2 * Copyright (c) 1992 Diomidis Spinellis.
3 * Copyright (c) 1992, 1993
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/types.h>
50 #include "freebsd-compat.h"
53 #define LHMASK (LHSZ - 1)
54 static struct labhash
{
55 struct labhash
*lh_next
;
57 struct s_command
*lh_cmd
;
61 static char *compile_addr(char *, struct s_addr
*);
62 static char *compile_ccl(char **, char *);
63 static char *compile_delimited(char *, char *, int);
64 static char *compile_flags(char *, struct s_subst
*);
65 static regex_t
*compile_re(char *, int);
66 static char *compile_subst(char *, struct s_subst
*);
67 static char *compile_text(void);
68 static char *compile_tr(char *, struct s_tr
**);
69 static struct s_command
70 **compile_stream(struct s_command
**);
71 static char *duptoeol(char *, const char *);
72 static void enterlabel(struct s_command
*);
73 static struct s_command
75 static void fixuplabel(struct s_command
*, struct s_command
*);
76 static void uselabel(void);
79 * Command specification. This is used to drive the command parser.
82 char code
; /* Command code */
83 int naddr
; /* Number of address args */
84 enum e_args args
; /* Argument type */
87 static struct s_format cmd_fmts
[] = {
119 /* The compiled program. */
120 struct s_command
*prog
;
123 * Compile the program into prog.
124 * Initialise appends.
129 *compile_stream(&prog
) = NULL
;
130 fixuplabel(prog
, NULL
);
134 else if ((appends
= malloc(sizeof(struct s_appends
) * appendnum
)) ==
137 if ((match
= malloc((maxnsub
+ 1) * sizeof(regmatch_t
))) == NULL
)
141 #define EATSPACE() do { \
143 while (*p && isspace((unsigned char)*p)) \
147 static struct s_command
**
148 compile_stream(struct s_command
**link
)
151 static char lbuf
[_POSIX2_LINE_MAX
+ 1]; /* To save stack */
152 struct s_command
*cmd
, *cmd2
, *stack
;
154 char re
[_POSIX2_LINE_MAX
+ 1];
155 int naddr
; /* Number of addresses */
159 if ((p
= cu_fgets(lbuf
, sizeof(lbuf
), NULL
)) == NULL
) {
161 errx(1, "%lu: %s: unexpected EOF (pending }'s)",
166 semicolon
: EATSPACE();
168 if (*p
== '#' || *p
== '\0')
170 else if (*p
== ';') {
175 if ((*link
= cmd
= malloc(sizeof(struct s_command
))) == NULL
)
178 cmd
->startline
= cmd
->nonsel
= 0;
179 /* First parse the addresses */
182 /* Valid characters to start an address */
183 #define addrchar(c) (strchr("0123456789/\\$", (c)))
186 if ((cmd
->a1
= malloc(sizeof(struct s_addr
))) == NULL
)
188 p
= compile_addr(p
, cmd
->a1
);
189 EATSPACE(); /* EXTENSION */
192 EATSPACE(); /* EXTENSION */
194 if ((cmd
->a2
= malloc(sizeof(struct s_addr
)))
197 p
= compile_addr(p
, cmd
->a2
);
202 cmd
->a1
= cmd
->a2
= NULL
;
204 nonsel
: /* Now parse the command */
206 errx(1, "%lu: %s: command expected", linenum
, fname
);
208 for (fp
= cmd_fmts
; fp
->code
; fp
++)
212 errx(1, "%lu: %s: invalid command code %c", linenum
, fname
, *p
);
213 if (naddr
> fp
->naddr
)
215 "%lu: %s: command %c expects up to %d address(es), found %d",
216 linenum
, fname
, *p
, fp
->naddr
, naddr
);
234 * Short-circuit command processing, since end of
235 * group is really just a noop.
239 errx(1, "%lu: %s: unexpected }", linenum
, fname
);
244 case EMPTY
: /* d D g G h H l n N p P q x = \0 */
253 errx(1, "%lu: %s: extra characters at the end of %c command",
254 linenum
, fname
, cmd
->code
);
256 case TEXT
: /* a c i */
261 "%lu: %s: command %c expects \\ followed by text", linenum
, fname
, cmd
->code
);
266 "%lu: %s: extra characters after \\ at the end of %c command",
267 linenum
, fname
, cmd
->code
);
268 cmd
->t
= compile_text();
270 case COMMENT
: /* \0 # */
276 errx(1, "%lu: %s: filename expected", linenum
, fname
);
277 cmd
->t
= duptoeol(p
, "w command");
280 else if ((cmd
->u
.fd
= open(p
,
281 O_WRONLY
|O_APPEND
|O_CREAT
|O_TRUNC
,
289 errx(1, "%lu: %s: filename expected", linenum
, fname
);
291 cmd
->t
= duptoeol(p
, "read command");
293 case BRANCH
: /* b t */
299 cmd
->t
= duptoeol(p
, "branch");
304 cmd
->t
= duptoeol(p
, "label");
306 errx(1, "%lu: %s: empty label", linenum
, fname
);
311 if (*p
== '\0' || *p
== '\\')
313 "%lu: %s: substitute pattern can not be delimited by newline or backslash",
315 if ((cmd
->u
.s
= calloc(1, sizeof(struct s_subst
))) == NULL
)
317 p
= compile_delimited(p
, re
, 0);
320 "%lu: %s: unterminated substitute pattern", linenum
, fname
);
322 /* Compile RE with no case sensitivity temporarily */
326 cmd
->u
.s
->re
= compile_re(re
, 0);
328 p
= compile_subst(p
, cmd
->u
.s
);
329 p
= compile_flags(p
, cmd
->u
.s
);
331 /* Recompile RE with case sensitivity from "I" flag if any */
335 cmd
->u
.s
->re
= compile_re(re
, cmd
->u
.s
->icase
);
345 p
= compile_tr(p
, &cmd
->u
.y
);
354 "%lu: %s: extra text at the end of a transform command", linenum
, fname
);
361 * Get a delimited string. P points to the delimiter of the string; d points
362 * to a buffer area. Newline and delimiter escapes are processed; other
363 * escapes are ignored.
365 * Returns a pointer to the first character after the final delimiter or NULL
366 * in the case of a non-terminated string. The character array d is filled
367 * with the processed string.
370 compile_delimited(char *p
, char *d
, int is_tr
)
378 errx(1, "%lu: %s: \\ can not be used as a string delimiter",
381 errx(1, "%lu: %s: newline can not be used as a string delimiter",
384 if (*p
== '[' && *p
!= c
) {
385 if ((d
= compile_ccl(&p
, d
)) == NULL
)
386 errx(1, "%lu: %s: unbalanced brackets ([])", linenum
, fname
);
388 } else if (*p
== '\\' && p
[1] == '[') {
390 } else if (*p
== '\\' && p
[1] == c
)
392 else if (*p
== '\\' && p
[1] == 'n') {
396 } else if (*p
== '\\' && p
[1] == '\\') {
401 } else if (*p
== c
) {
411 /* compile_ccl: expand a POSIX character class */
413 compile_ccl(char **sp
, char *t
)
423 for (; *s
&& (*t
= *s
) != ']'; s
++, t
++)
424 if (*s
== '[' && ((d
= *(s
+1)) == '.' || d
== ':' || d
== '=')) {
425 *++t
= *++s
, t
++, s
++;
426 for (c
= *s
; (*t
= *s
) != ']' || c
!= d
; s
++, t
++)
427 if ((c
= *s
) == '\0')
430 return (*s
== ']') ? *sp
= ++s
, ++t
: NULL
;
434 * Compiles the regular expression in RE and returns a pointer to the compiled
435 * regular expression.
436 * Cflags are passed to regcomp.
439 compile_re(char *re
, int case_insensitive
)
446 if (case_insensitive
)
448 if ((rep
= malloc(sizeof(regex_t
))) == NULL
)
450 if ((eval
= regcomp(rep
, re
, flags
)) != 0)
451 errx(1, "%lu: %s: RE error: %s",
452 linenum
, fname
, strregerror(eval
, rep
));
453 if (maxnsub
< rep
->re_nsub
)
454 maxnsub
= rep
->re_nsub
;
459 * Compile the substitution string of a regular expression and set res to
460 * point to a saved copy of it. Nsub is the number of parenthesized regular
464 compile_subst(char *p
, struct s_subst
*s
)
466 static char lbuf
[_POSIX2_LINE_MAX
+ 1];
469 char c
, *text
, *op
, *sp
;
470 int more
= 1, sawesc
= 0;
472 c
= *p
++; /* Terminator character */
477 s
->linenum
= linenum
;
478 asize
= 2 * _POSIX2_LINE_MAX
+ 1;
479 if ((text
= malloc(asize
)) == NULL
)
483 op
= sp
= text
+ size
;
485 if (*p
== '\\' || sawesc
) {
487 * If this is a continuation from the last
488 * buffer, we won't have a character to
498 * This escaped character is continued
499 * in the next part of the line. Note
500 * this fact, then cause the loop to
501 * exit w/ normal EOL case and reenter
502 * above with the new buffer.
507 } else if (strchr("123456789", *p
) != NULL
) {
511 ref
> s
->re
->re_nsub
)
512 errx(1, "%lu: %s: \\%c not defined in the RE",
514 if (s
->maxbref
< ref
)
516 } else if (*p
== '&' || *p
== '\\')
518 } else if (*p
== c
) {
519 if (*++p
== '\0' && more
) {
520 if (cu_fgets(lbuf
, sizeof(lbuf
), &more
))
525 if ((s
->new = realloc(text
, size
)) == NULL
)
528 } else if (*p
== '\n') {
530 "%lu: %s: unescaped newline inside substitute pattern", linenum
, fname
);
536 if (asize
- size
< _POSIX2_LINE_MAX
+ 1) {
538 if ((text
= realloc(text
, asize
)) == NULL
)
541 } while (cu_fgets(p
= lbuf
, sizeof(lbuf
), &more
) != NULL
);
542 errx(1, "%lu: %s: unterminated substitute in regular expression",
548 * Compile the flags of the s command
551 compile_flags(char *p
, struct s_subst
*s
)
553 int gn
; /* True if we have seen g or n */
555 char wfile
[_POSIX2_LINE_MAX
+ 1], *q
, *eq
;
557 s
->n
= 1; /* Default */
563 EATSPACE(); /* EXTENSION */
568 "%lu: %s: more than one number or 'g' in substitute flags", linenum
, fname
);
583 case '1': case '2': case '3':
584 case '4': case '5': case '6':
585 case '7': case '8': case '9':
588 "%lu: %s: more than one number or 'g' in substitute flags", linenum
, fname
);
591 nval
= strtol(p
, &p
, 10);
592 if (errno
== ERANGE
|| nval
> INT_MAX
)
594 "%lu: %s: overflow in the 'N' substitute flag", linenum
, fname
);
600 #ifdef HISTORIC_PRACTICE
602 warnx("%lu: %s: space missing before w wfile", linenum
, fname
);
608 eq
= wfile
+ sizeof(wfile
) - 1;
613 err(1, "wfile too long");
618 errx(1, "%lu: %s: no wfile specified", linenum
, fname
);
619 s
->wfile
= strdup(wfile
);
620 if (!aflag
&& (s
->wfd
= open(wfile
,
621 O_WRONLY
|O_APPEND
|O_CREAT
|O_TRUNC
,
626 errx(1, "%lu: %s: bad flag in substitute command: '%c'",
635 * Compile a translation set of strings into a lookup table.
638 compile_tr(char *p
, struct s_tr
**py
)
643 char old
[_POSIX2_LINE_MAX
+ 1];
644 char new[_POSIX2_LINE_MAX
+ 1];
645 size_t oclen
, oldlen
, nclen
, newlen
;
646 mbstate_t mbs1
, mbs2
;
648 if ((*py
= y
= malloc(sizeof(*y
))) == NULL
)
653 if (*p
== '\0' || *p
== '\\')
655 "%lu: %s: transform pattern can not be delimited by newline or backslash",
657 p
= compile_delimited(p
, old
, 1);
659 errx(1, "%lu: %s: unterminated transform source string",
661 p
= compile_delimited(p
- 1, new, 1);
663 errx(1, "%lu: %s: unterminated transform target string",
667 oldlen
= mbsrtowcs(NULL
, &op
, 0, NULL
);
668 if (oldlen
== (size_t)-1)
671 newlen
= mbsrtowcs(NULL
, &np
, 0, NULL
);
672 if (newlen
== (size_t)-1)
674 if (newlen
!= oldlen
)
675 errx(1, "%lu: %s: transform strings are not the same length",
677 if (MB_CUR_MAX
== 1) {
679 * The single-byte encoding case is easy: generate a
682 for (i
= 0; i
<= UCHAR_MAX
; i
++)
683 y
->bytetab
[i
] = (char)i
;
684 for (; *op
; op
++, np
++)
685 y
->bytetab
[(u_char
)*op
] = *np
;
688 * Multi-byte encoding case: generate a lookup table as
689 * above, but only for single-byte characters. The first
690 * bytes of multi-byte characters have their lookup table
691 * entries set to 0, which causes do_tr() to search through
692 * an auxiliary vector of multi-byte mappings.
694 memset(&mbs1
, 0, sizeof(mbs1
));
695 memset(&mbs2
, 0, sizeof(mbs2
));
696 for (i
= 0; i
<= UCHAR_MAX
; i
++)
697 y
->bytetab
[i
] = (btowc(i
) != WEOF
) ? i
: 0;
698 while (*op
!= '\0') {
699 oclen
= mbrlen(op
, MB_LEN_MAX
, &mbs1
);
700 if (oclen
== (size_t)-1 || oclen
== (size_t)-2)
701 errc(1, EILSEQ
, NULL
);
702 nclen
= mbrlen(np
, MB_LEN_MAX
, &mbs2
);
703 if (nclen
== (size_t)-1 || nclen
== (size_t)-2)
704 errc(1, EILSEQ
, NULL
);
705 if (oclen
== 1 && nclen
== 1)
706 y
->bytetab
[(u_char
)*op
] = *np
;
708 y
->bytetab
[(u_char
)*op
] = 0;
709 y
->multis
= reallocarray(y
->multis
,
710 y
->nmultis
+ 1, sizeof(*y
->multis
));
711 if (y
->multis
== NULL
)
714 y
->multis
[i
].fromlen
= oclen
;
715 memcpy(y
->multis
[i
].from
, op
, oclen
);
716 y
->multis
[i
].tolen
= nclen
;
717 memcpy(y
->multis
[i
].to
, np
, nclen
);
727 * Compile the text following an a, c, or i command.
732 int asize
, esc_nl
, size
;
733 char *text
, *p
, *op
, *s
;
734 char lbuf
[_POSIX2_LINE_MAX
+ 1];
736 asize
= 2 * _POSIX2_LINE_MAX
+ 1;
737 if ((text
= malloc(asize
)) == NULL
)
740 while (cu_fgets(lbuf
, sizeof(lbuf
), NULL
) != NULL
) {
741 op
= s
= text
+ size
;
743 #ifdef LEGACY_BSDSED_COMPAT
746 for (esc_nl
= 0; *p
!= '\0'; p
++) {
747 if (*p
== '\\' && p
[1] != '\0' && *++p
== '\n')
756 if (asize
- size
< _POSIX2_LINE_MAX
+ 1) {
758 if ((text
= realloc(text
, asize
)) == NULL
)
763 if ((p
= realloc(text
, size
+ 1)) == NULL
)
769 * Get an address and return a pointer to the first character after
770 * it. Fill the structure pointed to according to the address.
773 compile_addr(char *p
, struct s_addr
*a
)
775 char *end
, re
[_POSIX2_LINE_MAX
+ 1];
782 case '\\': /* Context address */
785 case '/': /* Context address */
786 p
= compile_delimited(p
, re
, 0);
788 errx(1, "%lu: %s: unterminated regular expression", linenum
, fname
);
789 /* Check for case insensitive regexp flag */
797 a
->u
.r
= compile_re(re
, icase
);
801 case '$': /* Last line */
805 case '+': /* Relative line number */
806 a
->type
= AT_RELLINE
;
810 case '0': case '1': case '2': case '3': case '4':
811 case '5': case '6': case '7': case '8': case '9':
814 a
->u
.l
= strtol(p
, &end
, 10);
817 errx(1, "%lu: %s: expected context address", linenum
, fname
);
824 * Return a copy of all the characters up to \n or \0.
827 duptoeol(char *s
, const char *ctype
)
834 for (start
= s
; *s
!= '\0' && *s
!= '\n'; ++s
)
835 ws
= isspace((unsigned char)*s
);
838 warnx("%lu: %s: whitespace after %s", linenum
, fname
, ctype
);
840 if ((p
= malloc(len
)) == NULL
)
842 return (memmove(p
, start
, len
));
846 * Convert goto label names to addresses, and count a and r commands, in
847 * the given subset of the script. Free the memory used by labels in b
848 * and t commands (but not by :).
850 * TODO: Remove } nodes
853 fixuplabel(struct s_command
*cp
, struct s_command
*end
)
856 for (; cp
!= end
; cp
= cp
->next
)
864 /* Resolve branch target. */
869 if ((cp
->u
.c
= findlabel(cp
->t
)) == NULL
)
870 errx(1, "%lu: %s: undefined label '%s'", linenum
, fname
, cp
->t
);
874 /* Do interior commands. */
875 fixuplabel(cp
->u
.c
, cp
->next
);
881 * Associate the given command label for later lookup.
884 enterlabel(struct s_command
*cp
)
886 struct labhash
**lhp
, *lh
;
890 for (h
= 0, p
= (u_char
*)cp
->t
; (c
= *p
) != 0; p
++)
891 h
= (h
<< 5) + h
+ c
;
892 lhp
= &labels
[h
& LHMASK
];
893 for (lh
= *lhp
; lh
!= NULL
; lh
= lh
->lh_next
)
894 if (lh
->lh_hash
== h
&& strcmp(cp
->t
, lh
->lh_cmd
->t
) == 0)
895 errx(1, "%lu: %s: duplicate label '%s'", linenum
, fname
, cp
->t
);
896 if ((lh
= malloc(sizeof *lh
)) == NULL
)
906 * Find the label contained in the command l in the command linked
907 * list cp. L is excluded from the search. Return NULL if not found.
909 static struct s_command
*
910 findlabel(char *name
)
916 for (h
= 0, p
= (u_char
*)name
; (c
= *p
) != 0; p
++)
917 h
= (h
<< 5) + h
+ c
;
918 for (lh
= labels
[h
& LHMASK
]; lh
!= NULL
; lh
= lh
->lh_next
) {
919 if (lh
->lh_hash
== h
&& strcmp(name
, lh
->lh_cmd
->t
) == 0) {
928 * Warn about any unused labels. As a side effect, release the label hash
934 struct labhash
*lh
, *next
;
937 for (i
= 0; i
< LHSZ
; i
++) {
938 for (lh
= labels
[i
]; lh
!= NULL
; lh
= next
) {
941 warnx("%lu: %s: unused label '%s'",
942 linenum
, fname
, lh
->lh_cmd
->t
);