Looks good. Just make sure there isn't a take in ArchSpec.cpp that converts COFF cpu type/subtype that relies on the value being 0 instead of LLDB_INVALID_CPUTYPE (I also can't remember if LLDB_INVALID_CPUTYPE is zero?). On Nov 21, 2013, at 2:30 PM, Colin Riley <colin at codeplay.com> wrote: > <ObjectFilePECOFF.cpp.patch>