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
33 * @(#)process.c 8.6 (Berkeley) 4/20/94
34 * $FreeBSD: src/usr.bin/sed/process.c,v 1.49 2008/02/09 09:12:02 dwmalone Exp $
35 * $DragonFly: src/usr.bin/sed/process.c,v 1.6 2008/04/08 13:23:38 swildner Exp $
38 #include <sys/types.h>
40 #include <sys/ioctl.h>
59 static SPACE HS
, PS
, SS
, YS
;
66 static __inline
int applies(struct s_command
*);
67 static void do_tr(struct s_tr
*);
68 static void flush_appends(void);
69 static void lputs(char *, size_t);
70 static __inline
int regexec_e(regex_t
*, const char *, int, int, size_t);
71 static void regsub(SPACE
*, char *, char *);
72 static int substitute(struct s_command
*);
74 struct s_appends
*appends
; /* Array of pointers to strings to append. */
75 static int appendx
; /* Index into appends array. */
76 int appendnum
; /* Size of appends array. */
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
;
85 #define OUT() do {fwrite(ps, 1, psl, outfile); fputc('\n', outfile);} while (0)
97 for (linenum
= 0; mf_fgets(&PS
, REPLACE
);) {
112 if (appendx
>= appendnum
)
113 if ((appends
= realloc(appends
,
114 sizeof(struct s_appends
) *
115 (appendnum
*= 2))) == NULL
)
117 appends
[appendx
].type
= AP_STRING
;
118 appends
[appendx
].s
= cp
->t
;
119 appends
[appendx
].len
= strlen(cp
->t
);
128 if (cp
->a2
== NULL
|| lastaddr
|| lastline())
129 (void)fprintf(outfile
, "%s", cp
->t
);
138 (p
= memchr(ps
, '\n', psl
)) == NULL
) {
143 memmove(ps
, p
+ 1, psl
);
147 cspace(&PS
, hs
, hsl
, REPLACE
);
150 cspace(&PS
, "\n", 1, APPEND
);
151 cspace(&PS
, hs
, hsl
, APPEND
);
154 cspace(&HS
, ps
, psl
, REPLACE
);
157 cspace(&HS
, "\n", 1, APPEND
);
158 cspace(&HS
, ps
, psl
, APPEND
);
161 (void)fprintf(outfile
, "%s", cp
->t
);
170 if (!mf_fgets(&PS
, REPLACE
))
176 cspace(&PS
, "\n", 1, APPEND
);
177 if (!mf_fgets(&PS
, APPEND
))
188 if ((p
= memchr(ps
, '\n', psl
)) != NULL
) {
202 if (appendx
>= appendnum
)
203 if ((appends
= realloc(appends
,
204 sizeof(struct s_appends
) *
205 (appendnum
*= 2))) == NULL
)
207 appends
[appendx
].type
= AP_FILE
;
208 appends
[appendx
].s
= cp
->t
;
209 appends
[appendx
].len
= strlen(cp
->t
);
213 sdone
|= substitute(cp
);
225 if (cp
->u
.fd
== -1 && (cp
->u
.fd
= open(cp
->t
,
226 O_WRONLY
|O_APPEND
|O_CREAT
|O_TRUNC
,
229 if (write(cp
->u
.fd
, ps
, psl
) != (ssize_t
)psl
||
230 write(cp
->u
.fd
, "\n", 1) != 1)
235 * If the hold space is null, make it empty
236 * but not null. Otherwise the pattern space
237 * will become null after the swap, which is
238 * an abnormal condition.
241 cspace(&HS
, "", 0, REPLACE
);
255 (void)fprintf(outfile
, "%lu\n", linenum
);
260 new: if (!nflag
&& !pd
)
263 } /* for all lines */
267 * TRUE if the address passed matches the current program state
268 * (lastline, linenumber, ps).
271 ((a)->type == AT_RE ? regexec_e((a)->u.r, ps, 0, 1, psl) : \
272 (a)->type == AT_LINE ? linenum == (a)->u.l : lastline())
275 * Return TRUE if the command applies to the current line. Sets the inrange
276 * flag to process ranges. Interprets the non-select (``!'') flag.
279 applies(struct s_command
*cp
)
284 if (cp
->a1
== NULL
&& cp
->a2
== NULL
)
292 } else if (cp
->a2
->type
== AT_LINE
&&
293 linenum
> cp
->a2
->u
.l
) {
295 * We missed the 2nd address due to a branch,
296 * so just close the range and return false.
302 } else if (MATCH(cp
->a1
)) {
304 * If the second address is a number less than or
305 * equal to the line number first selected, only
306 * one line shall be selected.
309 if (cp
->a2
->type
== AT_LINE
&&
310 linenum
>= cp
->a2
->u
.l
)
319 return (cp
->nonsel
? ! r
: r
);
323 * Reset the sed processor to its initial state.
328 struct s_command
*cp
;
331 * Reset all inrange markers.
333 for (cp
= prog
; cp
; cp
= cp
->code
== '{' ? cp
->u
.c
: cp
->next
)
338 * Clear out the hold space.
340 cspace(&HS
, "", 0, REPLACE
);
345 * Do substitutions in the pattern space. Currently, we build a
346 * copy of the new pattern space in the substitute space structure
347 * and then swap them.
350 substitute(struct s_command
*cp
)
354 regoff_t re_off
, slen
;
361 if (defpreg
!= NULL
&& cp
->u
.s
->maxbref
> defpreg
->re_nsub
) {
362 linenum
= cp
->u
.s
->linenum
;
363 errx(1, "%lu: %s: \\%u not defined in the RE",
364 linenum
, fname
, cp
->u
.s
->maxbref
);
367 if (!regexec_e(re
, s
, 0, 0, psl
))
370 SS
.len
= 0; /* Clean substitute space. */
378 if (lastempty
|| match
[0].rm_so
!= match
[0].rm_eo
) {
379 /* Locate start of replaced string. */
380 re_off
= match
[0].rm_so
;
381 /* Copy leading retained string. */
382 cspace(&SS
, s
, re_off
, APPEND
);
383 /* Add in regular expression. */
384 regsub(&SS
, s
, cp
->u
.s
->new);
387 /* Move past this match. */
388 if (match
[0].rm_so
!= match
[0].rm_eo
) {
390 slen
-= match
[0].rm_eo
;
393 if (match
[0].rm_so
< slen
)
394 cspace(&SS
, s
+ match
[0].rm_so
, 1,
396 s
+= match
[0].rm_so
+ 1;
397 slen
-= match
[0].rm_so
+ 1;
400 } while (slen
>= 0 && regexec_e(re
, s
, REG_NOTBOL
, 0, slen
));
401 /* Copy trailing retained string. */
403 cspace(&SS
, s
, slen
, APPEND
);
405 default: /* Nth occurrence */
407 if (match
[0].rm_eo
== match
[0].rm_so
)
408 match
[0].rm_eo
= match
[0].rm_so
+ 1;
410 slen
-= match
[0].rm_eo
;
413 if (!regexec_e(re
, s
, REG_NOTBOL
, 0, slen
))
417 case 1: /* 1st occurrence */
418 /* Locate start of replaced string. */
419 re_off
= match
[0].rm_so
+ (s
- ps
);
420 /* Copy leading retained string. */
421 cspace(&SS
, ps
, re_off
, APPEND
);
422 /* Add in regular expression. */
423 regsub(&SS
, s
, cp
->u
.s
->new);
424 /* Copy trailing retained string. */
426 slen
-= match
[0].rm_eo
;
427 cspace(&SS
, s
, slen
, APPEND
);
432 * Swap the substitute space and the pattern space, and make sure
433 * that any leftover pointers into stdio memory get lost.
440 /* Handle the 'p' flag. */
444 /* Handle the 'w' flag. */
445 if (cp
->u
.s
->wfile
&& !pd
) {
446 if (cp
->u
.s
->wfd
== -1 && (cp
->u
.s
->wfd
= open(cp
->u
.s
->wfile
,
447 O_WRONLY
|O_APPEND
|O_CREAT
|O_TRUNC
, DEFFILEMODE
)) == -1)
448 err(1, "%s", cp
->u
.s
->wfile
);
449 if (write(cp
->u
.s
->wfd
, ps
, psl
) != (ssize_t
)psl
||
450 write(cp
->u
.s
->wfd
, "\n", 1) != 1)
451 err(1, "%s", cp
->u
.s
->wfile
);
458 * Perform translation ('y' command) in the pattern space.
461 do_tr(struct s_tr
*y
)
468 if (MB_CUR_MAX
== 1) {
470 * Single-byte encoding: perform in-place translation
471 * of the pattern space.
473 for (p
= ps
; p
< &ps
[psl
]; p
++)
474 *p
= y
->bytetab
[(u_char
)*p
];
477 * Multi-byte encoding: perform translation into the
478 * translation space, then swap the translation and
481 /* Clean translation space. */
483 for (p
= ps
, left
= psl
; left
> 0; p
+= clen
, left
-= clen
) {
484 if ((c
= y
->bytetab
[(u_char
)*p
]) != '\0') {
485 cspace(&YS
, &c
, 1, APPEND
);
489 for (i
= 0; i
< y
->nmultis
; i
++)
490 if (left
>= y
->multis
[i
].fromlen
&&
491 memcmp(p
, y
->multis
[i
].from
,
492 y
->multis
[i
].fromlen
) == 0)
494 if (i
< y
->nmultis
) {
495 cspace(&YS
, y
->multis
[i
].to
,
496 y
->multis
[i
].tolen
, APPEND
);
497 clen
= y
->multis
[i
].fromlen
;
499 cspace(&YS
, p
, 1, APPEND
);
503 /* Swap the translation space and the pattern space. */
512 * Flush append requests. Always called before reading a line,
513 * therefore it also resets the substitution done (sdone) flag.
522 for (i
= 0; i
< appendx
; i
++)
523 switch (appends
[i
].type
) {
525 fwrite(appends
[i
].s
, sizeof(char), appends
[i
].len
,
530 * Read files probably shouldn't be cached. Since
531 * it's not an error to read a non-existent file,
532 * it's possible that another program is interacting
533 * with the sed script through the filesystem. It
534 * would be truly bizarre, but possible. It's probably
535 * not that big a performance win, anyhow.
537 if ((f
= fopen(appends
[i
].s
, "r")) == NULL
)
539 while ((count
= fread(buf
, sizeof(char), sizeof(buf
), f
)))
540 (void)fwrite(buf
, sizeof(char), count
, outfile
);
545 errx(1, "%s: %s", outfname
, strerror(errno
? errno
: EIO
));
550 lputs(char *s
, size_t len
)
552 static const char escapes
[] = "\\\a\b\f\r\t\v";
556 static int termwidth
= -1;
561 if (outfile
!= stdout
)
563 if (termwidth
== -1) {
564 if ((p
= getenv("COLUMNS")) && *p
!= '\0')
566 else if (ioctl(STDOUT_FILENO
, TIOCGWINSZ
, &win
) == 0 &&
568 termwidth
= win
.ws_col
;
575 memset(&mbs
, 0, sizeof(mbs
));
578 clen
= mbrtowc(&wc
, s
, len
, &mbs
);
581 if (clen
== (size_t)-1 || clen
== (size_t)-2) {
582 wc
= (unsigned char)*s
;
584 memset(&mbs
, 0, sizeof(mbs
));
587 if (col
+ 1 >= termwidth
)
588 fprintf(outfile
, "\\\n");
590 fputc('\n', outfile
);
592 } else if (iswprint(wc
)) {
594 if (col
+ width
>= termwidth
) {
595 fprintf(outfile
, "\\\n");
598 fwrite(s
, 1, clen
, outfile
);
600 } else if (wc
!= L
'\0' && (c
= wctob(wc
)) != EOF
&&
601 (p
= strchr(escapes
, c
)) != NULL
) {
602 if (col
+ 2 >= termwidth
) {
603 fprintf(outfile
, "\\\n");
606 fprintf(outfile
, "\\%c", "\\abfrtv"[p
- escapes
]);
609 if (col
+ 4 * clen
>= (unsigned)termwidth
) {
610 fprintf(outfile
, "\\\n");
613 for (i
= 0; i
< clen
; i
++)
614 fprintf(outfile
, "\\%03o",
615 (int)(unsigned char)s
[i
]);
621 if (col
+ 1 >= termwidth
)
622 fprintf(outfile
, "\\\n");
623 (void)fputc('$', outfile
);
624 (void)fputc('\n', outfile
);
626 errx(1, "%s: %s", outfname
, strerror(errno
? errno
: EIO
));
630 regexec_e(regex_t
*preg
, const char *string
, int eflags
, int nomatch
,
637 errx(1, "first RE may not be empty");
643 match
[0].rm_eo
= slen
;
645 eval
= regexec(defpreg
, string
,
646 nomatch
? 0 : maxnsub
+ 1, match
, eflags
| REG_STARTEND
);
653 errx(1, "RE error: %s", strregerror(eval
, defpreg
));
658 * regsub - perform substitutions after a regexp match
659 * Based on a routine by Henry Spencer
662 regsub(SPACE
*sp
, char *string
, char *src
)
667 #define NEEDSP(reqlen) \
668 /* XXX What is the +1 for? */ \
669 if (sp->len + (reqlen) + 1 >= sp->blen) { \
670 sp->blen += (reqlen) + 1024; \
671 if ((sp->space = sp->back = realloc(sp->back, sp->blen)) \
674 dst = sp->space + sp->len; \
677 dst
= sp
->space
+ sp
->len
;
678 while ((c
= *src
++) != '\0') {
681 else if (c
== '\\' && isdigit((unsigned char)*src
))
685 if (no
< 0) { /* Ordinary character. */
686 if (c
== '\\' && (*src
== '\\' || *src
== '&'))
691 } else if (match
[no
].rm_so
!= -1 && match
[no
].rm_eo
!= -1) {
692 len
= match
[no
].rm_eo
- match
[no
].rm_so
;
694 memmove(dst
, string
+ match
[no
].rm_so
, len
);
705 * Concatenate space: append the source space to the destination space,
706 * allocating new space as necessary.
709 cspace(SPACE
*sp
, const char *p
, size_t len
, enum e_spflag spflag
)
713 /* Make sure SPACE has enough memory and ramp up quickly. */
714 tlen
= sp
->len
+ len
+ 1;
715 if (tlen
> sp
->blen
) {
716 sp
->blen
= tlen
+ 1024;
717 if ((sp
->space
= sp
->back
= realloc(sp
->back
, sp
->blen
)) ==
722 if (spflag
== REPLACE
)
725 memmove(sp
->space
+ sp
->len
, p
, len
);
727 sp
->space
[sp
->len
+= len
] = '\0';
731 * Close all cached opened files and report any errors
734 cfclose(struct s_command
*cp
, struct s_command
*end
)
737 for (; cp
!= end
; cp
= cp
->next
)
740 if (cp
->u
.s
->wfd
!= -1 && close(cp
->u
.s
->wfd
))
741 err(1, "%s", cp
->u
.s
->wfile
);
745 if (cp
->u
.fd
!= -1 && close(cp
->u
.fd
))
750 cfclose(cp
->u
.c
, cp
->next
);