* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
 * This source file is part of SableVM.                            *
 *                                                                 *
 * See the file "LICENSE" for the copyright information and for    *
 * the terms and conditions for copying, distribution and          *
 * modification of this source file.                               *
 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

HANDLING SIGNALS in SableVM (current implementation which is fine):
  static void
  _svmf_signal_throw (_svmt_JNIEnv *env, jint signal_code)
  {
    assert (env->signal_code == SVM_SIGNAL_NONE);
    if (env->signal_handler == NULL)
      {
        _svmt_JavaVM *vm = env->vm;
        if (vm->initialization == NULL)
   {
     return;  /* delegate the signal */
   }
        _svmf_initialization_unrecoverable_exception (env);
      }
    env->signal_code = signal_code;
    siglongjmp (*(env->signal_handler), 1);
  }

  This is how SableVM gets out of the signal handler.
  Important to understand: SIGSEGV and other hardware interrupts caused by
  executing code (e.g. div by zero, access to protected memory, bus
  misalignment [e.g. trying to store a long with wrong alignment in memory])
  are delivered to a signal handler on the same execution thread.  This make
  the longjump() call possible. Those interrupts, if not handled, usually
  result in the OS killing the application (core dump).