From @TimToady

Using a normal closure works​:

lucca​:~/rc 1244> cat foo
my @​isprime = False,False; # 0 and 1 are not prime by definition

say (for 1 .. 10 -> $i {
  $i if @​isprime[$i] //= so $i %% none 2 ...^ { $_ > $i.sqrt.floor };
})
lucca​:~/rc 1245> p foo
(2 3 5 7)

Using a WhateverCode doesn't work​:

lucca​:~/rc 1246> cat bar
my @​isprime = False,False; # 0 and 1 are not prime by definition

say (for 1 .. 10 -> $i {
  $i if @​isprime[$i] //= so $i %% none 2 ...^ * > $i.sqrt.floor;
})
lucca​:~/rc 1247> p bar
(2 3 4 5 6 7 8 9 10)

But it works again if we expand out the //= to its desugar​:

lucca​:~/rc 1250> cat baz
my @​isprime = False,False; # 0 and 1 are not prime by definition

say (for 1 .. 10 -> $i {
  $i if @​isprime[$i] = @​isprime[$i] // so $i %% none 2 ...^ * >
$i.sqrt.floor;
})
lucca​:~/rc 1251> p baz
(2 3 5 7)

From @dogbert17

This seems to have been fixed with commit rakudo/rakudo@d1c2e76. There were tests added at the same time, i.e. Raku/roast@e797f92 but I'm unsure if those tests cover this RT. If they do then this ticket can probably be marked as resolved.

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

From @zoffixznet

Tests​: Raku/roast@2d35dd49497723ba5bdcc8ef

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