Currently, errors are thrown across the JNI boundary into Java.
However, when that happens apparently the C++ stack trace is lost, we'd like to
avoid that, perhaps using one of the signals defined in Signal.h (SIGINT?
SIGSTOP?).
If enabled, this should be an optional behavior, we still want to be able to
handle those errors as Java exceptions in most cases.
Original issue reported on code.google.com by javier.barreiro@gmail.com on 1 Dec 2010 at 11:58
Original issue reported on code.google.com by
javier.barreiro@gmail.com
on 1 Dec 2010 at 11:58