third_party/harfbuzz/tests/linebreaking/main.cpp

   1 /*
   2  * Copyright (C) 2008 Nokia Corporation and/or its subsidiary(-ies)
   3  *
   4  * This is part of HarfBuzz, an OpenType Layout engine library.
   5  *
   6  * Permission is hereby granted, without written agreement and without
   7  * license or royalty fees, to use, copy, modify, and distribute this
   8  * software and its documentation for any purpose, provided that the
   9  * above copyright notice and the following two paragraphs appear in
  10  * all copies of this software.
  11  *
  12  * IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
  13  * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
  14  * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
  15  * IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
  16  * DAMAGE.
  17  *
  18  * THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
  19  * BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
  20  * FITNESS FOR A PARTICULAR PURPOSE.  THE SOFTWARE PROVIDED HEREUNDER IS
  21  * ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
  22  * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
  23  */
  24
  25 /*
  26     !!!!!! Warning !!!!!
  27     Please don't save this file in emacs. It contains utf8 text sequences emacs will
  28     silently convert to a series of question marks.
  29  */
  30 #include <QtTest/QtTest>
  31 #include <QtCore/qdebug.h>
  32
  33 #include <harfbuzz-shaper.h>
  34
  35 static QVector<HB_CharAttributes> getCharAttributes(const QString &str, HB_Script script = HB_Script_Common)
  36 {
  37     QVector<HB_CharAttributes> attrs(str.length());
  38     HB_ScriptItem item;
  39     item.pos = 0;
  40     item.length = str.length();
  41     item.script = script;
  42     HB_GetCharAttributes(str.utf16(), str.length(),
  43                          &item, 1,
  44                          attrs.data());
  45     return attrs;
  46 }
  47
  48 class tst_CharAttributes : public QObject
  49 {
  50     Q_OBJECT
  51
  52 public:
  53     tst_CharAttributes();
  54     virtual ~tst_CharAttributes();
  55
  56 public slots:
  57     void init();
  58     void cleanup();
  59 private slots:
  60     void lineBreaking();
  61     void charWordStopOnLineSeparator();
  62     void charStopForSurrogatePairs();
  63     void thaiWordBreak();
  64 };
  65
  66
  67 tst_CharAttributes::tst_CharAttributes()
  68 {
  69 }
  70
  71 tst_CharAttributes::~tst_CharAttributes()
  72 {
  73 }
  74
  75 void tst_CharAttributes::init()
  76 {
  77 }
  78
  79 void tst_CharAttributes::cleanup()
  80 {
  81 }
  82
  83
  84 void tst_CharAttributes::lineBreaking()
  85 {
  86     struct Breaks {
  87         const char *utf8;
  88         uchar breaks[32];
  89     };
  90     Breaks brks[] = {
  91         { "11", { false, 0xff } },
  92         { "aa", { false, 0xff } },
  93         { "++", { false, 0xff } },
  94         { "--", { false, 0xff } },
  95         { "((", { false, 0xff } },
  96         { "))", { false, 0xff } },
  97         { "..", { false, 0xff } },
  98         { "\"\"", { false, 0xff } },
  99         { "$$", { false, 0xff } },
 100         { "!!", { false, 0xff } },
 101         { "??", { false, 0xff } },
 102         { ",,", { false, 0xff } },
 103
 104         { ")()", { true, false, 0xff } },
 105         { "?!?", { false, false, 0xff } },
 106         { ".,.", { false, false, 0xff } },
 107         { "+-+", { false, false, 0xff } },
 108         { "+=+", { false, false, 0xff } },
 109         { "+(+", { false, false, 0xff } },
 110         { "+)+", { false, false, 0xff } },
 111
 112         { "a b", { false, true, 0xff } },
 113         { "a(b", { false, false, 0xff } },
 114         { "a)b", { false, false, 0xff } },
 115         { "a-b", { false, true, 0xff } },
 116         { "a.b", { false, false, 0xff } },
 117         { "a+b", { false, false, 0xff } },
 118         { "a?b", { false, false, 0xff } },
 119         { "a!b", { false, false, 0xff } },
 120         { "a$b", { false, false, 0xff } },
 121         { "a,b", { false, false, 0xff } },
 122         { "a/b", { false, false, 0xff } },
 123         { "1/2", { false, false, 0xff } },
 124         { "./.", { false, false, 0xff } },
 125         { ",/,", { false, false, 0xff } },
 126         { "!/!", { false, false, 0xff } },
 127         { "\\/\\", { false, false, 0xff } },
 128         { "1 2", { false, true, 0xff } },
 129         { "1(2", { false, false, 0xff } },
 130         { "1)2", { false, false, 0xff } },
 131         { "1-2", { false, false, 0xff } },
 132         { "1.2", { false, false, 0xff } },
 133         { "1+2", { false, false, 0xff } },
 134         { "1?2", { false, true, 0xff } },
 135         { "1!2", { false, true, 0xff } },
 136         { "1$2", { false, false, 0xff } },
 137         { "1,2", { false, false, 0xff } },
 138         { "1/2", { false, false, 0xff } },
 139         { "\330\260\331\216\331\204\331\220\331\203\331\216", { false, false, false, false, false, 0xff } },
 140         { "\330\247\331\204\331\205 \330\247\331\204\331\205", { false, false, false, true, false, false, 0xff } },
 141         { "1#2", { false, false, 0xff } },
 142         { "!#!", { false, false, 0xff } },
 143         { 0, {} }
 144     };
 145     Breaks *b = brks;
 146     while (b->utf8) {
 147         QString str = QString::fromUtf8(b->utf8);
 148
 149         QVector<HB_CharAttributes> attrs = getCharAttributes(str);
 150
 151         int i;
 152         for (i = 0; i < (int)str.length() - 1; ++i) {
 153             QVERIFY(b->breaks[i] != 0xff);
 154             if ( (attrs[i].lineBreakType != HB_NoBreak) != (bool)b->breaks[i] ) {
 155                 qDebug("test case \"%s\" failed at char %d; break type: %d", b->utf8, i, attrs[i].lineBreakType);
 156                 QCOMPARE( (attrs[i].lineBreakType != HB_NoBreak), (bool)b->breaks[i] );
 157             }
 158         }
 159         QVERIFY(attrs[i].lineBreakType == HB_ForcedBreak);
 160         QCOMPARE(b->breaks[i], (uchar)0xff);
 161         ++b;
 162     }
 163 }
 164
 165 void tst_CharAttributes::charWordStopOnLineSeparator()
 166 {
 167     const QChar lineSeparator(QChar::LineSeparator);
 168     QString txt;
 169     txt.append(lineSeparator);
 170     txt.append(lineSeparator);
 171     QVector<HB_CharAttributes> attrs = getCharAttributes(txt);
 172     QVERIFY(attrs[1].charStop);
 173 }
 174
 175 void tst_CharAttributes::charStopForSurrogatePairs()
 176 {
 177     QString txt;
 178     txt.append("a");
 179     txt.append(0xd87e);
 180     txt.append(0xdc25);
 181     txt.append("b");
 182     QVector<HB_CharAttributes> attrs = getCharAttributes(txt);
 183     QVERIFY(attrs[0].charStop);
 184     QVERIFY(attrs[1].charStop);
 185     QVERIFY(!attrs[2].charStop);
 186     QVERIFY(attrs[3].charStop);
 187 }
 188
 189 void tst_CharAttributes::thaiWordBreak()
 190 {
 191     // สวัสดีครับ นี่เป็นการงทดสอบตัวเอ
 192     QTextCodec *codec = QTextCodec::codecForMib(2259);
 193     QString txt = codec->toUnicode(QByteArray("\xca\xc7\xd1\xca\xb4\xd5\xa4\xc3\xd1\xba\x20\xb9\xd5\xe8\xe0\xbb\xe7\xb9\xa1\xd2\xc3\xb7\xb4\xca\xcd\xba\xb5\xd1\xc7\xe0\xcd\xa7"));
 194
 195
 196     QCOMPARE(txt.length(), 32);
 197     QVector<HB_CharAttributes> attrs = getCharAttributes(txt, HB_Script_Thai);
 198     QVERIFY(attrs[0].lineBreakType == HB_NoBreak);
 199     QVERIFY(attrs[1].lineBreakType == HB_NoBreak);
 200     QVERIFY(attrs[2].lineBreakType == HB_NoBreak);
 201     QVERIFY(attrs[3].lineBreakType == HB_NoBreak);
 202     QVERIFY(attrs[4].lineBreakType == HB_NoBreak);
 203     QVERIFY(attrs[5].lineBreakType == HB_Break);
 204     QVERIFY(attrs[6].lineBreakType == HB_NoBreak);
 205     QVERIFY(attrs[7].lineBreakType == HB_NoBreak);
 206     QVERIFY(attrs[8].lineBreakType == HB_NoBreak);
 207     QVERIFY(attrs[9].lineBreakType == HB_NoBreak);
 208     QVERIFY(attrs[10].lineBreakType == HB_Break);
 209     QVERIFY(attrs[11].lineBreakType == HB_NoBreak);
 210     QVERIFY(attrs[12].lineBreakType == HB_NoBreak);
 211     QVERIFY(attrs[13].lineBreakType == HB_Break);
 212     QVERIFY(attrs[14].lineBreakType == HB_NoBreak);
 213     QVERIFY(attrs[15].lineBreakType == HB_NoBreak);
 214     QVERIFY(attrs[16].lineBreakType == HB_NoBreak);
 215     QVERIFY(attrs[17].lineBreakType == HB_Break);
 216     QVERIFY(attrs[18].lineBreakType == HB_NoBreak);
 217     QVERIFY(attrs[19].lineBreakType == HB_NoBreak);
 218     QVERIFY(attrs[20].lineBreakType == HB_Break);
 219     QVERIFY(attrs[21].lineBreakType == HB_NoBreak);
 220     QVERIFY(attrs[22].lineBreakType == HB_NoBreak);
 221     QVERIFY(attrs[23].lineBreakType == HB_NoBreak);
 222     QVERIFY(attrs[24].lineBreakType == HB_NoBreak);
 223     QVERIFY(attrs[25].lineBreakType == HB_Break);
 224     QVERIFY(attrs[26].lineBreakType == HB_NoBreak);
 225     for (int i = 27; i < 32; ++i)
 226         QVERIFY(attrs[i].lineBreakType == HB_NoBreak);
 227 }
 228
 229 QTEST_MAIN(tst_CharAttributes)
 230 #include "main.moc"