views:

261

answers:

1

I have a jni lib that crashes on a certain point. What I get is a hs_err_pid file with

[error occurred during error reporting (printing native stack), id 0xb]

content and a core file of this error creation. Therefore I would like to stop jvm from capturing these signals and just let the process dump core normally instead. Platform Debian/Linux, Java(TM) SE Runtime Environment (build 1.6.0_16-b01).

A: 

You can run the program under gdb, or attach gdb to the process before it crashes. Then when the SIGSEGV occurs, gdb will stop the program and allow you to inspect it before the signal handler executes.

mark4o
I had done that, somehow gdb captured all kind of other crashes when continuing.Done that now in my 32bit chroot and indeed I get a nice bt on the crash point.Thanks!