1 /**
2 *******************************************************************************
3 * Copyright (C) 1996-2006, International Business Machines Corporation and *
4 * others. All Rights Reserved. *
5 *******************************************************************************
6 *
7 *
8 *******************************************************************************
9 */
10 /*
11 * (C) Copyright IBM Corp. 2000 - All Rights Reserved
12 * A JNI wrapper to ICU native converter Interface
13 * @author: Ram Viswanadha
14 */
15
16 #define LOG_TAG "NativeConverter"
17
18 #include <stdlib.h>
19 #include <string.h>
20
21 #include <memory>
22 #include <vector>
23
24 #include <android/log.h>
25
26 #include "IcuUtilities.h"
27 #include "JNIHelp.h"
28 #include "JniConstants.h"
29 #include "JniException.h"
30 #include "ScopedLocalRef.h"
31 #include "ScopedPrimitiveArray.h"
32 #include "ScopedStringChars.h"
33 #include "ScopedUtfChars.h"
34 #include "toStringArray.h"
35 #include "unicode/ucnv.h"
36 #include "unicode/ucnv_cb.h"
37 #include "unicode/uniset.h"
38 #include "unicode/ustring.h"
39 #include "unicode/utypes.h"
40
41 #define NativeConverter_REPORT 0
42 #define NativeConverter_IGNORE 1
43 #define NativeConverter_REPLACE 2
44
45 #define MAX_REPLACEMENT_LENGTH 32 // equivalent to UCNV_ERROR_BUFFER_LENGTH
46
47 struct DecoderCallbackContext {
48 UChar replacementChars[MAX_REPLACEMENT_LENGTH];
49 size_t replacementCharCount;
50 UConverterToUCallback onUnmappableInput;
51 UConverterToUCallback onMalformedInput;
52 };
53
54 struct EncoderCallbackContext {
55 char replacementBytes[MAX_REPLACEMENT_LENGTH];
56 size_t replacementByteCount;
57 UConverterFromUCallback onUnmappableInput;
58 UConverterFromUCallback onMalformedInput;
59 };
60
toUConverter(jlong address)61 static UConverter* toUConverter(jlong address) {
62 return reinterpret_cast<UConverter*>(static_cast<uintptr_t>(address));
63 }
64
collectStandardNames(JNIEnv * env,const char * canonicalName,const char * standard,std::vector<std::string> & result)65 static bool collectStandardNames(JNIEnv* env, const char* canonicalName, const char* standard,
66 std::vector<std::string>& result) {
67 UErrorCode status = U_ZERO_ERROR;
68 icu::UStringEnumeration e(ucnv_openStandardNames(canonicalName, standard, &status));
69 if (maybeThrowIcuException(env, "ucnv_openStandardNames", status)) {
70 return false;
71 }
72
73 int32_t count = e.count(status);
74 if (maybeThrowIcuException(env, "StringEnumeration::count", status)) {
75 return false;
76 }
77
78 for (int32_t i = 0; i < count; ++i) {
79 const icu::UnicodeString* string = e.snext(status);
80 if (maybeThrowIcuException(env, "StringEnumeration::snext", status)) {
81 return false;
82 }
83 std::string s;
84 string->toUTF8String(s);
85 if (s.find_first_of("+,") == std::string::npos) {
86 result.push_back(s);
87 }
88 }
89
90 return true;
91 }
92
getICUCanonicalName(const char * name)93 static const char* getICUCanonicalName(const char* name) {
94 UErrorCode error = U_ZERO_ERROR;
95 const char* canonicalName = NULL;
96 if ((canonicalName = ucnv_getCanonicalName(name, "MIME", &error)) != NULL) {
97 return canonicalName;
98 } else if ((canonicalName = ucnv_getCanonicalName(name, "IANA", &error)) != NULL) {
99 return canonicalName;
100 } else if ((canonicalName = ucnv_getCanonicalName(name, "", &error)) != NULL) {
101 return canonicalName;
102 } else if ((canonicalName = ucnv_getAlias(name, 0, &error)) != NULL) {
103 // We have some aliases in the form x-blah .. match those first.
104 return canonicalName;
105 } else if (strstr(name, "x-") == name) {
106 // Check if the converter can be opened with the name given.
107 error = U_ZERO_ERROR;
108 icu::LocalUConverterPointer cnv(ucnv_open(name + 2, &error));
109 if (U_SUCCESS(error)) {
110 return name + 2;
111 }
112 }
113 return NULL;
114 }
115
116 // If a charset listed in the IANA Charset Registry is supported by an implementation
117 // of the Java platform then its canonical name must be the name listed in the registry.
118 // Many charsets are given more than one name in the registry, in which case the registry
119 // identifies one of the names as MIME-preferred. If a charset has more than one registry
120 // name then its canonical name must be the MIME-preferred name and the other names in
121 // the registry must be valid aliases. If a supported charset is not listed in the IANA
122 // registry then its canonical name must begin with one of the strings "X-" or "x-".
getJavaCanonicalName(JNIEnv * env,const char * icuCanonicalName)123 static jstring getJavaCanonicalName(JNIEnv* env, const char* icuCanonicalName) {
124 UErrorCode status = U_ZERO_ERROR;
125
126 // Check to see if this is a well-known MIME or IANA name.
127 const char* cName = NULL;
128 if ((cName = ucnv_getStandardName(icuCanonicalName, "MIME", &status)) != NULL) {
129 return env->NewStringUTF(cName);
130 } else if ((cName = ucnv_getStandardName(icuCanonicalName, "IANA", &status)) != NULL) {
131 return env->NewStringUTF(cName);
132 }
133
134 // Check to see if an alias already exists with "x-" prefix, if yes then
135 // make that the canonical name.
136 int32_t aliasCount = ucnv_countAliases(icuCanonicalName, &status);
137 for (int i = 0; i < aliasCount; ++i) {
138 const char* name = ucnv_getAlias(icuCanonicalName, i, &status);
139 if (name != NULL && name[0] == 'x' && name[1] == '-') {
140 return env->NewStringUTF(name);
141 }
142 }
143
144 // As a last resort, prepend "x-" to any alias and make that the canonical name.
145 status = U_ZERO_ERROR;
146 const char* name = ucnv_getStandardName(icuCanonicalName, "UTR22", &status);
147 if (name == NULL && strchr(icuCanonicalName, ',') != NULL) {
148 name = ucnv_getAlias(icuCanonicalName, 1, &status);
149 }
150 // If there is no UTR22 canonical name then just return the original name.
151 if (name == NULL) {
152 name = icuCanonicalName;
153 }
154 std::unique_ptr<char[]> result(new char[2 + strlen(name) + 1]);
155 strcpy(&result[0], "x-");
156 strcat(&result[0], name);
157 return env->NewStringUTF(&result[0]);
158 }
159
NativeConverter_openConverter(JNIEnv * env,jclass,jstring converterName)160 static jlong NativeConverter_openConverter(JNIEnv* env, jclass, jstring converterName) {
161 ScopedUtfChars converterNameChars(env, converterName);
162 if (converterNameChars.c_str() == NULL) {
163 return 0;
164 }
165 UErrorCode status = U_ZERO_ERROR;
166 UConverter* cnv = ucnv_open(converterNameChars.c_str(), &status);
167 maybeThrowIcuException(env, "ucnv_open", status);
168 return reinterpret_cast<uintptr_t>(cnv);
169 }
170
NativeConverter_closeConverter(JNIEnv *,jclass,jlong address)171 static void NativeConverter_closeConverter(JNIEnv*, jclass, jlong address) {
172 ucnv_close(toUConverter(address));
173 }
174
shouldCodecThrow(jboolean flush,UErrorCode error)175 static bool shouldCodecThrow(jboolean flush, UErrorCode error) {
176 if (flush) {
177 return (error != U_BUFFER_OVERFLOW_ERROR && error != U_TRUNCATED_CHAR_FOUND);
178 } else {
179 return (error != U_BUFFER_OVERFLOW_ERROR && error != U_INVALID_CHAR_FOUND && error != U_ILLEGAL_CHAR_FOUND);
180 }
181 }
182
NativeConverter_encode(JNIEnv * env,jclass,jlong address,jcharArray source,jint sourceEnd,jbyteArray target,jint targetEnd,jintArray data,jboolean flush)183 static jint NativeConverter_encode(JNIEnv* env, jclass, jlong address,
184 jcharArray source, jint sourceEnd, jbyteArray target, jint targetEnd,
185 jintArray data, jboolean flush) {
186
187 UConverter* cnv = toUConverter(address);
188 if (cnv == NULL) {
189 maybeThrowIcuException(env, "toUConverter", U_ILLEGAL_ARGUMENT_ERROR);
190 return U_ILLEGAL_ARGUMENT_ERROR;
191 }
192 ScopedCharArrayRO uSource(env, source);
193 if (uSource.get() == NULL) {
194 maybeThrowIcuException(env, "uSource", U_ILLEGAL_ARGUMENT_ERROR);
195 return U_ILLEGAL_ARGUMENT_ERROR;
196 }
197 ScopedByteArrayRW uTarget(env, target);
198 if (uTarget.get() == NULL) {
199 maybeThrowIcuException(env, "uTarget", U_ILLEGAL_ARGUMENT_ERROR);
200 return U_ILLEGAL_ARGUMENT_ERROR;
201 }
202 ScopedIntArrayRW myData(env, data);
203 if (myData.get() == NULL) {
204 maybeThrowIcuException(env, "myData", U_ILLEGAL_ARGUMENT_ERROR);
205 return U_ILLEGAL_ARGUMENT_ERROR;
206 }
207
208 // Do the conversion.
209 jint* sourceOffset = &myData[0];
210 jint* targetOffset = &myData[1];
211 const jchar* mySource = uSource.get() + *sourceOffset;
212 const UChar* mySourceLimit= uSource.get() + sourceEnd;
213 char* cTarget = reinterpret_cast<char*>(uTarget.get() + *targetOffset);
214 const char* cTargetLimit = reinterpret_cast<const char*>(uTarget.get() + targetEnd);
215 UErrorCode errorCode = U_ZERO_ERROR;
216 ucnv_fromUnicode(cnv , &cTarget, cTargetLimit, &mySource, mySourceLimit, NULL, (UBool) flush, &errorCode);
217 *sourceOffset = (mySource - uSource.get()) - *sourceOffset;
218 *targetOffset = (reinterpret_cast<jbyte*>(cTarget) - uTarget.get());
219
220 // If there was an error, count the problematic characters.
221 if (errorCode == U_ILLEGAL_CHAR_FOUND || errorCode == U_INVALID_CHAR_FOUND ||
222 errorCode == U_TRUNCATED_CHAR_FOUND) {
223 int8_t invalidUCharCount = 32;
224 UChar invalidUChars[32];
225 UErrorCode minorErrorCode = U_ZERO_ERROR;
226 ucnv_getInvalidUChars(cnv, invalidUChars, &invalidUCharCount, &minorErrorCode);
227 if (U_SUCCESS(minorErrorCode)) {
228 myData[2] = invalidUCharCount;
229 }
230 }
231
232 // Managed code handles some cases; throw all other errors.
233 if (shouldCodecThrow(flush, errorCode)) {
234 maybeThrowIcuException(env, "ucnv_fromUnicode", errorCode);
235 }
236 return errorCode;
237 }
238
NativeConverter_decode(JNIEnv * env,jclass,jlong address,jbyteArray source,jint sourceEnd,jcharArray target,jint targetEnd,jintArray data,jboolean flush)239 static jint NativeConverter_decode(JNIEnv* env, jclass, jlong address,
240 jbyteArray source, jint sourceEnd, jcharArray target, jint targetEnd,
241 jintArray data, jboolean flush) {
242
243 UConverter* cnv = toUConverter(address);
244 if (cnv == NULL) {
245 maybeThrowIcuException(env, "toUConverter", U_ILLEGAL_ARGUMENT_ERROR);
246 return U_ILLEGAL_ARGUMENT_ERROR;
247 }
248 ScopedByteArrayRO uSource(env, source);
249 if (uSource.get() == NULL) {
250 maybeThrowIcuException(env, "uSource", U_ILLEGAL_ARGUMENT_ERROR);
251 return U_ILLEGAL_ARGUMENT_ERROR;
252 }
253 ScopedCharArrayRW uTarget(env, target);
254 if (uTarget.get() == NULL) {
255 maybeThrowIcuException(env, "uTarget", U_ILLEGAL_ARGUMENT_ERROR);
256 return U_ILLEGAL_ARGUMENT_ERROR;
257 }
258 ScopedIntArrayRW myData(env, data);
259 if (myData.get() == NULL) {
260 maybeThrowIcuException(env, "myData", U_ILLEGAL_ARGUMENT_ERROR);
261 return U_ILLEGAL_ARGUMENT_ERROR;
262 }
263
264 // Do the conversion.
265 jint* sourceOffset = &myData[0];
266 jint* targetOffset = &myData[1];
267 const char* mySource = reinterpret_cast<const char*>(uSource.get() + *sourceOffset);
268 const char* mySourceLimit = reinterpret_cast<const char*>(uSource.get() + sourceEnd);
269 UChar* cTarget = uTarget.get() + *targetOffset;
270 const UChar* cTargetLimit = uTarget.get() + targetEnd;
271 UErrorCode errorCode = U_ZERO_ERROR;
272 ucnv_toUnicode(cnv, &cTarget, cTargetLimit, &mySource, mySourceLimit, NULL, flush, &errorCode);
273 *sourceOffset = mySource - reinterpret_cast<const char*>(uSource.get()) - *sourceOffset;
274 *targetOffset = cTarget - uTarget.get() - *targetOffset;
275
276 // If there was an error, count the problematic bytes.
277 if (errorCode == U_ILLEGAL_CHAR_FOUND || errorCode == U_INVALID_CHAR_FOUND ||
278 errorCode == U_TRUNCATED_CHAR_FOUND) {
279 int8_t invalidByteCount = 32;
280 char invalidBytes[32] = {'\0'};
281 UErrorCode minorErrorCode = U_ZERO_ERROR;
282 ucnv_getInvalidChars(cnv, invalidBytes, &invalidByteCount, &minorErrorCode);
283 if (U_SUCCESS(minorErrorCode)) {
284 myData[2] = invalidByteCount;
285 }
286 }
287
288 // Managed code handles some cases; throw all other errors.
289 if (shouldCodecThrow(flush, errorCode)) {
290 maybeThrowIcuException(env, "ucnv_toUnicode", errorCode);
291 }
292 return errorCode;
293 }
294
NativeConverter_resetByteToChar(JNIEnv *,jclass,jlong address)295 static void NativeConverter_resetByteToChar(JNIEnv*, jclass, jlong address) {
296 UConverter* cnv = toUConverter(address);
297 if (cnv) {
298 ucnv_resetToUnicode(cnv);
299 }
300 }
301
NativeConverter_resetCharToByte(JNIEnv *,jclass,jlong address)302 static void NativeConverter_resetCharToByte(JNIEnv*, jclass, jlong address) {
303 UConverter* cnv = toUConverter(address);
304 if (cnv) {
305 ucnv_resetFromUnicode(cnv);
306 }
307 }
308
NativeConverter_getMaxBytesPerChar(JNIEnv *,jclass,jlong address)309 static jint NativeConverter_getMaxBytesPerChar(JNIEnv*, jclass, jlong address) {
310 UConverter* cnv = toUConverter(address);
311 return (cnv != NULL) ? ucnv_getMaxCharSize(cnv) : -1;
312 }
313
NativeConverter_getMinBytesPerChar(JNIEnv *,jclass,jlong address)314 static jint NativeConverter_getMinBytesPerChar(JNIEnv*, jclass, jlong address) {
315 UConverter* cnv = toUConverter(address);
316 return (cnv != NULL) ? ucnv_getMinCharSize(cnv) : -1;
317 }
318
NativeConverter_getAveBytesPerChar(JNIEnv *,jclass,jlong address)319 static jfloat NativeConverter_getAveBytesPerChar(JNIEnv*, jclass, jlong address) {
320 UConverter* cnv = toUConverter(address);
321 return (cnv != NULL) ? ((ucnv_getMaxCharSize(cnv) + ucnv_getMinCharSize(cnv)) / 2.0) : -1;
322 }
323
NativeConverter_getAvailableCharsetNames(JNIEnv * env,jclass)324 static jobjectArray NativeConverter_getAvailableCharsetNames(JNIEnv* env, jclass) {
325 int32_t num = ucnv_countAvailable();
326 jobjectArray result = env->NewObjectArray(num, JniConstants::stringClass, NULL);
327 if (result == NULL) {
328 return NULL;
329 }
330 for (int i = 0; i < num; ++i) {
331 const char* name = ucnv_getAvailableName(i);
332 ScopedLocalRef<jstring> javaCanonicalName(env, getJavaCanonicalName(env, name));
333 if (javaCanonicalName.get() == NULL) {
334 return NULL;
335 }
336 env->SetObjectArrayElement(result, i, javaCanonicalName.get());
337 if (env->ExceptionCheck()) {
338 return NULL;
339 }
340 }
341 return result;
342 }
343
CHARSET_ENCODER_CALLBACK(const void * rawContext,UConverterFromUnicodeArgs * args,const UChar * codeUnits,int32_t length,UChar32 codePoint,UConverterCallbackReason reason,UErrorCode * status)344 static void CHARSET_ENCODER_CALLBACK(const void* rawContext, UConverterFromUnicodeArgs* args,
345 const UChar* codeUnits, int32_t length, UChar32 codePoint, UConverterCallbackReason reason,
346 UErrorCode* status) {
347 if (!rawContext) {
348 return;
349 }
350 const EncoderCallbackContext* ctx = reinterpret_cast<const EncoderCallbackContext*>(rawContext);
351 switch(reason) {
352 case UCNV_UNASSIGNED:
353 ctx->onUnmappableInput(ctx, args, codeUnits, length, codePoint, reason, status);
354 return;
355 case UCNV_ILLEGAL:
356 case UCNV_IRREGULAR:
357 ctx->onMalformedInput(ctx, args, codeUnits, length, codePoint, reason, status);
358 return;
359 case UCNV_CLOSE:
360 delete ctx;
361 return;
362 default:
363 *status = U_ILLEGAL_ARGUMENT_ERROR;
364 return;
365 }
366 }
367
encoderReplaceCallback(const void * rawContext,UConverterFromUnicodeArgs * fromArgs,const UChar *,int32_t,UChar32,UConverterCallbackReason,UErrorCode * err)368 static void encoderReplaceCallback(const void* rawContext,
369 UConverterFromUnicodeArgs* fromArgs, const UChar*, int32_t, UChar32,
370 UConverterCallbackReason, UErrorCode * err) {
371 if (rawContext == NULL) {
372 return;
373 }
374 const EncoderCallbackContext* context = reinterpret_cast<const EncoderCallbackContext*>(rawContext);
375 *err = U_ZERO_ERROR;
376 ucnv_cbFromUWriteBytes(fromArgs, context->replacementBytes, context->replacementByteCount, 0, err);
377 }
378
getFromUCallback(int32_t mode)379 static UConverterFromUCallback getFromUCallback(int32_t mode) {
380 switch(mode) {
381 case NativeConverter_IGNORE: return UCNV_FROM_U_CALLBACK_SKIP;
382 case NativeConverter_REPLACE: return encoderReplaceCallback;
383 case NativeConverter_REPORT: return UCNV_FROM_U_CALLBACK_STOP;
384 }
385 abort();
386 }
387
NativeConverter_setCallbackEncode(JNIEnv * env,jclass,jlong address,jint onMalformedInput,jint onUnmappableInput,jbyteArray javaReplacement)388 static void NativeConverter_setCallbackEncode(JNIEnv* env, jclass, jlong address,
389 jint onMalformedInput, jint onUnmappableInput, jbyteArray javaReplacement) {
390 UConverter* cnv = toUConverter(address);
391 if (cnv == NULL) {
392 maybeThrowIcuException(env, "toUConverter", U_ILLEGAL_ARGUMENT_ERROR);
393 return;
394 }
395
396 UConverterFromUCallback oldCallback = NULL;
397 const void* oldCallbackContext = NULL;
398 ucnv_getFromUCallBack(cnv, &oldCallback, const_cast<const void**>(&oldCallbackContext));
399
400 EncoderCallbackContext* callbackContext = const_cast<EncoderCallbackContext*>(
401 reinterpret_cast<const EncoderCallbackContext*>(oldCallbackContext));
402 if (callbackContext == NULL) {
403 callbackContext = new EncoderCallbackContext;
404 }
405
406 callbackContext->onMalformedInput = getFromUCallback(onMalformedInput);
407 callbackContext->onUnmappableInput = getFromUCallback(onUnmappableInput);
408
409 ScopedByteArrayRO replacementBytes(env, javaReplacement);
410 if (replacementBytes.get() == NULL) {
411 maybeThrowIcuException(env, "replacementBytes", U_ILLEGAL_ARGUMENT_ERROR);
412 return;
413 }
414 memcpy(callbackContext->replacementBytes, replacementBytes.get(), replacementBytes.size());
415 callbackContext->replacementByteCount = replacementBytes.size();
416
417 UErrorCode errorCode = U_ZERO_ERROR;
418 ucnv_setFromUCallBack(cnv, CHARSET_ENCODER_CALLBACK, callbackContext, NULL, NULL, &errorCode);
419 maybeThrowIcuException(env, "ucnv_setFromUCallBack", errorCode);
420 }
421
decoderIgnoreCallback(const void *,UConverterToUnicodeArgs *,const char *,int32_t,UConverterCallbackReason,UErrorCode * err)422 static void decoderIgnoreCallback(const void*, UConverterToUnicodeArgs*, const char*, int32_t, UConverterCallbackReason, UErrorCode* err) {
423 // The icu4c UCNV_FROM_U_CALLBACK_SKIP callback requires that the context is NULL, which is
424 // never true for us.
425 *err = U_ZERO_ERROR;
426 }
427
decoderReplaceCallback(const void * rawContext,UConverterToUnicodeArgs * toArgs,const char *,int32_t,UConverterCallbackReason,UErrorCode * err)428 static void decoderReplaceCallback(const void* rawContext,
429 UConverterToUnicodeArgs* toArgs, const char*, int32_t, UConverterCallbackReason,
430 UErrorCode* err) {
431 if (!rawContext) {
432 return;
433 }
434 const DecoderCallbackContext* context = reinterpret_cast<const DecoderCallbackContext*>(rawContext);
435 *err = U_ZERO_ERROR;
436 ucnv_cbToUWriteUChars(toArgs,context->replacementChars, context->replacementCharCount, 0, err);
437 }
438
getToUCallback(int32_t mode)439 static UConverterToUCallback getToUCallback(int32_t mode) {
440 switch (mode) {
441 case NativeConverter_IGNORE: return decoderIgnoreCallback;
442 case NativeConverter_REPLACE: return decoderReplaceCallback;
443 case NativeConverter_REPORT: return UCNV_TO_U_CALLBACK_STOP;
444 }
445 abort();
446 }
447
CHARSET_DECODER_CALLBACK(const void * rawContext,UConverterToUnicodeArgs * args,const char * codeUnits,int32_t length,UConverterCallbackReason reason,UErrorCode * status)448 static void CHARSET_DECODER_CALLBACK(const void* rawContext, UConverterToUnicodeArgs* args,
449 const char* codeUnits, int32_t length,
450 UConverterCallbackReason reason, UErrorCode* status) {
451 if (!rawContext) {
452 return;
453 }
454 const DecoderCallbackContext* ctx = reinterpret_cast<const DecoderCallbackContext*>(rawContext);
455 switch(reason) {
456 case UCNV_UNASSIGNED:
457 ctx->onUnmappableInput(ctx, args, codeUnits, length, reason, status);
458 return;
459 case UCNV_ILLEGAL:
460 case UCNV_IRREGULAR:
461 ctx->onMalformedInput(ctx, args, codeUnits, length, reason, status);
462 return;
463 case UCNV_CLOSE:
464 delete ctx;
465 return;
466 default:
467 *status = U_ILLEGAL_ARGUMENT_ERROR;
468 return;
469 }
470 }
471
NativeConverter_setCallbackDecode(JNIEnv * env,jclass,jlong address,jint onMalformedInput,jint onUnmappableInput,jstring javaReplacement)472 static void NativeConverter_setCallbackDecode(JNIEnv* env, jclass, jlong address,
473 jint onMalformedInput, jint onUnmappableInput, jstring javaReplacement) {
474 UConverter* cnv = toUConverter(address);
475 if (cnv == NULL) {
476 maybeThrowIcuException(env, "toConverter", U_ILLEGAL_ARGUMENT_ERROR);
477 return;
478 }
479
480 UConverterToUCallback oldCallback;
481 const void* oldCallbackContext;
482 ucnv_getToUCallBack(cnv, &oldCallback, &oldCallbackContext);
483
484 DecoderCallbackContext* callbackContext = const_cast<DecoderCallbackContext*>(
485 reinterpret_cast<const DecoderCallbackContext*>(oldCallbackContext));
486 if (callbackContext == NULL) {
487 callbackContext = new DecoderCallbackContext;
488 }
489
490 callbackContext->onMalformedInput = getToUCallback(onMalformedInput);
491 callbackContext->onUnmappableInput = getToUCallback(onUnmappableInput);
492
493 ScopedStringChars replacement(env, javaReplacement);
494 if (replacement.get() == NULL) {
495 maybeThrowIcuException(env, "replacement", U_ILLEGAL_ARGUMENT_ERROR);
496 return;
497 }
498 u_strncpy(callbackContext->replacementChars, replacement.get(), replacement.size());
499 callbackContext->replacementCharCount = replacement.size();
500
501 UErrorCode errorCode = U_ZERO_ERROR;
502 ucnv_setToUCallBack(cnv, CHARSET_DECODER_CALLBACK, callbackContext, NULL, NULL, &errorCode);
503 maybeThrowIcuException(env, "ucnv_setToUCallBack", errorCode);
504 }
505
NativeConverter_getAveCharsPerByte(JNIEnv * env,jclass,jlong handle)506 static jfloat NativeConverter_getAveCharsPerByte(JNIEnv* env, jclass, jlong handle) {
507 return (1 / (jfloat) NativeConverter_getMaxBytesPerChar(env, NULL, handle));
508 }
509
NativeConverter_getSubstitutionBytes(JNIEnv * env,jclass,jlong address)510 static jbyteArray NativeConverter_getSubstitutionBytes(JNIEnv* env, jclass, jlong address) {
511 UConverter* cnv = toUConverter(address);
512 if (cnv == NULL) {
513 return NULL;
514 }
515 UErrorCode status = U_ZERO_ERROR;
516 char replacementBytes[MAX_REPLACEMENT_LENGTH];
517 int8_t len = sizeof(replacementBytes);
518 ucnv_getSubstChars(cnv, replacementBytes, &len, &status);
519 if (!U_SUCCESS(status)) {
520 return env->NewByteArray(0);
521 }
522 jbyteArray result = env->NewByteArray(len);
523 if (result == NULL) {
524 return NULL;
525 }
526 env->SetByteArrayRegion(result, 0, len, reinterpret_cast<jbyte*>(replacementBytes));
527 return result;
528 }
529
NativeConverter_contains(JNIEnv * env,jclass,jstring name1,jstring name2)530 static jboolean NativeConverter_contains(JNIEnv* env, jclass, jstring name1, jstring name2) {
531 ScopedUtfChars name1Chars(env, name1);
532 if (name1Chars.c_str() == NULL) {
533 return JNI_FALSE;
534 }
535 ScopedUtfChars name2Chars(env, name2);
536 if (name2Chars.c_str() == NULL) {
537 return JNI_FALSE;
538 }
539
540 UErrorCode errorCode = U_ZERO_ERROR;
541 icu::LocalUConverterPointer converter1(ucnv_open(name1Chars.c_str(), &errorCode));
542 icu::UnicodeSet set1;
543 ucnv_getUnicodeSet(&*converter1, set1.toUSet(), UCNV_ROUNDTRIP_SET, &errorCode);
544
545 icu::LocalUConverterPointer converter2(ucnv_open(name2Chars.c_str(), &errorCode));
546 icu::UnicodeSet set2;
547 ucnv_getUnicodeSet(&*converter2, set2.toUSet(), UCNV_ROUNDTRIP_SET, &errorCode);
548
549 return U_SUCCESS(errorCode) && set1.containsAll(set2);
550 }
551
NativeConverter_charsetForName(JNIEnv * env,jclass,jstring charsetName)552 static jobject NativeConverter_charsetForName(JNIEnv* env, jclass, jstring charsetName) {
553 ScopedUtfChars charsetNameChars(env, charsetName);
554 if (charsetNameChars.c_str() == NULL) {
555 return NULL;
556 }
557
558 // Get ICU's canonical name for this charset.
559 const char* icuCanonicalName = getICUCanonicalName(charsetNameChars.c_str());
560 if (icuCanonicalName == NULL) {
561 return NULL;
562 }
563
564 // Get Java's canonical name for this charset.
565 jstring javaCanonicalName = getJavaCanonicalName(env, icuCanonicalName);
566 if (env->ExceptionCheck()) {
567 return NULL;
568 }
569
570 // Check that this charset is supported.
571 {
572 // ICU doesn't offer any "isSupported", so we just open and immediately close.
573 UErrorCode error = U_ZERO_ERROR;
574 icu::LocalUConverterPointer cnv(ucnv_open(icuCanonicalName, &error));
575 if (!U_SUCCESS(error)) {
576 return NULL;
577 }
578 }
579
580 // Get the aliases for this charset.
581 std::vector<std::string> aliases;
582 if (!collectStandardNames(env, icuCanonicalName, "IANA", aliases)) {
583 return NULL;
584 }
585 if (!collectStandardNames(env, icuCanonicalName, "MIME", aliases)) {
586 return NULL;
587 }
588 if (!collectStandardNames(env, icuCanonicalName, "JAVA", aliases)) {
589 return NULL;
590 }
591 if (!collectStandardNames(env, icuCanonicalName, "WINDOWS", aliases)) {
592 return NULL;
593 }
594 jobjectArray javaAliases = toStringArray(env, aliases);
595 if (env->ExceptionCheck()) {
596 return NULL;
597 }
598
599 // Construct the CharsetICU object.
600 static jmethodID charsetConstructor = env->GetMethodID(JniConstants::charsetICUClass, "<init>",
601 "(Ljava/lang/String;Ljava/lang/String;[Ljava/lang/String;)V");
602 if (env->ExceptionCheck()) {
603 return NULL;
604 }
605
606 jstring icuCanonicalNameStr = env->NewStringUTF(icuCanonicalName);
607 if (env->ExceptionCheck()) {
608 return NULL;
609 }
610
611 return env->NewObject(JniConstants::charsetICUClass, charsetConstructor,
612 javaCanonicalName, icuCanonicalNameStr, javaAliases);
613 }
614
FreeNativeConverter(void * converter)615 static void FreeNativeConverter(void *converter) {
616 ucnv_close(reinterpret_cast<UConverter*>(converter));
617 }
618
NativeConverter_getNativeFinalizer(JNIEnv *,jclass)619 static jlong NativeConverter_getNativeFinalizer(JNIEnv*, jclass) {
620 return reinterpret_cast<jlong>(&FreeNativeConverter);
621 }
622
623
NativeConverter_getNativeSize(JNIEnv *,jclass,jstring)624 static jlong NativeConverter_getNativeSize(JNIEnv*, jclass, jstring) {
625 // TODO: Improve estimate.
626 return 200;
627 }
628
629 static JNINativeMethod gMethods[] = {
630 NATIVE_METHOD(NativeConverter, charsetForName, "(Ljava/lang/String;)Ljava/nio/charset/Charset;"),
631 NATIVE_METHOD(NativeConverter, closeConverter, "(J)V"),
632 NATIVE_METHOD(NativeConverter, contains, "(Ljava/lang/String;Ljava/lang/String;)Z"),
633 NATIVE_METHOD(NativeConverter, decode, "(J[BI[CI[IZ)I"),
634 NATIVE_METHOD(NativeConverter, encode, "(J[CI[BI[IZ)I"),
635 NATIVE_METHOD(NativeConverter, getAvailableCharsetNames, "()[Ljava/lang/String;"),
636 NATIVE_METHOD(NativeConverter, getAveBytesPerChar, "(J)F"),
637 NATIVE_METHOD(NativeConverter, getAveCharsPerByte, "(J)F"),
638 NATIVE_METHOD(NativeConverter, getMaxBytesPerChar, "(J)I"),
639 NATIVE_METHOD(NativeConverter, getMinBytesPerChar, "(J)I"),
640 NATIVE_METHOD(NativeConverter, getSubstitutionBytes, "(J)[B"),
641 NATIVE_METHOD(NativeConverter, openConverter, "(Ljava/lang/String;)J"),
642 NATIVE_METHOD(NativeConverter, resetByteToChar, "(J)V"),
643 NATIVE_METHOD(NativeConverter, resetCharToByte, "(J)V"),
644 NATIVE_METHOD(NativeConverter, setCallbackDecode, "(JIILjava/lang/String;)V"),
645 NATIVE_METHOD(NativeConverter, setCallbackEncode, "(JII[B)V"),
646 NATIVE_METHOD(NativeConverter, getNativeFinalizer, "()J"),
647 NATIVE_METHOD(NativeConverter, getNativeSize, "()J")
648 };
register_libcore_icu_NativeConverter(JNIEnv * env)649 void register_libcore_icu_NativeConverter(JNIEnv* env) {
650 jniRegisterNativeMethods(env, "libcore/icu/NativeConverter", gMethods, NELEM(gMethods));
651 }
652