A stack trace occurs when the following sequence is executed: (fhpd) load ../pkglibdir/funit-hello [0.0] Loaded executable file: /home/rmoseley/frysk-git1/build/frysk-core/frysk/bindir/../pkglibdir/funit-hello (fhpd) load ../pkglibdir/funit-threads-looper [1.0] Loaded executable file: /home/rmoseley/frysk-git1/build/frysk-core/frysk/bindir/../pkglibdir/funit-threads-looper (fhpd) focus Target set pid id path-to-executable [0.0] 0 0 /home/rmoseley/frysk-git1/build/frysk-core/frysk/bindir/../pkglibdir/funit-hello [1.0] 0 0 /home/rmoseley/frysk-git1/build/frysk-core/frysk/bindir/../pkglibdir/funit-threads-looper (fhpd) run Attached to process 15662 Attached to process 15663 running with this command: ../pkglibdir/funit-hello Running process 15662 running with this command: ../pkglibdir/funit-threads-looper Running process 15663 (fhpd) Task 15662 is terminating from signal SIGSEGV(11) focus Target set pid id path-to-executable [0.0] 15662 15662 /home/rmoseley/frysk-git1/build/frysk-core/frysk/pkglibdir/funit-hello [1.0] 15663 15663 /home/rmoseley/frysk-git1/build/frysk-core/frysk/pkglibdir/funit-threads-looper [1.1] 15663 15664 (same) (fhpd) run Killing process 15662 that was created from /home/rmoseley/frysk-git1/build/frysk-core/frysk/pkglibdir/funit-hello Killing process 15663 that was created from /home/rmoseley/frysk-git1/build/frysk-core/frysk/pkglibdir/funit-threads-looper Exception in thread "main" java.lang.NullPointerException at frysk.sys.Fork.spawn(Fork.cxx:134) at frysk.sys.Fork.ptrace(Fork.java:97) at frysk.proc.live.LinuxPtraceHost$4.execute(LinuxPtraceHost.java:273) at frysk.event.EventLoop.runEventLoop(EventLoop.java:365) at frysk.event.EventLoop.run(EventLoop.java:482) at frysk.bindir.fhpd.main(fhpd.java:185)
Ends up not being related to sysroot. frysk.proc.live.LinuxPtraceHost.requestCreateAttachedProc sometimes arrives with args == null. So you end up in frysk.proc.live.LinuxPtraceHost$4.execute() then frysk.sys.Fork.ptrace then spawn() which checks: int argc = JvGetArrayLength (args); which cuases the failure
This bug has been fixed, as Stan found out it was null args being passed up to the code that activates a process. This was fixed when bz's 550/5501 were fixed.