/src/icu/icu4c/source/common/emojiprops.cpp
Line | Count | Source (jump to first uncovered line) |
1 | | // © 2021 and later: Unicode, Inc. and others. |
2 | | // License & terms of use: https://www.unicode.org/copyright.html |
3 | | |
4 | | // emojiprops.cpp |
5 | | // created: 2021sep04 Markus W. Scherer |
6 | | |
7 | | #include "unicode/utypes.h" |
8 | | #include "unicode/uchar.h" |
9 | | #include "unicode/ucharstrie.h" |
10 | | #include "unicode/ucptrie.h" |
11 | | #include "unicode/udata.h" |
12 | | #include "unicode/ustringtrie.h" |
13 | | #include "unicode/utf16.h" |
14 | | #include "emojiprops.h" |
15 | | #include "ucln.h" |
16 | | #include "ucln_cmn.h" |
17 | | #include "umutex.h" |
18 | | #include "uset_imp.h" |
19 | | |
20 | | U_NAMESPACE_BEGIN |
21 | | |
22 | | namespace { |
23 | | |
24 | | EmojiProps *singleton = nullptr; |
25 | | icu::UInitOnce emojiInitOnce {}; |
26 | | |
27 | 0 | UBool U_CALLCONV emojiprops_cleanup() { |
28 | 0 | delete singleton; |
29 | 0 | singleton = nullptr; |
30 | 0 | emojiInitOnce.reset(); |
31 | 0 | return true; |
32 | 0 | } |
33 | | |
34 | 0 | void U_CALLCONV initSingleton(UErrorCode &errorCode) { |
35 | 0 | if (U_FAILURE(errorCode)) { return; } |
36 | 0 | singleton = new EmojiProps(errorCode); |
37 | 0 | if (singleton == nullptr) { |
38 | 0 | errorCode = U_MEMORY_ALLOCATION_ERROR; |
39 | 0 | } else if (U_FAILURE(errorCode)) { |
40 | 0 | delete singleton; |
41 | 0 | singleton = nullptr; |
42 | 0 | } |
43 | 0 | ucln_common_registerCleanup(UCLN_COMMON_EMOJIPROPS, emojiprops_cleanup); |
44 | 0 | } |
45 | | |
46 | | // TODO: turn this into a shared helper function |
47 | | // Requires the major version to match, and then requires at least the minor version. |
48 | | UBool udata_isAcceptableMajorMinor( |
49 | 0 | const UDataInfo &info, const char16_t *dataFormat, uint8_t major, uint8_t minor) { |
50 | 0 | return |
51 | 0 | info.size >= 20 && |
52 | 0 | info.isBigEndian == U_IS_BIG_ENDIAN && |
53 | 0 | info.charsetFamily == U_CHARSET_FAMILY && |
54 | 0 | info.dataFormat[0] == dataFormat[0] && |
55 | 0 | info.dataFormat[1] == dataFormat[1] && |
56 | 0 | info.dataFormat[2] == dataFormat[2] && |
57 | 0 | info.dataFormat[3] == dataFormat[3] && |
58 | 0 | info.formatVersion[0] == major && |
59 | 0 | info.formatVersion[1] >= minor; |
60 | 0 | } |
61 | | |
62 | | } // namespace |
63 | | |
64 | 0 | EmojiProps::~EmojiProps() { |
65 | 0 | udata_close(memory); |
66 | 0 | ucptrie_close(cpTrie); |
67 | 0 | } |
68 | | |
69 | | const EmojiProps * |
70 | 0 | EmojiProps::getSingleton(UErrorCode &errorCode) { |
71 | 0 | if (U_FAILURE(errorCode)) { return nullptr; } |
72 | 0 | umtx_initOnce(emojiInitOnce, &initSingleton, errorCode); |
73 | 0 | return singleton; |
74 | 0 | } |
75 | | |
76 | | UBool U_CALLCONV |
77 | | EmojiProps::isAcceptable(void * /*context*/, const char * /*type*/, const char * /*name*/, |
78 | 0 | const UDataInfo *pInfo) { |
79 | 0 | return udata_isAcceptableMajorMinor(*pInfo, u"Emoj", 1, 0); |
80 | 0 | } |
81 | | |
82 | | void |
83 | 0 | EmojiProps::load(UErrorCode &errorCode) { |
84 | 0 | memory = udata_openChoice(nullptr, "icu", "uemoji", isAcceptable, this, &errorCode); |
85 | 0 | if (U_FAILURE(errorCode)) { return; } |
86 | 0 | const uint8_t* inBytes = static_cast<const uint8_t*>(udata_getMemory(memory)); |
87 | 0 | const int32_t* inIndexes = reinterpret_cast<const int32_t*>(inBytes); |
88 | 0 | int32_t indexesLength = inIndexes[IX_CPTRIE_OFFSET] / 4; |
89 | 0 | if (indexesLength <= IX_RGI_EMOJI_ZWJ_SEQUENCE_TRIE_OFFSET) { |
90 | 0 | errorCode = U_INVALID_FORMAT_ERROR; // Not enough indexes. |
91 | 0 | return; |
92 | 0 | } |
93 | | |
94 | 0 | int32_t i = IX_CPTRIE_OFFSET; |
95 | 0 | int32_t offset = inIndexes[i++]; |
96 | 0 | int32_t nextOffset = inIndexes[i]; |
97 | 0 | cpTrie = ucptrie_openFromBinary(UCPTRIE_TYPE_FAST, UCPTRIE_VALUE_BITS_8, |
98 | 0 | inBytes + offset, nextOffset - offset, nullptr, &errorCode); |
99 | 0 | if (U_FAILURE(errorCode)) { |
100 | 0 | return; |
101 | 0 | } |
102 | | |
103 | 0 | for (i = IX_BASIC_EMOJI_TRIE_OFFSET; i <= IX_RGI_EMOJI_ZWJ_SEQUENCE_TRIE_OFFSET; ++i) { |
104 | 0 | offset = inIndexes[i]; |
105 | 0 | nextOffset = inIndexes[i + 1]; |
106 | | // Set/leave nullptr if there is no UCharsTrie. |
107 | 0 | const char16_t* p = nextOffset > offset ? reinterpret_cast<const char16_t*>(inBytes + offset) : nullptr; |
108 | 0 | stringTries[getStringTrieIndex(i)] = p; |
109 | 0 | } |
110 | 0 | } |
111 | | |
112 | | void |
113 | 0 | EmojiProps::addPropertyStarts(const USetAdder *sa, UErrorCode & /*errorCode*/) const { |
114 | | // Add the start code point of each same-value range of the trie. |
115 | 0 | UChar32 start = 0, end; |
116 | 0 | uint32_t value; |
117 | 0 | while ((end = ucptrie_getRange(cpTrie, start, UCPMAP_RANGE_NORMAL, 0, |
118 | 0 | nullptr, nullptr, &value)) >= 0) { |
119 | 0 | sa->add(sa->set, start); |
120 | 0 | start = end + 1; |
121 | 0 | } |
122 | 0 | } |
123 | | |
124 | | UBool |
125 | 0 | EmojiProps::hasBinaryProperty(UChar32 c, UProperty which) { |
126 | 0 | UErrorCode errorCode = U_ZERO_ERROR; |
127 | 0 | const EmojiProps *ep = getSingleton(errorCode); |
128 | 0 | return U_SUCCESS(errorCode) && ep->hasBinaryPropertyImpl(c, which); |
129 | 0 | } |
130 | | |
131 | | UBool |
132 | 0 | EmojiProps::hasBinaryPropertyImpl(UChar32 c, UProperty which) const { |
133 | 0 | if (which < UCHAR_EMOJI || UCHAR_RGI_EMOJI < which) { |
134 | 0 | return false; |
135 | 0 | } |
136 | | // Note: UCHAR_REGIONAL_INDICATOR is a single, hardcoded range implemented elsewhere. |
137 | 0 | static constexpr int8_t bitFlags[] = { |
138 | 0 | BIT_EMOJI, // UCHAR_EMOJI=57 |
139 | 0 | BIT_EMOJI_PRESENTATION, // UCHAR_EMOJI_PRESENTATION=58 |
140 | 0 | BIT_EMOJI_MODIFIER, // UCHAR_EMOJI_MODIFIER=59 |
141 | 0 | BIT_EMOJI_MODIFIER_BASE, // UCHAR_EMOJI_MODIFIER_BASE=60 |
142 | 0 | BIT_EMOJI_COMPONENT, // UCHAR_EMOJI_COMPONENT=61 |
143 | 0 | -1, // UCHAR_REGIONAL_INDICATOR=62 |
144 | 0 | -1, // UCHAR_PREPENDED_CONCATENATION_MARK=63 |
145 | 0 | BIT_EXTENDED_PICTOGRAPHIC, // UCHAR_EXTENDED_PICTOGRAPHIC=64 |
146 | 0 | BIT_BASIC_EMOJI, // UCHAR_BASIC_EMOJI=65 |
147 | 0 | -1, // UCHAR_EMOJI_KEYCAP_SEQUENCE=66 |
148 | 0 | -1, // UCHAR_RGI_EMOJI_MODIFIER_SEQUENCE=67 |
149 | 0 | -1, // UCHAR_RGI_EMOJI_FLAG_SEQUENCE=68 |
150 | 0 | -1, // UCHAR_RGI_EMOJI_TAG_SEQUENCE=69 |
151 | 0 | -1, // UCHAR_RGI_EMOJI_ZWJ_SEQUENCE=70 |
152 | 0 | BIT_BASIC_EMOJI, // UCHAR_RGI_EMOJI=71 |
153 | 0 | }; |
154 | 0 | int32_t bit = bitFlags[which - UCHAR_EMOJI]; |
155 | 0 | if (bit < 0) { |
156 | 0 | return false; // not a property that we support in this function |
157 | 0 | } |
158 | 0 | uint8_t bits = UCPTRIE_FAST_GET(cpTrie, UCPTRIE_8, c); |
159 | 0 | return (bits >> bit) & 1; |
160 | 0 | } |
161 | | |
162 | | UBool |
163 | 0 | EmojiProps::hasBinaryProperty(const char16_t *s, int32_t length, UProperty which) { |
164 | 0 | UErrorCode errorCode = U_ZERO_ERROR; |
165 | 0 | const EmojiProps *ep = getSingleton(errorCode); |
166 | 0 | return U_SUCCESS(errorCode) && ep->hasBinaryPropertyImpl(s, length, which); |
167 | 0 | } |
168 | | |
169 | | UBool |
170 | 0 | EmojiProps::hasBinaryPropertyImpl(const char16_t *s, int32_t length, UProperty which) const { |
171 | 0 | if (s == nullptr && length != 0) { return false; } |
172 | 0 | if (length <= 0 && (length == 0 || *s == 0)) { return false; } // empty string |
173 | | // The caller should have delegated single code points to hasBinaryProperty(c, which). |
174 | 0 | if (which < UCHAR_BASIC_EMOJI || UCHAR_RGI_EMOJI < which) { |
175 | 0 | return false; |
176 | 0 | } |
177 | 0 | UProperty firstProp = which, lastProp = which; |
178 | 0 | if (which == UCHAR_RGI_EMOJI) { |
179 | | // RGI_Emoji is the union of the other emoji properties of strings. |
180 | 0 | firstProp = UCHAR_BASIC_EMOJI; |
181 | 0 | lastProp = UCHAR_RGI_EMOJI_ZWJ_SEQUENCE; |
182 | 0 | } |
183 | 0 | for (int32_t prop = firstProp; prop <= lastProp; ++prop) { |
184 | 0 | const char16_t *trieUChars = stringTries[prop - UCHAR_BASIC_EMOJI]; |
185 | 0 | if (trieUChars != nullptr) { |
186 | 0 | UCharsTrie trie(trieUChars); |
187 | 0 | UStringTrieResult result = trie.next(s, length); |
188 | 0 | if (USTRINGTRIE_HAS_VALUE(result)) { |
189 | 0 | return true; |
190 | 0 | } |
191 | 0 | } |
192 | 0 | } |
193 | 0 | return false; |
194 | 0 | } |
195 | | |
196 | | void |
197 | 0 | EmojiProps::addStrings(const USetAdder *sa, UProperty which, UErrorCode &errorCode) const { |
198 | 0 | if (U_FAILURE(errorCode)) { return; } |
199 | 0 | if (which < UCHAR_BASIC_EMOJI || UCHAR_RGI_EMOJI < which) { |
200 | 0 | return; |
201 | 0 | } |
202 | 0 | UProperty firstProp = which, lastProp = which; |
203 | 0 | if (which == UCHAR_RGI_EMOJI) { |
204 | | // RGI_Emoji is the union of the other emoji properties of strings. |
205 | 0 | firstProp = UCHAR_BASIC_EMOJI; |
206 | 0 | lastProp = UCHAR_RGI_EMOJI_ZWJ_SEQUENCE; |
207 | 0 | } |
208 | 0 | for (int32_t prop = firstProp; prop <= lastProp; ++prop) { |
209 | 0 | const char16_t *trieUChars = stringTries[prop - UCHAR_BASIC_EMOJI]; |
210 | 0 | if (trieUChars != nullptr) { |
211 | 0 | UCharsTrie::Iterator iter(trieUChars, 0, errorCode); |
212 | 0 | while (iter.next(errorCode)) { |
213 | 0 | const UnicodeString &s = iter.getString(); |
214 | 0 | sa->addString(sa->set, s.getBuffer(), s.length()); |
215 | 0 | } |
216 | 0 | } |
217 | 0 | } |
218 | 0 | } |
219 | | |
220 | | U_NAMESPACE_END |