gecko-dev/js/ref/jsnum.c
1998-06-11 20:27:22 +00:00

828 lines
20 KiB
C

/* -*- Mode: C; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 4 -*-
*
* The contents of this file are subject to the Netscape Public License
* Version 1.0 (the "NPL"); you may not use this file except in
* compliance with the NPL. You may obtain a copy of the NPL at
* http://www.mozilla.org/NPL/
*
* Software distributed under the NPL is distributed on an "AS IS" basis,
* WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
* for the specific language governing rights and limitations under the
* NPL.
*
* The Initial Developer of this code under the NPL is Netscape
* Communications Corporation. Portions created by Netscape are
* Copyright (C) 1998 Netscape Communications Corporation. All Rights
* Reserved.
*/
/*
* JS number type and wrapper class.
*/
#include "jsstddef.h"
#include <errno.h>
#ifdef XP_PC
#include <float.h>
#endif
#include <math.h>
#include <stdlib.h>
#include <string.h>
#include "prtypes.h"
#include "prassert.h"
#include "prdtoa.h"
#include "prprintf.h"
#include "jsapi.h"
#include "jsatom.h"
#include "jscntxt.h"
#include "jsconfig.h"
#include "jsgc.h"
#include "jsinterp.h"
#include "jsnum.h"
#include "jsobj.h"
#include "jsopcode.h"
#include "jsstr.h"
union dpun {
struct {
#ifdef IS_LITTLE_ENDIAN
uint32 lo, hi;
#else
uint32 hi, lo;
#endif
} s;
jsdouble d;
};
static JSBool
num_isNaN(JSContext *cx, JSObject *obj, uintN argc, jsval *argv, jsval *rval)
{
jsdouble x;
if (!js_ValueToNumber(cx, argv[0], &x))
return JS_FALSE;
*rval = BOOLEAN_TO_JSVAL(JSDOUBLE_IS_NaN(x));
return JS_TRUE;
}
static JSBool
num_isFinite(JSContext *cx, JSObject *obj, uintN argc, jsval *argv, jsval *rval)
{
jsdouble x;
if (!js_ValueToNumber(cx, argv[0], &x))
return JS_FALSE;
*rval = BOOLEAN_TO_JSVAL(JSDOUBLE_IS_FINITE(x));
return JS_TRUE;
}
static JSBool
num_parseFloat(JSContext *cx, JSObject *obj, uintN argc, jsval *argv, jsval *rval)
{
JSString *str;
jsdouble d;
const jschar *ep;
str = js_ValueToString(cx, argv[0]);
if (!str)
return JS_FALSE;
if (!js_strtod(cx, str->chars, &ep, &d))
return JS_FALSE;
if (ep == str->chars) {
*rval = DOUBLE_TO_JSVAL(cx->runtime->jsNaN);
return JS_TRUE;
}
return js_NewNumberValue(cx, d, rval);
}
/* See ECMA 15.1.2.2. */
static JSBool
num_parseInt(JSContext *cx, JSObject *obj, uintN argc, jsval *argv, jsval *rval)
{
JSString *str;
jsint radix;
jsdouble d;
const jschar *ep;
str = js_ValueToString(cx, argv[0]);
if (!str)
return JS_FALSE;
if (argc > 1) {
if (!js_ValueToECMAInt32(cx, argv[1], &radix))
return JS_FALSE;
} else
radix = 0;
if (radix != 0 && (radix < 2 || radix > 36)) {
*rval = DOUBLE_TO_JSVAL(cx->runtime->jsNaN);
return JS_TRUE;
}
if (!js_strtointeger(cx, str->chars, &ep, radix, &d))
return JS_FALSE;
if (ep == str->chars) {
*rval = DOUBLE_TO_JSVAL(cx->runtime->jsNaN);
return JS_TRUE;
}
return js_NewNumberValue(cx, d, rval);
}
static JSFunctionSpec number_functions[] = {
{"isNaN", num_isNaN, 1},
{"isFinite", num_isFinite, 1},
{"parseFloat", num_parseFloat, 1},
{"parseInt", num_parseInt, 2},
{0}
};
static JSClass number_class = {
"Number",
JSCLASS_HAS_PRIVATE,
JS_PropertyStub, JS_PropertyStub, JS_PropertyStub, JS_PropertyStub,
JS_EnumerateStub, JS_ResolveStub, JS_ConvertStub, JS_FinalizeStub
};
static JSBool
Number(JSContext *cx, JSObject *obj, uintN argc, jsval *argv, jsval *rval)
{
jsdouble d;
jsval v;
if (argc != 0) {
if (!js_ValueToNumber(cx, argv[0], &d))
return JS_FALSE;
} else {
d = 0.0;
}
if (!js_NewNumberValue(cx, d, &v))
return JS_FALSE;
if (!cx->fp->constructing) {
*rval = v;
return JS_TRUE;
}
OBJ_SET_SLOT(cx, obj, JSSLOT_PRIVATE, v);
return JS_TRUE;
}
#if JS_HAS_TOSOURCE
static JSBool
num_toSource(JSContext *cx, JSObject *obj, uintN argc, jsval *argv, jsval *rval)
{
jsval v;
jsdouble d;
size_t i;
char buf[64];
JSString *str;
if (!JS_InstanceOf(cx, obj, &number_class, argv))
return JS_FALSE;
v = OBJ_GET_SLOT(cx, obj, JSSLOT_PRIVATE);
if (!JSVAL_IS_NUMBER(v))
return js_obj_toSource(cx, obj, argc, argv, rval);
d = JSVAL_IS_INT(v) ? (jsdouble)JSVAL_TO_INT(v) : *JSVAL_TO_DOUBLE(v);
i = PR_snprintf(buf, sizeof buf, "(new %s(", number_class.name);
PR_cnvtf(buf + i, sizeof buf - i, 20, d);
i = strlen(buf);
PR_snprintf(buf + i, sizeof buf - i, "))");
str = JS_NewStringCopyZ(cx, buf);
if (!str)
return JS_FALSE;
*rval = STRING_TO_JSVAL(str);
return JS_TRUE;
}
#endif
static JSBool
num_toString(JSContext *cx, JSObject *obj, uintN argc, jsval *argv, jsval *rval)
{
jsval v;
jsdouble d;
jsint base, ival, dval;
char *bp, buf[32];
JSString *str;
if (!JS_InstanceOf(cx, obj, &number_class, argv))
return JS_FALSE;
v = OBJ_GET_SLOT(cx, obj, JSSLOT_PRIVATE);
if (!JSVAL_IS_NUMBER(v))
return js_obj_toString(cx, obj, argc, argv, rval);
d = JSVAL_IS_INT(v) ? (jsdouble)JSVAL_TO_INT(v) : *JSVAL_TO_DOUBLE(v);
if (argc != 0) {
if (!js_ValueToECMAInt32(cx, argv[0], &base))
return JS_FALSE;
if (base < 2 || base > 36) {
JS_ReportError(cx, "illegal radix %d", base);
return JS_FALSE;
}
if (base != 10 && JSDOUBLE_IS_FINITE(d)) {
ival = (jsint) js_DoubleToInteger(d);
bp = buf + sizeof buf;
for (*--bp = '\0'; ival != 0 && --bp >= buf; ival /= base) {
dval = ival % base;
*bp = (char)((dval >= 10) ? 'a' - 10 + dval : '0' + dval);
}
if (*bp == '\0')
*--bp = '0';
str = JS_NewStringCopyZ(cx, bp);
} else {
str = js_NumberToString(cx, d);
}
} else {
str = js_NumberToString(cx, d);
}
if (!str)
return JS_FALSE;
*rval = STRING_TO_JSVAL(str);
return JS_TRUE;
}
static JSBool
num_valueOf(JSContext *cx, JSObject *obj, uintN argc, jsval *argv, jsval *rval)
{
if (!JS_InstanceOf(cx, obj, &number_class, argv))
return JS_FALSE;
*rval = OBJ_GET_SLOT(cx, obj, JSSLOT_PRIVATE);
return JS_TRUE;
}
static JSFunctionSpec number_methods[] = {
#if JS_HAS_TOSOURCE
{js_toSource_str, num_toSource, 0},
#endif
{js_toString_str, num_toString, 0},
{js_valueOf_str, num_valueOf, 0},
{0}
};
/* NB: Keep this in synch with number_constants[]. */
enum nc_slot {
NC_NaN,
NC_POSITIVE_INFINITY,
NC_NEGATIVE_INFINITY,
NC_MAX_VALUE,
NC_MIN_VALUE,
NC_LIMIT
};
/*
* Some to most C compilers forbid spelling these at compile time, or barf
* if you try, so all but MAX_VALUE are set at runtime by js_InitNumberClass
* using union dpun.
*/
static JSConstDoubleSpec number_constants[] = {
{0, "NaN"},
{0, "POSITIVE_INFINITY"},
{0, "NEGATIVE_INFINITY"},
{1.7976931348623157E+308, "MAX_VALUE"},
{0, "MIN_VALUE"},
{0}
};
static jsdouble NaN;
JSObject *
js_InitNumberClass(JSContext *cx, JSObject *obj)
{
JSRuntime *rt;
union dpun u;
JSObject *proto, *ctor;
rt = cx->runtime;
if (!rt->jsNaN) {
#ifdef XP_PC
#ifdef XP_OS2
/*DSR071597 - I have no idea what this really does other than mucking with the floating */
/*point unit, but it does fix a "floating point underflow" exception I am getting, and there*/
/*is similar code in the Hursley java. Making sure we have the same code in Javascript */
/*where Netscape was calling control87 on Windows... */
_control87(MCW_EM+PC_53+RC_NEAR,MCW_EM+MCW_PC+MCW_RC);
#else
_control87(MCW_EM, MCW_EM);
#endif
#endif
u.s.hi = JSDOUBLE_HI32_EXPMASK | JSDOUBLE_HI32_MANTMASK;
u.s.lo = 0xffffffff;
number_constants[NC_NaN].dval = NaN = u.d;
rt->jsNaN = js_NewDouble(cx, NaN);
if (!rt->jsNaN || !js_LockGCThing(cx, rt->jsNaN))
return NULL;
u.s.hi = JSDOUBLE_HI32_EXPMASK;
u.s.lo = 0x00000000;
number_constants[NC_POSITIVE_INFINITY].dval = u.d;
rt->jsPositiveInfinity = js_NewDouble(cx, u.d);
if (!rt->jsPositiveInfinity ||
!js_LockGCThing(cx, rt->jsPositiveInfinity)) {
return NULL;
}
u.s.hi = JSDOUBLE_HI32_SIGNBIT | JSDOUBLE_HI32_EXPMASK;
u.s.lo = 0x00000000;
number_constants[NC_NEGATIVE_INFINITY].dval = u.d;
rt->jsNegativeInfinity = js_NewDouble(cx, u.d);
if (!rt->jsNegativeInfinity ||
!js_LockGCThing(cx, rt->jsNegativeInfinity)) {
return NULL;
}
u.s.hi = 0;
u.s.lo = 1;
number_constants[NC_MIN_VALUE].dval = u.d;
}
if (!JS_DefineFunctions(cx, obj, number_functions))
return NULL;
proto = JS_InitClass(cx, obj, NULL, &number_class, Number, 1,
NULL, number_methods, NULL, NULL);
if (!proto || !(ctor = JS_GetConstructor(cx, proto)))
return NULL;
OBJ_SET_SLOT(cx, proto, JSSLOT_PRIVATE, JSVAL_ZERO);
if (!JS_DefineConstDoubles(cx, ctor, number_constants))
return NULL;
/* ECMA 15.1.1.1 */
if (!JS_DefineProperty(cx, obj, "NaN", DOUBLE_TO_JSVAL(rt->jsNaN),
NULL, NULL, 0)) {
return NULL;
}
/* ECMA 15.1.1.2 */
if (!JS_DefineProperty(cx, obj, "Infinity",
DOUBLE_TO_JSVAL(rt->jsPositiveInfinity),
NULL, NULL, 0)) {
return NULL;
}
return proto;
}
jsdouble *
js_NewDouble(JSContext *cx, jsdouble d)
{
jsdouble *dp;
dp = js_AllocGCThing(cx, GCX_DOUBLE);
if (!dp)
return NULL;
*dp = d;
return dp;
}
void
js_FinalizeDouble(JSContext *cx, jsdouble *dp)
{
*dp = NaN;
}
JSBool
js_NewDoubleValue(JSContext *cx, jsdouble d, jsval *rval)
{
jsdouble *dp;
dp = js_NewDouble(cx, d);
if (!dp)
return JS_FALSE;
*rval = DOUBLE_TO_JSVAL(dp);
return JS_TRUE;
}
JSBool
js_NewNumberValue(JSContext *cx, jsdouble d, jsval *rval)
{
jsint i;
i = (jsint)d;
if (JSDOUBLE_IS_INT(d, i) && INT_FITS_IN_JSVAL(i)) {
*rval = INT_TO_JSVAL(i);
} else {
if (!js_NewDoubleValue(cx, d, rval))
return JS_FALSE;
}
return JS_TRUE;
}
JSObject *
js_NumberToObject(JSContext *cx, jsdouble d)
{
JSObject *obj;
jsval v;
obj = js_NewObject(cx, &number_class, NULL, NULL);
if (!obj)
return NULL;
if (!js_NewNumberValue(cx, d, &v)) {
cx->newborn[GCX_OBJECT] = NULL;
return NULL;
}
OBJ_SET_SLOT(cx, obj, JSSLOT_PRIVATE, v);
return obj;
}
/* XXXbe rewrite me to be ECMA-based! */
JSString *
js_NumberToString(JSContext *cx, jsdouble d)
{
jsint i;
char buf[32];
i = (jsint)d;
if (JSDOUBLE_IS_INT(d, i)) {
PR_snprintf(buf, sizeof buf, "%ld", (long)i);
} else {
PR_cnvtf(buf, sizeof buf, 20, d);
}
return JS_NewStringCopyZ(cx, buf);
}
JSBool
js_ValueToNumber(JSContext *cx, jsval v, jsdouble *dp)
{
JSObject *obj;
JSString *str;
const jschar *ep;
jsdouble d;
if (JSVAL_IS_OBJECT(v)) {
obj = JSVAL_TO_OBJECT(v);
if (!obj) {
*dp = 0;
return JS_TRUE;
}
if (!OBJ_DEFAULT_VALUE(cx, obj, JSTYPE_NUMBER, &v))
return JS_FALSE;
}
if (JSVAL_IS_INT(v)) {
*dp = (jsdouble)JSVAL_TO_INT(v);
} else if (JSVAL_IS_DOUBLE(v)) {
*dp = *JSVAL_TO_DOUBLE(v);
} else if (JSVAL_IS_STRING(v)) {
str = JSVAL_TO_STRING(v);
errno = 0;
/* Note that ECMAScript doesn't treat numbers beginning with a zero as octal numbers here.
* This works because all such numbers will be interpreted as decimal by js_strtod and
* will never get passed to js_strtointeger, which would interpret them as octal. */
if ((!js_strtod(cx, str->chars, &ep, &d) || js_SkipWhiteSpace(ep) != str->chars + str->length) &&
(!js_strtointeger(cx, str->chars, &ep, 0, &d) || js_SkipWhiteSpace(ep) != str->chars + str->length)) {
goto badstr;
}
*dp = d;
} else if (JSVAL_IS_BOOLEAN(v)) {
*dp = JSVAL_TO_BOOLEAN(v) ? 1 : 0;
} else {
#if JS_BUG_FALLIBLE_TONUM
str = js_DecompileValueGenerator(cx, v, NULL);
badstr:
if (str) {
JS_ReportError(cx, "%s is not a number",
JS_GetStringBytes(str));
}
return JS_FALSE;
#else
badstr:
*dp = *cx->runtime->jsNaN;
#endif
}
return JS_TRUE;
}
JSBool
js_ValueToECMAInt32(JSContext *cx, jsval v, int32 *ip)
{
jsdouble d;
if (!js_ValueToNumber(cx, v, &d))
return JS_FALSE;
return js_DoubleToECMAInt32(cx, d, ip);
}
JSBool
js_DoubleToECMAInt32(JSContext *cx, jsdouble d, int32 *ip)
{
jsdouble two32 = 4294967296.0;
jsdouble two31 = 2147483648.0;
if (!JSDOUBLE_IS_FINITE(d) || d == 0) {
*ip = 0;
return JS_TRUE;
}
d = fmod(d, two32);
d = d >= 0 ? d : d + two32;
if (d >= two31)
*ip = (int32)(d - two32);
else
*ip = (int32)d;
return JS_TRUE;
}
JSBool
js_ValueToECMAUint32(JSContext *cx, jsval v, uint32 *ip)
{
jsdouble d;
if (!js_ValueToNumber(cx, v, &d))
return JS_FALSE;
return js_DoubleToECMAUint32(cx, d, ip);
}
JSBool
js_DoubleToECMAUint32(JSContext *cx, jsdouble d, uint32 *ip)
{
JSBool neg;
jsdouble two32 = 4294967296.0;
if (!JSDOUBLE_IS_FINITE(d) || d == 0) {
*ip = 0;
return JS_TRUE;
}
neg = (d < 0);
d = floor(neg ? -d : d);
d = neg ? -d : d;
d = fmod(d, two32);
d = d >= 0 ? d : d + two32;
*ip = (uint32)d;
return JS_TRUE;
}
JSBool
js_ValueToInt32(JSContext *cx, jsval v, int32 *ip)
{
jsdouble d;
JSString *str;
if (!js_ValueToNumber(cx, v, &d))
return JS_FALSE;
if (JSDOUBLE_IS_NaN(d) || d <= -2147483649.0 || 2147483648.0 <= d) {
str = js_DecompileValueGenerator(cx, v, NULL);
if (str) {
JS_ReportError(cx, "can't convert %s to an integer",
JS_GetStringBytes(str));
}
return JS_FALSE;
}
*ip = (int32)floor(d + 0.5); /* Round to nearest */
return JS_TRUE;
}
JSBool
js_ValueToUint16(JSContext *cx, jsval v, uint16 *ip)
{
jsdouble d;
jsuint i, m;
JSBool neg;
if (!js_ValueToNumber(cx, v, &d))
return JS_FALSE;
if (d == 0 || !JSDOUBLE_IS_FINITE(d)) {
*ip = 0;
return JS_TRUE;
}
i = (jsuint)d;
if ((jsdouble)i == d) {
*ip = (uint16)i;
return JS_TRUE;
}
neg = (d < 0);
d = floor(neg ? -d : d);
d = neg ? -d : d;
m = PR_BIT(16);
d = fmod(d, m);
if (d < 0)
d += m;
*ip = (uint16) d;
return JS_TRUE;
}
jsdouble
js_DoubleToInteger(jsdouble d)
{
JSBool neg;
if (d == 0)
return d;
if (!JSDOUBLE_IS_FINITE(d)) {
if (JSDOUBLE_IS_NaN(d))
return 0;
return d;
}
neg = (d < 0);
d = floor(neg ? -d : d);
return neg ? -d : d;
}
JSBool
js_strtod(JSContext *cx, const jschar *s, const jschar **ep, jsdouble *dp)
{
size_t i;
char *cstr, *istr, *estr;
JSBool negative;
jsdouble d;
const jschar *s1 = js_SkipWhiteSpace(s);
size_t length = js_strlen(s1);
cstr = malloc(length + 1);
if (!cstr)
return JS_FALSE;
for (i = 0; i <= length; i++) {
if (s1[i] >> 8) {
cstr[i] = 0;
break;
}
cstr[i] = (char)s1[i];
}
istr = cstr;
if ((negative = (*istr == '-')) != 0 || *istr == '+')
istr++;
if (!strncmp(istr, "Infinity", 8)) {
d = *(negative ? cx->runtime->jsNegativeInfinity : cx->runtime->jsPositiveInfinity);
estr = istr + 8;
} else {
errno = 0;
d = PR_strtod(cstr, &estr);
if (errno == ERANGE)
if (d == HUGE_VAL)
d = *cx->runtime->jsPositiveInfinity;
else if (d == -HUGE_VAL)
d = *cx->runtime->jsNegativeInfinity;
}
free(cstr);
i = estr - cstr;
*ep = i ? s1 + i : s;
*dp = d;
return JS_TRUE;
}
struct BinaryDigitReader
{
uintN base; /* Base of number; must be a power of 2 */
uintN digit; /* Current digit value in radix given by base */
uintN digitMask; /* Mask to extract the next bit from digit */
const jschar *digits; /* Pointer to the remaining digits */
const jschar *end; /* Pointer to first non-digit */
};
/* Return the next binary digit from the number or -1 if done */
static intN GetNextBinaryDigit(struct BinaryDigitReader *bdr)
{
intN bit;
if (bdr->digitMask == 0) {
uintN c;
if (bdr->digits == bdr->end)
return -1;
c = *bdr->digits++;
if ('0' <= c && c <= '9')
bdr->digit = c - '0';
else if ('a' <= c && c <= 'z')
bdr->digit = c - 'a' + 10;
else bdr->digit = c - 'A' + 10;
bdr->digitMask = bdr->base >> 1;
}
bit = (bdr->digit & bdr->digitMask) != 0;
bdr->digitMask >>= 1;
return bit;
}
JSBool
js_strtointeger(JSContext *cx, const jschar *s, const jschar **ep, jsint base, jsdouble *dp)
{
JSBool negative;
jsdouble value;
const jschar *start;
const jschar *s1 = js_SkipWhiteSpace(s);
if ((negative = (*s1 == '-')) != 0 || *s1 == '+')
s1++;
if (base == 0)
/* No base supplied, or some base that evaluated to 0. */
if (*s1 == '0')
/* It's either hex or octal; only increment char if str isn't '0' */
if (s1[1] == 'X' || s1[1] == 'x') { /* Hex */
s1 += 2;
base = 16;
} else /* Octal */
base = 8;
else
base = 10; /* Default to decimal. */
else if (base == 16 && *s1 == '0' && (s1[1] == 'X' || s1[1] == 'x'))
/* If base is 16, ignore hex prefix. */
s1 += 2;
/* Done with the preliminaries; find some prefix of the string that's
* a number in the given base.
*/
start = s1; /* Mark - if string is empty, we return NaN. */
value = 0.0;
while (1) {
uintN digit;
jschar c = *s1;
if ('0' <= c && c <= '9')
digit = c - '0';
else if ('a' <= c && c <= 'z')
digit = c - 'a' + 10;
else if ('A' <= c && c <= 'Z')
digit = c - 'A' + 10;
else
break;
if (digit >= (uintN)base)
break;
value = value*base + digit;
s1++;
}
if (value >= 9007199254740992.0)
if (base == 10) {
/* If we're accumulating a decimal number and the number is >= 2^53, then
* the result from the repeated multiply-add above may be inaccurate. Call
* PR_strtod to get the correct answer.
*/
size_t i;
size_t length = s1 - start;
char *cstr = malloc(length + 1);
char *estr;
if (!cstr)
return JS_FALSE;
for (i = 0; i != length; i++)
cstr[i] = (char)start[i];
cstr[length] = 0;
errno = 0;
value = PR_strtod(cstr, &estr);
if (errno == ERANGE && value == HUGE_VAL)
value = *cx->runtime->jsPositiveInfinity;
free(cstr);
} else if (base == 2 || base == 4 || base == 8 || base == 16 || base == 32) {
/* The number may also be inaccurate for one of these bases. This
* happens if the addition in value*base + digit causes a round-down
* to an even least significant mantissa bit when the first dropped bit
* is a one. If any of the following digits in the number (which haven't
* been added in yet) are nonzero then the correct action would have
* been to round up instead of down. An example of this occurs when
* reading the number 0x1000000000000081, which rounds to 0x1000000000000000
* instead of 0x1000000000000100.
*/
struct BinaryDigitReader bdr;
intN bit, bit2;
intN j;
bdr.base = base;
bdr.digitMask = 0;
bdr.digits = start;
bdr.end = s1;
value = 0.0;
/* Skip leading zeros. */
do {
bit = GetNextBinaryDigit(&bdr);
} while (bit == 0);
if (bit == 1) {
/* Gather the 53 significant bits (including the leading 1) */
value = 1.0;
for (j = 52; j; j--) {
bit = GetNextBinaryDigit(&bdr);
if (bit < 0)
goto done;
value = value*2 + bit;
}
/* bit2 is the 54th bit (the first dropped from the mantissa) */
bit2 = GetNextBinaryDigit(&bdr);
if (bit2 >= 0) {
jsdouble factor = 2.0;
intN sticky = 0; /* sticky is 1 if any bit beyond the 54th is 1 */
intN bit3;
while ((bit3 = GetNextBinaryDigit(&bdr)) >= 0) {
sticky |= bit3;
factor *= 2;
}
value += bit2 & (bit | sticky);
value *= factor;
}
done:;
}
}
/* We don't worry about inaccurate numbers for any other base. */
if (s1 == start) {
*dp = 0.0;
*ep = s;
} else {
*dp = negative ? -value : value;
*ep = s1;
}
return JS_TRUE;
}