we do not want to shift by the block size, which is much larger than
[dragonfly.git] / sys / kern / subr_scanf.c
blob5060490e1c0a537e542bf57ba3abfcbdfcf4bef0
1 /*-
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
6 * Chris Torek.
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
10 * are met:
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. All advertising materials mentioning features or use of this software
17 * must display the following acknowledgement:
18 * This product includes software developed by the University of
19 * California, Berkeley and its contributors.
20 * 4. Neither the name of the University nor the names of its contributors
21 * may be used to endorse or promote products derived from this software
22 * without specific prior written permission.
24 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
25 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
28 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
29 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
30 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
33 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 * SUCH DAMAGE.
36 * $FreeBSD: src/sys/kern/subr_scanf.c,v 1.13 1999/11/24 01:03:01 archie Exp $
37 * $DragonFly: src/sys/kern/subr_scanf.c,v 1.4 2006/12/13 21:58:50 dillon Exp $
38 * From: Id: vfscanf.c,v 1.13 1998/09/25 12:20:27 obrien Exp
39 * From: static char sccsid[] = "@(#)strtol.c 8.1 (Berkeley) 6/4/93";
40 * From: static char sccsid[] = "@(#)strtoul.c 8.1 (Berkeley) 6/4/93";
43 #include <sys/param.h>
44 #include <sys/systm.h>
45 #include <sys/ctype.h>
46 #include <machine/limits.h>
49 * Note that stdarg.h and the ANSI style va_start macro is used for both
50 * ANSI and traditional C compilers.
52 #include <machine/stdarg.h>
54 #define BUF 32 /* Maximum length of numeric string. */
57 * Flags used during conversion.
59 #define LONG 0x01 /* l: long or double */
60 #define SHORT 0x04 /* h: short */
61 #define SUPPRESS 0x08 /* suppress assignment */
62 #define POINTER 0x10 /* weird %p pointer (`fake hex') */
63 #define NOSKIP 0x20 /* do not skip blanks */
64 #define QUAD 0x400
67 * The following are used in numeric conversions only:
68 * SIGNOK, NDIGITS, DPTOK, and EXPOK are for floating point;
69 * SIGNOK, NDIGITS, PFXOK, and NZDIGITS are for integral.
71 #define SIGNOK 0x40 /* +/- is (still) legal */
72 #define NDIGITS 0x80 /* no digits detected */
74 #define DPTOK 0x100 /* (float) decimal point is still legal */
75 #define EXPOK 0x200 /* (float) exponent (e+3, etc) still legal */
77 #define PFXOK 0x100 /* 0x prefix is (still) legal */
78 #define NZDIGITS 0x200 /* no zero digits detected */
81 * Conversion types.
83 #define CT_CHAR 0 /* %c conversion */
84 #define CT_CCL 1 /* %[...] conversion */
85 #define CT_STRING 2 /* %s conversion */
86 #define CT_INT 3 /* integer, i.e., strtoq or strtouq */
87 typedef u_quad_t (*ccfntype)(const char *, char **, int);
89 static const u_char *__sccl(char *, const u_char *);
91 int
92 ksscanf(const char *ibuf, const char *fmt, ...)
94 __va_list ap;
95 int ret;
97 __va_start(ap, fmt);
98 ret = kvsscanf(ibuf, fmt, ap);
99 __va_end(ap);
100 return(ret);
104 kvsscanf(const char *inp, char const *fmt0, __va_list ap)
106 int inr;
107 const u_char *fmt = (const u_char *)fmt0;
108 int c; /* character from format, or conversion */
109 size_t width; /* field width, or 0 */
110 char *p; /* points into all kinds of strings */
111 int n; /* handy integer */
112 int flags; /* flags as defined above */
113 char *p0; /* saves original value of p when necessary */
114 int nassigned; /* number of fields assigned */
115 int nconversions; /* number of conversions */
116 int nread; /* number of characters consumed from fp */
117 int base; /* base argument to strtoq/strtouq */
118 ccfntype ccfn; /* conversion function (strtoq/strtouq) */
119 char ccltab[256]; /* character class table for %[...] */
120 char buf[BUF]; /* buffer for numeric conversions */
122 /* `basefix' is used to avoid `if' tests in the integer scanner */
123 static short basefix[17] =
124 { 10, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 };
126 inr = strlen(inp);
128 nassigned = 0;
129 nconversions = 0;
130 nread = 0;
131 base = 0; /* XXX just to keep gcc happy */
132 ccfn = NULL; /* XXX just to keep gcc happy */
133 for (;;) {
134 c = *fmt++;
135 if (c == 0)
136 return (nassigned);
137 if (isspace(c)) {
138 while (inr > 0 && isspace(*inp))
139 nread++, inr--, inp++;
140 continue;
142 if (c != '%')
143 goto literal;
144 width = 0;
145 flags = 0;
147 * switch on the format. continue if done;
148 * break once format type is derived.
150 again: c = *fmt++;
151 switch (c) {
152 case '%':
153 literal:
154 if (inr <= 0)
155 goto input_failure;
156 if (*inp != c)
157 goto match_failure;
158 inr--, inp++;
159 nread++;
160 continue;
162 case '*':
163 flags |= SUPPRESS;
164 goto again;
165 case 'l':
166 flags |= LONG;
167 goto again;
168 case 'q':
169 flags |= QUAD;
170 goto again;
171 case 'h':
172 flags |= SHORT;
173 goto again;
175 case '0': case '1': case '2': case '3': case '4':
176 case '5': case '6': case '7': case '8': case '9':
177 width = width * 10 + c - '0';
178 goto again;
181 * Conversions.
184 case 'd':
185 c = CT_INT;
186 ccfn = (ccfntype)strtoq;
187 base = 10;
188 break;
190 case 'i':
191 c = CT_INT;
192 ccfn = (ccfntype)strtoq;
193 base = 0;
194 break;
196 case 'o':
197 c = CT_INT;
198 ccfn = strtouq;
199 base = 8;
200 break;
202 case 'u':
203 c = CT_INT;
204 ccfn = strtouq;
205 base = 10;
206 break;
208 case 'x':
209 flags |= PFXOK; /* enable 0x prefixing */
210 c = CT_INT;
211 ccfn = strtouq;
212 base = 16;
213 break;
215 case 's':
216 c = CT_STRING;
217 break;
219 case '[':
220 fmt = __sccl(ccltab, fmt);
221 flags |= NOSKIP;
222 c = CT_CCL;
223 break;
225 case 'c':
226 flags |= NOSKIP;
227 c = CT_CHAR;
228 break;
230 case 'p': /* pointer format is like hex */
231 flags |= POINTER | PFXOK;
232 c = CT_INT;
233 ccfn = strtouq;
234 base = 16;
235 break;
237 case 'n':
238 nconversions++;
239 if (flags & SUPPRESS) /* ??? */
240 continue;
241 if (flags & SHORT)
242 *__va_arg(ap, short *) = nread;
243 else if (flags & LONG)
244 *__va_arg(ap, long *) = nread;
245 else if (flags & QUAD)
246 *__va_arg(ap, quad_t *) = nread;
247 else
248 *__va_arg(ap, int *) = nread;
249 continue;
253 * We have a conversion that requires input.
255 if (inr <= 0)
256 goto input_failure;
259 * Consume leading white space, except for formats
260 * that suppress this.
262 if ((flags & NOSKIP) == 0) {
263 while (isspace(*inp)) {
264 nread++;
265 if (--inr > 0)
266 inp++;
267 else
268 goto input_failure;
271 * Note that there is at least one character in
272 * the buffer, so conversions that do not set NOSKIP
273 * can no longer result in an input failure.
278 * Do the conversion.
280 switch (c) {
282 case CT_CHAR:
283 /* scan arbitrary characters (sets NOSKIP) */
284 if (width == 0)
285 width = 1;
286 if (flags & SUPPRESS) {
287 size_t sum = 0;
288 for (;;) {
289 if ((n = inr) < width) {
290 sum += n;
291 width -= n;
292 inp += n;
293 if (sum == 0)
294 goto input_failure;
295 break;
296 } else {
297 sum += width;
298 inr -= width;
299 inp += width;
300 break;
303 nread += sum;
304 } else {
305 bcopy(inp, __va_arg(ap, char *), width);
306 inr -= width;
307 inp += width;
308 nread += width;
309 nassigned++;
311 nconversions++;
312 break;
314 case CT_CCL:
315 /* scan a (nonempty) character class (sets NOSKIP) */
316 if (width == 0)
317 width = (size_t)~0; /* `infinity' */
318 /* take only those things in the class */
319 if (flags & SUPPRESS) {
320 n = 0;
321 while (ccltab[(unsigned char)*inp]) {
322 n++, inr--, inp++;
323 if (--width == 0)
324 break;
325 if (inr <= 0) {
326 if (n == 0)
327 goto input_failure;
328 break;
331 if (n == 0)
332 goto match_failure;
333 } else {
334 p0 = p = __va_arg(ap, char *);
335 while (ccltab[(unsigned char)*inp]) {
336 inr--;
337 *p++ = *inp++;
338 if (--width == 0)
339 break;
340 if (inr <= 0) {
341 if (p == p0)
342 goto input_failure;
343 break;
346 n = p - p0;
347 if (n == 0)
348 goto match_failure;
349 *p = 0;
350 nassigned++;
352 nread += n;
353 nconversions++;
354 break;
356 case CT_STRING:
357 /* like CCL, but zero-length string OK, & no NOSKIP */
358 if (width == 0)
359 width = (size_t)~0;
360 if (flags & SUPPRESS) {
361 n = 0;
362 while (!isspace(*inp)) {
363 n++, inr--, inp++;
364 if (--width == 0)
365 break;
366 if (inr <= 0)
367 break;
369 nread += n;
370 } else {
371 p0 = p = __va_arg(ap, char *);
372 while (!isspace(*inp)) {
373 inr--;
374 *p++ = *inp++;
375 if (--width == 0)
376 break;
377 if (inr <= 0)
378 break;
380 *p = 0;
381 nread += p - p0;
382 nassigned++;
384 nconversions++;
385 continue;
387 case CT_INT:
388 /* scan an integer as if by strtoq/strtouq */
389 #ifdef hardway
390 if (width == 0 || width > sizeof(buf) - 1)
391 width = sizeof(buf) - 1;
392 #else
393 /* size_t is unsigned, hence this optimisation */
394 if (--width > sizeof(buf) - 2)
395 width = sizeof(buf) - 2;
396 width++;
397 #endif
398 flags |= SIGNOK | NDIGITS | NZDIGITS;
399 for (p = buf; width; width--) {
400 c = *inp;
402 * Switch on the character; `goto ok'
403 * if we accept it as a part of number.
405 switch (c) {
408 * The digit 0 is always legal, but is
409 * special. For %i conversions, if no
410 * digits (zero or nonzero) have been
411 * scanned (only signs), we will have
412 * base==0. In that case, we should set
413 * it to 8 and enable 0x prefixing.
414 * Also, if we have not scanned zero digits
415 * before this, do not turn off prefixing
416 * (someone else will turn it off if we
417 * have scanned any nonzero digits).
419 case '0':
420 if (base == 0) {
421 base = 8;
422 flags |= PFXOK;
424 if (flags & NZDIGITS)
425 flags &= ~(SIGNOK|NZDIGITS|NDIGITS);
426 else
427 flags &= ~(SIGNOK|PFXOK|NDIGITS);
428 goto ok;
430 /* 1 through 7 always legal */
431 case '1': case '2': case '3':
432 case '4': case '5': case '6': case '7':
433 base = basefix[base];
434 flags &= ~(SIGNOK | PFXOK | NDIGITS);
435 goto ok;
437 /* digits 8 and 9 ok iff decimal or hex */
438 case '8': case '9':
439 base = basefix[base];
440 if (base <= 8)
441 break; /* not legal here */
442 flags &= ~(SIGNOK | PFXOK | NDIGITS);
443 goto ok;
445 /* letters ok iff hex */
446 case 'A': case 'B': case 'C':
447 case 'D': case 'E': case 'F':
448 case 'a': case 'b': case 'c':
449 case 'd': case 'e': case 'f':
450 /* no need to fix base here */
451 if (base <= 10)
452 break; /* not legal here */
453 flags &= ~(SIGNOK | PFXOK | NDIGITS);
454 goto ok;
456 /* sign ok only as first character */
457 case '+': case '-':
458 if (flags & SIGNOK) {
459 flags &= ~SIGNOK;
460 goto ok;
462 break;
464 /* x ok iff flag still set & 2nd char */
465 case 'x': case 'X':
466 if (flags & PFXOK && p == buf + 1) {
467 base = 16; /* if %i */
468 flags &= ~PFXOK;
469 goto ok;
471 break;
475 * If we got here, c is not a legal character
476 * for a number. Stop accumulating digits.
478 break;
481 * c is legal: store it and look at the next.
483 *p++ = c;
484 if (--inr > 0)
485 inp++;
486 else
487 break; /* end of input */
490 * If we had only a sign, it is no good; push
491 * back the sign. If the number ends in `x',
492 * it was [sign] '0' 'x', so push back the x
493 * and treat it as [sign] '0'.
495 if (flags & NDIGITS) {
496 if (p > buf) {
497 inp--;
498 inr++;
500 goto match_failure;
502 c = ((u_char *)p)[-1];
503 if (c == 'x' || c == 'X') {
504 --p;
505 inp--;
506 inr++;
508 if ((flags & SUPPRESS) == 0) {
509 u_quad_t res;
511 *p = 0;
512 res = (*ccfn)(buf, (char **)NULL, base);
513 if (flags & POINTER)
514 *__va_arg(ap, void **) =
515 (void *)(uintptr_t)res;
516 else if (flags & SHORT)
517 *__va_arg(ap, short *) = res;
518 else if (flags & LONG)
519 *__va_arg(ap, long *) = res;
520 else if (flags & QUAD)
521 *__va_arg(ap, quad_t *) = res;
522 else
523 *__va_arg(ap, int *) = res;
524 nassigned++;
526 nread += p - buf;
527 nconversions++;
528 break;
532 input_failure:
533 return (nconversions != 0 ? nassigned : -1);
534 match_failure:
535 return (nassigned);
539 * Fill in the given table from the scanset at the given format
540 * (just after `['). Return a pointer to the character past the
541 * closing `]'. The table has a 1 wherever characters should be
542 * considered part of the scanset.
544 static const u_char *
545 __sccl(char *tab, const u_char *fmt)
547 int c, n, v;
549 /* first `clear' the whole table */
550 c = *fmt++; /* first char hat => negated scanset */
551 if (c == '^') {
552 v = 1; /* default => accept */
553 c = *fmt++; /* get new first char */
554 } else
555 v = 0; /* default => reject */
557 /* XXX: Will not work if sizeof(tab*) > sizeof(char) */
558 for (n = 0; n < 256; n++)
559 tab[n] = v; /* memset(tab, v, 256) */
561 if (c == 0)
562 return (fmt - 1);/* format ended before closing ] */
565 * Now set the entries corresponding to the actual scanset
566 * to the opposite of the above.
568 * The first character may be ']' (or '-') without being special;
569 * the last character may be '-'.
571 v = 1 - v;
572 for (;;) {
573 tab[c] = v; /* take character c */
574 doswitch:
575 n = *fmt++; /* and examine the next */
576 switch (n) {
578 case 0: /* format ended too soon */
579 return (fmt - 1);
581 case '-':
583 * A scanset of the form
584 * [01+-]
585 * is defined as `the digit 0, the digit 1,
586 * the character +, the character -', but
587 * the effect of a scanset such as
588 * [a-zA-Z0-9]
589 * is implementation defined. The V7 Unix
590 * scanf treats `a-z' as `the letters a through
591 * z', but treats `a-a' as `the letter a, the
592 * character -, and the letter a'.
594 * For compatibility, the `-' is not considerd
595 * to define a range if the character following
596 * it is either a close bracket (required by ANSI)
597 * or is not numerically greater than the character
598 * we just stored in the table (c).
600 n = *fmt;
601 if (n == ']' || n < c) {
602 c = '-';
603 break; /* resume the for(;;) */
605 fmt++;
606 /* fill in the range */
607 do {
608 tab[++c] = v;
609 } while (c < n);
610 c = n;
612 * Alas, the V7 Unix scanf also treats formats
613 * such as [a-c-e] as `the letters a through e'.
614 * This too is permitted by the standard....
616 goto doswitch;
617 break;
619 case ']': /* end of scanset */
620 return (fmt);
622 default: /* just another character */
623 c = n;
624 break;
627 /* NOTREACHED */