1 /*
2 * Copyright (C) 2011 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 "string-inl.h"
18
19 #include "arch/memcmp16.h"
20 #include "array.h"
21 #include "class-inl.h"
22 #include "gc/accounting/card_table-inl.h"
23 #include "intern_table.h"
24 #include "object-inl.h"
25 #include "runtime.h"
26 #include "handle_scope-inl.h"
27 #include "thread.h"
28 #include "utf-inl.h"
29
30 namespace art {
31 namespace mirror {
32
33 // TODO: get global references for these
34 GcRoot<Class> String::java_lang_String_;
35
FastIndexOf(int32_t ch,int32_t start)36 int32_t String::FastIndexOf(int32_t ch, int32_t start) {
37 int32_t count = GetLength();
38 if (start < 0) {
39 start = 0;
40 } else if (start > count) {
41 start = count;
42 }
43 const uint16_t* chars = GetCharArray()->GetData() + GetOffset();
44 const uint16_t* p = chars + start;
45 const uint16_t* end = chars + count;
46 while (p < end) {
47 if (*p++ == ch) {
48 return (p - 1) - chars;
49 }
50 }
51 return -1;
52 }
53
SetClass(Class * java_lang_String)54 void String::SetClass(Class* java_lang_String) {
55 CHECK(java_lang_String_.IsNull());
56 CHECK(java_lang_String != NULL);
57 java_lang_String_ = GcRoot<Class>(java_lang_String);
58 }
59
ResetClass()60 void String::ResetClass() {
61 CHECK(!java_lang_String_.IsNull());
62 java_lang_String_ = GcRoot<Class>(nullptr);
63 }
64
ComputeHashCode()65 int32_t String::ComputeHashCode() {
66 const int32_t hash_code = ComputeUtf16Hash(GetCharArray(), GetOffset(), GetLength());
67 SetHashCode(hash_code);
68 return hash_code;
69 }
70
GetUtfLength()71 int32_t String::GetUtfLength() {
72 return CountUtf8Bytes(GetCharArray()->GetData() + GetOffset(), GetLength());
73 }
74
AllocFromUtf16(Thread * self,int32_t utf16_length,const uint16_t * utf16_data_in,int32_t hash_code)75 String* String::AllocFromUtf16(Thread* self,
76 int32_t utf16_length,
77 const uint16_t* utf16_data_in,
78 int32_t hash_code) {
79 CHECK(utf16_data_in != nullptr || utf16_length == 0);
80 String* string = Alloc(self, utf16_length);
81 if (UNLIKELY(string == nullptr)) {
82 return nullptr;
83 }
84 CharArray* array = const_cast<CharArray*>(string->GetCharArray());
85 if (UNLIKELY(array == nullptr)) {
86 return nullptr;
87 }
88 memcpy(array->GetData(), utf16_data_in, utf16_length * sizeof(uint16_t));
89 if (hash_code != 0) {
90 DCHECK_EQ(hash_code, ComputeUtf16Hash(utf16_data_in, utf16_length));
91 string->SetHashCode(hash_code);
92 } else {
93 string->ComputeHashCode();
94 }
95 return string;
96 }
97
AllocFromModifiedUtf8(Thread * self,const char * utf)98 String* String::AllocFromModifiedUtf8(Thread* self, const char* utf) {
99 DCHECK(utf != nullptr);
100 size_t char_count = CountModifiedUtf8Chars(utf);
101 return AllocFromModifiedUtf8(self, char_count, utf);
102 }
103
AllocFromModifiedUtf8(Thread * self,int32_t utf16_length,const char * utf8_data_in)104 String* String::AllocFromModifiedUtf8(Thread* self, int32_t utf16_length,
105 const char* utf8_data_in) {
106 String* string = Alloc(self, utf16_length);
107 if (UNLIKELY(string == nullptr)) {
108 return nullptr;
109 }
110 uint16_t* utf16_data_out =
111 const_cast<uint16_t*>(string->GetCharArray()->GetData());
112 ConvertModifiedUtf8ToUtf16(utf16_data_out, utf8_data_in);
113 string->ComputeHashCode();
114 return string;
115 }
116
Alloc(Thread * self,int32_t utf16_length)117 String* String::Alloc(Thread* self, int32_t utf16_length) {
118 StackHandleScope<1> hs(self);
119 Handle<CharArray> array(hs.NewHandle(CharArray::Alloc(self, utf16_length)));
120 if (UNLIKELY(array.Get() == nullptr)) {
121 return nullptr;
122 }
123 return Alloc(self, array);
124 }
125
Alloc(Thread * self,Handle<CharArray> array)126 String* String::Alloc(Thread* self, Handle<CharArray> array) {
127 // Hold reference in case AllocObject causes GC.
128 String* string = down_cast<String*>(GetJavaLangString()->AllocObject(self));
129 if (LIKELY(string != nullptr)) {
130 string->SetArray(array.Get());
131 string->SetCount(array->GetLength());
132 }
133 return string;
134 }
135
Equals(String * that)136 bool String::Equals(String* that) {
137 if (this == that) {
138 // Quick reference equality test
139 return true;
140 } else if (that == NULL) {
141 // Null isn't an instanceof anything
142 return false;
143 } else if (this->GetLength() != that->GetLength()) {
144 // Quick length inequality test
145 return false;
146 } else {
147 // Note: don't short circuit on hash code as we're presumably here as the
148 // hash code was already equal
149 for (int32_t i = 0; i < that->GetLength(); ++i) {
150 if (this->CharAt(i) != that->CharAt(i)) {
151 return false;
152 }
153 }
154 return true;
155 }
156 }
157
Equals(const uint16_t * that_chars,int32_t that_offset,int32_t that_length)158 bool String::Equals(const uint16_t* that_chars, int32_t that_offset, int32_t that_length) {
159 if (this->GetLength() != that_length) {
160 return false;
161 } else {
162 for (int32_t i = 0; i < that_length; ++i) {
163 if (this->CharAt(i) != that_chars[that_offset + i]) {
164 return false;
165 }
166 }
167 return true;
168 }
169 }
170
Equals(const char * modified_utf8)171 bool String::Equals(const char* modified_utf8) {
172 for (int32_t i = 0; i < GetLength(); ++i) {
173 uint16_t ch = GetUtf16FromUtf8(&modified_utf8);
174 if (ch == '\0' || ch != CharAt(i)) {
175 return false;
176 }
177 }
178 return *modified_utf8 == '\0';
179 }
180
Equals(const StringPiece & modified_utf8)181 bool String::Equals(const StringPiece& modified_utf8) {
182 const char* p = modified_utf8.data();
183 for (int32_t i = 0; i < GetLength(); ++i) {
184 uint16_t ch = GetUtf16FromUtf8(&p);
185 if (ch != CharAt(i)) {
186 return false;
187 }
188 }
189 return true;
190 }
191
192 // Create a modified UTF-8 encoded std::string from a java/lang/String object.
ToModifiedUtf8()193 std::string String::ToModifiedUtf8() {
194 const uint16_t* chars = GetCharArray()->GetData() + GetOffset();
195 size_t byte_count = GetUtfLength();
196 std::string result(byte_count, static_cast<char>(0));
197 ConvertUtf16ToModifiedUtf8(&result[0], chars, GetLength());
198 return result;
199 }
200
CompareTo(String * rhs)201 int32_t String::CompareTo(String* rhs) {
202 // Quick test for comparison of a string with itself.
203 String* lhs = this;
204 if (lhs == rhs) {
205 return 0;
206 }
207 // TODO: is this still true?
208 // The annoying part here is that 0x00e9 - 0xffff != 0x00ea,
209 // because the interpreter converts the characters to 32-bit integers
210 // *without* sign extension before it subtracts them (which makes some
211 // sense since "char" is unsigned). So what we get is the result of
212 // 0x000000e9 - 0x0000ffff, which is 0xffff00ea.
213 int32_t lhsCount = lhs->GetLength();
214 int32_t rhsCount = rhs->GetLength();
215 int32_t countDiff = lhsCount - rhsCount;
216 int32_t minCount = (countDiff < 0) ? lhsCount : rhsCount;
217 const uint16_t* lhsChars = lhs->GetCharArray()->GetData() + lhs->GetOffset();
218 const uint16_t* rhsChars = rhs->GetCharArray()->GetData() + rhs->GetOffset();
219 int32_t otherRes = MemCmp16(lhsChars, rhsChars, minCount);
220 if (otherRes != 0) {
221 return otherRes;
222 }
223 return countDiff;
224 }
225
VisitRoots(RootCallback * callback,void * arg)226 void String::VisitRoots(RootCallback* callback, void* arg) {
227 java_lang_String_.VisitRootIfNonNull(callback, arg, RootInfo(kRootStickyClass));
228 }
229
230 } // namespace mirror
231 } // namespace art
232