masak (>): > <masak> rakudo: say (my @a).end > <p6eval> rakudo 142d22: OUTPUT«-1» > <masak> S32 says: "Returns the final subscript of the first dimension; > for a one-dimensional array this simply the index of the final > element." > * masak submits rakudobug > <masak> huh. I didn't even see the missing "is" in that second > sentence before :) > <jnthn> masak: huh. If there's nothing in there, what do you want it > to return? > <jnthn> 0 would be wrong because that'd imply there's one thing in > there. > <masak> jnthn: I don't know. but it's not to spec right now. > <masak> jnthn: maybe Failure. > <lue> but -1 could be in a list as well. > <masak> lue: if you say that, you've misunderstood the purpose of .end > (just like I did first time I saw it) > <jnthn> masak: I think you maybe need to explain it then. :) > <jnthn> masak: 'cus I suspect I'm missing something also... > <masak> well, first off, let's agree on what S32 actually *says*. > <masak> it says "returns the final *subscript*. > <masak> an empty array has no final subscript. > <masak> hence -1 is *wrong*. > <jnthn> Hm > <jnthn> True > <jnthn> It may be a handy sentinel if you planeed to do some math on > it. > <masak> if S32 said "returns less than the length of the array", -1 > would have been fine. > <jnthn> But yeah, I can see flagging it up as a Failure is probably > better. > <masak> but it doesn't. > <masak> so it isn't. > <masak> this is why we have a spec, for precision and clarity. > <masak> in this case, it's precise and clear. hence, bug. > <masak> if you want -1, please use C<+...@a - 1> :) > <jnthn> fairy nuff.
Spec has now been changed to explicitly mention -1, which means Rakudo's behavior is right. Rejecting ticket.
