From @masak

<moritz> on a completely different topic, S12 says that 'self' is
always assumed to be in item context
<moritz> why is that the case? that's the role of the $ sigil, which
isn't there on 'self'
<masak> the role of the $ sigil is forcing item context.
<masak> 'self', by current spec, has that built-in.
<moritz> yes, but why?
<masak> I don't know. maybe because of some psychological expectation
that 'self' is a single thing?
<moritz> we've extra eliminiated the $ from (|$foo) variables, because
it doesn't fit the "$ means item" philosphy
<masak> do you have a use case where 'self' being in item context is
cumbersome or unfortunate?
<masak> can you *guarantee* that a lot of existing Perl 6 code (in
CORE, for example) doesn't rely on the assumption that 'self' is in
item context? :)
<moritz> r​: my @​a = 1, 2; .say for @​a.push​: 3
<p6eval> rakudo 1a468d​: OUTPUT«1␤2␤3␤»
<moritz> push simply returns self. Very convenient, and only possible
because rakudo doesn't enforce the contnext
<masak> moritz​: so you're saying that Rakudo isn't to spec in that regard?
<moritz> masak​: correct
* masak submits rakudobug
* masak grins
<moritz> nom​: class A is Array { method identity { self } }; my @​a :=
A.new; @​a = 3, 4, 5; .say for @​a.identity
<p6eval> rakudo 1a468d​: OUTPUT«3␤4␤5␤»
<moritz> there you go, without reference to any core thing
<benabik> r​: my @​a = 1,2; .say for $(@​a.push​: 3)
<p6eval> rakudo 1a468d​: OUTPUT«1 2 3␤»

From @jnthn

On Wed Mar 21 08​:38​:33 2012, masak wrote​:

<moritz> on a completely different topic, S12 says that 'self' is
always assumed to be in item context
<moritz> why is that the case? that's the role of the $ sigil, which
isn't there on 'self'
<masak> the role of the $ sigil is forcing item context.
<masak> 'self', by current spec, has that built-in.
<moritz> yes, but why?
<masak> I don't know. maybe because of some psychological expectation
that 'self' is a single thing?
<moritz> we've extra eliminiated the $ from (|$foo) variables, because
it doesn't fit the "$ means item" philosphy
<masak> do you have a use case where 'self' being in item context is
cumbersome or unfortunate?
<masak> can you *guarantee* that a lot of existing Perl 6 code (in
CORE, for example) doesn't rely on the assumption that 'self' is in
item context? :)
<moritz> r​: my @​a = 1, 2; .say for @​a.push​: 3
<p6eval> rakudo 1a468d​: OUTPUT«1␤2␤3␤»
<moritz> push simply returns self. Very convenient, and only possible
because rakudo doesn't enforce the contnext
<masak> moritz​: so you're saying that Rakudo isn't to spec in that regard?
<moritz> masak​: correct
* masak submits rakudobug
* masak grins
<moritz> nom​: class A is Array { method identity { self } }; my @​a :=
A.new; @​a = 3, 4, 5; .say for @​a.identity
<p6eval> rakudo 1a468d​: OUTPUT«3␤4␤5␤»
<moritz> there you go, without reference to any core thing
<benabik> r​: my @​a = 1,2; .say for $(@​a.push​: 3)
<p6eval> rakudo 1a468d​: OUTPUT«1 2 3␤»

As of spec commit 8284cda, the current Rakudo behavior is now correct. There's also a test in roast commit 0ffea18.

/jnthn

The RT System itself - Status changed from 'new' to 'open'

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