This is interesting:
> 'pf_sbyte_IntPtr_IntPtr_pjvalue' to type 'CallBooleanMethod'.
The pointer to JNI method CallBooleanMethod is received from call to
JNI_CreateJavaVM on jvm.dll.
The jni4net is trying to find the DLL in JAVA_HOME
http://code.google.com/p/jni4net/source/browse/trunk/jni4net.n/src/jni/JNI.cs
I never seen this before, so I'm just guessing:
- JAVA_HOME is incorrect
- or wrong jvm.dll is on PATH and wins the binding
- something is corrupted, memory or files
> As far as we can tell the 32bit vs 64bit hint is incorrect.
It's just hint.
BTW: what version of CLR your friend have ?
Cheers
Pavel
Hi Nick, hi all
I thought what would be most appropriate message for it. The current
one might be misleading.
My current idea is
"Can't initialize jni4net. Talk to your software vendor or see
http://jni4net.sf.net/troubleshoot.shtml"
What you guys think ?
Also we could possibly improve the JVM detection. Find proper one in
"Program Files" vs "Program Files (x86)"is not that difficult.
See FindJvmDir()
http://code.google.com/p/jni4net/source/browse/trunk/jni4net.n/src/jni/JNI.cs
Also as far as I can remember the inner exception is some specific
type, when we hit 64 vs 32, we could do better there.
Shall we autofix wrong JAVA_HOME ? I don't think so, but would like to
hear from folks.
Anyone interested to contribute it ?
Thanks
Pavel
Improved JVM detection would certainly be a boon. I might even volunteer.
And let's not fix wrong JAVA_HOME; not our problem - and probably Oracle/Sun
should take care of it.
Hi Nick, hi all
Thanks
Pavel
--
You received this message because you are subscribed to
jni...@googlegroups.com
http://groups.google.com/group/jni4net?hl=en-GB?hl=en-GB
http://jni4net.sf.net/