On Tue, Jun 9, 2009 at 10:07 AM, Mark Click <mark...@gmail.com> wrote:
On Tue, Jun 9, 2009 at 6:13 AM, Timothy Wall <twal...@dev.java.net>wrote:
On Jun 8, 2009, at 6:09 PM, Mark Click wrote:
On Mon, Jun 8, 2009 at 2:17 PM, Timothy Wall <twal...@dev.java.net>
wrote:
Also make sure you use the same Java test program to test both scenarios,
so that you've ensured that the Java environment is otherwise identical.
Examine libjnidispatch.so and your 3rd-party shared library using ldd;
see if there are any dependent library conflicts (e.g. libc.so).
Here's the ldd output for the libraries, http://pastebin.com/m3c041b92
It looks like everything is normal/correct?
What happens when your JNI code dynamically loads your 3rd-party library,
e.g.
void * handle = dlopen("/path/to/lib3rdparty.so", RTLD_LAZY);
void (*func)(wchar_t*) = (void (*)(wchar_t*))dlsym("entryPoint");
(*func)(NULL);
dlclose(handle);
Everything works still, just as the JNI without it.
Does JNA pin memory allocated by native code as JNI does? There's no way
memory used for the deserialization could be getting garbage collected
before use? Sorry, grasping at straws at this point.
- Mark
33 messages in net.java.dev.jna.usersRe: [jna-users] Linux SIGSEGV under J...
.txt
