2 * Copyright (c) 1990, 1993
3 * The Regents of the University of California. All rights reserved.
5 * This code is derived from software contributed to Berkeley by
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in the
15 * documentation and/or other materials provided with the distribution.
16 * 3. Neither the name of the University nor the names of its contributors
17 * may be used to endorse or promote products derived from this software
18 * without specific prior written permission.
20 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
21 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
24 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 * $FreeBSD: src/sys/kern/subr_scanf.c,v 1.13 1999/11/24 01:03:01 archie Exp $
33 * $DragonFly: src/sys/kern/subr_scanf.c,v 1.4 2006/12/13 21:58:50 dillon Exp $
34 * From: Id: vfscanf.c,v 1.13 1998/09/25 12:20:27 obrien Exp
35 * From: static char sccsid[] = "@(#)strtol.c 8.1 (Berkeley) 6/4/93";
36 * From: static char sccsid[] = "@(#)strtoul.c 8.1 (Berkeley) 6/4/93";
39 #include <sys/param.h>
40 #include <sys/systm.h>
41 #include <sys/ctype.h>
42 #include <machine/limits.h>
45 * Note that stdarg.h and the ANSI style va_start macro is used for both
46 * ANSI and traditional C compilers.
48 #include <machine/stdarg.h>
50 #define BUF 32 /* Maximum length of numeric string. */
53 * Flags used during conversion.
55 #define LONG 0x01 /* l: long or double */
56 #define SHORT 0x04 /* h: short */
57 #define SUPPRESS 0x08 /* suppress assignment */
58 #define POINTER 0x10 /* weird %p pointer (`fake hex') */
59 #define NOSKIP 0x20 /* do not skip blanks */
61 #define SHORTSHORT 0x4000 /* hh: char */
64 * The following are used in numeric conversions only:
65 * SIGNOK, NDIGITS, DPTOK, and EXPOK are for floating point;
66 * SIGNOK, NDIGITS, PFXOK, and NZDIGITS are for integral.
68 #define SIGNOK 0x40 /* +/- is (still) legal */
69 #define NDIGITS 0x80 /* no digits detected */
71 #define DPTOK 0x100 /* (float) decimal point is still legal */
72 #define EXPOK 0x200 /* (float) exponent (e+3, etc) still legal */
74 #define PFXOK 0x100 /* 0x prefix is (still) legal */
75 #define NZDIGITS 0x200 /* no zero digits detected */
80 #define CT_CHAR 0 /* %c conversion */
81 #define CT_CCL 1 /* %[...] conversion */
82 #define CT_STRING 2 /* %s conversion */
83 #define CT_INT 3 /* integer, i.e., strtoq or strtouq */
84 typedef u_quad_t (*ccfntype
)(const char *, char **, int);
86 static const u_char
*__sccl(char *, const u_char
*);
89 ksscanf(const char *ibuf
, const char *fmt
, ...)
95 ret
= kvsscanf(ibuf
, fmt
, ap
);
101 kvsscanf(const char *inp
, char const *fmt0
, __va_list ap
)
104 const u_char
*fmt
= (const u_char
*)fmt0
;
105 int c
; /* character from format, or conversion */
106 size_t width
; /* field width, or 0 */
107 char *p
; /* points into all kinds of strings */
108 int n
; /* handy integer */
109 int flags
; /* flags as defined above */
110 char *p0
; /* saves original value of p when necessary */
111 int nassigned
; /* number of fields assigned */
112 int nconversions
; /* number of conversions */
113 int nread
; /* number of characters consumed from fp */
114 int base
; /* base argument to strtoq/strtouq */
115 ccfntype ccfn
; /* conversion function (strtoq/strtouq) */
116 char ccltab
[256]; /* character class table for %[...] */
117 char buf
[BUF
]; /* buffer for numeric conversions */
119 /* `basefix' is used to avoid `if' tests in the integer scanner */
120 static short basefix
[17] =
121 { 10, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 };
128 base
= 0; /* XXX just to keep gcc happy */
129 ccfn
= NULL
; /* XXX just to keep gcc happy */
135 while (inr
> 0 && isspace(*inp
))
136 nread
++, inr
--, inp
++;
144 * switch on the format. continue if done;
145 * break once format type is derived.
176 case '0': case '1': case '2': case '3': case '4':
177 case '5': case '6': case '7': case '8': case '9':
178 width
= width
* 10 + c
- '0';
187 ccfn
= (ccfntype
)strtoq
;
193 ccfn
= (ccfntype
)strtoq
;
210 flags
|= PFXOK
; /* enable 0x prefixing */
221 fmt
= __sccl(ccltab
, fmt
);
231 case 'p': /* pointer format is like hex */
232 flags
|= POINTER
| PFXOK
;
240 if (flags
& SUPPRESS
) /* ??? */
242 if (flags
& SHORTSHORT
)
243 *__va_arg(ap
, char *) = nread
;
244 else if (flags
& SHORT
)
245 *__va_arg(ap
, short *) = nread
;
246 else if (flags
& LONG
)
247 *__va_arg(ap
, long *) = nread
;
248 else if (flags
& QUAD
)
249 *__va_arg(ap
, quad_t
*) = nread
;
251 *__va_arg(ap
, int *) = nread
;
256 * We have a conversion that requires input.
262 * Consume leading white space, except for formats
263 * that suppress this.
265 if ((flags
& NOSKIP
) == 0) {
266 while (isspace(*inp
)) {
274 * Note that there is at least one character in
275 * the buffer, so conversions that do not set NOSKIP
276 * can no longer result in an input failure.
286 /* scan arbitrary characters (sets NOSKIP) */
289 if (flags
& SUPPRESS
) {
292 if ((n
= inr
) < width
) {
308 bcopy(inp
, __va_arg(ap
, char *), width
);
318 /* scan a (nonempty) character class (sets NOSKIP) */
320 width
= (size_t)~0; /* `infinity' */
321 /* take only those things in the class */
322 if (flags
& SUPPRESS
) {
324 while (ccltab
[(unsigned char)*inp
]) {
337 p0
= p
= __va_arg(ap
, char *);
338 while (ccltab
[(unsigned char)*inp
]) {
360 /* like CCL, but zero-length string OK, & no NOSKIP */
363 if (flags
& SUPPRESS
) {
365 while (!isspace(*inp
)) {
374 p0
= p
= __va_arg(ap
, char *);
375 while (!isspace(*inp
)) {
391 /* scan an integer as if by strtoq/strtouq */
393 if (width
== 0 || width
> sizeof(buf
) - 1)
394 width
= sizeof(buf
) - 1;
396 /* size_t is unsigned, hence this optimisation */
397 if (--width
> sizeof(buf
) - 2)
398 width
= sizeof(buf
) - 2;
401 flags
|= SIGNOK
| NDIGITS
| NZDIGITS
;
402 for (p
= buf
; width
; width
--) {
405 * Switch on the character; `goto ok'
406 * if we accept it as a part of number.
411 * The digit 0 is always legal, but is
412 * special. For %i conversions, if no
413 * digits (zero or nonzero) have been
414 * scanned (only signs), we will have
415 * base==0. In that case, we should set
416 * it to 8 and enable 0x prefixing.
417 * Also, if we have not scanned zero digits
418 * before this, do not turn off prefixing
419 * (someone else will turn it off if we
420 * have scanned any nonzero digits).
427 if (flags
& NZDIGITS
)
428 flags
&= ~(SIGNOK
|NZDIGITS
|NDIGITS
);
430 flags
&= ~(SIGNOK
|PFXOK
|NDIGITS
);
433 /* 1 through 7 always legal */
434 case '1': case '2': case '3':
435 case '4': case '5': case '6': case '7':
436 base
= basefix
[base
];
437 flags
&= ~(SIGNOK
| PFXOK
| NDIGITS
);
440 /* digits 8 and 9 ok iff decimal or hex */
442 base
= basefix
[base
];
444 break; /* not legal here */
445 flags
&= ~(SIGNOK
| PFXOK
| NDIGITS
);
448 /* letters ok iff hex */
449 case 'A': case 'B': case 'C':
450 case 'D': case 'E': case 'F':
451 case 'a': case 'b': case 'c':
452 case 'd': case 'e': case 'f':
453 /* no need to fix base here */
455 break; /* not legal here */
456 flags
&= ~(SIGNOK
| PFXOK
| NDIGITS
);
459 /* sign ok only as first character */
461 if (flags
& SIGNOK
) {
467 /* x ok iff flag still set & 2nd char */
469 if (flags
& PFXOK
&& p
== buf
+ 1) {
470 base
= 16; /* if %i */
478 * If we got here, c is not a legal character
479 * for a number. Stop accumulating digits.
484 * c is legal: store it and look at the next.
490 break; /* end of input */
493 * If we had only a sign, it is no good; push
494 * back the sign. If the number ends in `x',
495 * it was [sign] '0' 'x', so push back the x
496 * and treat it as [sign] '0'.
498 if (flags
& NDIGITS
) {
505 c
= ((u_char
*)p
)[-1];
506 if (c
== 'x' || c
== 'X') {
511 if ((flags
& SUPPRESS
) == 0) {
515 res
= (*ccfn
)(buf
, NULL
, base
);
517 *__va_arg(ap
, void **) =
518 (void *)(uintptr_t)res
;
519 else if (flags
& SHORTSHORT
)
520 *__va_arg(ap
, char *) = res
;
521 else if (flags
& SHORT
)
522 *__va_arg(ap
, short *) = res
;
523 else if (flags
& LONG
)
524 *__va_arg(ap
, long *) = res
;
525 else if (flags
& QUAD
)
526 *__va_arg(ap
, quad_t
*) = res
;
528 *__va_arg(ap
, int *) = res
;
538 return (nconversions
!= 0 ? nassigned
: -1);
544 * Fill in the given table from the scanset at the given format
545 * (just after `['). Return a pointer to the character past the
546 * closing `]'. The table has a 1 wherever characters should be
547 * considered part of the scanset.
549 static const u_char
*
550 __sccl(char *tab
, const u_char
*fmt
)
554 /* first `clear' the whole table */
555 c
= *fmt
++; /* first char hat => negated scanset */
557 v
= 1; /* default => accept */
558 c
= *fmt
++; /* get new first char */
560 v
= 0; /* default => reject */
562 /* XXX: Will not work if sizeof(tab*) > sizeof(char) */
563 for (n
= 0; n
< 256; n
++)
564 tab
[n
] = v
; /* memset(tab, v, 256) */
567 return (fmt
- 1);/* format ended before closing ] */
570 * Now set the entries corresponding to the actual scanset
571 * to the opposite of the above.
573 * The first character may be ']' (or '-') without being special;
574 * the last character may be '-'.
578 tab
[c
] = v
; /* take character c */
580 n
= *fmt
++; /* and examine the next */
583 case 0: /* format ended too soon */
588 * A scanset of the form
590 * is defined as `the digit 0, the digit 1,
591 * the character +, the character -', but
592 * the effect of a scanset such as
594 * is implementation defined. The V7 Unix
595 * scanf treats `a-z' as `the letters a through
596 * z', but treats `a-a' as `the letter a, the
597 * character -, and the letter a'.
599 * For compatibility, the `-' is not considerd
600 * to define a range if the character following
601 * it is either a close bracket (required by ANSI)
602 * or is not numerically greater than the character
603 * we just stored in the table (c).
606 if (n
== ']' || n
< c
) {
608 break; /* resume the for(;;) */
611 /* fill in the range */
617 * Alas, the V7 Unix scanf also treats formats
618 * such as [a-c-e] as `the letters a through e'.
619 * This too is permitted by the standard....
624 case ']': /* end of scanset */
627 default: /* just another character */