1/*
2 * Copyright (C) 2015 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#include <UnicodeUtils.h>
19#include <minikin/GraphemeBreak.h>
20
21using namespace android;
22
23bool IsBreak(const char* src) {
24    const size_t BUF_SIZE = 256;
25    uint16_t buf[BUF_SIZE];
26    size_t offset;
27    size_t size;
28    ParseUnicode(buf, BUF_SIZE, src, &size, &offset);
29    return GraphemeBreak::isGraphemeBreak(buf, 0, size, offset);
30}
31
32TEST(GraphemeBreak, utf16) {
33    EXPECT_FALSE(IsBreak("U+D83C | U+DC31"));  // emoji, U+1F431
34
35    // tests for invalid UTF-16
36    EXPECT_TRUE(IsBreak("U+D800 | U+D800"));  // two leading surrogates
37    EXPECT_TRUE(IsBreak("U+DC00 | U+DC00"));  // two trailing surrogates
38    EXPECT_TRUE(IsBreak("'a' | U+D800"));  // lonely leading surrogate
39    EXPECT_TRUE(IsBreak("U+DC00 | 'a'"));  // lonely trailing surrogate
40    EXPECT_TRUE(IsBreak("U+D800 | 'a'"));  // leading surrogate followed by non-surrogate
41    EXPECT_TRUE(IsBreak("'a' | U+DC00"));  // non-surrogate followed by trailing surrogate
42}
43
44TEST(GraphemeBreak, rules) {
45    // Rule GB1, sot ÷; Rule GB2, ÷ eot
46    EXPECT_TRUE(IsBreak("| 'a'"));
47    EXPECT_TRUE(IsBreak("'a' |"));
48
49    // Rule GB3, CR x LF
50    EXPECT_FALSE(IsBreak("U+000D | U+000A"));  // CR x LF
51
52    // Rule GB4, (Control | CR | LF) ÷
53    EXPECT_TRUE(IsBreak("'a' | U+2028"));  // Line separator
54    EXPECT_TRUE(IsBreak("'a' | U+000D"));  // LF
55    EXPECT_TRUE(IsBreak("'a' | U+000A"));  // CR
56
57    // Rule GB5, ÷ (Control | CR | LF)
58    EXPECT_TRUE(IsBreak("U+2028 | 'a'"));  // Line separator
59    EXPECT_TRUE(IsBreak("U+000D | 'a'"));  // LF
60    EXPECT_TRUE(IsBreak("U+000A | 'a'"));  // CR
61
62    // Rule GB6, L x ( L | V | LV | LVT )
63    EXPECT_FALSE(IsBreak("U+1100 | U+1100"));  // L x L
64    EXPECT_FALSE(IsBreak("U+1100 | U+1161"));  // L x V
65    EXPECT_FALSE(IsBreak("U+1100 | U+AC00"));  // L x LV
66    EXPECT_FALSE(IsBreak("U+1100 | U+AC01"));  // L x LVT
67
68    // Rule GB7, ( LV | V ) x ( V | T )
69    EXPECT_FALSE(IsBreak("U+AC00 | U+1161"));  // LV x V
70    EXPECT_FALSE(IsBreak("U+1161 | U+1161"));  // V x V
71    EXPECT_FALSE(IsBreak("U+AC00 | U+11A8"));  // LV x T
72    EXPECT_FALSE(IsBreak("U+1161 | U+11A8"));  // V x T
73
74    // Rule GB8, ( LVT | T ) x T
75    EXPECT_FALSE(IsBreak("U+AC01 | U+11A8"));  // LVT x T
76    EXPECT_FALSE(IsBreak("U+11A8 | U+11A8"));  // T x T
77
78    // Other hangul pairs not counted above _are_ breaks (GB10)
79    EXPECT_TRUE(IsBreak("U+AC00 | U+1100"));  // LV x L
80    EXPECT_TRUE(IsBreak("U+AC01 | U+1100"));  // LVT x L
81    EXPECT_TRUE(IsBreak("U+11A8 | U+1100"));  // T x L
82    EXPECT_TRUE(IsBreak("U+11A8 | U+AC00"));  // T x LV
83    EXPECT_TRUE(IsBreak("U+11A8 | U+AC01"));  // T x LVT
84
85    // Rule GB8a, Regional_Indicator x Regional_Indicator
86    EXPECT_FALSE(IsBreak("U+1F1FA | U+1F1F8"));
87    EXPECT_TRUE(IsBreak("U+1F1FA U+1F1F8 | U+1F1FA U+1F1F8")); // Regional indicator pair (flag)
88    EXPECT_FALSE(IsBreak("U+1F1FA | U+1F1F8 U+1F1FA U+1F1F8")); // Regional indicator pair (flag)
89    EXPECT_FALSE(IsBreak("U+1F1FA U+1F1F8 U+1F1FA | U+1F1F8")); // Regional indicator pair (flag)
90
91    EXPECT_TRUE(IsBreak("U+1F1FA U+1F1F8 | U+1F1FA"));  // Regional indicator pair (flag)
92    EXPECT_FALSE(IsBreak("U+1F1FA | U+1F1F8 U+1F1FA"));  // Regional indicator pair (flag)
93
94    EXPECT_TRUE(IsBreak("'a' U+1F1FA U+1F1F8 | U+1F1FA"));  // Regional indicator pair (flag)
95    EXPECT_FALSE(IsBreak("'a' U+1F1FA | U+1F1F8 U+1F1FA"));  // Regional indicator pair (flag)
96
97    EXPECT_TRUE(
98            IsBreak("'a' U+1F1FA U+1F1F8 | U+1F1FA U+1F1F8"));  // Regional indicator pair (flag)
99    EXPECT_FALSE(
100            IsBreak("'a' U+1F1FA | U+1F1F8 U+1F1FA U+1F1F8"));  // Regional indicator pair (flag)
101    EXPECT_FALSE(
102            IsBreak("'a' U+1F1FA U+1F1F8 U+1F1FA | U+1F1F8"));  // Regional indicator pair (flag)
103
104    // Rule GB9, x Extend
105    EXPECT_FALSE(IsBreak("'a' | U+0301"));  // combining accent
106    // Rule GB9a, x SpacingMark
107    EXPECT_FALSE(IsBreak("U+0915 | U+093E"));  // KA, AA (spacing mark)
108    // Rule GB9b, Prepend x
109    // see tailoring test for prepend, as current ICU doesn't have any characters in the class
110
111    // Rule GB10, Any ÷ Any
112    EXPECT_TRUE(IsBreak("'a' | 'b'"));
113    EXPECT_TRUE(IsBreak("'f' | 'i'"));  // probable ligature
114    EXPECT_TRUE(IsBreak("U+0644 | U+0627"));  // probable ligature, lam + alef
115    EXPECT_TRUE(IsBreak("U+4E00 | U+4E00"));  // CJK ideographs
116    EXPECT_TRUE(IsBreak("'a' | U+1F1FA U+1F1F8"));  // Regional indicator pair (flag)
117    EXPECT_TRUE(IsBreak("U+1F1FA U+1F1F8 | 'a'"));  // Regional indicator pair (flag)
118}
119
120TEST(GraphemeBreak, tailoring) {
121    // control characters that we interpret as "extend"
122    EXPECT_FALSE(IsBreak("'a' | U+00AD"));  // soft hyphen
123    EXPECT_FALSE(IsBreak("'a' | U+200B"));  // zwsp
124    EXPECT_FALSE(IsBreak("'a' | U+200E"));  // lrm
125    EXPECT_FALSE(IsBreak("'a' | U+202A"));  // lre
126    EXPECT_FALSE(IsBreak("'a' | U+E0041"));  // tag character
127
128    // UTC-approved characters for the Prepend class
129    EXPECT_FALSE(IsBreak("U+06DD | U+0661"));  // arabic subtending mark + digit one
130
131    EXPECT_TRUE(IsBreak("U+0E01 | U+0E33"));  // Thai sara am
132
133    // virama is not a grapheme break, but "pure killer" is
134    EXPECT_FALSE(IsBreak("U+0915 | U+094D U+0915"));  // Devanagari ka+virama+ka
135    EXPECT_FALSE(IsBreak("U+0915 U+094D | U+0915"));  // Devanagari ka+virama+ka
136    EXPECT_FALSE(IsBreak("U+0E01 | U+0E3A U+0E01"));  // thai phinthu = pure killer
137    EXPECT_TRUE(IsBreak("U+0E01 U+0E3A | U+0E01"));  // thai phinthu = pure killer
138
139    // suppress grapheme breaks in zwj emoji sequences, see
140    // http://www.unicode.org/emoji/charts/emoji-zwj-sequences.html
141    EXPECT_FALSE(IsBreak("U+1F469 U+200D | U+2764 U+FE0F U+200D U+1F48B U+200D U+1F468"));
142    EXPECT_FALSE(IsBreak("U+1F469 U+200D U+2764 U+FE0F U+200D | U+1F48B U+200D U+1F468"));
143    EXPECT_FALSE(IsBreak("U+1F469 U+200D U+2764 U+FE0F U+200D U+1F48B U+200D | U+1F468"));
144    EXPECT_FALSE(IsBreak("U+1F468 U+200D | U+1F469 U+200D U+1F466"));
145    EXPECT_FALSE(IsBreak("U+1F468 U+200D U+1F469 U+200D | U+1F466"));
146    EXPECT_FALSE(IsBreak("U+1F469 U+200D | U+1F469 U+200D U+1F467 U+200D U+1F466"));
147    EXPECT_FALSE(IsBreak("U+1F469 U+200D U+1F469 U+200D | U+1F467 U+200D U+1F466"));
148    EXPECT_FALSE(IsBreak("U+1F469 U+200D U+1F469 U+200D U+1F467 U+200D | U+1F466"));
149    EXPECT_FALSE(IsBreak("U+1F441 U+200D | U+1F5E8"));
150
151    // Do not break before and after zwj with all kind of emoji characters.
152    EXPECT_FALSE(IsBreak("U+1F431 | U+200D U+1F464"));
153    EXPECT_FALSE(IsBreak("U+1F431 U+200D | U+1F464"));
154
155    // ARABIC LETTER BEH + ZWJ + heart, not a zwj emoji sequence, so we preserve the break
156    EXPECT_TRUE(IsBreak("U+0628 U+200D | U+2764"));
157}
158
159TEST(GraphemeBreak, emojiModifiers) {
160    EXPECT_FALSE(IsBreak("U+261D | U+1F3FB"));  // white up pointing index + modifier
161    EXPECT_FALSE(IsBreak("U+270C | U+1F3FB"));  // victory hand + modifier
162    EXPECT_FALSE(IsBreak("U+1F466 | U+1F3FB"));  // boy + modifier
163    EXPECT_FALSE(IsBreak("U+1F466 | U+1F3FC"));  // boy + modifier
164    EXPECT_FALSE(IsBreak("U+1F466 | U+1F3FD"));  // boy + modifier
165    EXPECT_FALSE(IsBreak("U+1F466 | U+1F3FE"));  // boy + modifier
166    EXPECT_FALSE(IsBreak("U+1F466 | U+1F3FF"));  // boy + modifier
167    EXPECT_FALSE(IsBreak("U+1F918 | U+1F3FF"));  // sign of the horns + modifier
168    EXPECT_FALSE(IsBreak("U+1F933 | U+1F3FF"));  // selfie (Unicode 9) + modifier
169
170    // adding emoji style variation selector doesn't affect grapheme cluster
171    EXPECT_TRUE(IsBreak("U+270C U+FE0E | U+1F3FB"));  // victory hand + text style + modifier
172    EXPECT_FALSE(IsBreak("U+270C U+FE0F | U+1F3FB"));  // heart + emoji style + modifier
173
174    // heart is not an emoji base
175    EXPECT_TRUE(IsBreak("U+2764 | U+1F3FB"));  // heart + modifier
176    EXPECT_TRUE(IsBreak("U+2764 U+FE0E | U+1F3FB"));  // heart + emoji style + modifier
177    EXPECT_TRUE(IsBreak("U+2764 U+FE0F | U+1F3FB"));  // heart + emoji style + modifier
178    EXPECT_TRUE(IsBreak("U+1F3FB | U+1F3FB"));  // modifier + modifier
179
180    // rat is not an emoji modifer
181    EXPECT_TRUE(IsBreak("U+1F466 | U+1F400"));  // boy + rat
182}
183
184TEST(GraphemeBreak, offsets) {
185    uint16_t string[] = { 0x0041, 0x06DD, 0x0045, 0x0301, 0x0049, 0x0301 };
186    EXPECT_TRUE(GraphemeBreak::isGraphemeBreak(string, 2, 3, 2));
187    EXPECT_FALSE(GraphemeBreak::isGraphemeBreak(string, 2, 3, 3));
188    EXPECT_TRUE(GraphemeBreak::isGraphemeBreak(string, 2, 3, 4));
189    EXPECT_TRUE(GraphemeBreak::isGraphemeBreak(string, 2, 3, 5));
190}
191