1 /*
2 * Copyright (C) 2006 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 #define LOG_TAG "NativeBreakIterator"
18
19 #include "IcuUtilities.h"
20 #include "JNIHelp.h"
21 #include "JniConstants.h"
22 #include "JniException.h"
23 #include "ScopedUtfChars.h"
24 #include "unicode/brkiter.h"
25 #include "unicode/putil.h"
26 #include <stdlib.h>
27
28 // ICU documentation: http://icu-project.org/apiref/icu4c/classBreakIterator.html
29
toBreakIterator(jint address)30 static BreakIterator* toBreakIterator(jint address) {
31 return reinterpret_cast<BreakIterator*>(static_cast<uintptr_t>(address));
32 }
33
34 /**
35 * We use ICU4C's BreakIterator class, but our input is on the Java heap and potentially moving
36 * around between calls. This wrapper class ensures that our RegexMatcher is always pointing at
37 * the current location of the char[]. Earlier versions of Android simply copied the data to the
38 * native heap, but that's wasteful and hides allocations from the garbage collector.
39 */
40 class BreakIteratorAccessor {
41 public:
BreakIteratorAccessor(JNIEnv * env,jint address,jstring javaInput,bool reset)42 BreakIteratorAccessor(JNIEnv* env, jint address, jstring javaInput, bool reset) {
43 init(env, address);
44 mJavaInput = javaInput;
45
46 if (mJavaInput == NULL) {
47 return;
48 }
49
50 mChars = env->GetStringChars(mJavaInput, NULL);
51 if (mChars == NULL) {
52 return;
53 }
54
55 mUText = utext_openUChars(NULL, mChars, env->GetStringLength(mJavaInput), &mStatus);
56 if (mUText == NULL) {
57 return;
58 }
59
60 if (reset) {
61 mBreakIterator->setText(mUText, mStatus);
62 } else {
63 mBreakIterator->refreshInputText(mUText, mStatus);
64 }
65 }
66
BreakIteratorAccessor(JNIEnv * env,jint address)67 BreakIteratorAccessor(JNIEnv* env, jint address) {
68 init(env, address);
69 }
70
~BreakIteratorAccessor()71 ~BreakIteratorAccessor() {
72 utext_close(mUText);
73 if (mJavaInput) {
74 mEnv->ReleaseStringChars(mJavaInput, mChars);
75 }
76 maybeThrowIcuException(mEnv, "utext_close", mStatus);
77 }
78
operator ->()79 BreakIterator* operator->() {
80 return mBreakIterator;
81 }
82
status()83 UErrorCode& status() {
84 return mStatus;
85 }
86
87 private:
init(JNIEnv * env,jint address)88 void init(JNIEnv* env, jint address) {
89 mEnv = env;
90 mJavaInput = NULL;
91 mBreakIterator = toBreakIterator(address);
92 mChars = NULL;
93 mStatus = U_ZERO_ERROR;
94 mUText = NULL;
95 }
96
97 JNIEnv* mEnv;
98 jstring mJavaInput;
99 BreakIterator* mBreakIterator;
100 const jchar* mChars;
101 UErrorCode mStatus;
102 UText* mUText;
103
104 // Disallow copy and assignment.
105 BreakIteratorAccessor(const BreakIteratorAccessor&);
106 void operator=(const BreakIteratorAccessor&);
107 };
108
109 #define MAKE_BREAK_ITERATOR_INSTANCE(F) \
110 UErrorCode status = U_ZERO_ERROR; \
111 const ScopedUtfChars localeChars(env, javaLocale); \
112 if (localeChars.c_str() == NULL) { \
113 return 0; \
114 } \
115 Locale locale(Locale::createFromName(localeChars.c_str())); \
116 BreakIterator* it = F(locale, status); \
117 if (maybeThrowIcuException(env, "ubrk_open", status)) { \
118 return 0; \
119 } \
120 return reinterpret_cast<uintptr_t>(it)
121
NativeBreakIterator_cloneImpl(JNIEnv * env,jclass,jint address)122 static jint NativeBreakIterator_cloneImpl(JNIEnv* env, jclass, jint address) {
123 BreakIteratorAccessor it(env, address);
124 return reinterpret_cast<uintptr_t>(it->clone());
125 }
126
NativeBreakIterator_closeImpl(JNIEnv *,jclass,jint address)127 static void NativeBreakIterator_closeImpl(JNIEnv*, jclass, jint address) {
128 delete toBreakIterator(address);
129 }
130
NativeBreakIterator_currentImpl(JNIEnv * env,jclass,jint address,jstring javaInput)131 static jint NativeBreakIterator_currentImpl(JNIEnv* env, jclass, jint address, jstring javaInput) {
132 BreakIteratorAccessor it(env, address, javaInput, false);
133 return it->current();
134 }
135
NativeBreakIterator_firstImpl(JNIEnv * env,jclass,jint address,jstring javaInput)136 static jint NativeBreakIterator_firstImpl(JNIEnv* env, jclass, jint address, jstring javaInput) {
137 BreakIteratorAccessor it(env, address, javaInput, false);
138 return it->first();
139 }
140
NativeBreakIterator_followingImpl(JNIEnv * env,jclass,jint address,jstring javaInput,jint offset)141 static jint NativeBreakIterator_followingImpl(JNIEnv* env, jclass, jint address, jstring javaInput, jint offset) {
142 BreakIteratorAccessor it(env, address, javaInput, false);
143 return it->following(offset);
144 }
145
NativeBreakIterator_getCharacterInstanceImpl(JNIEnv * env,jclass,jstring javaLocale)146 static jint NativeBreakIterator_getCharacterInstanceImpl(JNIEnv* env, jclass, jstring javaLocale) {
147 MAKE_BREAK_ITERATOR_INSTANCE(BreakIterator::createCharacterInstance);
148 }
149
NativeBreakIterator_getLineInstanceImpl(JNIEnv * env,jclass,jstring javaLocale)150 static jint NativeBreakIterator_getLineInstanceImpl(JNIEnv* env, jclass, jstring javaLocale) {
151 MAKE_BREAK_ITERATOR_INSTANCE(BreakIterator::createLineInstance);
152 }
153
NativeBreakIterator_getSentenceInstanceImpl(JNIEnv * env,jclass,jstring javaLocale)154 static jint NativeBreakIterator_getSentenceInstanceImpl(JNIEnv* env, jclass, jstring javaLocale) {
155 MAKE_BREAK_ITERATOR_INSTANCE(BreakIterator::createSentenceInstance);
156 }
157
NativeBreakIterator_getWordInstanceImpl(JNIEnv * env,jclass,jstring javaLocale)158 static jint NativeBreakIterator_getWordInstanceImpl(JNIEnv* env, jclass, jstring javaLocale) {
159 MAKE_BREAK_ITERATOR_INSTANCE(BreakIterator::createWordInstance);
160 }
161
NativeBreakIterator_isBoundaryImpl(JNIEnv * env,jclass,jint address,jstring javaInput,jint offset)162 static jboolean NativeBreakIterator_isBoundaryImpl(JNIEnv* env, jclass, jint address, jstring javaInput, jint offset) {
163 BreakIteratorAccessor it(env, address, javaInput, false);
164 return it->isBoundary(offset);
165 }
166
NativeBreakIterator_lastImpl(JNIEnv * env,jclass,jint address,jstring javaInput)167 static jint NativeBreakIterator_lastImpl(JNIEnv* env, jclass, jint address, jstring javaInput) {
168 BreakIteratorAccessor it(env, address, javaInput, false);
169 return it->last();
170 }
171
NativeBreakIterator_nextImpl(JNIEnv * env,jclass,jint address,jstring javaInput,jint n)172 static jint NativeBreakIterator_nextImpl(JNIEnv* env, jclass, jint address, jstring javaInput, jint n) {
173 BreakIteratorAccessor it(env, address, javaInput, false);
174 if (n < 0) {
175 while (n++ < -1) {
176 it->previous();
177 }
178 return it->previous();
179 } else if (n == 0) {
180 return it->current();
181 } else {
182 while (n-- > 1) {
183 it->next();
184 }
185 return it->next();
186 }
187 return -1;
188 }
189
NativeBreakIterator_precedingImpl(JNIEnv * env,jclass,jint address,jstring javaInput,jint offset)190 static jint NativeBreakIterator_precedingImpl(JNIEnv* env, jclass, jint address, jstring javaInput, jint offset) {
191 BreakIteratorAccessor it(env, address, javaInput, false);
192 return it->preceding(offset);
193 }
194
NativeBreakIterator_previousImpl(JNIEnv * env,jclass,jint address,jstring javaInput)195 static jint NativeBreakIterator_previousImpl(JNIEnv* env, jclass, jint address, jstring javaInput) {
196 BreakIteratorAccessor it(env, address, javaInput, false);
197 return it->previous();
198 }
199
NativeBreakIterator_setTextImpl(JNIEnv * env,jclass,jint address,jstring javaInput)200 static void NativeBreakIterator_setTextImpl(JNIEnv* env, jclass, jint address, jstring javaInput) {
201 BreakIteratorAccessor it(env, address, javaInput, true);
202 }
203
204 static JNINativeMethod gMethods[] = {
205 NATIVE_METHOD(NativeBreakIterator, cloneImpl, "(I)I"),
206 NATIVE_METHOD(NativeBreakIterator, closeImpl, "(I)V"),
207 NATIVE_METHOD(NativeBreakIterator, currentImpl, "(ILjava/lang/String;)I"),
208 NATIVE_METHOD(NativeBreakIterator, firstImpl, "(ILjava/lang/String;)I"),
209 NATIVE_METHOD(NativeBreakIterator, followingImpl, "(ILjava/lang/String;I)I"),
210 NATIVE_METHOD(NativeBreakIterator, getCharacterInstanceImpl, "(Ljava/lang/String;)I"),
211 NATIVE_METHOD(NativeBreakIterator, getLineInstanceImpl, "(Ljava/lang/String;)I"),
212 NATIVE_METHOD(NativeBreakIterator, getSentenceInstanceImpl, "(Ljava/lang/String;)I"),
213 NATIVE_METHOD(NativeBreakIterator, getWordInstanceImpl, "(Ljava/lang/String;)I"),
214 NATIVE_METHOD(NativeBreakIterator, isBoundaryImpl, "(ILjava/lang/String;I)Z"),
215 NATIVE_METHOD(NativeBreakIterator, lastImpl, "(ILjava/lang/String;)I"),
216 NATIVE_METHOD(NativeBreakIterator, nextImpl, "(ILjava/lang/String;I)I"),
217 NATIVE_METHOD(NativeBreakIterator, precedingImpl, "(ILjava/lang/String;I)I"),
218 NATIVE_METHOD(NativeBreakIterator, previousImpl, "(ILjava/lang/String;)I"),
219 NATIVE_METHOD(NativeBreakIterator, setTextImpl, "(ILjava/lang/String;)V"),
220 };
register_libcore_icu_NativeBreakIterator(JNIEnv * env)221 void register_libcore_icu_NativeBreakIterator(JNIEnv* env) {
222 jniRegisterNativeMethods(env, "libcore/icu/NativeBreakIterator", gMethods, NELEM(gMethods));
223 }
224