jdk.nashorn.internal.runtime.regexp.joni.EncodingHelper Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of nashorn Show documentation
Show all versions of nashorn Show documentation
A maven release for nashorn. Changed nothing, just put it to maven.
/*
* Permission is hereby granted, free of charge, to any person obtaining a copy of
* this software and associated documentation files (the "Software"), to deal in
* the Software without restriction, including without limitation the rights to
* use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies
* of the Software, and to permit persons to whom the Software is furnished to do
* so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in all
* copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
package jdk.nashorn.internal.runtime.regexp.joni;
import java.util.Arrays;
import jdk.nashorn.internal.runtime.regexp.joni.encoding.CharacterType;
import jdk.nashorn.internal.runtime.regexp.joni.encoding.IntHolder;
@SuppressWarnings("javadoc")
public final class EncodingHelper {
final static int NEW_LINE = 0x000a;
final static int RETURN = 0x000d;
final static int LINE_SEPARATOR = 0x2028;
final static int PARAGRAPH_SEPARATOR = 0x2029;
final static char[] EMPTYCHARS = new char[0];
final static int[][] codeRanges = new int[15][];
public static int digitVal(final int code) {
return code - '0';
}
public static int odigitVal(final int code) {
return digitVal(code);
}
public static boolean isXDigit(final int code) {
return Character.isDigit(code) || (code >= 'a' && code <= 'f') || (code >= 'A' && code <= 'F');
}
public static int xdigitVal(final int code) {
if (Character.isDigit(code)) {
return code - '0';
} else if (code >= 'a' && code <= 'f') {
return code - 'a' + 10;
} else {
return code - 'A' + 10;
}
}
public static boolean isDigit(final int code) {
return code >= '0' && code <= '9';
}
public static boolean isWord(final int code) {
// letter, digit, or '_'
return (1 << Character.getType(code) & CharacterType.WORD_MASK) != 0;
}
public static boolean isNewLine(final int code) {
return code == NEW_LINE || code == RETURN || code == LINE_SEPARATOR || code == PARAGRAPH_SEPARATOR;
}
public static boolean isNewLine(final char[] chars, final int p, final int end) {
return p < end && isNewLine(chars[p]);
}
// Encoding.prevCharHead
public static int prevCharHead(final int p, final int s) {
return s <= p ? -1 : s - 1;
}
/* onigenc_get_right_adjust_char_head_with_prev */
public static int rightAdjustCharHeadWithPrev(final int s, final IntHolder prev) {
if (prev != null) {
prev.value = -1; /* Sorry */
}
return s;
}
// Encoding.stepBack
public static int stepBack(final int p, final int sp, final int np) {
int s = sp, n = np;
while (s != -1 && n-- > 0) {
if (s <= p) {
return -1;
}
s--;
}
return s;
}
public static int mbcodeStartPosition() {
return 0x80;
}
public static char[] caseFoldCodesByString(final int flag, final char c) {
char[] codes = EMPTYCHARS;
final char upper = toUpperCase(c);
if (upper != toLowerCase(upper)) {
int count = 0;
char ch = 0;
do {
final char u = toUpperCase(ch);
if (u == upper && ch != c) {
// Almost all characters will return array of length 1, very few 2 or 3, so growing by one is fine.
codes = count == 0 ? new char[1] : Arrays.copyOf(codes, count + 1);
codes[count++] = ch;
}
} while (ch++ < 0xffff);
}
return codes;
}
public static void applyAllCaseFold(final int flag, final ApplyCaseFold fun, final Object arg) {
for (int c = 0; c < 0xffff; c++) {
if (Character.isLowerCase(c)) {
final int upper = toUpperCase(c);
if (upper != c) {
ApplyCaseFold.apply(c, upper, arg);
}
}
}
// Some characters have multiple lower case variants, hence we need to do a second run
for (int c = 0; c < 0xffff; c++) {
if (Character.isLowerCase(c)) {
final int upper = toUpperCase(c);
if (upper != c) {
ApplyCaseFold.apply(upper, c, arg);
}
}
}
}
public static char toLowerCase(final char c) {
return (char)toLowerCase((int)c);
}
public static int toLowerCase(final int c) {
if (c < 128) {
return ('A' <= c && c <= 'Z') ? (c + ('a' - 'A')) : c;
}
// Do not convert non-ASCII upper case character to ASCII lower case.
final int lower = Character.toLowerCase(c);
return (lower < 128) ? c : lower;
}
public static char toUpperCase(final char c) {
return (char)toUpperCase((int)c);
}
public static int toUpperCase(final int c) {
if (c < 128) {
return ('a' <= c && c <= 'z') ? c + ('A' - 'a') : c;
}
// Do not convert non-ASCII lower case character to ASCII upper case.
final int upper = Character.toUpperCase(c);
return (upper < 128) ? c : upper;
}
public static int[] ctypeCodeRange(final int ctype, final IntHolder sbOut) {
sbOut.value = 0x100; // use bitset for codes smaller than 256
int[] range = null;
if (ctype < codeRanges.length) {
range = codeRanges[ctype];
if (range == null) {
// format: [numberOfRanges, rangeStart, rangeEnd, ...]
range = new int[16];
int rangeCount = 0;
int lastCode = -2;
for (int code = 0; code <= 0xffff; code++) {
if (isCodeCType(code, ctype)) {
if (lastCode < code -1) {
if (rangeCount * 2 + 2 >= range.length) {
range = Arrays.copyOf(range, range.length * 2);
}
range[rangeCount * 2 + 1] = code;
rangeCount++;
}
range[rangeCount * 2] = lastCode = code;
}
}
if (rangeCount * 2 + 1 < range.length) {
range = Arrays.copyOf(range, rangeCount * 2 + 1);
}
range[0] = rangeCount;
codeRanges[ctype] = range;
}
}
return range;
}
// CodeRange.isInCodeRange
public static boolean isInCodeRange(final int[] p, final int offset, final int code) {
int low = 0;
final int n = p[offset];
int high = n ;
while (low < high) {
final int x = (low + high) >> 1;
if (code > p[(x << 1) + 2 + offset]) {
low = x + 1;
} else {
high = x;
}
}
return low < n && code >= p[(low << 1) + 1 + offset];
}
/**
* @see http://www.geocities.jp/kosako3/oniguruma/doc/RE.txt
*
* @param code code
* @param ctype ctype
*
* @return isCodeCType
*/
public static boolean isCodeCType(final int code, final int ctype) {
int type;
switch (ctype) {
case CharacterType.NEWLINE:
return isNewLine(code);
case CharacterType.ALPHA:
return (1 << Character.getType(code) & CharacterType.ALPHA_MASK) != 0;
case CharacterType.BLANK:
return code == 0x09 || Character.getType(code) == Character.SPACE_SEPARATOR;
case CharacterType.CNTRL:
type = Character.getType(code);
return (1 << type & CharacterType.CNTRL_MASK) != 0 || type == Character.UNASSIGNED;
case CharacterType.DIGIT:
return EncodingHelper.isDigit(code);
case CharacterType.GRAPH:
switch (code) {
case 0x09:
case 0x0a:
case 0x0b:
case 0x0c:
case 0x0d:
return false;
default:
type = Character.getType(code);
return (1 << type & CharacterType.GRAPH_MASK) == 0 && type != Character.UNASSIGNED;
}
case CharacterType.LOWER:
return Character.isLowerCase(code);
case CharacterType.PRINT:
type = Character.getType(code);
return (1 << type & CharacterType.PRINT_MASK) == 0 && type != Character.UNASSIGNED;
case CharacterType.PUNCT:
return (1 << Character.getType(code) & CharacterType.PUNCT_MASK) != 0;
case CharacterType.SPACE:
// ECMA 7.2 and 7.3
switch (code) {
case 0x09:
case 0x0a:
case 0x0b:
case 0x0c:
case 0x0d:
return true;
default:
// true if Unicode separator or BOM
return (1 << Character.getType(code) & CharacterType.SPACE_MASK) != 0 || code == 0xfeff;
}
case CharacterType.UPPER:
return Character.isUpperCase(code);
case CharacterType.XDIGIT:
return EncodingHelper.isXDigit(code);
case CharacterType.WORD:
return (1 << Character.getType(code) & CharacterType.WORD_MASK) != 0;
case CharacterType.ALNUM:
return (1 << Character.getType(code) & CharacterType.ALNUM_MASK) != 0;
case CharacterType.ASCII:
return code < 0x80;
default:
throw new RuntimeException("illegal character type: " + ctype);
}
}
}