From @japhb

09​:09 < japhb__> r​: class A { has Int @​.ints }; my A $one .= new(ints
=> [1, 2]); say $one.ints; my $two = $one.clone(ints => [3, 4, 5]);
say $two.ints;
09​:09 < camelia> rakudo 0808ac​: OUTPUT«1 2␤Type check failed in
assignment to '@​!ints'; expected 'Int' but got 'Array'␤ in method
REIFY at src/gen/CORE.setting​:8106␤ in method reify at
src/gen/CORE.setting​:7027␤ in method gimme at
src/gen/CORE.setting​:7441␤ in method sink at src/gen/CORE.setting…»

09​:10 < jnthn> japhb__​: Something's inconsistent there...

09​:11 < japhb__> Yeah. It feels like there's just something missing
in one code path, or the .clone code bitrotted a bit, or somesuch

09​:12 < jnthn> japhb__​: Please RT it if you didn't already

From @usev6

I added a test (fudged 'skip') to S12-attributes/instance.t with commit Raku/roast@dbf53cb4b1

AFAIU this inconsistency stems from using nqp​::decont on attributes given to .new (in src/core/Mu.pm), but not doing so on attributes given to .clone. I added a PR (rakudo/rakudo#443) which would make the added test pass.

@usev6 - Status changed from 'new' to 'open'

From @usev6

The test passes now. I'm closing this ticket as 'resolved'.

@usev6 - Status changed from 'open' to 'resolved'