I know we can use command line argument like –X:base:mc=true or –X:opt:mc=true to dump out machine code.

And I notice that for Intel architecture, the dump work is done in VM_Lister.java. It seems fine in BaseBaseSemiSpace

configuration. But in FastAdaptiveSemiSpace configuration, VM_Lister.java is created from VM_Lister.fake in

rvmRoot/rvm/src/vm/arch/intel/compilers/optimizing/ir/conversions/mir2mc/. And the body of each method is empty.

How does the real dump work do? And under which reason do you choose to use seperate VM_Lister.java file?