On Tue, 2 Jun 2020 at 00:32, Richard Henderson <richard.hender...@linaro.org> wrote: > It might be clearer to use dup_const for each case, which would more closely > match the pseudocode. E.g. here, > > return dup_const(MO_16, imm << 8); > > > + imm |= (imm << 8) | (imm << 16) | (imm << 24); > > return dup_const(MO_8, imm);
Yeah, I did think about this, but figured that keeping the existing code structure was clearer for purposes of reviewing this refactoring series. thanks -- PMM
