From pjablons@conncoll.edu

10​:37 ren1us m​: sub a(@​list is copy, $l = @​list.elems) { say "Elems​: " ~ @​list.elems; say "l​: $l"; }; a(<a b c d e f g>); a(<a b c>);
10​:37 camelia rakudo-moar cce7f9​: OUTPUT«Elems​: 7␤l​: 7␤Elems​: 3␤l​: 7␤»
10​:37 ren1us m​: sub a(@​list, $l = @​list.elems) { say "Elems​: " ~ @​list.elems; say "l​: $l"; }; a(<a b c d e f g>); a(<a b c>);
10​:37 camelia rakudo-moar cce7f9​: OUTPUT«Elems​: 7␤l​: 7␤Elems​: 3␤l​: 3␤»
10​:45 colomon_ sure looks like a rakudobug
10​:45 * ren1us submits
10​:47 jnthn What the thunk...
10​:48 jnthn I think the first one goes through the slow-path binder thanks to the "is copy"
10​:49 ren1us m​: class A { method test(@​list is copy, $l = @​list.elems) { say "Elems​: " ~ @​list.elems; say "l​: $l"; }; }; my $a = A.new; $a.test(<a b c d e f g>); $a.test(<a b c>);
10​:49 camelia rakudo-moar cce7f9​: OUTPUT«Elems​: 7␤l​: 7␤Elems​: 3␤l​: 7␤»
10​:49 jnthn So the bug is probably in there.

From @usev6

This works now as expected (on 'nom' and 'glr')​:

$ perl6-m -e 'sub a(@​list is copy, $l = @​list.elems) { say "Elems​: " ~ @​list.elems; say "l​: $l"; }; a(<a b c d e f g>); a(<a b c>)'
Elems​: 7
l​: 7
Elems​: 3
l​: 3

I added a test to S06-traits/is-copy.t with commit Raku/roast@4747cd49fc

I'm closing this ticket as 'resolved'.

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