I agree it's a code _smell_. But a "smell" doesn't mean that something is necessarily wrong, just that there's a clue that it might be wrong. > in that sense I

I agree it's a code _smell_. But a "smell" doesn't mean that something is necessarily wrong, just that there's a clue that it might be wrong. > in that sense I disagree that the duplicated version is more readable I didn't say it is, I agreed it's _less_ readable. I said it's trading off readability for 8 bytes of memory at runtime. > If you consistently deduplicate code that is supposed to do the same and evolve the same, then [...] I agree with all this. I'm not saying to consistently go for the deduplicated approach (I don't think anyone would say that), I'm saying it's a reasonable trade-off in this specific case (each branch is still trivial, and the code won't evolve much if at all). > About possible performance reasons, those need an explanatory comment, exactly for the above reason. Agreed. > if performance reasons warrant complicating the code, then it isn’t “as simple as it gets” any more. I was commenting because I disagreed with that latter characterization. Also agreed.