1 /* $Header: /p/tcsh/cvsroot/tcsh/sh.dol.c,v 3.70 2006/09/14 18:30:16 christos Exp $ */
3 * sh.dol.c: Variable substitutions
6 * Copyright (c) 1980, 1991 The Regents of the University of California.
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 * 3. 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
35 RCSID("$tcsh: sh.dol.c,v 3.70 2006/09/14 18:30:16 christos Exp $")
42 * These routines perform variable substitution and quoting via ' and ".
43 * To this point these constructs have been preserved in the divided
44 * input words. Here we expand variables and turn quoting via ' and " into
45 * QUOTE bits on characters (which prevent further interpretation).
46 * If the `:q' modifier was applied during history expansion, then
47 * some QUOTEing may have occurred already, so we dont "trim()" here.
50 static Char Dpeekc
; /* Peek for DgetC */
51 static eChar Dpeekrd
; /* Peek for Dreadc */
52 static Char
*Dcp
, *const *Dvp
; /* Input vector for Dreadc */
56 #define unDgetC(c) Dpeekc = c
58 #define QUOTES (_QF|_QB|_ESC) /* \ ' " ` */
61 * The following variables give the information about the current
62 * $ expansion, recording the current word position, the remaining
63 * words within this expansion, the count of remaining words, and the
64 * information about any : modifier which is being applied.
66 static Char
*dolp
; /* Remaining chars from this word */
67 static Char
**dolnxt
; /* Further words */
68 static int dolcnt
; /* Count of further words */
69 static struct Strbuf dolmod
; /* = Strbuf_INIT; : modifier characters */
70 static int dolmcnt
; /* :gx -> INT_MAX, else 1 */
71 static int dol_flag_a
; /* :ax -> 1, else 0 */
73 static Char
**Dfix2 (Char
*const *);
74 static int Dpack (struct Strbuf
*);
75 static int Dword (struct blk_buf
*);
76 static void dolerror (Char
*);
77 static eChar
DgetC (int);
78 static void Dgetdol (void);
79 static void fixDolMod (void);
80 static void setDolp (Char
*);
81 static void unDredc (eChar
);
82 static eChar
Dredc (void);
83 static void Dtestq (Char
);
86 * Fix up the $ expansions and quotations in the
87 * argument list to command t.
90 Dfix(struct command
*t
)
97 /* Note that t_dcom isn't trimmed thus !...:q's aren't lost */
98 for (pp
= t
->t_dcom
; (p
= *pp
++) != NULL
;) {
100 if (cmap(*p
, _DOL
| QUOTES
)) { /* $, \, ', ", ` */
103 expanded
= Dfix2(t
->t_dcom
); /* found one */
105 t
->t_dcom
= expanded
;
113 * $ substitute one word, for i/o redirection
118 Char
*Dv
[2], **expanded
;
124 expanded
= Dfix2(Dv
);
125 if (expanded
[0] == NULL
|| expanded
[1] != NULL
) {
127 setname(short2str(cp
));
128 stderror(ERR_NAME
| ERR_AMBIG
);
130 cp
= Strsave(expanded
[0]);
136 * Subroutine to do actual fixing after state initialization.
139 Dfix2(Char
*const *v
)
141 struct blk_buf bb
= BLK_BUF_INIT
;
144 Dcp
= STRNULL
; /* Setup input vector for Dreadc */
146 unDredc(0); /* Clear out any old peeks (at error) */
148 dolcnt
= 0; /* Clear out residual $ expands (...) */
149 cleanup_push(&bb
, bb_cleanup
);
154 return bb_finish(&bb
);
158 * Pack up more characters in this word
161 Dpack(struct Strbuf
*wbuf
)
182 if (cmap(c
, _SP
| _NL
| _QF
| _QB
)) { /* sp \t\n'"` */
188 Strbuf_append1(wbuf
, (Char
) c
);
193 * Get a word. This routine is analogous to the routine
194 * word() in sh.lex.c for the main lexical input. One difference
195 * here is that we don't get a newline to terminate our expansion.
196 * Rather, DgetC will return a DEOF when we hit the end-of-input.
199 Dword(struct blk_buf
*bb
)
202 struct Strbuf wbuf
= Strbuf_INIT
;
206 cleanup_push(&wbuf
, Strbuf_cleanup
);
213 cleanup_until(&wbuf
);
216 /* finish this word and catch the code above the next time */
228 /* We preserve ` quotations which are done yet later */
229 Strbuf_append1(&wbuf
, (Char
) c
);
234 * Note that DgetC never returns a QUOTES character from an
235 * expansion, so only true input quotes will get us here or out.
238 dolflg
= c1
== '"' ? DODOL
: 0;
243 if (c
== '\n' || c
== DEOF
)
244 stderror(ERR_UNMATCHED
, (int)c1
);
245 if ((c
& (QUOTE
| TRIM
)) == ('\n' | QUOTE
)) {
246 if (wbuf
.len
!= 0 && (wbuf
.s
[wbuf
.len
- 1] & TRIM
) == '\\')
253 * Leave any `s alone for later. Other chars are all
254 * quoted, thus `...` can tell it was within "...".
256 Strbuf_append1(&wbuf
, c
== '`' ? '`' : c
| QUOTE
);
260 /* Prevent all further interpretation */
261 Strbuf_append1(&wbuf
, c
| QUOTE
);
265 /* Leave all text alone for later */
266 Strbuf_append1(&wbuf
, (Char
) c
);
274 Strbuf_append1(&wbuf
, '`');
276 if (Dpack(&wbuf
) != 0)
281 c
= DgetC(0); /* No $ subst! */
282 if (c
== '\n' || c
== DEOF
)
292 if (Dpack(&wbuf
) != 0)
297 cleanup_ignore(&wbuf
);
298 cleanup_until(&wbuf
);
299 bb_append(bb
, Strbuf_finish(&wbuf
));
305 * Get a character, performing $ substitution unless flag is 0.
306 * Any QUOTES character which is returned from a $ expansion is
307 * QUOTEd so that it will not be recognized above.
315 if ((c
= Dpeekc
) != 0) {
319 if (lap
< labuf
.len
) {
320 c
= labuf
.s
[lap
++] & (QUOTE
| TRIM
);
327 if ((c
= *dolp
++ & (QUOTE
| TRIM
)) != 0)
342 if (c
== '$' && flag
) {
349 static Char
*nulvec
[] = { NULL
};
350 static struct varent nulargv
= {nulvec
, STRargv
, VAR_READWRITE
,
351 { NULL
, NULL
, NULL
}, 0 };
356 setname(short2str(s
));
357 stderror(ERR_NAME
| ERR_RANGE
);
361 * Handle the multitudinous $ expansion forms.
368 struct varent
*vp
= NULL
;
369 struct Strbuf name
= Strbuf_INIT
;
371 int subscr
= 0, lwb
= 1, upb
= 0;
372 int dimen
= 0, bitset
= 0, length
= 0;
373 static Char
*dolbang
= NULL
;
375 cleanup_push(&name
, Strbuf_cleanup
);
376 dolmod
.len
= dolmcnt
= dol_flag_a
= 0;
379 stderror(ERR_SYNTAX
);
383 c
= DgetC(0); /* sc is { to take } later */
384 if ((c
& TRIM
) == '#')
385 dimen
++, c
= DgetC(0); /* $# takes dimension */
387 bitset
++, c
= DgetC(0); /* $? tests existence */
389 length
++, c
= DgetC(0); /* $% returns length in chars */
393 if (dimen
|| bitset
|| length
)
394 stderror(ERR_SYNTAX
);
397 setDolp(dolbang
= putn(backpid
));
399 cleanup_until(&name
);
403 if (dimen
|| bitset
|| length
)
404 stderror(ERR_SYNTAX
);
406 cleanup_until(&name
);
410 static struct Strbuf wbuf
; /* = Strbuf_INIT; */
413 stderror(ERR_NOTALLOWED
, "$?<");
415 stderror(ERR_NOTALLOWED
, "$#<");
417 stderror(ERR_NOTALLOWED
, "$%<");
420 char cbuf
[MB_LEN_MAX
];
422 int old_pintr_disabled
;
429 pintr_push_enable(&old_pintr_disabled
);
430 res
= force_read(OLDSTD
, cbuf
+ cbp
, 1);
431 cleanup_until(&old_pintr_disabled
);
435 len
= normal_mbtowc(&wc
, cbuf
, cbp
);
438 if (cbp
< MB_LEN_MAX
)
439 continue; /* Maybe a partial character */
440 wc
= (unsigned char)*cbuf
| INVALID_BYTE
;
444 if (cbp
!= (size_t)len
)
445 memmove(cbuf
, cbuf
+ len
, cbp
- len
);
449 Strbuf_append1(&wbuf
, wc
);
455 len
= normal_mbtowc(&wc
, cbuf
, cbp
);
458 wc
= (unsigned char)*cbuf
| INVALID_BYTE
;
462 if (cbp
!= (size_t)len
)
463 memmove(cbuf
, cbuf
+ len
, cbp
- len
);
467 Strbuf_append1(&wbuf
, wc
);
469 Strbuf_terminate(&wbuf
);
473 setDolp(wbuf
.s
); /* Kept allocated until next $< expansion */
474 cleanup_until(&name
);
479 Strbuf_append(&name
, STRargv
);
480 Strbuf_terminate(&name
);
482 subscr
= -1; /* Prevent eating [...] */
487 np
= dimen
? STRargv
: (bitset
? STRstatus
: NULL
);
490 Strbuf_append(&name
, np
);
491 Strbuf_terminate(&name
);
493 subscr
= -1; /* Prevent eating [...] */
498 stderror(ERR_SYNTAX
);
504 stderror(ERR_NOTALLOWED
, "$#<num>");
507 subscr
= subscr
* 10 + c
- '0';
509 } while (c
!= DEOF
&& Isdigit(c
));
515 dolp
= dolzero
? STR1
: STR0
;
516 cleanup_until(&name
);
520 stderror(ERR_DOLZERO
);
522 length
= Strlen(ffile
);
529 cleanup_until(&name
);
534 stderror(ERR_NOTALLOWED
, "$?<num>");
536 stderror(ERR_NOTALLOWED
, "$%<num>");
541 cleanup_until(&name
);
546 if (c
== DEOF
|| !alnum(c
)) {
547 np
= dimen
? STRargv
: (bitset
? STRstatus
: NULL
);
550 Strbuf_append(&name
, np
);
551 Strbuf_terminate(&name
);
553 subscr
= -1; /* Prevent eating [...] */
558 stderror(ERR_VARALNUM
);
561 Strbuf_append1(&name
, (Char
) c
);
563 if (c
== DEOF
|| !alnum(c
))
566 Strbuf_terminate(&name
);
571 dolp
= (vp
|| getenv(short2str(name
.s
))) ? STR1
: STR0
;
572 cleanup_until(&name
);
575 if (vp
== NULL
|| vp
->vec
== NULL
) {
576 np
= str2short(getenv(short2str(name
.s
)));
578 static Char
*env_val
; /* = NULL; */
580 cleanup_until(&name
);
583 env_val
= Strsave(np
);
590 cleanup_until(&name
);
592 upb
= blklen(vp
->vec
);
593 if (dimen
== 0 && subscr
== 0 && c
== '[') {
595 cleanup_push(&name
, Strbuf_cleanup
);
598 c
= DgetC(DODOL
); /* Allow $ expand within [ ] */
601 if (c
== '\n' || c
== DEOF
)
603 Strbuf_append1(&name
, (Char
) c
);
605 Strbuf_terminate(&name
);
607 if (dolp
|| dolcnt
) /* $ exp must end before ] */
608 stderror(ERR_EXPORD
);
610 stderror(ERR_SYNTAX
);
614 for (i
= 0; Isdigit(*np
); i
= i
* 10 + *np
++ - '0')
616 if ((i
< 0 || i
> upb
) && !any("-*", *np
)) {
617 cleanup_until(&name
);
618 dolerror(vp
->v_name
);
623 upb
= lwb
, np
= STRstar
;
628 stderror(ERR_MISSING
, '-');
636 i
= i
* 10 + *np
++ - '0';
637 if (i
< 0 || i
> upb
) {
638 cleanup_until(&name
);
639 dolerror(vp
->v_name
);
650 cleanup_until(&name
);
651 dolerror(vp
->v_name
);
657 stderror(ERR_SYNTAX
);
658 cleanup_until(&name
);
670 /* this is a kludge. It prevents Dgetdol() from */
671 /* pushing erroneous ${#<error> values into the labuf. */
675 stderror(ERR_MISSING
, '}');
678 addla(putn(upb
- lwb
+ 1));
683 for (i
= lwb
- 1, length
= 0; i
< upb
; i
++)
684 length
+= Strlen(vp
->vec
[i
]);
686 /* We don't want that, since we can always compute it by adding $#xxx */
687 length
+= i
- 1; /* Add the number of spaces in */
694 dolnxt
= &vp
->vec
[lwb
- 1];
695 dolcnt
= upb
- lwb
+ 1;
701 stderror(ERR_MISSING
, '}');
713 c
= DgetC(0), dolmcnt
= 1, dol_flag_a
= 0;
714 if (c
== 'g' || c
== 'a') {
721 if ((c
== 'g' && dolmcnt
!= INT_MAX
) ||
722 (c
== 'a' && dol_flag_a
== 0)) {
730 if (c
== 's') { /* [eichin:19910926.0755EST] */
732 eChar delim
= DgetC(0);
733 Strbuf_append1(&dolmod
, (Char
) c
);
734 Strbuf_append1(&dolmod
, (Char
) delim
);
736 if (delim
== DEOF
|| !delim
|| letter(delim
)
737 || Isdigit(delim
) || any(" \t\n", delim
)) {
738 seterror(ERR_BADSUBST
);
741 while ((c
= DgetC(0)) != DEOF
) {
742 Strbuf_append1(&dolmod
, (Char
) c
);
743 if(c
== delim
) delimcnt
--;
747 seterror(ERR_BADSUBST
);
752 if (!any("luhtrqxes", c
))
753 stderror(ERR_BADMOD
, (int)c
);
754 Strbuf_append1(&dolmod
, (Char
) c
);
758 while ((c
= DgetC(0)) == ':');
771 if (dolmod
.len
== 0 || dolmcnt
== 0) {
776 for (i
= 0; i
< dolmod
.len
; i
++) {
779 /* handle s// [eichin:19910926.0510EST] */
780 if(dolmod
.s
[i
] == 's') {
782 Char
*lhsub
, *rhsub
, *np
;
783 size_t lhlen
= 0, rhlen
= 0;
785 delim
= dolmod
.s
[++i
];
786 if (!delim
|| letter(delim
)
787 || Isdigit(delim
) || any(" \t\n", delim
)) {
788 seterror(ERR_BADSUBST
);
791 lhsub
= &dolmod
.s
[++i
];
792 while(dolmod
.s
[i
] != delim
&& dolmod
.s
[++i
]) {
796 rhsub
= &dolmod
.s
[++i
];
797 while(dolmod
.s
[i
] != delim
&& dolmod
.s
[++i
]) {
806 dp
= Strstr(dp
, lhsub
);
808 ptrdiff_t diff
= dp
- cp
;
809 np
= xmalloc((Strlen(cp
) + 1 - lhlen
+ rhlen
) *
811 (void) Strncpy(np
, cp
, diff
);
812 (void) Strcpy(np
+ diff
, rhsub
);
813 (void) Strcpy(np
+ diff
+ rhlen
, dp
+ lhlen
);
820 /* should this do a seterror? */
824 while (dol_flag_a
!= 0);
826 * restore dolmod for additional words
828 dolmod
.s
[i
] = rhsub
[-1] = (Char
) delim
;
832 if ((dp
= domod(cp
, dolmod
.s
[i
])) != NULL
) {
834 if (Strcmp(cp
, dp
) == 0) {
847 while (dol_flag_a
!= 0);
849 if (didmod
&& dolmcnt
!= INT_MAX
)
876 if ((c
= Dpeekrd
) != 0) {
880 if (Dcp
&& (c
= *Dcp
++))
881 return (c
& (QUOTE
| TRIM
));
901 inheredoc_cleanup(void *dummy
)
908 * Form a shell temporary file (in unit 0) from the words
909 * of the shell input up to EOF or a line the same as "term".
910 * Unit 0 should have been closed before this call.
917 struct Strbuf lbuf
= Strbuf_INIT
, mbuf
= Strbuf_INIT
;
918 Char obuf
[BUFSIZE
+ 1];
919 #define OBUF_END (obuf + sizeof(obuf) / sizeof (*obuf) - 1)
920 Char
*lbp
, *obp
, *mbp
;
928 #endif /* WINNT_NATIVE */
929 tmp
= short2str(shtemp
);
932 if (xcreat(tmp
, 0600) < 0)
933 stderror(ERR_SYSTEM
, tmp
, strerror(errno
));
937 # define O_TEMPORARY 0
942 if (xopen(tmp
, O_RDWR
|O_CREAT
|O_EXCL
|O_TEMPORARY
|O_LARGEFILE
, 0600) ==
946 if (errno
== EEXIST
) {
947 if (unlink(tmp
) == -1) {
948 (void) gettimeofday(&tv
, NULL
);
950 mbp
= putn((((int)tv
.tv_sec
) ^
951 ((int)tv
.tv_usec
) ^ ((int)getpid())) & 0x00ffffff);
952 shtemp
= Strspl(STRtmpsh
, mbp
);
957 #endif /* WINNT_NATIVE */
960 stderror(ERR_SYSTEM
, tmp
, strerror(errno
));
962 (void) unlink(tmp
); /* 0 0 inode! */
972 cleanup_push(&inheredoc
, inheredoc_cleanup
);
975 #endif /* WINNT_NATIVE */
980 setmode(0, O_BINARY
);
982 cleanup_push(&lbuf
, Strbuf_cleanup
);
983 cleanup_push(&mbuf
, Strbuf_cleanup
);
992 c
= readc(1); /* 1 -> Want EOF returns */
993 if (c
== CHAR_ERR
|| c
== '\n')
995 if ((c
&= TRIM
) != 0)
996 Strbuf_append1(&lbuf
, (Char
) c
);
998 Strbuf_terminate(&lbuf
);
1001 * Check for EOF or compare to terminator -- before expansion
1003 if (c
== CHAR_ERR
|| eq(lbuf
.s
, term
))
1007 * If term was quoted or -n just pass it on
1009 if (quoted
|| noexec
) {
1010 Strbuf_append1(&lbuf
, '\n');
1011 Strbuf_terminate(&lbuf
);
1012 for (lbp
= lbuf
.s
; (c
= *lbp
++) != 0;) {
1014 if (obp
== OBUF_END
) {
1015 tmp
= short2str(obuf
);
1016 (void) xwrite(0, tmp
, strlen (tmp
));
1024 * Term wasn't quoted so variable and then command expand the input
1034 if ((c
&= TRIM
) == 0)
1036 /* \ quotes \ $ ` here */
1039 if (!any("$\\`", c
))
1040 unDgetC(c
| QUOTE
), c
= '\\';
1044 Strbuf_append1(&mbuf
, (Char
) c
);
1046 Strbuf_terminate(&mbuf
);
1049 * If any ` in line do command substitution
1052 if (Strchr(mbp
, '`') != NULL
) {
1054 * 1 arg to dobackp causes substitution to be literal. Words are
1055 * broken only at newlines so that all blanks and tabs are
1056 * preserved. Blank lines (null words) are not discarded.
1058 words
= dobackp(mbp
, 1);
1061 /* Setup trivial vector similar to return of dobackp */
1062 Dv
[0] = mbp
, Dv
[1] = NULL
, words
= Dv
;
1065 * Resurrect the words from the command substitution each separated by
1066 * a newline. Note that the last newline of a command substitution
1067 * will have been discarded, but we put a newline after the last word
1068 * because this represents the newline after the last input line!
1070 for (vp
= words
; *vp
; vp
++) {
1071 for (mbp
= *vp
; *mbp
; mbp
++) {
1072 *obp
++ = *mbp
& TRIM
;
1073 if (obp
== OBUF_END
) {
1074 tmp
= short2str(obuf
);
1075 (void) xwrite(0, tmp
, strlen (tmp
));
1080 if (obp
== OBUF_END
) {
1081 tmp
= short2str(obuf
);
1082 (void) xwrite(0, tmp
, strlen (tmp
));
1090 tmp
= short2str(obuf
);
1091 (void) xwrite(0, tmp
, strlen (tmp
));
1092 (void) lseek(0, (off_t
) 0, L_SET
);
1093 cleanup_until(&inheredoc
);