test/jdk/sun/nio/cs/TestStringCodingUTF8.java
author jiangli
Wed, 27 Sep 2017 17:55:20 -0400
changeset 47548 664b9d44db74
parent 47216 71c04702a3d5
permissions -rw-r--r--
8068314: "Java fields that are currently set during shared space dumping" comment is incorrect Summary: CDS dump time should also initialize preallocated out_of_memory error messages. Reviewed-by: iklam, hseigel
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
9547
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
     1
/*
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
     2
 * Copyright (c) 2011, Oracle and/or its affiliates. All rights reserved.
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
     3
 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
     4
 *
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
     5
 * This code is free software; you can redistribute it and/or modify it
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
     6
 * under the terms of the GNU General Public License version 2 only, as
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
     7
 * published by the Free Software Foundation.
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
     8
 *
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
     9
 * This code is distributed in the hope that it will be useful, but WITHOUT
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    10
 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    11
 * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    12
 * version 2 for more details (a copy is included in the LICENSE file that
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    13
 * accompanied this code).
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    14
 *
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    15
 * You should have received a copy of the GNU General Public License version
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    16
 * 2 along with this work; if not, write to the Free Software Foundation,
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    17
 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    18
 *
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    19
 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    20
 * or visit www.oracle.com if you need additional information or have any
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    21
 * questions.
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    22
 */
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    23
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    24
/* @test
33663
2cd62a4bd471 8141132: JEP 254: Compact Strings
thartmann
parents: 30046
diff changeset
    25
   @bug 7040220 8054307
9547
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    26
   @summary Test if StringCoding and NIO result have the same de/encoding result for UTF-8
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    27
 * @run main/othervm/timeout=2000 TestStringCodingUTF8
30046
cf2c86e1819e 8078334: Mark regression tests using randomness
darcy
parents: 10898
diff changeset
    28
 * @key randomness
9547
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    29
 */
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    30
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    31
import java.util.*;
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    32
import java.nio.*;
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    33
import java.nio.charset.*;
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    34
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    35
public class TestStringCodingUTF8 {
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    36
    public static void main(String[] args) throws Throwable {
10898
01a232fa2ddb 7096080: UTF8 update and new CESU-8 charset
sherman
parents: 9547
diff changeset
    37
        test("UTF-8");
01a232fa2ddb 7096080: UTF8 update and new CESU-8 charset
sherman
parents: 9547
diff changeset
    38
        test("CESU-8");
9547
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    39
        // security manager on
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    40
        System.setSecurityManager(new PermissiveSecurityManger());
10898
01a232fa2ddb 7096080: UTF8 update and new CESU-8 charset
sherman
parents: 9547
diff changeset
    41
        test("UTF-8");
01a232fa2ddb 7096080: UTF8 update and new CESU-8 charset
sherman
parents: 9547
diff changeset
    42
        test("CESU-8");
9547
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    43
    }
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    44
10898
01a232fa2ddb 7096080: UTF8 update and new CESU-8 charset
sherman
parents: 9547
diff changeset
    45
    static void test(String csn) throws Throwable {
01a232fa2ddb 7096080: UTF8 update and new CESU-8 charset
sherman
parents: 9547
diff changeset
    46
        Charset cs = Charset.forName(csn);
9547
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    47
        char[] bmp = new char[0x10000];
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    48
        for (int i = 0; i < 0x10000; i++) {
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    49
            bmp[i] = (char)i;
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    50
        }
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    51
        test(cs, bmp, 0, bmp.length);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    52
33663
2cd62a4bd471 8141132: JEP 254: Compact Strings
thartmann
parents: 30046
diff changeset
    53
        char[] ascii = new char[0x80];
2cd62a4bd471 8141132: JEP 254: Compact Strings
thartmann
parents: 30046
diff changeset
    54
        for (int i = 0; i < 0x80; i++) {
2cd62a4bd471 8141132: JEP 254: Compact Strings
thartmann
parents: 30046
diff changeset
    55
            ascii[i] = (char)i;
2cd62a4bd471 8141132: JEP 254: Compact Strings
thartmann
parents: 30046
diff changeset
    56
        }
2cd62a4bd471 8141132: JEP 254: Compact Strings
thartmann
parents: 30046
diff changeset
    57
        test(cs, ascii, 0, ascii.length);
2cd62a4bd471 8141132: JEP 254: Compact Strings
thartmann
parents: 30046
diff changeset
    58
2cd62a4bd471 8141132: JEP 254: Compact Strings
thartmann
parents: 30046
diff changeset
    59
        char[] latin1 = new char[0x100];
2cd62a4bd471 8141132: JEP 254: Compact Strings
thartmann
parents: 30046
diff changeset
    60
        for (int i = 0; i < 0x100; i++) {
2cd62a4bd471 8141132: JEP 254: Compact Strings
thartmann
parents: 30046
diff changeset
    61
            latin1[i] = (char)i;
2cd62a4bd471 8141132: JEP 254: Compact Strings
thartmann
parents: 30046
diff changeset
    62
        }
2cd62a4bd471 8141132: JEP 254: Compact Strings
thartmann
parents: 30046
diff changeset
    63
        test(cs, latin1, 0, latin1.length);
2cd62a4bd471 8141132: JEP 254: Compact Strings
thartmann
parents: 30046
diff changeset
    64
9547
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    65
        ArrayList<Integer> list = new ArrayList<>(0x20000);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    66
        for (int i = 0; i < 0x20000; i++) {
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    67
            list.add(i, i);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    68
        }
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    69
        Collections.shuffle(list);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    70
        int j = 0;
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    71
        char[] bmpsupp = new char[0x30000];
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    72
        for (int i = 0; i < 0x20000; i++) {
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    73
            j += Character.toChars(list.get(i), bmpsupp, j);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    74
        }
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    75
        assert (j == bmpsupp.length);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    76
        test(cs, bmpsupp, 0, bmpsupp.length);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    77
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    78
        // randomed "off" and "len" on shuffled data
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    79
        Random rnd = new Random();
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    80
        int maxlen = 1000;
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    81
        int itr = 5000;
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    82
        for (int i = 0; i < itr; i++) {
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    83
            int off = rnd.nextInt(bmpsupp.length - maxlen);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    84
            int len = rnd.nextInt(maxlen);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    85
            test(cs, bmpsupp, off, len);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    86
        }
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    87
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    88
        // random length of bytes, test the edge corner case
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    89
        for (int i = 0; i < itr; i++) {
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    90
            byte[] ba = new byte[rnd.nextInt(maxlen)];
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    91
            rnd.nextBytes(ba);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    92
            //new String(csn);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    93
            if (!new String(ba, cs.name()).equals(
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    94
                 new String(decode(cs, ba, 0, ba.length))))
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    95
                throw new RuntimeException("new String(csn) failed");
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    96
            //new String(cs);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    97
            if (!new String(ba, cs).equals(
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    98
                 new String(decode(cs, ba, 0, ba.length))))
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
    99
                throw new RuntimeException("new String(cs) failed");
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   100
        }
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   101
        System.out.println("done!");
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   102
    }
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   103
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   104
    static void test(Charset cs, char[] ca, int off, int len) throws Throwable {
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   105
        String str = new String(ca, off, len);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   106
        byte[] ba = encode(cs, ca, off, len);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   107
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   108
        //getBytes(csn);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   109
        byte[] baStr = str.getBytes(cs.name());
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   110
        if (!Arrays.equals(ba, baStr))
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   111
            throw new RuntimeException("getBytes(csn) failed");
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   112
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   113
        //getBytes(cs);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   114
        baStr = str.getBytes(cs);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   115
        if (!Arrays.equals(ba, baStr))
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   116
            throw new RuntimeException("getBytes(cs) failed");
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   117
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   118
        //new String(csn);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   119
        if (!new String(ba, cs.name()).equals(new String(decode(cs, ba, 0, ba.length))))
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   120
            throw new RuntimeException("new String(csn) failed");
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   121
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   122
        //new String(cs);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   123
        if (!new String(ba, cs).equals(new String(decode(cs, ba, 0, ba.length))))
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   124
            throw new RuntimeException("new String(cs) failed");
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   125
    }
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   126
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   127
    // copy/paste of the StringCoding.decode()
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   128
    static char[] decode(Charset cs, byte[] ba, int off, int len) {
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   129
        CharsetDecoder cd = cs.newDecoder();
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   130
        int en = (int)(len * cd.maxCharsPerByte());
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   131
        char[] ca = new char[en];
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   132
        if (len == 0)
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   133
            return ca;
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   134
        cd.onMalformedInput(CodingErrorAction.REPLACE)
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   135
          .onUnmappableCharacter(CodingErrorAction.REPLACE)
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   136
          .reset();
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   137
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   138
        ByteBuffer bb = ByteBuffer.wrap(ba, off, len);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   139
        CharBuffer cb = CharBuffer.wrap(ca);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   140
        try {
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   141
            CoderResult cr = cd.decode(bb, cb, true);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   142
            if (!cr.isUnderflow())
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   143
                cr.throwException();
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   144
            cr = cd.flush(cb);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   145
            if (!cr.isUnderflow())
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   146
                cr.throwException();
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   147
        } catch (CharacterCodingException x) {
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   148
            throw new Error(x);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   149
        }
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   150
        return Arrays.copyOf(ca, cb.position());
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   151
    }
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   152
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   153
    // copy/paste of the StringCoding.encode()
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   154
    static byte[] encode(Charset cs, char[] ca, int off, int len) {
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   155
        CharsetEncoder ce = cs.newEncoder();
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   156
        int en = (int)(len * ce.maxBytesPerChar());
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   157
        byte[] ba = new byte[en];
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   158
        if (len == 0)
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   159
            return ba;
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   160
        ce.onMalformedInput(CodingErrorAction.REPLACE)
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   161
          .onUnmappableCharacter(CodingErrorAction.REPLACE)
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   162
          .reset();
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   163
        ByteBuffer bb = ByteBuffer.wrap(ba);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   164
        CharBuffer cb = CharBuffer.wrap(ca, off, len);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   165
        try {
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   166
            CoderResult cr = ce.encode(cb, bb, true);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   167
            if (!cr.isUnderflow())
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   168
                cr.throwException();
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   169
            cr = ce.flush(bb);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   170
            if (!cr.isUnderflow())
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   171
                cr.throwException();
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   172
        } catch (CharacterCodingException x) {
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   173
            throw new Error(x);
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   174
        }
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   175
        return Arrays.copyOf(ba, bb.position());
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   176
    }
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   177
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   178
    static class PermissiveSecurityManger extends SecurityManager {
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   179
        @Override public void checkPermission(java.security.Permission p) {}
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   180
    }
454881baaca0 7040220: java/char_encodin Optimize UTF-8 charset for String.getBytes()/new String(byte[])
sherman
parents:
diff changeset
   181
}