< prev index next >
src/java.base/share/native/libjava/jni_util.c
Print this page
rev 17266 : 8181147: JNI_GetStringPlatformChars should have a fast path for UTF-8
Reviewed-by: shade, chegar, erikj
@@ -442,13 +442,12 @@
return obj;
}
/* Optimized for char set ISO_8559_1 */
static jstring
-newString8859_1(JNIEnv *env, const char *str)
+newSizedString8859_1(JNIEnv *env, const char *str, int len)
{
- int len = (int)strlen(str);
jchar buf[512];
jchar *str1;
jstring result;
int i;
@@ -467,10 +466,17 @@
if (str1 != buf)
free(str1);
return result;
}
+static jstring
+newString8859_1(JNIEnv *env, const char *str)
+{
+ int len = (int)strlen(str);
+ return newSizedString8859_1(env, str, len);
+}
+
static const char*
getString8859_1Chars(JNIEnv *env, jstring jstr)
{
int i;
char *result;
@@ -498,11 +504,10 @@
result[len] = 0;
(*env)->ReleaseStringCritical(env, jstr, str);
return result;
}
-
/* Optimized for char set ISO646-US (us-ascii) */
static jstring
newString646_US(JNIEnv *env, const char *str)
{
int len = (int)strlen(str);
@@ -669,12 +674,40 @@
/* Cached method IDs */
static jmethodID String_init_ID; /* String(byte[], enc) */
static jmethodID String_getBytes_ID; /* String.getBytes(enc) */
-int getFastEncoding() {
- return fastEncoding;
+/* Optimized for char set UTF-8 */
+static jstring
+newStringUTF8(JNIEnv *env, const char *str)
+{
+ jboolean isAscii = JNI_TRUE;
+ jstring result;
+ jbyteArray hab = NULL;
+ int len = 0;
+ char b;
+ for (b = str[len]; b != '\0'; len++, b = str[len]) {
+ if (isAscii && b & 0x80) {
+ isAscii = JNI_FALSE;
+ }
+ }
+
+ if (isAscii) {
+ return newSizedString8859_1(env, str, len);
+ }
+
+ hab = (*env)->NewByteArray(env, len);
+ if (hab != 0) {
+ jclass strClazz = JNU_ClassString(env);
+ CHECK_NULL_RETURN(strClazz, 0);
+ (*env)->SetByteArrayRegion(env, hab, 0, len, (jbyte *)str);
+ result = (*env)->NewObject(env, strClazz,
+ String_init_ID, hab, jnuEncoding);
+ (*env)->DeleteLocalRef(env, hab);
+ return result;
+ }
+ return NULL;
}
/* Initialize the fast encoding. If the "sun.jnu.encoding" property
* has not yet been set, we leave fastEncoding == NO_ENCODING_YET.
*/
@@ -716,21 +749,24 @@
* "en_UK" locale -> "ISO8859-1" (on 2.6)
*/
if ((strcmp(encname, "8859_1") == 0) ||
(strcmp(encname, "ISO8859-1") == 0) ||
(strcmp(encname, "ISO8859_1") == 0) ||
- (strcmp(encname, "ISO-8859-1") == 0))
+ (strcmp(encname, "ISO-8859-1") == 0)) {
fastEncoding = FAST_8859_1;
- else if (strcmp(encname, "ISO646-US") == 0)
+ } else if (strcmp(encname, "UTF-8") == 0) {
+ fastEncoding = FAST_UTF_8;
+ jnuEncoding = (jstring)(*env)->NewGlobalRef(env, enc);
+ } else if (strcmp(encname, "ISO646-US") == 0) {
fastEncoding = FAST_646_US;
- else if (strcmp(encname, "Cp1252") == 0 ||
+ } else if (strcmp(encname, "Cp1252") == 0 ||
/* This is a temporary fix until we move */
/* to wide character versions of all Windows */
/* calls. */
- strcmp(encname, "utf-16le") == 0)
+ strcmp(encname, "utf-16le") == 0) {
fastEncoding = FAST_CP1252;
- else {
+ } else {
fastEncoding = NO_FAST_ENCODING;
jnuEncoding = (jstring)(*env)->NewGlobalRef(env, enc);
}
(*env)->ReleaseStringUTFChars(env, enc, encname);
}
@@ -790,10 +826,12 @@
return newString8859_1(env, str);
if (fastEncoding == FAST_646_US)
return newString646_US(env, str);
if (fastEncoding == FAST_CP1252)
return newStringCp1252(env, str);
+ if (fastEncoding == FAST_UTF_8)
+ return newStringUTF8(env, str);
if ((*env)->EnsureLocalCapacity(env, 2) < 0)
return NULL;
len = (int)strlen(str);
@@ -828,31 +866,14 @@
GetStringPlatformChars(JNIEnv *env, jstring jstr, jboolean *isCopy)
{
return JNU_GetStringPlatformChars(env, jstr, isCopy);
}
-JNIEXPORT const char * JNICALL
-JNU_GetStringPlatformChars(JNIEnv *env, jstring jstr, jboolean *isCopy)
-{
+static const char* getStringBytes(JNIEnv *env, jstring jstr) {
char *result = NULL;
jbyteArray hab = 0;
- if (isCopy)
- *isCopy = JNI_TRUE;
-
- if (fastEncoding == NO_ENCODING_YET) {
- initializeEncoding(env);
- JNU_CHECK_EXCEPTION_RETURN(env, 0);
- }
-
- if ((fastEncoding == FAST_8859_1) || (fastEncoding == NO_ENCODING_YET))
- return getString8859_1Chars(env, jstr);
- if (fastEncoding == FAST_646_US)
- return getString646_USChars(env, jstr);
- if (fastEncoding == FAST_CP1252)
- return getStringCp1252Chars(env, jstr);
-
if ((*env)->EnsureLocalCapacity(env, 2) < 0)
return 0;
if (jnuEncodingSupported(env)) {
hab = (*env)->CallObjectMethod(env, jstr, String_getBytes_ID, jnuEncoding);
@@ -881,10 +902,78 @@
(*env)->DeleteLocalRef(env, hab);
return result;
}
+static const char*
+getStringUTF8(JNIEnv *env, jstring jstr)
+{
+ int i;
+ char *result;
+ jbyte encoding = 0;
+ jint len = 0;
+ const jbyte *str = (*env)->GetStringBytesCritical(env, jstr, &len, &encoding, NULL);
+
+ if (str == NULL) {
+ return NULL;
+ }
+
+ // only support ASCII encoding natively.
+ if (encoding == 0) {
+ result = (char*)str;
+ for (i = 0; i < len; i++) {
+ if (str[i] < 0) {
+ // not ASCII, abort to slow path
+ str = NULL;
+ break;
+ }
+ }
+ }
+
+
+ if (str == NULL) {
+ (*env)->ReleaseStringBytesCritical(env, jstr, str);
+ return getStringBytes(env, jstr);
+ }
+
+ result = MALLOC_MIN4(len);
+ if (result == 0) {
+ (*env)->ReleaseStringBytesCritical(env, jstr, str);
+ JNU_ThrowOutOfMemoryError(env, 0);
+ return NULL;
+ }
+ for (i = 0; i < len; i++) {
+ result[i] = (char)str[i];
+ }
+ result[len] = '\0';
+ (*env)->ReleaseStringBytesCritical(env, jstr, str);
+ return result;
+}
+
+JNIEXPORT const char * JNICALL
+JNU_GetStringPlatformChars(JNIEnv *env, jstring jstr, jboolean *isCopy)
+{
+
+ if (isCopy)
+ *isCopy = JNI_TRUE;
+
+ if (fastEncoding == NO_ENCODING_YET) {
+ initializeEncoding(env);
+ JNU_CHECK_EXCEPTION_RETURN(env, 0);
+ }
+
+ if ((fastEncoding == FAST_8859_1) || (fastEncoding == NO_ENCODING_YET))
+ return getString8859_1Chars(env, jstr);
+ if (fastEncoding == FAST_646_US)
+ return getString646_USChars(env, jstr);
+ if (fastEncoding == FAST_CP1252)
+ return getStringCp1252Chars(env, jstr);
+ if (fastEncoding == FAST_UTF_8)
+ return getStringUTF8(env, jstr);
+ else
+ return getStringBytes(env, jstr);
+}
JNIEXPORT void JNICALL
JNU_ReleaseStringPlatformChars(JNIEnv *env, jstring jstr, const char *str)
{
free((void *)str);
}
< prev index next >