1 /*
2 * Copyright (C) 2010 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 "libcore_util_CharsetUtils.h"
18
19 #include <string.h>
20
21 #include "handle_scope-inl.h"
22 #include "jni/jni_internal.h"
23 #include "mirror/string-inl.h"
24 #include "mirror/string.h"
25 #include "native_util.h"
26 #include "nativehelper/scoped_primitive_array.h"
27 #include "nativehelper/jni_macros.h"
28 #include "scoped_fast_native_object_access-inl.h"
29 #include "unicode/utf16.h"
30
31 namespace art {
32
CharsetUtils_asciiBytesToChars(JNIEnv * env,jclass,jbyteArray javaBytes,jint offset,jint length,jcharArray javaChars)33 static void CharsetUtils_asciiBytesToChars(JNIEnv* env, jclass, jbyteArray javaBytes, jint offset,
34 jint length, jcharArray javaChars) {
35 ScopedByteArrayRO bytes(env, javaBytes);
36 if (bytes.get() == nullptr) {
37 return;
38 }
39 ScopedCharArrayRW chars(env, javaChars);
40 if (chars.get() == nullptr) {
41 return;
42 }
43
44 const jbyte* src = &bytes[offset];
45 jchar* dst = &chars[0];
46 static const jchar REPLACEMENT_CHAR = 0xfffd;
47 for (int i = length - 1; i >= 0; --i) {
48 jchar ch = static_cast<jchar>(*src++ & 0xff);
49 *dst++ = (ch <= 0x7f) ? ch : REPLACEMENT_CHAR;
50 }
51 }
52
CharsetUtils_isoLatin1BytesToChars(JNIEnv * env,jclass,jbyteArray javaBytes,jint offset,jint length,jcharArray javaChars)53 static void CharsetUtils_isoLatin1BytesToChars(JNIEnv* env, jclass, jbyteArray javaBytes,
54 jint offset, jint length, jcharArray javaChars) {
55 ScopedByteArrayRO bytes(env, javaBytes);
56 if (bytes.get() == nullptr) {
57 return;
58 }
59 ScopedCharArrayRW chars(env, javaChars);
60 if (chars.get() == nullptr) {
61 return;
62 }
63
64 const jbyte* src = &bytes[offset];
65 jchar* dst = &chars[0];
66 for (int i = length - 1; i >= 0; --i) {
67 *dst++ = static_cast<jchar>(*src++ & 0xff);
68 }
69 }
70
71 /**
72 * Translates the given characters to US-ASCII or ISO-8859-1 bytes, using the fact that
73 * Unicode code points between U+0000 and U+007f inclusive are identical to US-ASCII, while
74 * U+0000 to U+00ff inclusive are identical to ISO-8859-1.
75 */
charsToBytes(JNIEnv * env,jstring java_string,jint offset,jint length,jchar maxValidChar)76 static jbyteArray charsToBytes(JNIEnv* env, jstring java_string, jint offset, jint length,
77 jchar maxValidChar) {
78 ScopedFastNativeObjectAccess soa(env);
79 StackHandleScope<1> hs(soa.Self());
80 Handle<mirror::String> string(hs.NewHandle(soa.Decode<mirror::String>(java_string)));
81 if (string == nullptr) {
82 return nullptr;
83 }
84
85 ObjPtr<mirror::ByteArray> result = mirror::ByteArray::Alloc(soa.Self(), length);
86 if (result == nullptr) {
87 return nullptr;
88 }
89
90 if (string->IsCompressed()) {
91 // All characters in a compressed string are ASCII and therefore do not need a replacement.
92 DCHECK_GE(maxValidChar, 0x7f);
93 memcpy(result->GetData(), string->GetValueCompressed() + offset, length);
94 } else {
95 const uint16_t* src = string->GetValue() + offset;
96 auto clamp = [maxValidChar](uint16_t c) {
97 return static_cast<jbyte>(dchecked_integral_cast<uint8_t>((c > maxValidChar) ? '?' : c));
98 };
99 std::transform(src, src + length, result->GetData(), clamp);
100 }
101 return soa.AddLocalReference<jbyteArray>(result);
102 }
103
CharsetUtils_toAsciiBytes(JNIEnv * env,jclass,jstring java_string,jint offset,jint length)104 static jbyteArray CharsetUtils_toAsciiBytes(JNIEnv* env, jclass, jstring java_string, jint offset,
105 jint length) {
106 return charsToBytes(env, java_string, offset, length, 0x7f);
107 }
108
CharsetUtils_toIsoLatin1Bytes(JNIEnv * env,jclass,jstring java_string,jint offset,jint length)109 static jbyteArray CharsetUtils_toIsoLatin1Bytes(JNIEnv* env, jclass, jstring java_string,
110 jint offset, jint length) {
111 return charsToBytes(env, java_string, offset, length, 0xff);
112 }
113
CharsetUtils_toUtf8Bytes(JNIEnv * env,jclass,jstring java_string,jint offset,jint length)114 static jbyteArray CharsetUtils_toUtf8Bytes(JNIEnv* env, jclass, jstring java_string, jint offset,
115 jint length) {
116 ScopedFastNativeObjectAccess soa(env);
117 StackHandleScope<1> hs(soa.Self());
118 Handle<mirror::String> string(hs.NewHandle(soa.Decode<mirror::String>(java_string)));
119 if (string == nullptr) {
120 return nullptr;
121 }
122
123 DCHECK_GE(offset, 0);
124 DCHECK_LE(offset, string->GetLength());
125 DCHECK_GE(length, 0);
126 DCHECK_LE(length, string->GetLength() - offset);
127
128 auto visit_chars16 = [string, offset, length](auto append) REQUIRES_SHARED(Locks::mutator_lock_) {
129 const uint16_t* chars16 = string->GetValue() + offset;
130 for (int i = 0; i < length; ++i) {
131 jint ch = chars16[i];
132 if (ch < 0x80) {
133 // One byte.
134 append(ch);
135 } else if (ch < 0x800) {
136 // Two bytes.
137 append((ch >> 6) | 0xc0);
138 append((ch & 0x3f) | 0x80);
139 } else if (U16_IS_SURROGATE(ch)) {
140 // A supplementary character.
141 jchar high = static_cast<jchar>(ch);
142 jchar low = (i + 1 != length) ? chars16[i + 1] : 0;
143 if (!U16_IS_SURROGATE_LEAD(high) || !U16_IS_SURROGATE_TRAIL(low)) {
144 append('?');
145 continue;
146 }
147 // Now we know we have a *valid* surrogate pair, we can consume the low surrogate.
148 ++i;
149 ch = U16_GET_SUPPLEMENTARY(high, low);
150 // Four bytes.
151 append((ch >> 18) | 0xf0);
152 append(((ch >> 12) & 0x3f) | 0x80);
153 append(((ch >> 6) & 0x3f) | 0x80);
154 append((ch & 0x3f) | 0x80);
155 } else {
156 // Three bytes.
157 append((ch >> 12) | 0xe0);
158 append(((ch >> 6) & 0x3f) | 0x80);
159 append((ch & 0x3f) | 0x80);
160 }
161 }
162 };
163
164 bool compressed = string->IsCompressed();
165 size_t utf8_length = 0;
166 if (compressed) {
167 utf8_length = length;
168 } else {
169 visit_chars16([&utf8_length](jbyte c ATTRIBUTE_UNUSED) { ++utf8_length; });
170 }
171 ObjPtr<mirror::ByteArray> result =
172 mirror::ByteArray::Alloc(soa.Self(), dchecked_integral_cast<int32_t>(utf8_length));
173 if (result == nullptr) {
174 return nullptr;
175 }
176
177 if (compressed) {
178 memcpy(result->GetData(), string->GetValueCompressed() + offset, length);
179 } else {
180 int8_t* data = result->GetData();
181 visit_chars16([&data](jbyte c) { *data++ = c; });
182 }
183 return soa.AddLocalReference<jbyteArray>(result);
184 }
185
186 static JNINativeMethod gMethods[] = {
187 FAST_NATIVE_METHOD(CharsetUtils, asciiBytesToChars, "([BII[C)V"),
188 FAST_NATIVE_METHOD(CharsetUtils, isoLatin1BytesToChars, "([BII[C)V"),
189 FAST_NATIVE_METHOD(CharsetUtils, toAsciiBytes, "(Ljava/lang/String;II)[B"),
190 FAST_NATIVE_METHOD(CharsetUtils, toIsoLatin1Bytes, "(Ljava/lang/String;II)[B"),
191 FAST_NATIVE_METHOD(CharsetUtils, toUtf8Bytes, "(Ljava/lang/String;II)[B"),
192 };
193
register_libcore_util_CharsetUtils(JNIEnv * env)194 void register_libcore_util_CharsetUtils(JNIEnv* env) {
195 REGISTER_NATIVE_METHODS("libcore/util/CharsetUtils");
196 }
197
198 } // namespace art
199