--- a/jdk/src/share/classes/java/lang/Long.java Thu Jul 17 11:08:50 2014 -0700
+++ b/jdk/src/share/classes/java/lang/Long.java Sat Jul 12 01:36:25 2014 +0200
@@ -27,6 +27,7 @@
import java.lang.annotation.Native;
import java.math.*;
+import java.util.Objects;
/**
@@ -561,12 +562,9 @@
" greater than Character.MAX_RADIX");
}
- long result = 0;
boolean negative = false;
int i = 0, len = s.length();
long limit = -Long.MAX_VALUE;
- long multmin;
- int digit;
if (len > 0) {
char firstChar = s.charAt(0);
@@ -574,21 +572,21 @@
if (firstChar == '-') {
negative = true;
limit = Long.MIN_VALUE;
- } else if (firstChar != '+')
+ } else if (firstChar != '+') {
throw NumberFormatException.forInputString(s);
+ }
- if (len == 1) // Cannot have lone "+" or "-"
+ if (len == 1) { // Cannot have lone "+" or "-"
throw NumberFormatException.forInputString(s);
+ }
i++;
}
- multmin = limit / radix;
+ long multmin = limit / radix;
+ long result = 0;
while (i < len) {
// Accumulating negatively avoids surprises near MAX_VALUE
- digit = Character.digit(s.charAt(i++),radix);
- if (digit < 0) {
- throw NumberFormatException.forInputString(s);
- }
- if (result < multmin) {
+ int digit = Character.digit(s.charAt(i++),radix);
+ if (digit < 0 || result < multmin) {
throw NumberFormatException.forInputString(s);
}
result *= radix;
@@ -597,10 +595,126 @@
}
result -= digit;
}
+ return negative ? result : -result;
} else {
throw NumberFormatException.forInputString(s);
}
- return negative ? result : -result;
+ }
+
+ /**
+ * Parses the {@link CharSequence} argument as a signed {@code long} in
+ * the specified {@code radix}, beginning at the specified {@code beginIndex}
+ * and extending to the end of the sequence.
+ *
+ * <p>The method does not take steps to guard against the
+ * {@code CharSequence} being mutated while parsing.
+ *
+ * @param s the {@code CharSequence} containing the {@code long}
+ * representation to be parsed
+ * @param radix the radix to be used while parsing {@code s}.
+ * @param beginIndex the beginning index, inclusive.
+ * @return the signed {@code long} represented by the subsequence in
+ * the specified radix.
+ * @throws NullPointerException if {@code s} is null.
+ * @throws IndexOutOfBoundsException if {@code beginIndex} is
+ * negative, or if {@code beginIndex} is greater than
+ * {@code s.length()}.
+ * @throws NumberFormatException if the {@code CharSequence} does not
+ * contain a parsable {@code long} in the specified
+ * {@code radix}, or if {@code radix} is either smaller than
+ * {@link java.lang.Character#MIN_RADIX} or larger than
+ * {@link java.lang.Character#MAX_RADIX}.
+ * @since 1.9
+ */
+ public static long parseLong(CharSequence s, int radix, int beginIndex)
+ throws NumberFormatException {
+ // forces a null check of s
+ return parseLong(s, radix, beginIndex, s.length());
+ }
+
+ /**
+ * Parses the {@link CharSequence} argument as a signed {@code long} in
+ * the specified {@code radix}, beginning at the specified
+ * {@code beginIndex} and extending to {@code endIndex - 1}.
+ *
+ * <p>The method does not take steps to guard against the
+ * {@code CharSequence} being mutated while parsing.
+ *
+ * @param s the {@code CharSequence} containing the {@code long}
+ * representation to be parsed
+ * @param radix the radix to be used while parsing {@code s}.
+ * @param beginIndex the beginning index, inclusive.
+ * @param endIndex the ending index, exclusive.
+ * @return the signed {@code long} represented by the subsequence in
+ * the specified radix.
+ * @throws NullPointerException if {@code s} is null.
+ * @throws IndexOutOfBoundsException if {@code beginIndex} is
+ * negative, or if {@code beginIndex} is greater than
+ * {@code endIndex} or if {@code endIndex} is greater than
+ * {@code s.length()}.
+ * @throws NumberFormatException if the {@code CharSequence} does not
+ * contain a parsable {@code int} in the specified
+ * {@code radix}, or if {@code radix} is either smaller than
+ * {@link java.lang.Character#MIN_RADIX} or larger than
+ * {@link java.lang.Character#MAX_RADIX}.
+ * @since 1.9
+ */
+ public static long parseLong(CharSequence s, int radix, int beginIndex, int endIndex)
+ throws NumberFormatException {
+ s = Objects.requireNonNull(s);
+
+ if (beginIndex < 0 || beginIndex > endIndex || endIndex > s.length()) {
+ throw new IndexOutOfBoundsException();
+ }
+ if (radix < Character.MIN_RADIX) {
+ throw new NumberFormatException("radix " + radix +
+ " less than Character.MIN_RADIX");
+ }
+ if (radix > Character.MAX_RADIX) {
+ throw new NumberFormatException("radix " + radix +
+ " greater than Character.MAX_RADIX");
+ }
+
+ boolean negative = false;
+ int i = beginIndex;
+ long limit = -Long.MAX_VALUE;
+
+ if (i < endIndex) {
+ char firstChar = s.charAt(i);
+ if (firstChar < '0') { // Possible leading "+" or "-"
+ if (firstChar == '-') {
+ negative = true;
+ limit = Long.MIN_VALUE;
+ } else if (firstChar != '+') {
+ throw NumberFormatException.forCharSequence(s, beginIndex,
+ endIndex, i);
+ }
+ i++;
+ }
+ if (i >= endIndex) { // Cannot have lone "+", "-" or ""
+ throw NumberFormatException.forCharSequence(s, beginIndex,
+ endIndex, i);
+ }
+ long multmin = limit / radix;
+ long result = 0;
+ while (i < endIndex) {
+ // Accumulating negatively avoids surprises near MAX_VALUE
+ int digit = Character.digit(s.charAt(i++), radix);
+ if (digit < 0 || result < multmin) {
+ throw NumberFormatException.forCharSequence(s, beginIndex,
+ endIndex, i);
+ }
+ result *= radix;
+ if (result < limit + digit) {
+ throw NumberFormatException.forCharSequence(s, beginIndex,
+ endIndex, i);
+ }
+ result -= digit;
+ }
+ return negative ? result : -result;
+ } else {
+ throw new NumberFormatException("");
+ }
}
/**
@@ -694,7 +808,7 @@
}
// No need for range checks on len due to testing above.
- long first = parseLong(s.substring(0, len - 1), radix);
+ long first = parseLong(s, radix, 0, len - 1);
int second = Character.digit(s.charAt(len - 1), radix);
if (second < 0) {
throw new NumberFormatException("Bad digit at end of " + s);
@@ -764,6 +878,155 @@
}
/**
+ * Parses the {@link CharSequence} argument as an unsigned {@code long} in
+ * the specified {@code radix}, beginning at the specified
+ * {@code beginIndex} and extending to the end of the sequence.
+ *
+ * <p>The method does not take steps to guard against the
+ * {@code CharSequence} being mutated while parsing.
+ *
+ * @param s the {@code CharSequence} containing the unsigned
+ * {@code long} representation to be parsed
+ * @param radix the radix to be used while parsing {@code s}.
+ * @param beginIndex the beginning index, inclusive.
+ * @return the unsigned {@code long} represented by the subsequence in
+ * the specified radix.
+ * @throws NullPointerException if {@code s} is null.
+ * @throws IndexOutOfBoundsException if {@code beginIndex} is
+ * negative, or if {@code beginIndex} is greater than
+ * {@code s.length()}.
+ * @throws NumberFormatException if the {@code CharSequence} does not
+ * contain a parsable unsigned {@code long} in the specified
+ * {@code radix}, or if {@code radix} is either smaller than
+ * {@link java.lang.Character#MIN_RADIX} or larger than
+ * {@link java.lang.Character#MAX_RADIX}.
+ * @since 1.9
+ */
+ public static long parseUnsignedLong(CharSequence s, int radix, int beginIndex)
+ throws NumberFormatException {
+ // forces a null check of s
+ return parseUnsignedLong(s, radix, beginIndex, s.length());
+ }
+
+ /**
+ * Parses the {@link CharSequence} argument as an unsigned {@code long} in
+ * the specified {@code radix}, beginning at the specified
+ * {@code beginIndex} and extending to {@code endIndex - 1}.
+ *
+ * <p>The method does not take steps to guard against the
+ * {@code CharSequence} being mutated while parsing.
+ *
+ * @param s the {@code CharSequence} containing the unsigned
+ * {@code long} representation to be parsed
+ * @param radix the radix to be used while parsing {@code s}.
+ * @param beginIndex the beginning index, inclusive.
+ * @param endIndex the ending index, exclusive.
+ * @return the unsigned {@code long} represented by the subsequence in
+ * the specified radix.
+ * @throws NullPointerException if {@code s} is null.
+ * @throws IndexOutOfBoundsException if {@code beginIndex} is
+ * negative, or if {@code beginIndex} is greater than
+ * {@code endIndex} or if {@code endIndex} is greater than
+ * {@code s.length()}.
+ * @throws NumberFormatException if the {@code CharSequence} does not
+ * contain a parsable unsigned {@code long} in the specified
+ * {@code radix}, or if {@code radix} is either smaller than
+ * {@link java.lang.Character#MIN_RADIX} or larger than
+ * {@link java.lang.Character#MAX_RADIX}.
+ * @since 1.9
+ */
+ public static long parseUnsignedLong(CharSequence s, int radix, int beginIndex, int endIndex)
+ throws NumberFormatException {
+ s = Objects.requireNonNull(s);
+
+ if (beginIndex < 0 || beginIndex > endIndex || endIndex > s.length()) {
+ throw new IndexOutOfBoundsException();
+ }
+ int start = beginIndex, len = endIndex - beginIndex;
+
+ if (len > 0) {
+ char firstChar = s.charAt(start);
+ if (firstChar == '-') {
+ throw new NumberFormatException(String.format("Illegal leading minus sign " +
+ "on unsigned string %s.", s.subSequence(start, start + len)));
+ } else {
+ if (len <= 12 || // Long.MAX_VALUE in Character.MAX_RADIX is 13 digits
+ (radix == 10 && len <= 18) ) { // Long.MAX_VALUE in base 10 is 19 digits
+ return parseLong(s, radix, start, start + len);
+ }
+
+ // No need for range checks on end due to testing above.
+ long first = parseLong(s, radix, start, start + len - 1);
+ int second = Character.digit(s.charAt(start + len - 1), radix);
+ if (second < 0) {
+ throw new NumberFormatException("Bad digit at end of " +
+ s.subSequence(start, start + len));
+ }
+ long result = first * radix + second;
+
+ /*
+ * Test leftmost bits of multiprecision extension of first*radix
+ * for overflow. The number of bits needed is defined by
+ * GUARD_BIT = ceil(log2(Character.MAX_RADIX)) + 1 = 7. Then
+ * int guard = radix*(int)(first >>> (64 - GUARD_BIT)) and
+ * overflow is tested by splitting guard in the ranges
+ * guard < 92, 92 <= guard < 128, and 128 <= guard, where
+ * 92 = 128 - Character.MAX_RADIX. Note that guard cannot take
+ * on a value which does not include a prime factor in the legal
+ * radix range.
+ */
+ int guard = radix * (int) (first >>> 57);
+ if (guard >= 128 ||
+ (result >= 0 && guard >= 128 - Character.MAX_RADIX)) {
+ /*
+ * For purposes of exposition, the programmatic statements
+ * below should be taken to be multi-precision, i.e., not
+ * subject to overflow.
+ *
+ * A) Condition guard >= 128:
+ * If guard >= 128 then first*radix >= 2^7 * 2^57 = 2^64
+ * hence always overflow.
+ *
+ * B) Condition guard < 92:
+ * Define left7 = first >>> 57.
+ * Given first = (left7 * 2^57) + (first & (2^57 - 1)) then
+ * result <= (radix*left7)*2^57 + radix*(2^57 - 1) + second.
+ * Thus if radix*left7 < 92, radix <= 36, and second < 36,
+ * then result < 92*2^57 + 36*(2^57 - 1) + 36 = 2^64 hence
+ * never overflow.
+ *
+ * C) Condition 92 <= guard < 128:
+ * first*radix + second >= radix*left7*2^57 + second
+ * so that first*radix + second >= 92*2^57 + 0 > 2^63
+ *
+ * D) Condition guard < 128:
+ * radix*first <= (radix*left7) * 2^57 + radix*(2^57 - 1)
+ * so
+ * radix*first + second <= (radix*left7) * 2^57 + radix*(2^57 - 1) + 36
+ * thus
+ * radix*first + second < 128 * 2^57 + 36*2^57 - radix + 36
+ * whence
+ * radix*first + second < 2^64 + 2^6*2^57 = 2^64 + 2^63
+ *
+ * E) Conditions C, D, and result >= 0:
+ * C and D combined imply the mathematical result
+ * 2^63 < first*radix + second < 2^64 + 2^63. The lower
+ * bound is therefore negative as a signed long, but the
+ * upper bound is too small to overflow again after the
+ * signed long overflows to positive above 2^64 - 1. Hence
+ * result >= 0 implies overflow given C and D.
+ */
+ throw new NumberFormatException(String.format("String value %s exceeds " +
+ "range of unsigned long.", s.subSequence(start, start + len)));
+ }
+ return result;
+ }
+ } else {
+ throw NumberFormatException.forInputString("");
+ }
+ }
+
+ /**
* Parses the string argument as an unsigned decimal {@code long}. The
* characters in the string must all be decimal digits, except
* that the first character may be an an ASCII plus sign {@code