On Fri, May 16, 2025 at 07:05:30PM +0700, Robert Elz wrote: > How about we make it > &"lstat"[usestat] > then? > > Or even > &usestat["lstat"] > ? > > They all mean the same thing, but those might almost be considered obfuscated. > The first sounds as if it is what the clang warning was saying should be used > to avoid producing that warning. > > They all avoid the code needing to do 2 jumps in at least some cases > (or one in all), and having two strings (though a good compiler might > avoid that).
I don't like the idea of micro-optimizing an error path.
