From @zoffixznet

There is some combination of subset/multi that causes Cannot invoke
this object (REPR​: Null; VMNull) error to be displayed​:

<ZoffixW> m​: subset InfTan of Numeric where $_ == 1; multi tan
(InfTan) {}; say tan(2)
<camelia> rakudo-moar f7770e​: OUTPUT«Cannot invoke this object (REPR​:
Null; VMNull)␤ in any accepts_type at gen/moar/m-Metamodel.nqp line
3472␤ in sub tan at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
<ZoffixW> m​: subset InfTan of Num where $_ % (π/2) ≅ 0; multi tan
(InfTan) { ∞ }; say tan(π)
<camelia> rakudo-moar f7770e​: OUTPUT«Cannot invoke this object (REPR​:
Null; VMNull)␤ in any accepts_type at gen/moar/m-Metamodel.nqp line
3472␤ in sub tan at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»

Yet, slightly altered variations such as changing base type of subset,
the where condition, or multi-to-only do not produce that error​:

<ZoffixW> m​: subset InfTan of Num where $_ == 1; multi tan (InfTan)
{}; say tan(2)
<camelia> rakudo-moar f7770e​: OUTPUT«-2.18503986326152␤»
<ZoffixW> m​: subset InfTan of Num where $_ == 1; multi tan (InfTan)
{}; say tan(2)
<camelia> rakudo-moar f7770e​: OUTPUT«-2.18503986326152␤»
<ZoffixW> m​: multi tan ($ where * % (π/2) ≅ 0 ) { ∞ }; say tan(π)
<camelia> rakudo-moar f7770e​: OUTPUT«-1.22464679914735e-16␤»
<ZoffixW> m​: multi tan ($ where $_ % (π/2) ≅ 0 ) { ∞ }; say tan(π)
<camelia> rakudo-moar f7770e​: OUTPUT«-1.22464679914735e-16␤»
<ZoffixW> m​: subset InfTan of Num; multi tan (InfTan) { ∞ }; say tan(π)
<camelia> rakudo-moar f7770e​: OUTPUT«Inf␤»
<ZoffixW> m​: subset InfTan of Num where * % (π/2) ≅ 0; say tan(π) ~~ InfTan
<camelia> rakudo-moar f7770e​: OUTPUT«False␤»

From @zoffixznet

Seems an optimizer bug. Running with --optimize=off does not have the issue present.

From @zoffixznet

On Wed, 14 Dec 2016 15​:14​:27 -0800, cpan@​zoffix.com wrote​:

Seems an optimizer bug. Running with --optimize=off does not have the
issue present.

Fixed and tested as rakudo/rakudo#1465

@zoffixznet - Status changed from 'new' to 'resolved'