;
 359         n = cp->u.s->n;
 360         lastempty = 1;
 361
 362         do {
 363                 /* Copy the leading retained string. */
 364                 if (n <= 1 && (match[0].rm_so > le))
 365                         cspace(&SS, s, match[0].rm_so - le, APPEND);
 366
 367                 /* Skip zero-length matches right after other matches. */
 368                 if (lastempty || (match[0].rm_so - le) ||
 369                     match[0].rm_so != match[0].rm_eo) {
 370                         if (n <= 1) {
 371                                 /* Want this match: append replacement. */
 372                                 regsub(&SS, ps, cp->u.s->new);
 373                                 if (n == 1)
 374                                         n = -1;
 375                         } else {
 376                                 /* Want a later match: append original. */
 377                                 if (match[0].rm_eo - le)
 378                                         cspace(&SS, s, match[0].rm_eo - le,
 379                                             APPEND);
 380                                 n--;
 381                         }
 382                 }
 383
 384                 /* Move past this match. */
 385                 s = ps + match[0].rm_eo;
 386                 slen = psl - match[0].rm_eo;
 387                 le = match[0].rm_eo;
 388
 389                 /*
 390                  * After a zero-length match, advance one byte,
 391                  * and at the end of the line, terminate.
 392                  */
 393                 if (match[0].rm_so == match[0].rm_eo) {
 394                         if (*s == '\0' || *s == '\n')
 395                                 slen = -1;
 396                         else
 397                                 slen--;
 398                         if (*s != '\0') {
 399                                 cspace(&SS, s++, 1, APPEND);
 400                                 le++;
 401                         }
 402                         lastempty = 1;
 403                 } else
 404                         lastempty = 0;
 405
 406         } while (n >= 0 && slen >= 0 &&
 407             regexec_e(re, ps, REG_NOTBOL, 0, le, psl));
 408
 409         /* Did not find the requested number of matches. */
 410         if (n > 0)
 411                 return (0);
 412
 413         /* Copy the trailing retained string. */
 414         if (slen > 0)
 415                 cspace(&SS, s, slen, APPEND);
 416
 417         /*
 418          * Swap the substitute space and the pattern space, and make sure
 419          * that any leftover pointers into stdio memory get lost.
 420          */
 421         tspace = PS;
 422         PS = SS;
 423         psanl = tspace.append_newline;
 424         SS = tspace;
 425         SS.space = SS.back;
 426
 427         /* Handle the 'p' flag. */
 428         if (cp->u.s->p)
 429                 OUT();
 430
 431         /* Handle the 'w' flag. */
 432         if (cp->u.s->wfile && !pd) {
 433                 if (cp->u.s->wfd == -1 && (cp->u.s->wfd = open(cp->u.s->wfile,
 434                     O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, DEFFILEMODE)) == -1)
 435                         error(FATAL, "%s: %s", cp->u.s->wfile, strerror(errno));
 436                 if ((size_t)write(cp->u.s->wfd, ps, psl) != psl ||
 437                     write(cp->u.s->wfd, "\n", 1) != 1)
 438                         error(FATAL, "%s: %s", cp->u.s->wfile, strerror(errno));
 439         }
 440         return (1);
 441 }
 442
 443 /*
 444  * Flush append requests.  Always called before reading a line,
 445  * therefore it also resets the substitution done (sdone) flag.
 446  */
 447 static void
 448 flush_appends(void)
 449 {
 450         FILE *f;
 451         int count;
 452         size_t i;
 453         char buf[8 * 1024];
 454
 455         for (i = 0; i < appendx; i++)
 456                 switch (appends[i].type) {
 457                 case AP_STRING:
 458                         fwrite(appends[i].s, sizeof(char), appends[i].len,
 459                             outfile);
 460                         break;
 461                 case AP_FILE:
 462                         /*
 463                          * Read files probably shouldn't be cached.  Since
 464                          * it's not an error to read a non-existent file,
 465                          * it's possible that another program is interacting
 466                          * with the sed script through the filesystem.  It
 467                          * would be truly bizarre, but possible.  It's probably
 468                          * not that big a performance win, anyhow.
 469                          */
 470                         if ((f = fopen(appends[i].s, "r")) == NULL)
 471                                 break;
 472                         while ((count = fread(buf, sizeof(char), sizeof(buf), f)))
 473                                 (void)fwrite(buf, sizeof(char), count, outfile);
 474                         (void)fclose(f);
 475                         break;
 476                 }
 477         if (ferror(outfile))
 478                 error(FATAL, "%s: %s", outfname, strerror(errno ? errno : EIO));
 479         appendx = sdone = 0;
 480 }
 481
 482 static void
 483 lputs(char *s)
 484 {
 485         int count;
 486         extern int termwidth;
 487         const char *escapes;
 488         char *p;
 489
 490         for (count = 0; *s; ++s) {
 491                 if (count >= termwidth) {
 492                         (void)fprintf(outfile, "\\\n");
 493                         count = 0;
 494                 }
 495                 if (isascii((unsigned char)*s) && isprint((unsigned char)*s)
 496                     && *s != '\\') {
 497                         (void)fputc(*s, outfile);
 498                         count++;
 499                 } else {
 500                         escapes = "\\\a\b\f\n\r\t\v";
 501                         if (*s == '\n' && s[1] == '\0') {
 502                                 /* omit trailing newline */
 503                                 break;
 504                         }
 505                         (void)fputc('\\', outfile);
 506                         if ((p = strchr(escapes, *s))) {
 507                                 (void)fputc("\\abfnrtv"[p - escapes], outfile);
 508                                 count += 2;
 509                         } else {
 510                                 (void)fprintf(outfile, "%03o", *(u_char *)s);
 511                                 count += 4;
 512                         }
 513                 }
 514         }
 515         (void)fputc('$', outfile);
 516         (void)fputc('\n', outfile);
 517         if (ferror(outfile))
 518                 error(FATAL, "%s: %s", outfname, strerror(errno ? errno : EIO));
 519 }
 520
 521 static inline int
 522 regexec_e(regex_t *preg, const char *string, int eflags,
 523     int nomatch, size_t start, size_t stop)
 524 {
 525         int eval;
 526
 527         if (preg == NULL) {
 528                 if (defpreg == NULL)
 529                         error(FATAL, "first RE may not be empty");
 530         } else
 531                 defpreg = preg;
 532
 533         /* Set anchors */
 534         match[0].rm_so = start;
 535         match[0].rm_eo = stop;
 536
 537         eval = regexec(defpreg, string,
 538             nomatch ? 0 : maxnsub + 1, match, eflags | REG_STARTEND);
 539         switch (eval) {
 540         case 0:
 541                 return (1);
 542         case REG_NOMATCH:
 543                 return (0);
 544         }
 545         error(FATAL, "RE error: %s", strregerror(eval, defpreg));
 546 }
 547
 548 /*
 549  * regsub - perform substitutions after a regexp match
 550  * Based on a routine by Henry Spencer
 551  */
 552 static void
 553 regsub(SPACE *sp, char *string, char *src)
 554 {
 555         int len, no;
 556         char c, *dst;
 557
 558 #define NEEDSP(reqlen)                                                  \
 559         if (sp->len + (reqlen) + 1 >= sp->blen) {                       \
 560                 size_t newlen = sp->blen + (reqlen) + 1024;             \
 561                 sp->space = sp->back = xrealloc(sp->back, newlen);      \
 562                 sp->blen = newlen;                                      \
 563                 dst = sp->space + sp->len;                              \
 564         }
 565
 566         dst = sp->space + sp->len;
 567         while ((c = *src++) != '\0') {
 568                 if (c == '&')
 569                         no = 0;
 570                 else if (c == '\\' && isdigit((unsigned char)*src))
 571                         no = *src++ - '0';
 572                 else
 573                         no = -1;
 574                 if (no < 0) {           /* Ordinary character. */
 575                         if (c == '\\' && (*src == '\\' || *src == '&'))
 576                                 c = *src++;
 577                         NEEDSP(1);
 578                         *dst++ = c;
 579                         ++sp->len;
 580                 } else if (match[no].rm_so != -1 && match[no].rm_eo != -1) {
 581                         len = match[no].rm_eo - match[no].rm_so;
 582                         NEEDSP(len);
 583                         memmove(dst, string + match[no].rm_so, len);
 584                         dst += len;
 585                         sp->len += len;
 586                 }
 587         }
 588         NEEDSP(1);
 589         *dst = '\0';
 590 }
 591
 592 /*
 593  * aspace --
 594  *      Append the source space to the destination space, allocating new
 595  *      space as necessary.
 596  */
 597 void
 598 cspace(SPACE *sp, const char *p, size_t len, enum e_spflag spflag)
 599 {
 600         size_t tlen;
 601
 602         /* Make sure SPACE has enough memory and ramp up quickly. */
 603         tlen = sp->len + len + 1;
 604         if (tlen > sp->blen) {
 605                 size_t newlen = tlen + 1024;
 606                 sp->space = sp->back = xrealloc(sp->back, newlen);
 607                 sp->blen = newlen;
 608         }
 609
 610         if (spflag == REPLACE)
 611                 sp->len = 0;
 612
 613         memmove(sp->space + sp->len, p, len);
 614
 615         sp->space[sp->len += len] = '\0';
 616 }
 617
 618 /*
 619  * Close all cached opened files and report any errors
 620  */
 621 void
 622 cfclose(struct s_command *cp, struct s_command *end)
 623 {
 624
 625         for (; cp != end; cp = cp->next)
 626                 switch (cp->code) {
 627                 case 's':
 628                         if (cp->u.s->wfd != -1 && close(cp->u.s->wfd))
 629                                 error(FATAL,
 630                                     "%s: %s", cp->u.s->wfile, strerror(errno));
 631                         cp->u.s->wfd = -1;
 632                         break;
 633                 case 'w':
 634                         if (cp->u.fd != -1 && close(cp->u.fd))
 635                                 error(FATAL, "%s: %s", cp->t, strerror(errno));
 636                         cp->u.fd = -1;
 637                         break;
 638                 case '{':
 639                         cfclose(cp->u.c, cp->next);
 640                         break;
 641                 }
 642 }