From @masak

<masak> rakudo​: say ~(1, 1, &[+] ... *)[^10]
<p6eval> rakudo 11cbd4​: OUTPUT«get_attr_str() not implemented in class
'Perl6MultiSub'␤ in 'infix​:<...>' at line 1 [...]
* masak submits rakuodbug
<colomon> rakudo​: say ~(1, 1, &[+] ... 50)
<p6eval> rakudo 11cbd4​: OUTPUT«get_attr_str() not implemented in class
'Perl6MultiSub'␤ in 'infix​:<...>' at line 1 [...]
<colomon> rakudo​: say ~(1, 1, *+* ... 50)
<p6eval> rakudo 11cbd4​: OUTPUT«1 1 2 3 5 8 13 21 34␤»
<moritz_> I guess it's related to &[+] referring to a multi
<colomon> moritz_​: yes
<colomon> I'm not sure the &[+] can actually work here.
<masak> why not?
<colomon> ... needs to know how many arguments the closure takes.
<masak> ah.
<colomon> how many does &[+] take?
<masak> two.
<colomon> or zero
<masak> zero!?
<moritz_> for [+] to return 0
<colomon> rakudo​: say infix​:<+>()
<p6eval> rakudo 11cbd4​: OUTPUT«0␤»
<masak> huh.
<TimToady> &[+] is specced to choose the 2-arg case
<masak> that settles it. thanks.
<colomon> there's the bug, then. :)
<TimToady> S03​:4686

@coke - Status changed from 'new' to 'open'

From patrick.abisalloum@gmail.com

On Thu Jun 24 05​:18​:07 2010, masak wrote​:

<masak> rakudo​: say ~(1, 1, &[+] ... *)[^10]
<p6eval> rakudo 11cbd4​: OUTPUT«get_attr_str() not implemented in class
'Perl6MultiSub'␤ in 'infix​:<...>' at line 1 [...]
* masak submits rakuodbug
<colomon> rakudo​: say ~(1, 1, &[+] ... 50)
<p6eval> rakudo 11cbd4​: OUTPUT«get_attr_str() not implemented in class
'Perl6MultiSub'␤ in 'infix​:<...>' at line 1 [...]
<colomon> rakudo​: say ~(1, 1, *+* ... 50)
<p6eval> rakudo 11cbd4​: OUTPUT«1 1 2 3 5 8 13 21 34␤»
<moritz_> I guess it's related to &[+] referring to a multi
<colomon> moritz_​: yes
<colomon> I'm not sure the &[+] can actually work here.
<masak> why not?
<colomon> ... needs to know how many arguments the closure takes.
<masak> ah.
<colomon> how many does &[+] take?
<masak> two.
<colomon> or zero
<masak> zero!?
<moritz_> for [+] to return 0
<colomon> rakudo​: say infix​:<+>()
<p6eval> rakudo 11cbd4​: OUTPUT«0␤»
<masak> huh.
<TimToady> &[+] is specced to choose the 2-arg case
<masak> that settles it. thanks.
<colomon> there's the bug, then. :)
<TimToady> S03​:4686

This can be narrowed down to
[19​:50] <patrickas> rakudo​: my $foo = &prefix​:<!>;say $foo.signature;
[19​:51] <+p6eval> rakudo a47c45​: OUTPUT«get_attr_str() not implemented
in class 'Perl6MultiSub'␤ in main program body at line 1␤»

From @coke

On Thu Jun 24 05​:18​:07 2010, masak wrote​:

<masak> rakudo​: say ~(1, 1, &[+] ... *)[^10]
<p6eval> rakudo 11cbd4​: OUTPUT«get_attr_str() not implemented in class
'Perl6MultiSub'␤ in 'infix​:<...>' at line 1 [...]
* masak submits rakuodbug
<colomon> rakudo​: say ~(1, 1, &[+] ... 50)
<p6eval> rakudo 11cbd4​: OUTPUT«get_attr_str() not implemented in class
'Perl6MultiSub'␤ in 'infix​:<...>' at line 1 [...]
<colomon> rakudo​: say ~(1, 1, *+* ... 50)
<p6eval> rakudo 11cbd4​: OUTPUT«1 1 2 3 5 8 13 21 34␤»
<moritz_> I guess it's related to &[+] referring to a multi
<colomon> moritz_​: yes
<colomon> I'm not sure the &[+] can actually work here.
<masak> why not?
<colomon> ... needs to know how many arguments the closure takes.
<masak> ah.
<colomon> how many does &[+] take?
<masak> two.
<colomon> or zero
<masak> zero!?
<moritz_> for [+] to return 0
<colomon> rakudo​: say infix​:<+>()
<p6eval> rakudo 11cbd4​: OUTPUT«0␤»
<masak> huh.
<TimToady> &[+] is specced to choose the 2-arg case
<masak> that settles it. thanks.
<colomon> there's the bug, then. :)
<TimToady> S03​:4686

These are all happy now​:

22​:20 < [Coke]> rakudo​: say ~(1, 1, &[+] ... *)[^10]
22​:20 <+p6eval> rakudo 142c41​: OUTPUT«1 1 2 3 5 8 13 21 34 55␤»
22​:20 < [Coke]> rakudo​: say ~(1, 1, &[+] ... 50)
22​:21 <+p6eval> rakudo 142c41​: OUTPUT«(timeout)»
22​:21 < [Coke]> rakudo​: say ~(1, 1, &[+] ... 55)
22​:21 <+p6eval> rakudo 142c41​: OUTPUT«1 1 2 3 5 8 13 21 34 55␤»
22​:21 < [Coke]> rakudo​: my $foo = &prefix​:<!>;say $foo.signature;
22​:21 <+p6eval> rakudo 142c41​: OUTPUT«​:(Any)␤»

Closable with tests.

--
Will "Coke" Coleda

From @moritz

We've been running and passing tests for that in
t/spec/S03-sequence/basic.t for quite a while now, resolving...

@moritz - Status changed from 'open' to 'resolved'