| 1 | /**************************************************************************** |
| 2 | ** |
| 3 | ** Copyright (C) 2018 The Qt Company Ltd. |
| 4 | ** Copyright (C) 2018 Intel Corporation. |
| 5 | ** Contact: https://www.qt.io/licensing/ |
| 6 | ** |
| 7 | ** This file is part of the test suite of the Qt Toolkit. |
| 8 | ** |
| 9 | ** $QT_BEGIN_LICENSE:GPL-EXCEPT$ |
| 10 | ** Commercial License Usage |
| 11 | ** Licensees holding valid commercial Qt licenses may use this file in |
| 12 | ** accordance with the commercial license agreement provided with the |
| 13 | ** Software or, alternatively, in accordance with the terms contained in |
| 14 | ** a written agreement between you and The Qt Company. For licensing terms |
| 15 | ** and conditions see https://www.qt.io/terms-conditions. For further |
| 16 | ** information use the contact form at https://www.qt.io/contact-us. |
| 17 | ** |
| 18 | ** GNU General Public License Usage |
| 19 | ** Alternatively, this file may be used under the terms of the GNU |
| 20 | ** General Public License version 3 as published by the Free Software |
| 21 | ** Foundation with exceptions as appearing in the file LICENSE.GPL3-EXCEPT |
| 22 | ** included in the packaging of this file. Please review the following |
| 23 | ** information to ensure the GNU General Public License requirements will |
| 24 | ** be met: https://www.gnu.org/licenses/gpl-3.0.html. |
| 25 | ** |
| 26 | ** $QT_END_LICENSE$ |
| 27 | ** |
| 28 | ****************************************************************************/ |
| 29 | #include <QtTest/QtTest> |
| 30 | |
| 31 | void loadInvalidUtf8Rows() |
| 32 | { |
| 33 | // Wrong continuations |
| 34 | QTest::newRow(dataTag: "bad-continuation-1char" ) << QByteArray("\x80" ); |
| 35 | QTest::newRow(dataTag: "bad-continuation-2chars-1" ) << QByteArray("\xC2\xC0" ); |
| 36 | QTest::newRow(dataTag: "bad-continuation-2chars-2" ) << QByteArray("\xC3\xDF" ); |
| 37 | QTest::newRow(dataTag: "bad-continuation-2chars-3" ) << QByteArray("\xC7\xF0" ); |
| 38 | QTest::newRow(dataTag: "bad-continuation-3chars-1" ) << QByteArray("\xE0\xA0\xC0" ); |
| 39 | QTest::newRow(dataTag: "bad-continuation-3chars-2" ) << QByteArray("\xE0\xC0\xA0" ); |
| 40 | QTest::newRow(dataTag: "bad-continuation-4chars-1" ) << QByteArray("\xF0\x90\x80\xC0" ); |
| 41 | QTest::newRow(dataTag: "bad-continuation-4chars-2" ) << QByteArray("\xF0\x90\xC0\x80" ); |
| 42 | QTest::newRow(dataTag: "bad-continuation-4chars-3" ) << QByteArray("\xF0\xC0\x80\x80" ); |
| 43 | |
| 44 | // Too short |
| 45 | QTest::newRow(dataTag: "too-short-2chars" ) << QByteArray("\xC2" ); |
| 46 | QTest::newRow(dataTag: "too-short-3chars-1" ) << QByteArray("\xE0" ); |
| 47 | QTest::newRow(dataTag: "too-short-3chars-2" ) << QByteArray("\xE0\xA0" ); |
| 48 | QTest::newRow(dataTag: "too-short-4chars-1" ) << QByteArray("\xF0" ); |
| 49 | QTest::newRow(dataTag: "too-short-4chars-2" ) << QByteArray("\xF0\x90" ); |
| 50 | QTest::newRow(dataTag: "too-short-4chars-3" ) << QByteArray("\xF0\x90\x80" ); |
| 51 | |
| 52 | // Surrogate pairs must now be present either |
| 53 | // U+D800: 1101 10 0000 00 0000 |
| 54 | // encoding: xxxz:1101 xz10:0000 xz00:0000 |
| 55 | QTest::newRow(dataTag: "hi-surrogate" ) << QByteArray("\xED\xA0\x80" ); |
| 56 | // U+DC00: 1101 11 0000 00 0000 |
| 57 | // encoding: xxxz:1101 xz11:0000 xz00:0000 |
| 58 | QTest::newRow(dataTag: "lo-surrogate" ) << QByteArray("\xED\xB0\x80" ); |
| 59 | |
| 60 | // not even in pair: |
| 61 | QTest::newRow(dataTag: "surrogate-pair" ) << QByteArray("\xED\xA0\x80\xED\xB0\x80" ); |
| 62 | |
| 63 | // Characters outside the Unicode range: |
| 64 | // 0x110000: 00 0100 01 0000 00 0000 00 0000 |
| 65 | // encoding: xxxx:z100 xz01:0000 xz00:0000 xz00:0000 |
| 66 | QTest::newRow(dataTag: "non-unicode-1" ) << QByteArray("\xF4\x90\x80\x80" ); |
| 67 | // 0x200000: 00 1000 00 0000 00 0000 00 0000 |
| 68 | // encoding: xxxx:xz00 xz00:1000 xz00:0000 xz00:0000 xz00:0000 |
| 69 | QTest::newRow(dataTag: "non-unicode-2" ) << QByteArray("\xF8\x88\x80\x80\x80" ); |
| 70 | // 0x04000000: 0100 00 0000 00 0000 00 0000 00 0000 |
| 71 | // encoding: xxxx:xxz0 xz00:0100 xz00:0000 xz00:0000 xz00:0001 xz00:0001 |
| 72 | QTest::newRow(dataTag: "non-unicode-3" ) << QByteArray("\xFC\x84\x80\x80\x80\x80" ); |
| 73 | // 0x7fffffff: 1 11 1111 11 1111 11 1111 11 1111 11 1111 |
| 74 | // encoding: xxxx:xxz0 xz00:0100 xz00:0000 xz00:0000 xz00:0001 xz00:0001 |
| 75 | QTest::newRow(dataTag: "non-unicode-4" ) << QByteArray("\xFD\xBF\xBF\xBF\xBF\xBF" ); |
| 76 | |
| 77 | // As seen above, 0xFE and 0xFF never appear: |
| 78 | QTest::newRow(dataTag: "fe" ) << QByteArray("\xFE" ); |
| 79 | QTest::newRow(dataTag: "fe-bis" ) << QByteArray("\xFE\xBF\xBF\xBF\xBF\xBF\xBF" ); |
| 80 | QTest::newRow(dataTag: "ff" ) << QByteArray("\xFF" ); |
| 81 | QTest::newRow(dataTag: "ff-bis" ) << QByteArray("\xFF\xBF\xBF\xBF\xBF\xBF\xBF\xBF" ); |
| 82 | |
| 83 | // some combinations in UTF-8 are invalid even though they have the proper bits set |
| 84 | // these are known as overlong sequences |
| 85 | |
| 86 | // "A": U+0041: 01 00 0001 |
| 87 | // overlong 2: xxz0:0001 xz00:0001 |
| 88 | QTest::newRow(dataTag: "overlong-1-2" ) << QByteArray("\xC1\x81" ); |
| 89 | // overlong 3: xxxz:0000 xz00:0001 xz00:0001 |
| 90 | QTest::newRow(dataTag: "overlong-1-3" ) << QByteArray("\xE0\x81\x81" ); |
| 91 | // overlong 4: xxxx:z000 xz00:0000 xz00:0001 xz00:0001 |
| 92 | QTest::newRow(dataTag: "overlong-1-4" ) << QByteArray("\xF0\x80\x81\x81" ); |
| 93 | // overlong 5: xxxx:xz00 xz00:0000 xz00:0000 xz00:0001 xz00:0001 |
| 94 | QTest::newRow(dataTag: "overlong-1-5" ) << QByteArray("\xF8\x80\x80\x81\x81" ); |
| 95 | // overlong 6: xxxx:xxz0 xz00:0000 xz00:0000 xz00:0000 xz00:0001 xz00:0001 |
| 96 | QTest::newRow(dataTag: "overlong-1-6" ) << QByteArray("\xFC\x80\x80\x80\x81\x81" ); |
| 97 | |
| 98 | // U+0080: 10 00 0000 |
| 99 | // proper encoding: xxz0:0010 xz00:0000 |
| 100 | // overlong 3: xxxz:0000 xz00:0010 xz00:0000 |
| 101 | QTest::newRow(dataTag: "overlong-2-3" ) << QByteArray("\xE0\x82\x80" ); |
| 102 | // overlong 4: xxxx:z000 xz00:0000 xz00:0010 xz00:0000 |
| 103 | QTest::newRow(dataTag: "overlong-2-4" ) << QByteArray("\xF0\x80\x82\x80" ); |
| 104 | // overlong 5: xxxx:xz00 xz00:0000 xz00:0000 xz00:0010 xz00:0000 |
| 105 | QTest::newRow(dataTag: "overlong-2-5" ) << QByteArray("\xF8\x80\x80\x82\x80" ); |
| 106 | // overlong 6: xxxx:xxz0 xz00:0000 xz00:0000 xz00:0000 xz00:0010 xz00:0000 |
| 107 | QTest::newRow(dataTag: "overlong-2-6" ) << QByteArray("\xFC\x80\x80\x80\x82\x80" ); |
| 108 | |
| 109 | // U+0800: 10 0000 00 0000 |
| 110 | // proper encoding: xxxz:0000 xz10:0000 xz00:0000 |
| 111 | // overlong 4: xxxx:z000 xz00:0000 xz10:0000 xz00:0000 |
| 112 | QTest::newRow(dataTag: "overlong-3-4" ) << QByteArray("\xF0\x80\xA0\x80" ); |
| 113 | // overlong 5: xxxx:xz00 xz00:0000 xz00:0000 xz10:0000 xz00:0000 |
| 114 | QTest::newRow(dataTag: "overlong-3-5" ) << QByteArray("\xF8\x80\x80\xA0\x80" ); |
| 115 | // overlong 6: xxxx:xxz0 xz00:0000 xz00:0000 xz00:0000 xz10:0000 xz00:0000 |
| 116 | QTest::newRow(dataTag: "overlong-3-6" ) << QByteArray("\xFC\x80\x80\x80\xA0\x80" ); |
| 117 | |
| 118 | // U+010000: 00 0100 00 0000 00 0000 |
| 119 | // proper encoding: xxxx:z000 xz00:0100 xz00:0000 xz00:0000 |
| 120 | // overlong 5: xxxx:xz00 xz00:0000 xz00:0100 xz00:0000 xz00:0000 |
| 121 | QTest::newRow(dataTag: "overlong-4-5" ) << QByteArray("\xF8\x80\x84\x80\x80" ); |
| 122 | // overlong 6: xxxx:xxz0 xz00:0000 xz00:0000 xz00:0100 xz00:0000 xz00:0000 |
| 123 | QTest::newRow(dataTag: "overlong-4-6" ) << QByteArray("\xFC\x80\x80\x84\x80\x80" ); |
| 124 | |
| 125 | } |
| 126 | |
| 127 | void loadNonCharactersRows() |
| 128 | { |
| 129 | // Unicode has a couple of "non-characters" that one can use internally |
| 130 | // These characters are allowed for text-interchange (see http://www.unicode.org/versions/corrigendum9.html) |
| 131 | // |
| 132 | // Those are the last two entries each Unicode Plane (U+FFFE, U+FFFF, |
| 133 | // U+1FFFE, U+1FFFF, etc.) as well as the entries between U+FDD0 and |
| 134 | // U+FDEF (inclusive) |
| 135 | |
| 136 | // U+FDD0 through U+FDEF |
| 137 | for (int i = 0; i < 16; ++i) { |
| 138 | char utf8[] = { char(0357), char(0267), char(0220 + i), 0 }; |
| 139 | QString utf16 = QChar(0xfdd0 + i); |
| 140 | QTest::newRow(qPrintable(QString::number(0xfdd0 + i, 16))) << QByteArray(utf8) << utf16; |
| 141 | } |
| 142 | |
| 143 | // the last two in Planes 1 through 16 |
| 144 | for (uint plane = 1; plane <= 16; ++plane) { |
| 145 | for (uint lower = 0xfffe; lower < 0x10000; ++lower) { |
| 146 | uint ucs4 = (plane << 16) | lower; |
| 147 | char utf8[] = { char(0xf0 | uchar(ucs4 >> 18)), |
| 148 | char(0x80 | (uchar(ucs4 >> 12) & 0x3f)), |
| 149 | char(0x80 | (uchar(ucs4 >> 6) & 0x3f)), |
| 150 | char(0x80 | (uchar(ucs4) & 0x3f)), |
| 151 | 0 }; |
| 152 | ushort utf16[] = { QChar::highSurrogate(ucs4), QChar::lowSurrogate(ucs4), 0 }; |
| 153 | |
| 154 | QTest::newRow(qPrintable(QString::number(ucs4, 16))) << QByteArray(utf8) << QString::fromUtf16(utf16); |
| 155 | } |
| 156 | } |
| 157 | |
| 158 | QTest::newRow(dataTag: "fffe" ) << QByteArray("\xEF\xBF\xBE" ) << QString(QChar(0xfffe)); |
| 159 | QTest::newRow(dataTag: "ffff" ) << QByteArray("\xEF\xBF\xBF" ) << QString(QChar(0xffff)); |
| 160 | } |
| 161 | |