source: vendor/perl/5.8.8/numeric.c@ 3656

Last change on this file since 3656 was 3181, checked in by bird, 18 years ago

perl 5.8.8

File size: 30.7 KB
Line 
1/* numeric.c
2 *
3 * Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999,
4 * 2000, 2001, 2002, 2003, 2004, 2005, 2006, by Larry Wall and others
5 *
6 * You may distribute under the terms of either the GNU General Public
7 * License or the Artistic License, as specified in the README file.
8 *
9 */
10
11/*
12 * "That only makes eleven (plus one mislaid) and not fourteen, unless
13 * wizards count differently to other people."
14 */
15
16/*
17=head1 Numeric functions
18
19This file contains all the stuff needed by perl for manipulating numeric
20values, including such things as replacements for the OS's atof() function
21
22=cut
23
24*/
25
26#include "EXTERN.h"
27#define PERL_IN_NUMERIC_C
28#include "perl.h"
29
30U32
31Perl_cast_ulong(pTHX_ NV f)
32{
33 if (f < 0.0)
34 return f < I32_MIN ? (U32) I32_MIN : (U32)(I32) f;
35 if (f < U32_MAX_P1) {
36#if CASTFLAGS & 2
37 if (f < U32_MAX_P1_HALF)
38 return (U32) f;
39 f -= U32_MAX_P1_HALF;
40 return ((U32) f) | (1 + U32_MAX >> 1);
41#else
42 return (U32) f;
43#endif
44 }
45 return f > 0 ? U32_MAX : 0 /* NaN */;
46}
47
48I32
49Perl_cast_i32(pTHX_ NV f)
50{
51 if (f < I32_MAX_P1)
52 return f < I32_MIN ? I32_MIN : (I32) f;
53 if (f < U32_MAX_P1) {
54#if CASTFLAGS & 2
55 if (f < U32_MAX_P1_HALF)
56 return (I32)(U32) f;
57 f -= U32_MAX_P1_HALF;
58 return (I32)(((U32) f) | (1 + U32_MAX >> 1));
59#else
60 return (I32)(U32) f;
61#endif
62 }
63 return f > 0 ? (I32)U32_MAX : 0 /* NaN */;
64}
65
66IV
67Perl_cast_iv(pTHX_ NV f)
68{
69 if (f < IV_MAX_P1)
70 return f < IV_MIN ? IV_MIN : (IV) f;
71 if (f < UV_MAX_P1) {
72#if CASTFLAGS & 2
73 /* For future flexibility allowing for sizeof(UV) >= sizeof(IV) */
74 if (f < UV_MAX_P1_HALF)
75 return (IV)(UV) f;
76 f -= UV_MAX_P1_HALF;
77 return (IV)(((UV) f) | (1 + UV_MAX >> 1));
78#else
79 return (IV)(UV) f;
80#endif
81 }
82 return f > 0 ? (IV)UV_MAX : 0 /* NaN */;
83}
84
85UV
86Perl_cast_uv(pTHX_ NV f)
87{
88 if (f < 0.0)
89 return f < IV_MIN ? (UV) IV_MIN : (UV)(IV) f;
90 if (f < UV_MAX_P1) {
91#if CASTFLAGS & 2
92 if (f < UV_MAX_P1_HALF)
93 return (UV) f;
94 f -= UV_MAX_P1_HALF;
95 return ((UV) f) | (1 + UV_MAX >> 1);
96#else
97 return (UV) f;
98#endif
99 }
100 return f > 0 ? UV_MAX : 0 /* NaN */;
101}
102
103#if defined(HUGE_VAL) || (defined(USE_LONG_DOUBLE) && defined(HUGE_VALL))
104/*
105 * This hack is to force load of "huge" support from libm.a
106 * So it is in perl for (say) POSIX to use.
107 * Needed for SunOS with Sun's 'acc' for example.
108 */
109NV
110Perl_huge(void)
111{
112# if defined(USE_LONG_DOUBLE) && defined(HUGE_VALL)
113 return HUGE_VALL;
114# endif
115 return HUGE_VAL;
116}
117#endif
118
119/*
120=for apidoc grok_bin
121
122converts a string representing a binary number to numeric form.
123
124On entry I<start> and I<*len> give the string to scan, I<*flags> gives
125conversion flags, and I<result> should be NULL or a pointer to an NV.
126The scan stops at the end of the string, or the first invalid character.
127Unless C<PERL_SCAN_SILENT_ILLDIGIT> is set in I<*flags>, encountering an
128invalid character will also trigger a warning.
129On return I<*len> is set to the length of the scanned string,
130and I<*flags> gives output flags.
131
132If the value is <= C<UV_MAX> it is returned as a UV, the output flags are clear,
133and nothing is written to I<*result>. If the value is > UV_MAX C<grok_bin>
134returns UV_MAX, sets C<PERL_SCAN_GREATER_THAN_UV_MAX> in the output flags,
135and writes the value to I<*result> (or the value is discarded if I<result>
136is NULL).
137
138The binary number may optionally be prefixed with "0b" or "b" unless
139C<PERL_SCAN_DISALLOW_PREFIX> is set in I<*flags> on entry. If
140C<PERL_SCAN_ALLOW_UNDERSCORES> is set in I<*flags> then the binary
141number may use '_' characters to separate digits.
142
143=cut
144 */
145
146UV
147Perl_grok_bin(pTHX_ char *start, STRLEN *len_p, I32 *flags, NV *result) {
148 const char *s = start;
149 STRLEN len = *len_p;
150 UV value = 0;
151 NV value_nv = 0;
152
153 const UV max_div_2 = UV_MAX / 2;
154 const bool allow_underscores = (bool)(*flags & PERL_SCAN_ALLOW_UNDERSCORES);
155 bool overflowed = FALSE;
156 char bit;
157
158 if (!(*flags & PERL_SCAN_DISALLOW_PREFIX)) {
159 /* strip off leading b or 0b.
160 for compatibility silently suffer "b" and "0b" as valid binary
161 numbers. */
162 if (len >= 1) {
163 if (s[0] == 'b') {
164 s++;
165 len--;
166 }
167 else if (len >= 2 && s[0] == '0' && s[1] == 'b') {
168 s+=2;
169 len-=2;
170 }
171 }
172 }
173
174 for (; len-- && (bit = *s); s++) {
175 if (bit == '0' || bit == '1') {
176 /* Write it in this wonky order with a goto to attempt to get the
177 compiler to make the common case integer-only loop pretty tight.
178 With gcc seems to be much straighter code than old scan_bin. */
179 redo:
180 if (!overflowed) {
181 if (value <= max_div_2) {
182 value = (value << 1) | (bit - '0');
183 continue;
184 }
185 /* Bah. We're just overflowed. */
186 if (ckWARN_d(WARN_OVERFLOW))
187 Perl_warner(aTHX_ packWARN(WARN_OVERFLOW),
188 "Integer overflow in binary number");
189 overflowed = TRUE;
190 value_nv = (NV) value;
191 }
192 value_nv *= 2.0;
193 /* If an NV has not enough bits in its mantissa to
194 * represent a UV this summing of small low-order numbers
195 * is a waste of time (because the NV cannot preserve
196 * the low-order bits anyway): we could just remember when
197 * did we overflow and in the end just multiply value_nv by the
198 * right amount. */
199 value_nv += (NV)(bit - '0');
200 continue;
201 }
202 if (bit == '_' && len && allow_underscores && (bit = s[1])
203 && (bit == '0' || bit == '1'))
204 {
205 --len;
206 ++s;
207 goto redo;
208 }
209 if (!(*flags & PERL_SCAN_SILENT_ILLDIGIT) && ckWARN(WARN_DIGIT))
210 Perl_warner(aTHX_ packWARN(WARN_DIGIT),
211 "Illegal binary digit '%c' ignored", *s);
212 break;
213 }
214
215 if ( ( overflowed && value_nv > 4294967295.0)
216#if UVSIZE > 4
217 || (!overflowed && value > 0xffffffff )
218#endif
219 ) {
220 if (ckWARN(WARN_PORTABLE))
221 Perl_warner(aTHX_ packWARN(WARN_PORTABLE),
222 "Binary number > 0b11111111111111111111111111111111 non-portable");
223 }
224 *len_p = s - start;
225 if (!overflowed) {
226 *flags = 0;
227 return value;
228 }
229 *flags = PERL_SCAN_GREATER_THAN_UV_MAX;
230 if (result)
231 *result = value_nv;
232 return UV_MAX;
233}
234
235/*
236=for apidoc grok_hex
237
238converts a string representing a hex number to numeric form.
239
240On entry I<start> and I<*len> give the string to scan, I<*flags> gives
241conversion flags, and I<result> should be NULL or a pointer to an NV.
242The scan stops at the end of the string, or the first invalid character.
243Unless C<PERL_SCAN_SILENT_ILLDIGIT> is set in I<*flags>, encountering an
244invalid character will also trigger a warning.
245On return I<*len> is set to the length of the scanned string,
246and I<*flags> gives output flags.
247
248If the value is <= UV_MAX it is returned as a UV, the output flags are clear,
249and nothing is written to I<*result>. If the value is > UV_MAX C<grok_hex>
250returns UV_MAX, sets C<PERL_SCAN_GREATER_THAN_UV_MAX> in the output flags,
251and writes the value to I<*result> (or the value is discarded if I<result>
252is NULL).
253
254The hex number may optionally be prefixed with "0x" or "x" unless
255C<PERL_SCAN_DISALLOW_PREFIX> is set in I<*flags> on entry. If
256C<PERL_SCAN_ALLOW_UNDERSCORES> is set in I<*flags> then the hex
257number may use '_' characters to separate digits.
258
259=cut
260 */
261
262UV
263Perl_grok_hex(pTHX_ char *start, STRLEN *len_p, I32 *flags, NV *result) {
264 const char *s = start;
265 STRLEN len = *len_p;
266 UV value = 0;
267 NV value_nv = 0;
268
269 const UV max_div_16 = UV_MAX / 16;
270 const bool allow_underscores = (bool)(*flags & PERL_SCAN_ALLOW_UNDERSCORES);
271 bool overflowed = FALSE;
272
273 if (!(*flags & PERL_SCAN_DISALLOW_PREFIX)) {
274 /* strip off leading x or 0x.
275 for compatibility silently suffer "x" and "0x" as valid hex numbers.
276 */
277 if (len >= 1) {
278 if (s[0] == 'x') {
279 s++;
280 len--;
281 }
282 else if (len >= 2 && s[0] == '0' && s[1] == 'x') {
283 s+=2;
284 len-=2;
285 }
286 }
287 }
288
289 for (; len-- && *s; s++) {
290 const char *hexdigit = strchr(PL_hexdigit, *s);
291 if (hexdigit) {
292 /* Write it in this wonky order with a goto to attempt to get the
293 compiler to make the common case integer-only loop pretty tight.
294 With gcc seems to be much straighter code than old scan_hex. */
295 redo:
296 if (!overflowed) {
297 if (value <= max_div_16) {
298 value = (value << 4) | ((hexdigit - PL_hexdigit) & 15);
299 continue;
300 }
301 /* Bah. We're just overflowed. */
302 if (ckWARN_d(WARN_OVERFLOW))
303 Perl_warner(aTHX_ packWARN(WARN_OVERFLOW),
304 "Integer overflow in hexadecimal number");
305 overflowed = TRUE;
306 value_nv = (NV) value;
307 }
308 value_nv *= 16.0;
309 /* If an NV has not enough bits in its mantissa to
310 * represent a UV this summing of small low-order numbers
311 * is a waste of time (because the NV cannot preserve
312 * the low-order bits anyway): we could just remember when
313 * did we overflow and in the end just multiply value_nv by the
314 * right amount of 16-tuples. */
315 value_nv += (NV)((hexdigit - PL_hexdigit) & 15);
316 continue;
317 }
318 if (*s == '_' && len && allow_underscores && s[1]
319 && (hexdigit = strchr(PL_hexdigit, s[1])))
320 {
321 --len;
322 ++s;
323 goto redo;
324 }
325 if (!(*flags & PERL_SCAN_SILENT_ILLDIGIT) && ckWARN(WARN_DIGIT))
326 Perl_warner(aTHX_ packWARN(WARN_DIGIT),
327 "Illegal hexadecimal digit '%c' ignored", *s);
328 break;
329 }
330
331 if ( ( overflowed && value_nv > 4294967295.0)
332#if UVSIZE > 4
333 || (!overflowed && value > 0xffffffff )
334#endif
335 ) {
336 if (ckWARN(WARN_PORTABLE))
337 Perl_warner(aTHX_ packWARN(WARN_PORTABLE),
338 "Hexadecimal number > 0xffffffff non-portable");
339 }
340 *len_p = s - start;
341 if (!overflowed) {
342 *flags = 0;
343 return value;
344 }
345 *flags = PERL_SCAN_GREATER_THAN_UV_MAX;
346 if (result)
347 *result = value_nv;
348 return UV_MAX;
349}
350
351/*
352=for apidoc grok_oct
353
354converts a string representing an octal number to numeric form.
355
356On entry I<start> and I<*len> give the string to scan, I<*flags> gives
357conversion flags, and I<result> should be NULL or a pointer to an NV.
358The scan stops at the end of the string, or the first invalid character.
359Unless C<PERL_SCAN_SILENT_ILLDIGIT> is set in I<*flags>, encountering an
360invalid character will also trigger a warning.
361On return I<*len> is set to the length of the scanned string,
362and I<*flags> gives output flags.
363
364If the value is <= UV_MAX it is returned as a UV, the output flags are clear,
365and nothing is written to I<*result>. If the value is > UV_MAX C<grok_oct>
366returns UV_MAX, sets C<PERL_SCAN_GREATER_THAN_UV_MAX> in the output flags,
367and writes the value to I<*result> (or the value is discarded if I<result>
368is NULL).
369
370If C<PERL_SCAN_ALLOW_UNDERSCORES> is set in I<*flags> then the octal
371number may use '_' characters to separate digits.
372
373=cut
374 */
375
376UV
377Perl_grok_oct(pTHX_ char *start, STRLEN *len_p, I32 *flags, NV *result) {
378 const char *s = start;
379 STRLEN len = *len_p;
380 UV value = 0;
381 NV value_nv = 0;
382
383 const UV max_div_8 = UV_MAX / 8;
384 const bool allow_underscores = (bool)(*flags & PERL_SCAN_ALLOW_UNDERSCORES);
385 bool overflowed = FALSE;
386
387 for (; len-- && *s; s++) {
388 /* gcc 2.95 optimiser not smart enough to figure that this subtraction
389 out front allows slicker code. */
390 int digit = *s - '0';
391 if (digit >= 0 && digit <= 7) {
392 /* Write it in this wonky order with a goto to attempt to get the
393 compiler to make the common case integer-only loop pretty tight.
394 */
395 redo:
396 if (!overflowed) {
397 if (value <= max_div_8) {
398 value = (value << 3) | digit;
399 continue;
400 }
401 /* Bah. We're just overflowed. */
402 if (ckWARN_d(WARN_OVERFLOW))
403 Perl_warner(aTHX_ packWARN(WARN_OVERFLOW),
404 "Integer overflow in octal number");
405 overflowed = TRUE;
406 value_nv = (NV) value;
407 }
408 value_nv *= 8.0;
409 /* If an NV has not enough bits in its mantissa to
410 * represent a UV this summing of small low-order numbers
411 * is a waste of time (because the NV cannot preserve
412 * the low-order bits anyway): we could just remember when
413 * did we overflow and in the end just multiply value_nv by the
414 * right amount of 8-tuples. */
415 value_nv += (NV)digit;
416 continue;
417 }
418 if (digit == ('_' - '0') && len && allow_underscores
419 && (digit = s[1] - '0') && (digit >= 0 && digit <= 7))
420 {
421 --len;
422 ++s;
423 goto redo;
424 }
425 /* Allow \octal to work the DWIM way (that is, stop scanning
426 * as soon as non-octal characters are seen, complain only if
427 * someone seems to want to use the digits eight and nine). */
428 if (digit == 8 || digit == 9) {
429 if (!(*flags & PERL_SCAN_SILENT_ILLDIGIT) && ckWARN(WARN_DIGIT))
430 Perl_warner(aTHX_ packWARN(WARN_DIGIT),
431 "Illegal octal digit '%c' ignored", *s);
432 }
433 break;
434 }
435
436 if ( ( overflowed && value_nv > 4294967295.0)
437#if UVSIZE > 4
438 || (!overflowed && value > 0xffffffff )
439#endif
440 ) {
441 if (ckWARN(WARN_PORTABLE))
442 Perl_warner(aTHX_ packWARN(WARN_PORTABLE),
443 "Octal number > 037777777777 non-portable");
444 }
445 *len_p = s - start;
446 if (!overflowed) {
447 *flags = 0;
448 return value;
449 }
450 *flags = PERL_SCAN_GREATER_THAN_UV_MAX;
451 if (result)
452 *result = value_nv;
453 return UV_MAX;
454}
455
456/*
457=for apidoc scan_bin
458
459For backwards compatibility. Use C<grok_bin> instead.
460
461=for apidoc scan_hex
462
463For backwards compatibility. Use C<grok_hex> instead.
464
465=for apidoc scan_oct
466
467For backwards compatibility. Use C<grok_oct> instead.
468
469=cut
470 */
471
472NV
473Perl_scan_bin(pTHX_ char *start, STRLEN len, STRLEN *retlen)
474{
475 NV rnv;
476 I32 flags = *retlen ? PERL_SCAN_ALLOW_UNDERSCORES : 0;
477 const UV ruv = grok_bin (start, &len, &flags, &rnv);
478
479 *retlen = len;
480 return (flags & PERL_SCAN_GREATER_THAN_UV_MAX) ? rnv : (NV)ruv;
481}
482
483NV
484Perl_scan_oct(pTHX_ char *start, STRLEN len, STRLEN *retlen)
485{
486 NV rnv;
487 I32 flags = *retlen ? PERL_SCAN_ALLOW_UNDERSCORES : 0;
488 const UV ruv = grok_oct (start, &len, &flags, &rnv);
489
490 *retlen = len;
491 return (flags & PERL_SCAN_GREATER_THAN_UV_MAX) ? rnv : (NV)ruv;
492}
493
494NV
495Perl_scan_hex(pTHX_ char *start, STRLEN len, STRLEN *retlen)
496{
497 NV rnv;
498 I32 flags = *retlen ? PERL_SCAN_ALLOW_UNDERSCORES : 0;
499 const UV ruv = grok_hex (start, &len, &flags, &rnv);
500
501 *retlen = len;
502 return (flags & PERL_SCAN_GREATER_THAN_UV_MAX) ? rnv : (NV)ruv;
503}
504
505/*
506=for apidoc grok_numeric_radix
507
508Scan and skip for a numeric decimal separator (radix).
509
510=cut
511 */
512bool
513Perl_grok_numeric_radix(pTHX_ const char **sp, const char *send)
514{
515#ifdef USE_LOCALE_NUMERIC
516 if (PL_numeric_radix_sv && IN_LOCALE) {
517 STRLEN len;
518 const char* radix = SvPV(PL_numeric_radix_sv, len);
519 if (*sp + len <= send && memEQ(*sp, radix, len)) {
520 *sp += len;
521 return TRUE;
522 }
523 }
524 /* always try "." if numeric radix didn't match because
525 * we may have data from different locales mixed */
526#endif
527 if (*sp < send && **sp == '.') {
528 ++*sp;
529 return TRUE;
530 }
531 return FALSE;
532}
533
534/*
535=for apidoc grok_number
536
537Recognise (or not) a number. The type of the number is returned
538(0 if unrecognised), otherwise it is a bit-ORed combination of
539IS_NUMBER_IN_UV, IS_NUMBER_GREATER_THAN_UV_MAX, IS_NUMBER_NOT_INT,
540IS_NUMBER_NEG, IS_NUMBER_INFINITY, IS_NUMBER_NAN (defined in perl.h).
541
542If the value of the number can fit an in UV, it is returned in the *valuep
543IS_NUMBER_IN_UV will be set to indicate that *valuep is valid, IS_NUMBER_IN_UV
544will never be set unless *valuep is valid, but *valuep may have been assigned
545to during processing even though IS_NUMBER_IN_UV is not set on return.
546If valuep is NULL, IS_NUMBER_IN_UV will be set for the same cases as when
547valuep is non-NULL, but no actual assignment (or SEGV) will occur.
548
549IS_NUMBER_NOT_INT will be set with IS_NUMBER_IN_UV if trailing decimals were
550seen (in which case *valuep gives the true value truncated to an integer), and
551IS_NUMBER_NEG if the number is negative (in which case *valuep holds the
552absolute value). IS_NUMBER_IN_UV is not set if e notation was used or the
553number is larger than a UV.
554
555=cut
556 */
557int
558Perl_grok_number(pTHX_ const char *pv, STRLEN len, UV *valuep)
559{
560 const char *s = pv;
561 const char *send = pv + len;
562 const UV max_div_10 = UV_MAX / 10;
563 const char max_mod_10 = UV_MAX % 10;
564 int numtype = 0;
565 int sawinf = 0;
566 int sawnan = 0;
567
568 while (s < send && isSPACE(*s))
569 s++;
570 if (s == send) {
571 return 0;
572 } else if (*s == '-') {
573 s++;
574 numtype = IS_NUMBER_NEG;
575 }
576 else if (*s == '+')
577 s++;
578
579 if (s == send)
580 return 0;
581
582 /* next must be digit or the radix separator or beginning of infinity */
583 if (isDIGIT(*s)) {
584 /* UVs are at least 32 bits, so the first 9 decimal digits cannot
585 overflow. */
586 UV value = *s - '0';
587 /* This construction seems to be more optimiser friendly.
588 (without it gcc does the isDIGIT test and the *s - '0' separately)
589 With it gcc on arm is managing 6 instructions (6 cycles) per digit.
590 In theory the optimiser could deduce how far to unroll the loop
591 before checking for overflow. */
592 if (++s < send) {
593 int digit = *s - '0';
594 if (digit >= 0 && digit <= 9) {
595 value = value * 10 + digit;
596 if (++s < send) {
597 digit = *s - '0';
598 if (digit >= 0 && digit <= 9) {
599 value = value * 10 + digit;
600 if (++s < send) {
601 digit = *s - '0';
602 if (digit >= 0 && digit <= 9) {
603 value = value * 10 + digit;
604 if (++s < send) {
605 digit = *s - '0';
606 if (digit >= 0 && digit <= 9) {
607 value = value * 10 + digit;
608 if (++s < send) {
609 digit = *s - '0';
610 if (digit >= 0 && digit <= 9) {
611 value = value * 10 + digit;
612 if (++s < send) {
613 digit = *s - '0';
614 if (digit >= 0 && digit <= 9) {
615 value = value * 10 + digit;
616 if (++s < send) {
617 digit = *s - '0';
618 if (digit >= 0 && digit <= 9) {
619 value = value * 10 + digit;
620 if (++s < send) {
621 digit = *s - '0';
622 if (digit >= 0 && digit <= 9) {
623 value = value * 10 + digit;
624 if (++s < send) {
625 /* Now got 9 digits, so need to check
626 each time for overflow. */
627 digit = *s - '0';
628 while (digit >= 0 && digit <= 9
629 && (value < max_div_10
630 || (value == max_div_10
631 && digit <= max_mod_10))) {
632 value = value * 10 + digit;
633 if (++s < send)
634 digit = *s - '0';
635 else
636 break;
637 }
638 if (digit >= 0 && digit <= 9
639 && (s < send)) {
640 /* value overflowed.
641 skip the remaining digits, don't
642 worry about setting *valuep. */
643 do {
644 s++;
645 } while (s < send && isDIGIT(*s));
646 numtype |=
647 IS_NUMBER_GREATER_THAN_UV_MAX;
648 goto skip_value;
649 }
650 }
651 }
652 }
653 }
654 }
655 }
656 }
657 }
658 }
659 }
660 }
661 }
662 }
663 }
664 }
665 }
666 }
667 numtype |= IS_NUMBER_IN_UV;
668 if (valuep)
669 *valuep = value;
670
671 skip_value:
672 if (GROK_NUMERIC_RADIX(&s, send)) {
673 numtype |= IS_NUMBER_NOT_INT;
674 while (s < send && isDIGIT(*s)) /* optional digits after the radix */
675 s++;
676 }
677 }
678 else if (GROK_NUMERIC_RADIX(&s, send)) {
679 numtype |= IS_NUMBER_NOT_INT | IS_NUMBER_IN_UV; /* valuep assigned below */
680 /* no digits before the radix means we need digits after it */
681 if (s < send && isDIGIT(*s)) {
682 do {
683 s++;
684 } while (s < send && isDIGIT(*s));
685 if (valuep) {
686 /* integer approximation is valid - it's 0. */
687 *valuep = 0;
688 }
689 }
690 else
691 return 0;
692 } else if (*s == 'I' || *s == 'i') {
693 s++; if (s == send || (*s != 'N' && *s != 'n')) return 0;
694 s++; if (s == send || (*s != 'F' && *s != 'f')) return 0;
695 s++; if (s < send && (*s == 'I' || *s == 'i')) {
696 s++; if (s == send || (*s != 'N' && *s != 'n')) return 0;
697 s++; if (s == send || (*s != 'I' && *s != 'i')) return 0;
698 s++; if (s == send || (*s != 'T' && *s != 't')) return 0;
699 s++; if (s == send || (*s != 'Y' && *s != 'y')) return 0;
700 s++;
701 }
702 sawinf = 1;
703 } else if (*s == 'N' || *s == 'n') {
704 /* XXX TODO: There are signaling NaNs and quiet NaNs. */
705 s++; if (s == send || (*s != 'A' && *s != 'a')) return 0;
706 s++; if (s == send || (*s != 'N' && *s != 'n')) return 0;
707 s++;
708 sawnan = 1;
709 } else
710 return 0;
711
712 if (sawinf) {
713 numtype &= IS_NUMBER_NEG; /* Keep track of sign */
714 numtype |= IS_NUMBER_INFINITY | IS_NUMBER_NOT_INT;
715 } else if (sawnan) {
716 numtype &= IS_NUMBER_NEG; /* Keep track of sign */
717 numtype |= IS_NUMBER_NAN | IS_NUMBER_NOT_INT;
718 } else if (s < send) {
719 /* we can have an optional exponent part */
720 if (*s == 'e' || *s == 'E') {
721 /* The only flag we keep is sign. Blow away any "it's UV" */
722 numtype &= IS_NUMBER_NEG;
723 numtype |= IS_NUMBER_NOT_INT;
724 s++;
725 if (s < send && (*s == '-' || *s == '+'))
726 s++;
727 if (s < send && isDIGIT(*s)) {
728 do {
729 s++;
730 } while (s < send && isDIGIT(*s));
731 }
732 else
733 return 0;
734 }
735 }
736 while (s < send && isSPACE(*s))
737 s++;
738 if (s >= send)
739 return numtype;
740 if (len == 10 && memEQ(pv, "0 but true", 10)) {
741 if (valuep)
742 *valuep = 0;
743 return IS_NUMBER_IN_UV;
744 }
745 return 0;
746}
747
748STATIC NV
749S_mulexp10(NV value, I32 exponent)
750{
751 NV result = 1.0;
752 NV power = 10.0;
753 bool negative = 0;
754 I32 bit;
755
756 if (exponent == 0)
757 return value;
758 if (value == 0)
759 return (NV)0;
760
761 /* On OpenVMS VAX we by default use the D_FLOAT double format,
762 * and that format does not have *easy* capabilities [1] for
763 * overflowing doubles 'silently' as IEEE fp does. We also need
764 * to support G_FLOAT on both VAX and Alpha, and though the exponent
765 * range is much larger than D_FLOAT it still doesn't do silent
766 * overflow. Therefore we need to detect early whether we would
767 * overflow (this is the behaviour of the native string-to-float
768 * conversion routines, and therefore of native applications, too).
769 *
770 * [1] Trying to establish a condition handler to trap floating point
771 * exceptions is not a good idea. */
772
773 /* In UNICOS and in certain Cray models (such as T90) there is no
774 * IEEE fp, and no way at all from C to catch fp overflows gracefully.
775 * There is something you can do if you are willing to use some
776 * inline assembler: the instruction is called DFI-- but that will
777 * disable *all* floating point interrupts, a little bit too large
778 * a hammer. Therefore we need to catch potential overflows before
779 * it's too late. */
780
781#if ((defined(VMS) && !defined(__IEEE_FP)) || defined(_UNICOS)) && defined(NV_MAX_10_EXP)
782 STMT_START {
783 NV exp_v = log10(value);
784 if (exponent >= NV_MAX_10_EXP || exponent + exp_v >= NV_MAX_10_EXP)
785 return NV_MAX;
786 if (exponent < 0) {
787 if (-(exponent + exp_v) >= NV_MAX_10_EXP)
788 return 0.0;
789 while (-exponent >= NV_MAX_10_EXP) {
790 /* combination does not overflow, but 10^(-exponent) does */
791 value /= 10;
792 ++exponent;
793 }
794 }
795 } STMT_END;
796#endif
797
798 if (exponent < 0) {
799 negative = 1;
800 exponent = -exponent;
801 }
802 for (bit = 1; exponent; bit <<= 1) {
803 if (exponent & bit) {
804 exponent ^= bit;
805 result *= power;
806 /* Floating point exceptions are supposed to be turned off,
807 * but if we're obviously done, don't risk another iteration.
808 */
809 if (exponent == 0) break;
810 }
811 power *= power;
812 }
813 return negative ? value / result : value * result;
814}
815
816NV
817Perl_my_atof(pTHX_ const char* s)
818{
819 NV x = 0.0;
820#ifdef USE_LOCALE_NUMERIC
821 if (PL_numeric_local && IN_LOCALE) {
822 NV y;
823
824 /* Scan the number twice; once using locale and once without;
825 * choose the larger result (in absolute value). */
826 Perl_atof2(s, x);
827 SET_NUMERIC_STANDARD();
828 Perl_atof2(s, y);
829 SET_NUMERIC_LOCAL();
830 if ((y < 0.0 && y < x) || (y > 0.0 && y > x))
831 return y;
832 }
833 else
834 Perl_atof2(s, x);
835#else
836 Perl_atof2(s, x);
837#endif
838 return x;
839}
840
841char*
842Perl_my_atof2(pTHX_ const char* orig, NV* value)
843{
844 NV result[3] = {0.0, 0.0, 0.0};
845 const char* s = orig;
846#ifdef USE_PERL_ATOF
847 UV accumulator[2] = {0,0}; /* before/after dp */
848 bool negative = 0;
849 const char* send = s + strlen(orig) - 1;
850 bool seen_digit = 0;
851 I32 exp_adjust[2] = {0,0};
852 I32 exp_acc[2] = {-1, -1};
853 /* the current exponent adjust for the accumulators */
854 I32 exponent = 0;
855 I32 seen_dp = 0;
856 I32 digit = 0;
857 I32 old_digit = 0;
858 I32 sig_digits = 0; /* noof significant digits seen so far */
859
860/* There is no point in processing more significant digits
861 * than the NV can hold. Note that NV_DIG is a lower-bound value,
862 * while we need an upper-bound value. We add 2 to account for this;
863 * since it will have been conservative on both the first and last digit.
864 * For example a 32-bit mantissa with an exponent of 4 would have
865 * exact values in the set
866 * 4
867 * 8
868 * ..
869 * 17179869172
870 * 17179869176
871 * 17179869180
872 *
873 * where for the purposes of calculating NV_DIG we would have to discount
874 * both the first and last digit, since neither can hold all values from
875 * 0..9; but for calculating the value we must examine those two digits.
876 */
877#define MAX_SIG_DIGITS (NV_DIG+2)
878
879/* the max number we can accumulate in a UV, and still safely do 10*N+9 */
880#define MAX_ACCUMULATE ( (UV) ((UV_MAX - 9)/10))
881
882 /* leading whitespace */
883 while (isSPACE(*s))
884 ++s;
885
886 /* sign */
887 switch (*s) {
888 case '-':
889 negative = 1;
890 /* fall through */
891 case '+':
892 ++s;
893 }
894
895 /* punt to strtod for NaN/Inf; if no support for it there, tough luck */
896
897#ifdef HAS_STRTOD
898 if (*s == 'n' || *s == 'N' || *s == 'i' || *s == 'I') {
899 const char *p = negative ? s - 1 : s;
900 char *endp;
901 NV rslt;
902 rslt = strtod(p, &endp);
903 if (endp != p) {
904 *value = rslt;
905 return (char *)endp;
906 }
907 }
908#endif
909
910 /* we accumulate digits into an integer; when this becomes too
911 * large, we add the total to NV and start again */
912
913 while (1) {
914 if (isDIGIT(*s)) {
915 seen_digit = 1;
916 old_digit = digit;
917 digit = *s++ - '0';
918 if (seen_dp)
919 exp_adjust[1]++;
920
921 /* don't start counting until we see the first significant
922 * digit, eg the 5 in 0.00005... */
923 if (!sig_digits && digit == 0)
924 continue;
925
926 if (++sig_digits > MAX_SIG_DIGITS) {
927 /* limits of precision reached */
928 if (digit > 5) {
929 ++accumulator[seen_dp];
930 } else if (digit == 5) {
931 if (old_digit % 2) { /* round to even - Allen */
932 ++accumulator[seen_dp];
933 }
934 }
935 if (seen_dp) {
936 exp_adjust[1]--;
937 } else {
938 exp_adjust[0]++;
939 }
940 /* skip remaining digits */
941 while (isDIGIT(*s)) {
942 ++s;
943 if (! seen_dp) {
944 exp_adjust[0]++;
945 }
946 }
947 /* warn of loss of precision? */
948 }
949 else {
950 if (accumulator[seen_dp] > MAX_ACCUMULATE) {
951 /* add accumulator to result and start again */
952 result[seen_dp] = S_mulexp10(result[seen_dp],
953 exp_acc[seen_dp])
954 + (NV)accumulator[seen_dp];
955 accumulator[seen_dp] = 0;
956 exp_acc[seen_dp] = 0;
957 }
958 accumulator[seen_dp] = accumulator[seen_dp] * 10 + digit;
959 ++exp_acc[seen_dp];
960 }
961 }
962 else if (!seen_dp && GROK_NUMERIC_RADIX(&s, send)) {
963 seen_dp = 1;
964 if (sig_digits > MAX_SIG_DIGITS) {
965 ++s;
966 while (isDIGIT(*s)) {
967 ++s;
968 }
969 break;
970 }
971 }
972 else {
973 break;
974 }
975 }
976
977 result[0] = S_mulexp10(result[0], exp_acc[0]) + (NV)accumulator[0];
978 if (seen_dp) {
979 result[1] = S_mulexp10(result[1], exp_acc[1]) + (NV)accumulator[1];
980 }
981
982 if (seen_digit && (*s == 'e' || *s == 'E')) {
983 bool expnegative = 0;
984
985 ++s;
986 switch (*s) {
987 case '-':
988 expnegative = 1;
989 /* fall through */
990 case '+':
991 ++s;
992 }
993 while (isDIGIT(*s))
994 exponent = exponent * 10 + (*s++ - '0');
995 if (expnegative)
996 exponent = -exponent;
997 }
998
999
1000
1001 /* now apply the exponent */
1002
1003 if (seen_dp) {
1004 result[2] = S_mulexp10(result[0],exponent+exp_adjust[0])
1005 + S_mulexp10(result[1],exponent-exp_adjust[1]);
1006 } else {
1007 result[2] = S_mulexp10(result[0],exponent+exp_adjust[0]);
1008 }
1009
1010 /* now apply the sign */
1011 if (negative)
1012 result[2] = -result[2];
1013#endif /* USE_PERL_ATOF */
1014 *value = result[2];
1015 return (char *)s;
1016}
1017
1018#if ! defined(HAS_MODFL) && defined(HAS_AINTL) && defined(HAS_COPYSIGNL)
1019long double
1020Perl_my_modfl(long double x, long double *ip)
1021{
1022 *ip = aintl(x);
1023 return (x == *ip ? copysignl(0.0L, x) : x - *ip);
1024}
1025#endif
1026
1027#if ! defined(HAS_FREXPL) && defined(HAS_ILOGBL) && defined(HAS_SCALBNL)
1028long double
1029Perl_my_frexpl(long double x, int *e) {
1030 *e = x == 0.0L ? 0 : ilogbl(x) + 1;
1031 return (scalbnl(x, -*e));
1032}
1033#endif
1034
1035/*
1036 * Local variables:
1037 * c-indentation-style: bsd
1038 * c-basic-offset: 4
1039 * indent-tabs-mode: t
1040 * End:
1041 *
1042 * ex: set ts=8 sts=4 sw=4 noet:
1043 */
Note: See TracBrowser for help on using the repository browser.