A hole in $/.list is a Mu, but a hole in @a.List is a Nil / null object error

[zoffixznet]

IMO that Mu should be a Nil as well, as many routines don't take a Mu and so working with bare Mu is painful. In fact, if you index into that element, you get a Nil instead of a Mu

<Zoffix_> m: my @a = ^5; @a[2]:delete; dd @a.List
<camelia> rakudo-moar ecbf44fb5: OUTPUT: «(0, 1, Nil, 3, 4)␤»
<Zoffix_> m: "ac" ~~ /(.)(.)?(.)/; say $/.list
<camelia> rakudo-moar ecbf44fb5: OUTPUT: «(｢a｣ (Mu) ｢c｣)␤»
<Zoffix_> m: "ac" ~~ /(.)(.)?(.)/; say $/.list[1]
<camelia> rakudo-moar 0a32c6212: OUTPUT: «Nil␤»

Also, looks like it doesn't even get hllized in some cases:

<Zoffix_> m: "ac" ~~ /(.)(.)?(.)/; for $/.list { .say }
<camelia> rakudo-moar 0a32c6212: OUTPUT: «｢a｣␤Cannot call method 'say' on a null object␤  in block <unit> at <tmp> line 1␤␤»

[zoffixznet]

FYI: this started off from review of the test for RT#129299 that adds a proptest I committed in 2016 for a Mu in there.

I disagreed with 2016-Zoffix and threw out that test. It should be placed back (in potentially modified form) to re-cover RT#129299.

# RT #129299
subtest '$/.list returns all items even with optional captures in the middle', {
    "ac" ~~ /(.)(.)?(.)/;
    my ($one, $two, $three) = $/.list;
    is $one,       'a',  'first item';
    is $two.^name, 'Mu', 'second item'; # can't use is-deeply to test for Mu's
    is $three,     'c',  'third item';
}