About Bug 1742275: featurep(false, 'complex)



Harald,

I'm generally sympathetic to your views about featurep,
and I agree it would be cleaner to move the inferential stuff
into a separate function so that featurep only tests for
declared properties, but I'm hesitant to do so until the
inferential stuff finds a home in another existing or new function.

That said, to handle the case which is mentioned in the
bug report, I think featurep(false, foo) should be false for
all built-in properties foo, except for 'constant, same with
featurep(true, foo) and featurep("bar", foo), and without
changing the value returned by featurep for any other arguments.

best
Robert