# New Ticket Created by "Carl Mäsak"
# Please include the string: [perl #76242]
# in the subject line of all future correspondence about this issue.
# <URL: http://rt.perl.org/rt3/Ticket/Display.html?id=76242 >
<masak> rakudo: my $a = [1, 2, 3]; foo($a); bar($a); sub foo(@a) { say
@a.perl }; sub bar(@a is copy) { say @a.perl }
<p6eval> rakudo 05684c: OUTPUT«[1, 2, 3][[1, 2, 3]]»
<masak> that one's new, isn't it?
* masak submits rakudobug
<pmichaud> masak: that one is interesting. I'm not sure it's a bug.
<mathw> pmichaud: it is quite surprising...
<masak> pmichaud: oh, come on! :/
<mathw> although I suspect flattening would hide it in many cases
<pmichaud> my $a = [1,2,3]; my @b = $a; say @b.perl;
<masak> pmichaud: er. put differently, where does the spec say
"putting 'is copy' on an array gives you an extra layer of array"?
<moritz_> rakudo: sub bar(@a is copy) { say @a.perl }; say bar [1, 2, 3]
<p6eval> rakudo 05684c: OUTPUT«[[1, 2, 3]]1»
<moritz_> pmichaud: that's assignment
<moritz_> pmichaud: signature binding is *binding*
<pmichaud> moritz_: except for is_copy
<pmichaud> it's the same logic by which
<pmichaud> my $b = [1,2,3]; sub xyz($x is copy) { $x[1] = 'b'; };
xyz($b); say $b.perl;
<pmichaud> rakudo: my $b = [1,2,3]; sub xyz($x is copy) { $x[1] =
'b'; }; xyz($b); say $b.perl;
<p6eval> rakudo 05684c: OUTPUT«[1, "b", 3]»
<pmichaud> anyway, I can fix it.
<pmichaud> I agree that the other interpretation likely makes more sense.
<masak> rakudo: my $a = 5; sub bar(@a is copy) { say @a.perl }; bar($a)
<p6eval> rakudo 05684c: OUTPUT«Nominal type check failed for parameter
'@a'; expected Positional but got Int instead in 'bar' [...]
<masak> pmichaud: if the above doesn't bind, why does my first line
give double arrays?
<masak> seems very arbitrary.
<pmichaud> I just said I wasn't certain of it. :-)
<masak> just trying to make you even less certain :P
<pmichaud> the "is copy" logic is still not completely nailed down,
based on discussion with TimToady++ at yapc::na
<masak> I do see the parallel with attr arrays and :attr[]. by the way.
<pmichaud> anyway, the one thing I *am* certain of is that "is copy"
is not "binding". Indeed, the whole point of "is copy" is to avoid the
bind. :-)
<mathw> yes that's what I thought :)
<moritz_> I thought it copied, and then binds
<pmichaud> okay, one can think of it that way. :-)
<pmichaud> but it's not binding to the argument
<pmichaud> in the same way, my @b = $a; is a bind only in the sense
that we bind he symbol '@b' to an array before doing the assign
<pmichaud> anyway, I know the fix. Should be one-line :)
<pmichaud> did the ticket get filed yet?
