1 #include <QtCore/QScopedPointer>
2 #include <QtTest/QtTest>
3
4 #include <poppler-private.h>
5
6 #include <cstring>
7 #include <cstdint> // for uint16_t
8
9 #include "GlobalParams.h"
10 #include "UnicodeTypeTable.h"
11 #include "UTF.h"
12
13 class TestUTFConversion : public QObject
14 {
15 Q_OBJECT
16 public:
TestUTFConversion(QObject * parent=nullptr)17 explicit TestUTFConversion(QObject *parent = nullptr) : QObject(parent) { }
18 private slots:
19 void testUTF_data();
20 void testUTF();
21 void testUnicodeToAscii7();
22 void testUnicodeLittleEndian();
23 };
24
compare(const char * a,const char * b)25 static bool compare(const char *a, const char *b)
26 {
27 return strcmp(a, b) == 0;
28 }
29
compare(const uint16_t * a,const uint16_t * b)30 static bool compare(const uint16_t *a, const uint16_t *b)
31 {
32 while (*a && *b) {
33 if (*a++ != *b++)
34 return false;
35 }
36 return *a == *b;
37 }
38
compare(const Unicode * a,const char * b,int len)39 static bool compare(const Unicode *a, const char *b, int len)
40 {
41 for (int i = 0; i < len; i++) {
42 if (a[i] != (Unicode)b[i])
43 return false;
44 }
45
46 return true;
47 }
48
compare(const Unicode * a,const uint16_t * b,int len)49 static bool compare(const Unicode *a, const uint16_t *b, int len)
50 {
51 for (int i = 0; i < len; i++) {
52 if (a[i] != b[i])
53 return false;
54 }
55
56 return true;
57 }
58
testUTF_data()59 void TestUTFConversion::testUTF_data()
60 {
61 QTest::addColumn<QString>("s");
62
63 QTest::newRow("<empty>") << QString(QLatin1String(""));
64 QTest::newRow("a") << QStringLiteral("a");
65 QTest::newRow("abc") << QStringLiteral("abc");
66 QTest::newRow("Latin") << QStringLiteral("Vitrum edere possum; mihi non nocet");
67 QTest::newRow("Greek") << QStringLiteral("Μπορώ να φάω σπασμένα γυαλιά χωρίς να πάθω τίποτα");
68 QTest::newRow("Icelandic") << QStringLiteral("Ég get etið gler án þess að meiða mig");
69 QTest::newRow("Russian") << QStringLiteral("Я могу есть стекло, оно мне не вредит.");
70 QTest::newRow("Sanskrit") << QStringLiteral("काचं शक्नोम्यत्तुम् । नोपहिनस्ति माम् ॥");
71 QTest::newRow("Arabic") << QStringLiteral("أنا قادر على أكل الزجاج و هذا لا يؤلمني");
72 QTest::newRow("Chinese") << QStringLiteral("我能吞下玻璃而不伤身体。");
73 QTest::newRow("Thai") << QStringLiteral("ฉันกินกระจกได้ แต่มันไม่ทำให้ฉันเจ็บ");
74 QTest::newRow("non BMP") << QStringLiteral("");
75 }
76
testUTF()77 void TestUTFConversion::testUTF()
78 {
79 char utf8Buf[1000];
80 char *utf8String;
81 uint16_t utf16Buf[1000];
82 uint16_t *utf16String;
83 int len;
84
85 QFETCH(QString, s);
86 char *str = strdup(s.toUtf8().constData());
87
88 // UTF-8 to UTF-16
89
90 len = utf8CountUtf16CodeUnits(str);
91 QCOMPARE(len, s.size()); // QString size() returns number of code units, not code points
92 Q_ASSERT(len < (int)sizeof(utf16Buf)); // if this fails, make utf16Buf larger
93
94 len = utf8ToUtf16(str, utf16Buf);
95 QVERIFY(compare(utf16Buf, s.utf16()));
96 QCOMPARE(len, s.size());
97
98 utf16String = utf8ToUtf16(str);
99 QVERIFY(compare(utf16String, s.utf16()));
100 free(utf16String);
101
102 std::string sUtf8(str);
103 std::unique_ptr<GooString> gsUtf16_a(utf8ToUtf16WithBom(sUtf8));
104 std::unique_ptr<GooString> gsUtf16_b(Poppler::QStringToUnicodeGooString(s));
105 QCOMPARE(gsUtf16_a->cmp(gsUtf16_b.get()), 0);
106
107 // UTF-16 to UTF-8
108
109 len = utf16CountUtf8Bytes(s.utf16());
110 QCOMPARE(len, (int)strlen(str));
111 Q_ASSERT(len < (int)sizeof(utf8Buf)); // if this fails, make utf8Buf larger
112
113 len = utf16ToUtf8(s.utf16(), utf8Buf);
114 QVERIFY(compare(utf8Buf, str));
115 QCOMPARE(len, (int)strlen(str));
116
117 utf8String = utf16ToUtf8(s.utf16());
118 QVERIFY(compare(utf8String, str));
119 free(utf8String);
120
121 free(str);
122 }
123
testUnicodeToAscii7()124 void TestUTFConversion::testUnicodeToAscii7()
125 {
126 globalParams = std::make_unique<GlobalParams>();
127
128 // Test string is one 'Registered' and twenty 'Copyright' chars
129 // so it's long enough to reproduce the bug given that glibc
130 // malloc() always returns 8-byte aligned memory addresses.
131 GooString *goo = Poppler::QStringToUnicodeGooString(QString::fromUtf8("®©©©©©©©©©©©©©©©©©©©©")); // clazy:exclude=qstring-allocations
132
133 Unicode *in;
134 const int in_len = TextStringToUCS4(goo->toStr(), &in);
135
136 delete goo;
137
138 int in_norm_len;
139 int *in_norm_idx;
140 Unicode *in_norm = unicodeNormalizeNFKC(in, in_len, &in_norm_len, &in_norm_idx, true);
141
142 free(in);
143
144 Unicode *out;
145 int out_len;
146 int *out_ascii_idx;
147
148 unicodeToAscii7(in_norm, in_norm_len, &out, &out_len, in_norm_idx, &out_ascii_idx);
149
150 free(in_norm);
151 free(in_norm_idx);
152
153 // ascii7 conversion: ® -> (R) © -> (c)
154 const char *expected_ascii = (char *)"(R)(c)(c)(c)(c)(c)(c)(c)(c)(c)(c)(c)(c)(c)(c)(c)(c)(c)(c)(c)(c)";
155
156 QCOMPARE(out_len, (int)strlen(expected_ascii));
157 QVERIFY(compare(out, expected_ascii, out_len));
158
159 free(out);
160 free(out_ascii_idx);
161 }
162
testUnicodeLittleEndian()163 void TestUTFConversion::testUnicodeLittleEndian()
164 {
165 uint16_t UTF16LE_hi[5] { 0xFFFE, 0x4800, 0x4900, 0x2100, 0x1126 }; // UTF16-LE "HI!☑"
166 std::string GooUTF16LE(reinterpret_cast<const char *>(UTF16LE_hi), sizeof(UTF16LE_hi));
167
168 uint16_t UTF16BE_hi[5] { 0xFEFF, 0x0048, 0x0049, 0x0021, 0x2611 }; // UTF16-BE "HI!☑"
169 std::string GooUTF16BE(reinterpret_cast<const char *>(UTF16BE_hi), sizeof(UTF16BE_hi));
170
171 // Let's assert both GooString's are different
172 QVERIFY(GooUTF16LE != GooUTF16BE);
173
174 Unicode *UCS4fromLE, *UCS4fromBE;
175 const int len1 = TextStringToUCS4(GooUTF16LE, &UCS4fromLE);
176 const int len2 = TextStringToUCS4(GooUTF16BE, &UCS4fromBE);
177
178 // len is 4 because TextStringToUCS4() removes the two leading Byte Order Mark (BOM) code points
179 QCOMPARE(len1, len2);
180 QCOMPARE(len1, 4);
181
182 // Check that now after conversion, UCS4fromLE and UCS4fromBE are now the same
183 for (int i = 0; i < len1; i++) {
184 QCOMPARE(UCS4fromLE[i], UCS4fromBE[i]);
185 }
186
187 const QString expected = QString::fromUtf8("HI!☑"); // clazy:exclude=qstring-allocations
188
189 // Do some final verifications, checking the strings to be "HI!"
190 QVERIFY(*UCS4fromLE == *UCS4fromBE);
191 QVERIFY(compare(UCS4fromLE, expected.utf16(), len1));
192 QVERIFY(compare(UCS4fromBE, expected.utf16(), len1));
193 }
194
195 QTEST_GUILESS_MAIN(TestUTFConversion)
196 #include "check_utf_conversion.moc"
197