From @masak

<masak> n​: my @​x = <& a& &b>; s​:g/\&/\\\&/ for @​x; say @​x
<camelia> niecza v24-86-g39ab531​: OUTPUT«\& a\& \&b␤»
<masak> r​: my @​x = <& a& &b>; s​:g/\&/\\\&/ for @​x; say @​x
<masak> I expect to be able to change those things in an array. is
Rakudo wrong and Niecza right?
<masak> rn​: my @​x = <& a& &b>; for @​x { s​:g/\&/\\\&/ }; say @​x
<camelia> rakudo 29fb2b, niecza v24-86-g39ab531​: OUTPUT«\& a\& \&b␤»
<masak> $_ should be 'is rw' both in a normal for loop and in a
statement_mod for loop, right?
<jnthn> right
* masak submits rakudobug
<masak> r​: my @​x = <x x x>; $_ = 'foo' for @​x; say @​x
<camelia> rakudo 29fb2b​: OUTPUT«Cannot assign to a readonly variable
or a value␤ in block at /tmp/vbd0908EfA​:1␤␤»
<masak> n​: my @​x = <x x x>; $_ = 'foo' for @​x; say @​x
<camelia> niecza v24-86-g39ab531​: OUTPUT«foo foo foo␤»

From @usev6

The two forms of the "for" loop now behave the same​:

$ perl6 -e 'my @​x = <& a& &b>; for @​x { s​:g/\&/\\\&/ }; say @​x'
\& a\& \&b

$ perl6 -e 'my @​x = <& a& &b>; s​:g/\&/\\\&/ for @​x; say @​x'
\& a\& \&b

I added two tests to S04-statement-modifiers/for.t with commit Raku/roast@82d502c355.

I'm closing this ticket now.

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