string.cc revision 7940e44f4517de5e2634a7e07d58d0fb26160513
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.h" 18 19#include "array.h" 20#include "gc/accounting/card_table-inl.h" 21#include "intern_table.h" 22#include "object-inl.h" 23#include "runtime.h" 24#include "sirt_ref.h" 25#include "thread.h" 26#include "utf.h" 27 28namespace art { 29namespace mirror { 30 31const CharArray* String::GetCharArray() const { 32 return GetFieldObject<const CharArray*>(ValueOffset(), false); 33} 34 35void String::ComputeHashCode() SHARED_LOCKS_REQUIRED(Locks::mutator_lock_) { 36 SetHashCode(ComputeUtf16Hash(GetCharArray(), GetOffset(), GetLength())); 37} 38 39int32_t String::GetUtfLength() const { 40 return CountUtf8Bytes(GetCharArray()->GetData() + GetOffset(), GetLength()); 41} 42 43int32_t String::FastIndexOf(int32_t ch, int32_t start) const { 44 int32_t count = GetLength(); 45 if (start < 0) { 46 start = 0; 47 } else if (start > count) { 48 start = count; 49 } 50 const uint16_t* chars = GetCharArray()->GetData() + GetOffset(); 51 const uint16_t* p = chars + start; 52 const uint16_t* end = chars + count; 53 while (p < end) { 54 if (*p++ == ch) { 55 return (p - 1) - chars; 56 } 57 } 58 return -1; 59} 60 61void String::SetArray(CharArray* new_array) SHARED_LOCKS_REQUIRED(Locks::mutator_lock_) { 62 DCHECK(new_array != NULL); 63 SetFieldObject(OFFSET_OF_OBJECT_MEMBER(String, array_), new_array, false); 64} 65 66// TODO: get global references for these 67Class* String::java_lang_String_ = NULL; 68 69void String::SetClass(Class* java_lang_String) { 70 CHECK(java_lang_String_ == NULL); 71 CHECK(java_lang_String != NULL); 72 java_lang_String_ = java_lang_String; 73} 74 75void String::ResetClass() { 76 CHECK(java_lang_String_ != NULL); 77 java_lang_String_ = NULL; 78} 79 80String* String::Intern() { 81 return Runtime::Current()->GetInternTable()->InternWeak(this); 82} 83 84int32_t String::GetHashCode() { 85 int32_t result = GetField32(OFFSET_OF_OBJECT_MEMBER(String, hash_code_), false); 86 if (result == 0) { 87 ComputeHashCode(); 88 } 89 result = GetField32(OFFSET_OF_OBJECT_MEMBER(String, hash_code_), false); 90 DCHECK(result != 0 || ComputeUtf16Hash(GetCharArray(), GetOffset(), GetLength()) == 0) 91 << ToModifiedUtf8() << " " << result; 92 return result; 93} 94 95int32_t String::GetLength() const { 96 int32_t result = GetField32(OFFSET_OF_OBJECT_MEMBER(String, count_), false); 97 DCHECK(result >= 0 && result <= GetCharArray()->GetLength()); 98 return result; 99} 100 101uint16_t String::CharAt(int32_t index) const { 102 // TODO: do we need this? Equals is the only caller, and could 103 // bounds check itself. 104 if (index < 0 || index >= count_) { 105 Thread* self = Thread::Current(); 106 ThrowLocation throw_location = self->GetCurrentLocationForThrow(); 107 self->ThrowNewExceptionF(throw_location, "Ljava/lang/StringIndexOutOfBoundsException;", 108 "length=%i; index=%i", count_, index); 109 return 0; 110 } 111 return GetCharArray()->Get(index + GetOffset()); 112} 113 114String* String::AllocFromUtf16(Thread* self, 115 int32_t utf16_length, 116 const uint16_t* utf16_data_in, 117 int32_t hash_code) { 118 CHECK(utf16_data_in != NULL || utf16_length == 0); 119 String* string = Alloc(self, GetJavaLangString(), utf16_length); 120 if (string == NULL) { 121 return NULL; 122 } 123 // TODO: use 16-bit wide memset variant 124 CharArray* array = const_cast<CharArray*>(string->GetCharArray()); 125 if (array == NULL) { 126 return NULL; 127 } 128 for (int i = 0; i < utf16_length; i++) { 129 array->Set(i, utf16_data_in[i]); 130 } 131 if (hash_code != 0) { 132 string->SetHashCode(hash_code); 133 } else { 134 string->ComputeHashCode(); 135 } 136 return string; 137} 138 139 String* String::AllocFromModifiedUtf8(Thread* self, const char* utf) { 140 if (utf == NULL) { 141 return NULL; 142 } 143 size_t char_count = CountModifiedUtf8Chars(utf); 144 return AllocFromModifiedUtf8(self, char_count, utf); 145} 146 147String* String::AllocFromModifiedUtf8(Thread* self, int32_t utf16_length, 148 const char* utf8_data_in) { 149 String* string = Alloc(self, GetJavaLangString(), utf16_length); 150 if (string == NULL) { 151 return NULL; 152 } 153 uint16_t* utf16_data_out = 154 const_cast<uint16_t*>(string->GetCharArray()->GetData()); 155 ConvertModifiedUtf8ToUtf16(utf16_data_out, utf8_data_in); 156 string->ComputeHashCode(); 157 return string; 158} 159 160String* String::Alloc(Thread* self, Class* java_lang_String, int32_t utf16_length) { 161 SirtRef<CharArray> array(self, CharArray::Alloc(self, utf16_length)); 162 if (array.get() == NULL) { 163 return NULL; 164 } 165 return Alloc(self, java_lang_String, array.get()); 166} 167 168String* String::Alloc(Thread* self, Class* java_lang_String, CharArray* array) { 169 // Hold reference in case AllocObject causes GC. 170 SirtRef<CharArray> array_ref(self, array); 171 String* string = down_cast<String*>(java_lang_String->AllocObject(self)); 172 if (string == NULL) { 173 return NULL; 174 } 175 string->SetArray(array); 176 string->SetCount(array->GetLength()); 177 return string; 178} 179 180bool String::Equals(const String* that) const { 181 if (this == that) { 182 // Quick reference equality test 183 return true; 184 } else if (that == NULL) { 185 // Null isn't an instanceof anything 186 return false; 187 } else if (this->GetLength() != that->GetLength()) { 188 // Quick length inequality test 189 return false; 190 } else { 191 // Note: don't short circuit on hash code as we're presumably here as the 192 // hash code was already equal 193 for (int32_t i = 0; i < that->GetLength(); ++i) { 194 if (this->CharAt(i) != that->CharAt(i)) { 195 return false; 196 } 197 } 198 return true; 199 } 200} 201 202bool String::Equals(const uint16_t* that_chars, int32_t that_offset, int32_t that_length) const { 203 if (this->GetLength() != that_length) { 204 return false; 205 } else { 206 for (int32_t i = 0; i < that_length; ++i) { 207 if (this->CharAt(i) != that_chars[that_offset + i]) { 208 return false; 209 } 210 } 211 return true; 212 } 213} 214 215bool String::Equals(const char* modified_utf8) const { 216 for (int32_t i = 0; i < GetLength(); ++i) { 217 uint16_t ch = GetUtf16FromUtf8(&modified_utf8); 218 if (ch == '\0' || ch != CharAt(i)) { 219 return false; 220 } 221 } 222 return *modified_utf8 == '\0'; 223} 224 225bool String::Equals(const StringPiece& modified_utf8) const { 226 if (modified_utf8.size() != GetLength()) { 227 return false; 228 } 229 const char* p = modified_utf8.data(); 230 for (int32_t i = 0; i < GetLength(); ++i) { 231 uint16_t ch = GetUtf16FromUtf8(&p); 232 if (ch != CharAt(i)) { 233 return false; 234 } 235 } 236 return true; 237} 238 239// Create a modified UTF-8 encoded std::string from a java/lang/String object. 240std::string String::ToModifiedUtf8() const { 241 const uint16_t* chars = GetCharArray()->GetData() + GetOffset(); 242 size_t byte_count = GetUtfLength(); 243 std::string result(byte_count, static_cast<char>(0)); 244 ConvertUtf16ToModifiedUtf8(&result[0], chars, GetLength()); 245 return result; 246} 247 248#ifdef HAVE__MEMCMP16 249// "count" is in 16-bit units. 250extern "C" uint32_t __memcmp16(const uint16_t* s0, const uint16_t* s1, size_t count); 251#define MemCmp16 __memcmp16 252#else 253static uint32_t MemCmp16(const uint16_t* s0, const uint16_t* s1, size_t count) { 254 for (size_t i = 0; i < count; i++) { 255 if (s0[i] != s1[i]) { 256 return static_cast<int32_t>(s0[i]) - static_cast<int32_t>(s1[i]); 257 } 258 } 259 return 0; 260} 261#endif 262 263int32_t String::CompareTo(String* rhs) const { 264 // Quick test for comparison of a string with itself. 265 const String* lhs = this; 266 if (lhs == rhs) { 267 return 0; 268 } 269 // TODO: is this still true? 270 // The annoying part here is that 0x00e9 - 0xffff != 0x00ea, 271 // because the interpreter converts the characters to 32-bit integers 272 // *without* sign extension before it subtracts them (which makes some 273 // sense since "char" is unsigned). So what we get is the result of 274 // 0x000000e9 - 0x0000ffff, which is 0xffff00ea. 275 int lhsCount = lhs->GetLength(); 276 int rhsCount = rhs->GetLength(); 277 int countDiff = lhsCount - rhsCount; 278 int minCount = (countDiff < 0) ? lhsCount : rhsCount; 279 const uint16_t* lhsChars = lhs->GetCharArray()->GetData() + lhs->GetOffset(); 280 const uint16_t* rhsChars = rhs->GetCharArray()->GetData() + rhs->GetOffset(); 281 int otherRes = MemCmp16(lhsChars, rhsChars, minCount); 282 if (otherRes != 0) { 283 return otherRes; 284 } 285 return countDiff; 286} 287 288} // namespace mirror 289} // namespace art 290 291