On Sunday 21 November 2004 08:52, Alkis Evlogimenos wrote: > At line 115 of jni.h: typedef void* jobject; > > This type definition conflicts with any implementation dependent type > definition given by the vm when _JNI_VM_INTERNAL_TYPES_DEFINED. Should this > definition be in an else block when _JNI_VM_INTERNAL_TYPES_DEFINED? > > Here's the proposed patch:
The previous patch doesn't work with C++. This one works though: --- /tmp/classpath-0.12/include/jni.h 2004-11-14 17:41:18.000000000 -0600 +++ include/jni.h 2004-11-21 08:58:36.116468947 -0600 @@ -63,11 +63,11 @@ * declares the old versions of the types. */ #ifndef _JNI_VM_INTERNAL_TYPES_DEFINED -/* + typedef void *jobject; typedef void *jfieldID; typedef void *jmethodID; -*/ + struct _jfieldID; struct _jmethodID; typedef struct _jfieldID *jfieldID; @@ -112,7 +112,6 @@ #else /* __cplusplus */ -typedef void *jobject; typedef jobject jclass; typedef jobject jstring; typedef jobject jarray; -- Alkis _______________________________________________ Classpath mailing list [EMAIL PROTECTED] http://lists.gnu.org/mailman/listinfo/classpath
