Re: [RFC] Number of insn slots used for method compilation

[Tom Tromey <tromey at redhat dot com>]

>>>>> "Bryce" == Bryce McKinlay <mckinlay@redhat.com> writes:

Bryce> I don't think we use it at present. However, the JDWP ByteCodes
Bryce> command provides access to the original bytecode for the method, so we
Bryce> need access to code_length somehow in order to implement that.

Adding a field won't be a huge additional overhead if we're keeping
around the original bytes, etc.

I wonder if we should just disable the direct threading when JDWP is
in use.  That would mean refactoring a big chunk of
interpret.cc... but this would probably make it more readable.

Tom