From @masak

<masak> rakudo​: role A { has @​!foo; method bar() { @​!foo ~~ (1,2); }
}; class B does A { has $.baz; }; B.new().bar()
<p6eval> rakudo 6999e5​: OUTPUT«Null PMC access in isa()␤in method A[]​::bar [...]
* masak submits rakudobug
<masak> rakudo​: role A { has @​!foo; method bar() { say @​!foo } };
class B does A { has $.baz; }; B.new().bar()
<p6eval> rakudo 6999e5​: OUTPUT«Null PMC access in isa()␤in method A[]​::bar [...]

I've been testing the phenomenon pretty extensively, and it doesn't
seem to matter whether it's @​!foo, $!foo or %!foo -- the error occurs
in all those cases. Also doesn't matter if it's $.baz or $!baz.
Removing $.baz entirely makes the error go away. Defining &bar in B
instead of A also doesn't make any difference.

From @jnthn

On Tue Jul 28 16​:32​:52 2009, masak wrote​:

<masak> rakudo​: role A { has @​!foo; method bar() { @​!foo ~~ (1,2); }
}; class B does A { has $.baz; }; B.new().bar()
<p6eval> rakudo 6999e5​: OUTPUT«Null PMC access in isa()␤in method
A[]​::bar [...]
* masak submits rakudobug
<masak> rakudo​: role A { has @​!foo; method bar() { say @​!foo } };
class B does A { has $.baz; }; B.new().bar()
<p6eval> rakudo 6999e5​: OUTPUT«Null PMC access in isa()␤in method
A[]​::bar [...]

I've been testing the phenomenon pretty extensively, and it doesn't
seem to matter whether it's @​!foo, $!foo or %!foo -- the error occurs
in all those cases. Also doesn't matter if it's $.baz or $!baz.
Removing $.baz entirely makes the error go away. Defining &bar in B
instead of A also doesn't make any difference.

Fixed in git 6999e58 and test added to S14-role/attributes.t.

Thanks,

Jonathan

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

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