1 /*- 2 * Copyright (c) 1992 Diomidis Spinellis. 3 * Copyright (c) 1992, 1993, 1994 4 * The Regents of the University of California. All rights reserved. 5 * 6 * This code is derived from software contributed to Berkeley by 7 * Diomidis Spinellis of Imperial College, University of London. 8 * 9 * Redistribution and use in source and binary forms, with or without 10 * modification, are permitted provided that the following conditions 11 * are met: 12 * 1. Redistributions of source code must retain the above copyright 13 * notice, this list of conditions and the following disclaimer. 14 * 2. Redistributions in binary form must reproduce the above copyright 15 * notice, this list of conditions and the following disclaimer in the 16 * documentation and/or other materials provided with the distribution. 17 * 3. All advertising materials mentioning features or use of this software 18 * must display the following acknowledgement: 19 * This product includes software developed by the University of 20 * California, Berkeley and its contributors. 21 * 4. Neither the name of the University nor the names of its contributors 22 * may be used to endorse or promote products derived from this software 23 * without specific prior written permission. 24 * 25 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 26 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 27 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 28 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 29 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 30 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 31 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 32 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 33 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 34 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 35 * SUCH DAMAGE. 36 */ 37 38 #include <sys/cdefs.h> 39 __FBSDID("$FreeBSD$"); 40 41 #ifndef lint 42 static const char sccsid[] = "@(#)process.c 8.6 (Berkeley) 4/20/94"; 43 #endif 44 45 #include <sys/types.h> 46 #include <sys/stat.h> 47 #include <sys/ioctl.h> 48 #include <sys/uio.h> 49 50 #include <ctype.h> 51 #include <err.h> 52 #include <errno.h> 53 #include <fcntl.h> 54 #include <limits.h> 55 #include <regex.h> 56 #include <stdio.h> 57 #include <stdlib.h> 58 #include <string.h> 59 #include <unistd.h> 60 #include <wchar.h> 61 #include <wctype.h> 62 63 #include "defs.h" 64 #include "extern.h" 65 66 static SPACE HS, PS, SS; 67 #define pd PS.deleted 68 #define ps PS.space 69 #define psl PS.len 70 #define hs HS.space 71 #define hsl HS.len 72 73 static __inline int applies(struct s_command *); 74 static void flush_appends(void); 75 static void lputs(char *, size_t); 76 static __inline int regexec_e(regex_t *, const char *, int, int, size_t); 77 static void regsub(SPACE *, char *, char *); 78 static int substitute(struct s_command *); 79 80 struct s_appends *appends; /* Array of pointers to strings to append. */ 81 static int appendx; /* Index into appends array. */ 82 int appendnum; /* Size of appends array. */ 83 84 static int lastaddr; /* Set by applies if last address of a range. */ 85 static int sdone; /* If any substitutes since last line input. */ 86 /* Iov structure for 'w' commands. */ 87 static regex_t *defpreg; 88 size_t maxnsub; 89 regmatch_t *match; 90 91 #define OUT(s) { fwrite(s, sizeof(u_char), psl, outfile); fputc('\n', outfile); } 92 93 void 94 process(void) 95 { 96 struct s_command *cp; 97 SPACE tspace; 98 size_t len, oldpsl = 0; 99 char *p; 100 101 p = NULL; 102 103 for (linenum = 0; mf_fgets(&PS, REPLACE);) { 104 pd = 0; 105 top: 106 cp = prog; 107 redirect: 108 while (cp != NULL) { 109 if (!applies(cp)) { 110 cp = cp->next; 111 continue; 112 } 113 switch (cp->code) { 114 case '{': 115 cp = cp->u.c; 116 goto redirect; 117 case 'a': 118 if (appendx >= appendnum) 119 if ((appends = realloc(appends, 120 sizeof(struct s_appends) * 121 (appendnum *= 2))) == NULL) 122 err(1, "realloc"); 123 appends[appendx].type = AP_STRING; 124 appends[appendx].s = cp->t; 125 appends[appendx].len = strlen(cp->t); 126 appendx++; 127 break; 128 case 'b': 129 cp = cp->u.c; 130 goto redirect; 131 case 'c': 132 pd = 1; 133 psl = 0; 134 if (cp->a2 == NULL || lastaddr) 135 (void)fprintf(outfile, "%s", cp->t); 136 break; 137 case 'd': 138 pd = 1; 139 goto new; 140 case 'D': 141 if (pd) 142 goto new; 143 if (psl == 0 || 144 (p = memchr(ps, '\n', psl)) == NULL) { 145 pd = 1; 146 goto new; 147 } else { 148 psl -= (p + 1) - ps; 149 memmove(ps, p + 1, psl); 150 goto top; 151 } 152 case 'g': 153 cspace(&PS, hs, hsl, REPLACE); 154 break; 155 case 'G': 156 cspace(&PS, "\n", 1, 0); 157 cspace(&PS, hs, hsl, 0); 158 break; 159 case 'h': 160 cspace(&HS, ps, psl, REPLACE); 161 break; 162 case 'H': 163 cspace(&HS, "\n", 1, 0); 164 cspace(&HS, ps, psl, 0); 165 break; 166 case 'i': 167 (void)fprintf(outfile, "%s", cp->t); 168 break; 169 case 'l': 170 lputs(ps, psl); 171 break; 172 case 'n': 173 if (!nflag && !pd) 174 OUT(ps) 175 flush_appends(); 176 if (!mf_fgets(&PS, REPLACE)) 177 exit(0); 178 pd = 0; 179 break; 180 case 'N': 181 flush_appends(); 182 cspace(&PS, "\n", 1, 0); 183 if (!mf_fgets(&PS, 0)) 184 exit(0); 185 break; 186 case 'p': 187 if (pd) 188 break; 189 OUT(ps) 190 break; 191 case 'P': 192 if (pd) 193 break; 194 if (psl != 0 && 195 (p = memchr(ps, '\n', psl)) != NULL) { 196 oldpsl = psl; 197 psl = p - ps; 198 } 199 OUT(ps) 200 if (p != NULL) 201 psl = oldpsl; 202 break; 203 case 'q': 204 if (!nflag && !pd) 205 OUT(ps) 206 flush_appends(); 207 exit(0); 208 case 'r': 209 if (appendx >= appendnum) 210 if ((appends = realloc(appends, 211 sizeof(struct s_appends) * 212 (appendnum *= 2))) == NULL) 213 err(1, "realloc"); 214 appends[appendx].type = AP_FILE; 215 appends[appendx].s = cp->t; 216 appends[appendx].len = strlen(cp->t); 217 appendx++; 218 break; 219 case 's': 220 sdone |= substitute(cp); 221 break; 222 case 't': 223 if (sdone) { 224 sdone = 0; 225 cp = cp->u.c; 226 goto redirect; 227 } 228 break; 229 case 'w': 230 if (pd) 231 break; 232 if (cp->u.fd == -1 && (cp->u.fd = open(cp->t, 233 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, 234 DEFFILEMODE)) == -1) 235 err(1, "%s", cp->t); 236 if (write(cp->u.fd, ps, psl) != psl || 237 write(cp->u.fd, "\n", 1) != 1) 238 err(1, "%s", cp->t); 239 break; 240 case 'x': 241 if (hs == NULL) 242 cspace(&HS, "", 0, REPLACE); 243 tspace = PS; 244 PS = HS; 245 HS = tspace; 246 break; 247 case 'y': 248 if (pd || psl == 0) 249 break; 250 for (p = ps, len = psl; len--; ++p) 251 *p = cp->u.y[(unsigned char)*p]; 252 break; 253 case ':': 254 case '}': 255 break; 256 case '=': 257 (void)fprintf(outfile, "%lu\n", linenum); 258 } 259 cp = cp->next; 260 } /* for all cp */ 261 262 new: if (!nflag && !pd) 263 OUT(ps) 264 flush_appends(); 265 } /* for all lines */ 266 } 267 268 /* 269 * TRUE if the address passed matches the current program state 270 * (lastline, linenumber, ps). 271 */ 272 #define MATCH(a) \ 273 (a)->type == AT_RE ? regexec_e((a)->u.r, ps, 0, 1, psl) : \ 274 (a)->type == AT_LINE ? linenum == (a)->u.l : lastline() 275 276 /* 277 * Return TRUE if the command applies to the current line. Sets the inrange 278 * flag to process ranges. Interprets the non-select (``!'') flag. 279 */ 280 static __inline int 281 applies(struct s_command *cp) 282 { 283 int r; 284 285 lastaddr = 0; 286 if (cp->a1 == NULL && cp->a2 == NULL) 287 r = 1; 288 else if (cp->a2) 289 if (cp->inrange) { 290 if (MATCH(cp->a2)) { 291 cp->inrange = 0; 292 lastaddr = 1; 293 } 294 r = 1; 295 } else if (MATCH(cp->a1)) { 296 /* 297 * If the second address is a number less than or 298 * equal to the line number first selected, only 299 * one line shall be selected. 300 * -- POSIX 1003.2 301 */ 302 if (cp->a2->type == AT_LINE && 303 linenum >= cp->a2->u.l) 304 lastaddr = 1; 305 else 306 cp->inrange = 1; 307 r = 1; 308 } else 309 r = 0; 310 else 311 r = MATCH(cp->a1); 312 return (cp->nonsel ? ! r : r); 313 } 314 315 /* 316 * substitute -- 317 * Do substitutions in the pattern space. Currently, we build a 318 * copy of the new pattern space in the substitute space structure 319 * and then swap them. 320 */ 321 static int 322 substitute(struct s_command *cp) 323 { 324 SPACE tspace; 325 regex_t *re; 326 regoff_t re_off, slen; 327 int lastempty, n; 328 char *s; 329 330 s = ps; 331 re = cp->u.s->re; 332 if (re == NULL) { 333 if (defpreg != NULL && cp->u.s->maxbref > defpreg->re_nsub) { 334 linenum = cp->u.s->linenum; 335 errx(1, "%lu: %s: \\%d not defined in the RE", 336 linenum, fname, cp->u.s->maxbref); 337 } 338 } 339 if (!regexec_e(re, s, 0, 0, psl)) 340 return (0); 341 342 SS.len = 0; /* Clean substitute space. */ 343 slen = psl; 344 n = cp->u.s->n; 345 lastempty = 1; 346 347 switch (n) { 348 case 0: /* Global */ 349 do { 350 if (lastempty || match[0].rm_so != match[0].rm_eo) { 351 /* Locate start of replaced string. */ 352 re_off = match[0].rm_so; 353 /* Copy leading retained string. */ 354 cspace(&SS, s, re_off, APPEND); 355 /* Add in regular expression. */ 356 regsub(&SS, s, cp->u.s->new); 357 } 358 359 /* Move past this match. */ 360 if (match[0].rm_so != match[0].rm_eo) { 361 s += match[0].rm_eo; 362 slen -= match[0].rm_eo; 363 lastempty = 0; 364 } else { 365 if (match[0].rm_so < slen) 366 cspace(&SS, s + match[0].rm_so, 1, 367 APPEND); 368 s += match[0].rm_so + 1; 369 slen -= match[0].rm_so + 1; 370 lastempty = 1; 371 } 372 } while (slen >= 0 && regexec_e(re, s, REG_NOTBOL, 0, slen)); 373 /* Copy trailing retained string. */ 374 if (slen > 0) 375 cspace(&SS, s, slen, APPEND); 376 break; 377 default: /* Nth occurrence */ 378 while (--n) { 379 if (match[0].rm_eo == match[0].rm_so) 380 match[0].rm_eo = match[0].rm_so + 1; 381 s += match[0].rm_eo; 382 slen -= match[0].rm_eo; 383 if (slen < 0) 384 return (0); 385 if (!regexec_e(re, s, REG_NOTBOL, 0, slen)) 386 return (0); 387 } 388 /* FALLTHROUGH */ 389 case 1: /* 1st occurrence */ 390 /* Locate start of replaced string. */ 391 re_off = match[0].rm_so + (s - ps); 392 /* Copy leading retained string. */ 393 cspace(&SS, ps, re_off, APPEND); 394 /* Add in regular expression. */ 395 regsub(&SS, s, cp->u.s->new); 396 /* Copy trailing retained string. */ 397 s += match[0].rm_eo; 398 slen -= match[0].rm_eo; 399 cspace(&SS, s, slen, APPEND); 400 break; 401 } 402 403 /* 404 * Swap the substitute space and the pattern space, and make sure 405 * that any leftover pointers into stdio memory get lost. 406 */ 407 tspace = PS; 408 PS = SS; 409 SS = tspace; 410 SS.space = SS.back; 411 412 /* Handle the 'p' flag. */ 413 if (cp->u.s->p) 414 OUT(ps) 415 416 /* Handle the 'w' flag. */ 417 if (cp->u.s->wfile && !pd) { 418 if (cp->u.s->wfd == -1 && (cp->u.s->wfd = open(cp->u.s->wfile, 419 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, DEFFILEMODE)) == -1) 420 err(1, "%s", cp->u.s->wfile); 421 if (write(cp->u.s->wfd, ps, psl) != psl || 422 write(cp->u.s->wfd, "\n", 1) != 1) 423 err(1, "%s", cp->u.s->wfile); 424 } 425 return (1); 426 } 427 428 /* 429 * Flush append requests. Always called before reading a line, 430 * therefore it also resets the substitution done (sdone) flag. 431 */ 432 static void 433 flush_appends(void) 434 { 435 FILE *f; 436 int count, i; 437 char buf[8 * 1024]; 438 439 for (i = 0; i < appendx; i++) 440 switch (appends[i].type) { 441 case AP_STRING: 442 fwrite(appends[i].s, sizeof(char), appends[i].len, 443 outfile); 444 break; 445 case AP_FILE: 446 /* 447 * Read files probably shouldn't be cached. Since 448 * it's not an error to read a non-existent file, 449 * it's possible that another program is interacting 450 * with the sed script through the filesystem. It 451 * would be truly bizarre, but possible. It's probably 452 * not that big a performance win, anyhow. 453 */ 454 if ((f = fopen(appends[i].s, "r")) == NULL) 455 break; 456 while ((count = fread(buf, sizeof(char), sizeof(buf), f))) 457 (void)fwrite(buf, sizeof(char), count, outfile); 458 (void)fclose(f); 459 break; 460 } 461 if (ferror(outfile)) 462 errx(1, "%s: %s", outfname, strerror(errno ? errno : EIO)); 463 appendx = sdone = 0; 464 } 465 466 static void 467 lputs(char *s, size_t len) 468 { 469 static const char escapes[] = "\\\a\b\f\r\t\v"; 470 int c, col, width; 471 char *p; 472 struct winsize win; 473 static int termwidth = -1; 474 size_t clen, i; 475 wchar_t wc; 476 mbstate_t mbs; 477 478 if (outfile != stdout) 479 termwidth = 60; 480 if (termwidth == -1) { 481 if ((p = getenv("COLUMNS")) && *p != '\0') 482 termwidth = atoi(p); 483 else if (ioctl(STDOUT_FILENO, TIOCGWINSZ, &win) == 0 && 484 win.ws_col > 0) 485 termwidth = win.ws_col; 486 else 487 termwidth = 60; 488 } 489 490 memset(&mbs, 0, sizeof(mbs)); 491 col = 0; 492 while (len != 0) { 493 clen = mbrtowc(&wc, s, len, &mbs); 494 if (clen == 0) 495 clen = 1; 496 if (clen == (size_t)-1 || clen == (size_t)-2) { 497 wc = (unsigned char)*s; 498 clen = 1; 499 memset(&mbs, 0, sizeof(mbs)); 500 } 501 if (wc == '\n') { 502 if (col + 1 >= termwidth) 503 fprintf(outfile, "\\\n"); 504 fputc('$', outfile); 505 fputc('\n', outfile); 506 col = 0; 507 } else if (iswprint(wc)) { 508 width = wcwidth(wc); 509 if (col + width >= termwidth) { 510 fprintf(outfile, "\\\n"); 511 col = 0; 512 } 513 fwrite(s, 1, clen, outfile); 514 col += width; 515 } else if (wc != L'\0' && (c = wctob(wc)) != EOF && 516 (p = strchr(escapes, c)) != NULL) { 517 if (col + 2 >= termwidth) { 518 fprintf(outfile, "\\\n"); 519 col = 0; 520 } 521 fprintf(outfile, "\\%c", "\\abfrtv"[p - escapes]); 522 col += 2; 523 } else { 524 if (col + 4 * clen >= termwidth) { 525 fprintf(outfile, "\\\n"); 526 col = 0; 527 } 528 for (i = 0; i < clen; i++) 529 fprintf(outfile, "\\%03o", 530 (int)(unsigned char)s[i]); 531 col += 4 * clen; 532 } 533 s += clen; 534 len -= clen; 535 } 536 if (col + 1 >= termwidth) 537 fprintf(outfile, "\\\n"); 538 (void)fputc('$', outfile); 539 (void)fputc('\n', outfile); 540 if (ferror(outfile)) 541 errx(1, "%s: %s", outfname, strerror(errno ? errno : EIO)); 542 } 543 544 static __inline int 545 regexec_e(regex_t *preg, const char *string, int eflags, int nomatch, 546 size_t slen) 547 { 548 int eval; 549 550 if (preg == NULL) { 551 if (defpreg == NULL) 552 errx(1, "first RE may not be empty"); 553 } else 554 defpreg = preg; 555 556 /* Set anchors */ 557 match[0].rm_so = 0; 558 match[0].rm_eo = slen; 559 560 eval = regexec(defpreg, string, 561 nomatch ? 0 : maxnsub + 1, match, eflags | REG_STARTEND); 562 switch(eval) { 563 case 0: 564 return (1); 565 case REG_NOMATCH: 566 return (0); 567 } 568 errx(1, "RE error: %s", strregerror(eval, defpreg)); 569 /* NOTREACHED */ 570 } 571 572 /* 573 * regsub - perform substitutions after a regexp match 574 * Based on a routine by Henry Spencer 575 */ 576 static void 577 regsub(SPACE *sp, char *string, char *src) 578 { 579 int len, no; 580 char c, *dst; 581 582 #define NEEDSP(reqlen) \ 583 /* XXX What is the +1 for? */ \ 584 if (sp->len + (reqlen) + 1 >= sp->blen) { \ 585 sp->blen += (reqlen) + 1024; \ 586 if ((sp->space = sp->back = realloc(sp->back, sp->blen)) \ 587 == NULL) \ 588 err(1, "realloc"); \ 589 dst = sp->space + sp->len; \ 590 } 591 592 dst = sp->space + sp->len; 593 while ((c = *src++) != '\0') { 594 if (c == '&') 595 no = 0; 596 else if (c == '\\' && isdigit((unsigned char)*src)) 597 no = *src++ - '0'; 598 else 599 no = -1; 600 if (no < 0) { /* Ordinary character. */ 601 if (c == '\\' && (*src == '\\' || *src == '&')) 602 c = *src++; 603 NEEDSP(1); 604 *dst++ = c; 605 ++sp->len; 606 } else if (match[no].rm_so != -1 && match[no].rm_eo != -1) { 607 len = match[no].rm_eo - match[no].rm_so; 608 NEEDSP(len); 609 memmove(dst, string + match[no].rm_so, len); 610 dst += len; 611 sp->len += len; 612 } 613 } 614 NEEDSP(1); 615 *dst = '\0'; 616 } 617 618 /* 619 * aspace -- 620 * Append the source space to the destination space, allocating new 621 * space as necessary. 622 */ 623 void 624 cspace(SPACE *sp, const char *p, size_t len, enum e_spflag spflag) 625 { 626 size_t tlen; 627 628 /* Make sure SPACE has enough memory and ramp up quickly. */ 629 tlen = sp->len + len + 1; 630 if (tlen > sp->blen) { 631 sp->blen = tlen + 1024; 632 if ((sp->space = sp->back = realloc(sp->back, sp->blen)) == 633 NULL) 634 err(1, "realloc"); 635 } 636 637 if (spflag == REPLACE) 638 sp->len = 0; 639 640 memmove(sp->space + sp->len, p, len); 641 642 sp->space[sp->len += len] = '\0'; 643 } 644 645 /* 646 * Close all cached opened files and report any errors 647 */ 648 void 649 cfclose(struct s_command *cp, struct s_command *end) 650 { 651 652 for (; cp != end; cp = cp->next) 653 switch(cp->code) { 654 case 's': 655 if (cp->u.s->wfd != -1 && close(cp->u.s->wfd)) 656 err(1, "%s", cp->u.s->wfile); 657 cp->u.s->wfd = -1; 658 break; 659 case 'w': 660 if (cp->u.fd != -1 && close(cp->u.fd)) 661 err(1, "%s", cp->t); 662 cp->u.fd = -1; 663 break; 664 case '{': 665 cfclose(cp->u.c, cp->next); 666 break; 667 } 668 } 669