Re: [RFA] Insn handling methods for _Jv_InterpMethod

[Tom Tromey <tromey at redhat dot com>]

>>>>> "Keith" == Keith Seitz <keiths@redhat.com> writes:

Keith> The attached patch adds three methods to _Jv_InterpMethod which deal
Keith> with manipulating the method's insns. These three methods are used by
Keith> breakpoint management code to install and remove breakpoints.

Keith> Questions/comments/concerns?

Looks ok.  A couple questions though.

Keith> +pc_t
Keith> +_Jv_InterpMethod::install_break (jlong index)
Keith> +{
Keith> +#ifdef DIRECT_THREADED
Keith> +  static insn_slot breakpoint_insn = {0};
Keith> +  printf ("_Jv_InterpMethod::install_break at index %ld\n", (long) index);

printf...

Keith> +  return set_insn (index, &breakpoint_insn);
Keith> +#else
Keith> +  static pc_t breakpoint_insn = NULL;
Keith> +  return set_insn (index, breakpoint_insn);
Keith> +#endif

Is the 'breakpoint_insn' stuff a placeholder for something that will
come later?

Keith> +  code = reinterpret_cast<pc_t> (prepared);

Perhaps 'prepared' should simply be declared as pc_t?

I think in the past we didn't declare this in java-interp.h.  But now
that we do, being type-correct seems nicer.

Tom