2 * Copyright (c) 1992 Diomidis Spinellis.
3 * Copyright (c) 1992, 1993, 1994
4 * The Regents of the University of California. All rights reserved.
6 * This code is derived from software contributed to Berkeley by
7 * Diomidis Spinellis of Imperial College, University of London.
9 * Redistribution and use in source and binary forms, with or without
10 * modification, are permitted provided that the following conditions
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 * 4. Neither the name of the University nor the names of its contributors
18 * may be used to endorse or promote products derived from this software
19 * without specific prior written permission.
21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 #include <sys/cdefs.h>
35 __FBSDID("$FreeBSD$");
38 static const char sccsid[] = "@(#)process.c 8.6 (Berkeley) 4/20/94";
41 #include <sys/types.h>
43 #include <sys/ioctl.h>
62 static SPACE HS, PS, SS, YS;
69 static __inline int applies(struct s_command *);
70 static void do_tr(struct s_tr *);
71 static void flush_appends(void);
72 static void lputs(char *, size_t);
73 static __inline int regexec_e(regex_t *, const char *, int, int, size_t);
74 static void regsub(SPACE *, char *, char *);
75 static int substitute(struct s_command *);
77 struct s_appends *appends; /* Array of pointers to strings to append. */
78 static int appendx; /* Index into appends array. */
79 int appendnum; /* Size of appends array. */
81 static int lastaddr; /* Set by applies if last address of a range. */
82 static int sdone; /* If any substitutes since last line input. */
83 /* Iov structure for 'w' commands. */
84 static regex_t *defpreg;
88 #define OUT(s) { fwrite(s, sizeof(u_char), psl, outfile); fputc('\n', outfile); }
100 for (linenum = 0; mf_fgets(&PS, REPLACE);) {
115 if (appendx >= appendnum)
116 if ((appends = realloc(appends,
117 sizeof(struct s_appends) *
118 (appendnum *= 2))) == NULL)
120 appends[appendx].type = AP_STRING;
121 appends[appendx].s = cp->t;
122 appends[appendx].len = strlen(cp->t);
131 if (cp->a2 == NULL || lastaddr)
132 (void)fprintf(outfile, "%s", cp->t);
141 (p = memchr(ps, '\n', psl)) == NULL) {
146 memmove(ps, p + 1, psl);
150 cspace(&PS, hs, hsl, REPLACE);
153 cspace(&PS, "\n", 1, 0);
154 cspace(&PS, hs, hsl, 0);
157 cspace(&HS, ps, psl, REPLACE);
160 cspace(&HS, "\n", 1, 0);
161 cspace(&HS, ps, psl, 0);
164 (void)fprintf(outfile, "%s", cp->t);
173 if (!mf_fgets(&PS, REPLACE))
179 cspace(&PS, "\n", 1, 0);
180 if (!mf_fgets(&PS, 0))
191 if ((p = memchr(ps, '\n', psl)) != NULL) {
205 if (appendx >= appendnum)
206 if ((appends = realloc(appends,
207 sizeof(struct s_appends) *
208 (appendnum *= 2))) == NULL)
210 appends[appendx].type = AP_FILE;
211 appends[appendx].s = cp->t;
212 appends[appendx].len = strlen(cp->t);
216 sdone |= substitute(cp);
228 if (cp->u.fd == -1 && (cp->u.fd = open(cp->t,
229 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
232 if (write(cp->u.fd, ps, psl) != psl ||
233 write(cp->u.fd, "\n", 1) != 1)
238 cspace(&HS, "", 0, REPLACE);
252 (void)fprintf(outfile, "%lu\n", linenum);
257 new: if (!nflag && !pd)
260 } /* for all lines */
264 * TRUE if the address passed matches the current program state
265 * (lastline, linenumber, ps).
268 (a)->type == AT_RE ? regexec_e((a)->u.r, ps, 0, 1, psl) : \
269 (a)->type == AT_LINE ? linenum == (a)->u.l : lastline()
272 * Return TRUE if the command applies to the current line. Sets the inrange
273 * flag to process ranges. Interprets the non-select (``!'') flag.
276 applies(struct s_command *cp)
281 if (cp->a1 == NULL && cp->a2 == NULL)
290 } else if (MATCH(cp->a1)) {
292 * If the second address is a number less than or
293 * equal to the line number first selected, only
294 * one line shall be selected.
297 if (cp->a2->type == AT_LINE &&
298 linenum >= cp->a2->u.l)
307 return (cp->nonsel ? ! r : r);
312 * Do substitutions in the pattern space. Currently, we build a
313 * copy of the new pattern space in the substitute space structure
314 * and then swap them.
317 substitute(struct s_command *cp)
321 regoff_t re_off, slen;
328 if (defpreg != NULL && cp->u.s->maxbref > defpreg->re_nsub) {
329 linenum = cp->u.s->linenum;
330 errx(1, "%lu: %s: \\%d not defined in the RE",
331 linenum, fname, cp->u.s->maxbref);
334 if (!regexec_e(re, s, 0, 0, psl))
337 SS.len = 0; /* Clean substitute space. */
345 if (lastempty || match[0].rm_so != match[0].rm_eo) {
346 /* Locate start of replaced string. */
347 re_off = match[0].rm_so;
348 /* Copy leading retained string. */
349 cspace(&SS, s, re_off, APPEND);
350 /* Add in regular expression. */
351 regsub(&SS, s, cp->u.s->new);
354 /* Move past this match. */
355 if (match[0].rm_so != match[0].rm_eo) {
357 slen -= match[0].rm_eo;
360 if (match[0].rm_so < slen)
361 cspace(&SS, s + match[0].rm_so, 1,
363 s += match[0].rm_so + 1;
364 slen -= match[0].rm_so + 1;
367 } while (slen >= 0 && regexec_e(re, s, REG_NOTBOL, 0, slen));
368 /* Copy trailing retained string. */
370 cspace(&SS, s, slen, APPEND);
372 default: /* Nth occurrence */
374 if (match[0].rm_eo == match[0].rm_so)
375 match[0].rm_eo = match[0].rm_so + 1;
377 slen -= match[0].rm_eo;
380 if (!regexec_e(re, s, REG_NOTBOL, 0, slen))
384 case 1: /* 1st occurrence */
385 /* Locate start of replaced string. */
386 re_off = match[0].rm_so + (s - ps);
387 /* Copy leading retained string. */
388 cspace(&SS, ps, re_off, APPEND);
389 /* Add in regular expression. */
390 regsub(&SS, s, cp->u.s->new);
391 /* Copy trailing retained string. */
393 slen -= match[0].rm_eo;
394 cspace(&SS, s, slen, APPEND);
399 * Swap the substitute space and the pattern space, and make sure
400 * that any leftover pointers into stdio memory get lost.
407 /* Handle the 'p' flag. */
411 /* Handle the 'w' flag. */
412 if (cp->u.s->wfile && !pd) {
413 if (cp->u.s->wfd == -1 && (cp->u.s->wfd = open(cp->u.s->wfile,
414 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, DEFFILEMODE)) == -1)
415 err(1, "%s", cp->u.s->wfile);
416 if (write(cp->u.s->wfd, ps, psl) != psl ||
417 write(cp->u.s->wfd, "\n", 1) != 1)
418 err(1, "%s", cp->u.s->wfile);
425 * Perform translation ('y' command) in the pattern space.
428 do_tr(struct s_tr *y)
435 if (MB_CUR_MAX == 1) {
437 * Single-byte encoding: perform in-place translation
438 * of the pattern space.
440 for (p = ps; p < &ps[psl]; p++)
441 *p = y->bytetab[(u_char)*p];
444 * Multi-byte encoding: perform translation into the
445 * translation space, then swap the translation and
448 /* Clean translation space. */
450 for (p = ps, left = psl; left > 0; p += clen, left -= clen) {
451 if ((c = y->bytetab[(u_char)*p]) != '\0') {
452 cspace(&YS, &c, 1, APPEND);
456 for (i = 0; i < y->nmultis; i++)
457 if (left >= y->multis[i].fromlen &&
458 memcmp(p, y->multis[i].from,
459 y->multis[i].fromlen) == 0)
461 if (i < y->nmultis) {
462 cspace(&YS, y->multis[i].to,
463 y->multis[i].tolen, APPEND);
464 clen = y->multis[i].fromlen;
466 cspace(&YS, p, 1, APPEND);
470 /* Swap the translation space and the pattern space. */
479 * Flush append requests. Always called before reading a line,
480 * therefore it also resets the substitution done (sdone) flag.
489 for (i = 0; i < appendx; i++)
490 switch (appends[i].type) {
492 fwrite(appends[i].s, sizeof(char), appends[i].len,
497 * Read files probably shouldn't be cached. Since
498 * it's not an error to read a non-existent file,
499 * it's possible that another program is interacting
500 * with the sed script through the filesystem. It
501 * would be truly bizarre, but possible. It's probably
502 * not that big a performance win, anyhow.
504 if ((f = fopen(appends[i].s, "r")) == NULL)
506 while ((count = fread(buf, sizeof(char), sizeof(buf), f)))
507 (void)fwrite(buf, sizeof(char), count, outfile);
512 errx(1, "%s: %s", outfname, strerror(errno ? errno : EIO));
517 lputs(char *s, size_t len)
519 static const char escapes[] = "\\\a\b\f\r\t\v";
523 static int termwidth = -1;
528 if (outfile != stdout)
530 if (termwidth == -1) {
531 if ((p = getenv("COLUMNS")) && *p != '\0')
533 else if (ioctl(STDOUT_FILENO, TIOCGWINSZ, &win) == 0 &&
535 termwidth = win.ws_col;
540 memset(&mbs, 0, sizeof(mbs));
543 clen = mbrtowc(&wc, s, len, &mbs);
546 if (clen == (size_t)-1 || clen == (size_t)-2) {
547 wc = (unsigned char)*s;
549 memset(&mbs, 0, sizeof(mbs));
552 if (col + 1 >= termwidth)
553 fprintf(outfile, "\\\n");
555 fputc('\n', outfile);
557 } else if (iswprint(wc)) {
559 if (col + width >= termwidth) {
560 fprintf(outfile, "\\\n");
563 fwrite(s, 1, clen, outfile);
565 } else if (wc != L'\0' && (c = wctob(wc)) != EOF &&
566 (p = strchr(escapes, c)) != NULL) {
567 if (col + 2 >= termwidth) {
568 fprintf(outfile, "\\\n");
571 fprintf(outfile, "\\%c", "\\abfrtv"[p - escapes]);
574 if (col + 4 * clen >= termwidth) {
575 fprintf(outfile, "\\\n");
578 for (i = 0; i < clen; i++)
579 fprintf(outfile, "\\%03o",
580 (int)(unsigned char)s[i]);
586 if (col + 1 >= termwidth)
587 fprintf(outfile, "\\\n");
588 (void)fputc('$', outfile);
589 (void)fputc('\n', outfile);
591 errx(1, "%s: %s", outfname, strerror(errno ? errno : EIO));
595 regexec_e(regex_t *preg, const char *string, int eflags, int nomatch,
602 errx(1, "first RE may not be empty");
608 match[0].rm_eo = slen;
610 eval = regexec(defpreg, string,
611 nomatch ? 0 : maxnsub + 1, match, eflags | REG_STARTEND);
618 errx(1, "RE error: %s", strregerror(eval, defpreg));
623 * regsub - perform substitutions after a regexp match
624 * Based on a routine by Henry Spencer
627 regsub(SPACE *sp, char *string, char *src)
632 #define NEEDSP(reqlen) \
633 /* XXX What is the +1 for? */ \
634 if (sp->len + (reqlen) + 1 >= sp->blen) { \
635 sp->blen += (reqlen) + 1024; \
636 if ((sp->space = sp->back = realloc(sp->back, sp->blen)) \
639 dst = sp->space + sp->len; \
642 dst = sp->space + sp->len;
643 while ((c = *src++) != '\0') {
646 else if (c == '\\' && isdigit((unsigned char)*src))
650 if (no < 0) { /* Ordinary character. */
651 if (c == '\\' && (*src == '\\' || *src == '&'))
656 } else if (match[no].rm_so != -1 && match[no].rm_eo != -1) {
657 len = match[no].rm_eo - match[no].rm_so;
659 memmove(dst, string + match[no].rm_so, len);
670 * Append the source space to the destination space, allocating new
671 * space as necessary.
674 cspace(SPACE *sp, const char *p, size_t len, enum e_spflag spflag)
678 /* Make sure SPACE has enough memory and ramp up quickly. */
679 tlen = sp->len + len + 1;
680 if (tlen > sp->blen) {
681 sp->blen = tlen + 1024;
682 if ((sp->space = sp->back = realloc(sp->back, sp->blen)) ==
687 if (spflag == REPLACE)
690 memmove(sp->space + sp->len, p, len);
692 sp->space[sp->len += len] = '\0';
696 * Close all cached opened files and report any errors
699 cfclose(struct s_command *cp, struct s_command *end)
702 for (; cp != end; cp = cp->next)
705 if (cp->u.s->wfd != -1 && close(cp->u.s->wfd))
706 err(1, "%s", cp->u.s->wfile);
710 if (cp->u.fd != -1 && close(cp->u.fd))
715 cfclose(cp->u.c, cp->next);