1 /*
2 * Copyright (C) 2023 The Android Open Source Project
3 *
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
7 *
8 * http://www.apache.org/licenses/LICENSE-2.0
9 *
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
15 */
16
17 #include <gtest/gtest.h>
18
19 #include "ScriptUtils.h"
20 #include "UnicodeUtils.h"
21
22 namespace minikin {
23 namespace {
24
25 struct Result {
Resultminikin::__anon433396a90111::Result26 Result(int start, int end, hb_script_t script) : start(start), end(end), script(script) {}
27 int start;
28 int end;
29 hb_script_t script;
30 };
31
operator ==(const Result & l,const Result & r)32 bool operator==(const Result& l, const Result& r) {
33 return l.start == r.start && l.end == r.end && l.script == r.script;
34 }
35
operator <<(std::ostream & os,const Result & r)36 std::ostream& operator<<(std::ostream& os, const Result& r) {
37 char buf[5] = {};
38 buf[0] = static_cast<char>((r.script >> 24) & 0xFF);
39 buf[1] = static_cast<char>((r.script >> 16) & 0xFF);
40 buf[2] = static_cast<char>((r.script >> 8) & 0xFF);
41 buf[3] = static_cast<char>((r.script) & 0xFF);
42 return os << "(" << r.start << "," << r.end << "): " << buf;
43 }
44
splitByScript(const std::vector<uint16_t> & text,uint32_t start,uint32_t end)45 std::vector<Result> splitByScript(const std::vector<uint16_t>& text, uint32_t start, uint32_t end) {
46 std::vector<Result> result;
47 for (const auto [range, script] : ScriptText(text, start, end)) {
48 result.emplace_back(range.getStart(), range.getEnd(), script);
49 }
50 return result;
51 }
52
splitByScript(const std::string & text,uint32_t start,uint32_t end)53 std::vector<Result> splitByScript(const std::string& text, uint32_t start, uint32_t end) {
54 std::vector<uint16_t> utf16 = utf8ToUtf16(text);
55 return splitByScript(utf16, start, end);
56 }
57
TEST(ScriptUtilsTest,Latin)58 TEST(ScriptUtilsTest, Latin) {
59 auto result = splitByScript("abcde", 0, 5);
60 ASSERT_EQ(1u, result.size());
61 EXPECT_EQ(Result(0, 5, HB_SCRIPT_LATIN), result[0]);
62
63 result = splitByScript("abcde", 0, 3);
64 ASSERT_EQ(1u, result.size());
65 EXPECT_EQ(Result(0, 3, HB_SCRIPT_LATIN), result[0]);
66
67 result = splitByScript("abcde", 2, 5);
68 ASSERT_EQ(1u, result.size());
69 EXPECT_EQ(Result(2, 5, HB_SCRIPT_LATIN), result[0]);
70
71 result = splitByScript("abcde", 2, 3);
72 ASSERT_EQ(1u, result.size());
73 EXPECT_EQ(Result(2, 3, HB_SCRIPT_LATIN), result[0]);
74 }
75
TEST(ScriptUtilsTest,Arabic)76 TEST(ScriptUtilsTest, Arabic) {
77 auto result = splitByScript("\u0645\u0631\u062D\u0628\u064B\u0627", 0, 6);
78 ASSERT_EQ(1u, result.size());
79 EXPECT_EQ(Result(0, 6, HB_SCRIPT_ARABIC), result[0]);
80
81 result = splitByScript("\u0645\u0631\u062D\u0628\u064B\u0627", 0, 3);
82 ASSERT_EQ(1u, result.size());
83 EXPECT_EQ(Result(0, 3, HB_SCRIPT_ARABIC), result[0]);
84
85 result = splitByScript("\u0645\u0631\u062D\u0628\u064B\u0627", 2, 5);
86 ASSERT_EQ(1u, result.size());
87 EXPECT_EQ(Result(2, 5, HB_SCRIPT_ARABIC), result[0]);
88
89 result = splitByScript("\u0645\u0631\u062D\u0628\u064B\u0627", 2, 3);
90 ASSERT_EQ(1u, result.size());
91 EXPECT_EQ(Result(2, 3, HB_SCRIPT_ARABIC), result[0]);
92 }
93
TEST(ScriptUtilsTest,Common)94 TEST(ScriptUtilsTest, Common) {
95 auto result = splitByScript(" ", 0, 5);
96 ASSERT_EQ(1u, result.size());
97 EXPECT_EQ(Result(0, 5, HB_SCRIPT_COMMON), result[0]);
98
99 result = splitByScript(" ", 0, 3);
100 ASSERT_EQ(1u, result.size());
101 EXPECT_EQ(Result(0, 3, HB_SCRIPT_COMMON), result[0]);
102
103 result = splitByScript(" ", 2, 5);
104 ASSERT_EQ(1u, result.size());
105 EXPECT_EQ(Result(2, 5, HB_SCRIPT_COMMON), result[0]);
106
107 result = splitByScript(" ", 2, 3);
108 ASSERT_EQ(1u, result.size());
109 EXPECT_EQ(Result(2, 3, HB_SCRIPT_COMMON), result[0]);
110 }
111
TEST(ScriptUtilsTest,InheritOrCommon)112 TEST(ScriptUtilsTest, InheritOrCommon) {
113 // Parens are inherit which is inherit from the previous script. If there is no character
114 // before, use the next non-inherit type of script.
115 auto result = splitByScript("(abc)", 0, 5);
116 ASSERT_EQ(1u, result.size());
117 EXPECT_EQ(Result(0, 5, HB_SCRIPT_LATIN), result[0]);
118
119 result = splitByScript("[(b)]", 0, 5);
120 ASSERT_EQ(1u, result.size());
121 EXPECT_EQ(Result(0, 5, HB_SCRIPT_LATIN), result[0]);
122
123 result = splitByScript("[(b)]", 0, 2);
124 ASSERT_EQ(1u, result.size());
125 EXPECT_EQ(Result(0, 2, HB_SCRIPT_COMMON), result[0]);
126 }
127
TEST(ScriptUtilsTest,MultiScript_InheritOrCommon)128 TEST(ScriptUtilsTest, MultiScript_InheritOrCommon) {
129 auto result = splitByScript("a(\u0645)e", 0, 5);
130 EXPECT_EQ(Result(0, 2, HB_SCRIPT_LATIN), result[0]);
131 EXPECT_EQ(Result(2, 4, HB_SCRIPT_ARABIC), result[1]);
132 EXPECT_EQ(Result(4, 5, HB_SCRIPT_LATIN), result[2]);
133 }
134
TEST(ScriptUtilsTest,MultiScript_NoInheritOrCommon)135 TEST(ScriptUtilsTest, MultiScript_NoInheritOrCommon) {
136 auto result = splitByScript("a\u0645b\u0631c", 0, 5);
137 EXPECT_EQ(Result(0, 1, HB_SCRIPT_LATIN), result[0]);
138 EXPECT_EQ(Result(1, 2, HB_SCRIPT_ARABIC), result[1]);
139 EXPECT_EQ(Result(2, 3, HB_SCRIPT_LATIN), result[2]);
140 EXPECT_EQ(Result(3, 4, HB_SCRIPT_ARABIC), result[3]);
141 EXPECT_EQ(Result(4, 5, HB_SCRIPT_LATIN), result[4]);
142 }
143
TEST(ScriptUtilsTest,SurrogatePair)144 TEST(ScriptUtilsTest, SurrogatePair) {
145 auto result = splitByScript(std::vector<uint16_t>({0xD83C, 0xDFF3, 0xD83C, 0xDFF3}), 0, 4);
146 ASSERT_EQ(1u, result.size());
147 EXPECT_EQ(Result(0, 4, HB_SCRIPT_COMMON), result[0]);
148
149 result = splitByScript(std::vector<uint16_t>({0xD83C, 0xDFF3, 0xD83C, 0xDFF3}), 0, 3);
150 ASSERT_EQ(1u, result.size());
151 EXPECT_EQ(Result(0, 3, HB_SCRIPT_COMMON), result[0]);
152
153 result = splitByScript(std::vector<uint16_t>({0xD83C, 0xDFF3, 0xD83C, 0xDFF3}), 1, 4);
154 ASSERT_EQ(1u, result.size());
155 EXPECT_EQ(Result(1, 4, HB_SCRIPT_COMMON), result[0]);
156
157 result = splitByScript(std::vector<uint16_t>({0xD83C, 0xDFF3, 0xD83C, 0xDFF3}), 1, 3);
158 ASSERT_EQ(1u, result.size());
159 EXPECT_EQ(Result(1, 3, HB_SCRIPT_COMMON), result[0]);
160 }
161 } // namespace
162 } // namespace minikin
163