Might be because there's a prefix form of + with a wider type? "Every DWIM has a WAT"
On Sun, Sep 30, 2018 at 9:55 AM Fernando Santagata < nando.santag...@gmail.com> wrote: > What I was pointing out is that the '*' operator outputs an error, and I > was expecting that, while the '+' doesn't. > Besides, the '+' delivers the correct result, which apparently indicates > that the two operators are treated differently. > > I was trying to understand whether this behavior is inherent to the Perl 6 > type system or it's just a bug. > > On Sun, Sep 30, 2018, 12:27 JJ Merelo <jjmer...@gmail.com> wrote: > >> This works: >> >> my int32 $a = 3; my int32 $b = 7; say $a * $b # OUTPUT: «21» >> >> El dom., 30 sept. 2018 a las 11:28, Fernando Santagata (< >> nando.santag...@gmail.com>) escribió: >> >>> Hi, >>> >>> I was hunting for a bug in my code, this one: >>> >>> my @a := CArray[int32].new: 6,3; >>> say [+] @a; # it should be "say [+] @a.list;" >>> >>> That statement prints "9", while changing the '+' with a '*' it >>> generates an error, as it should: >>> >>> Cannot resolve caller Numeric(NativeCall::Types::CArray[int32]: ); none >>> of these signatures match: >>> (Mu:U \v: *%_) >>> >>> Is this a bug, an inconsistency, or a necessary evil? >>> >> >> It's difficult to say. Some conversion is taking place somewhere to >> Numeric, and it's getting stuff it does not like. The type is >> NativeCall::Types::CArray[int32].new and it's being considered a single >> item; it's attempting conversion to Numeric. The thing here is that CArray >> is actually not an Array; it inherits directly from Any: >> >> say @a.^mro; # OUTPUT: «((CArray[int32]) (CArray) (Any) (Mu))» >> >> It does include a `list` for convenience, but it's not, by itself, a >> list. That operator does not do the conversion, but expects it to be a >> list, hence the error. >> >> Hope this helps >> >> JJ >> > -- brandon s allbery kf8nh allber...@gmail.com
