From @lizmat

[23​:44​:41] <lizmat> r​: my %e; (%e{$_.WHICH} //= ($_ => 0)).value++ for <a b b c c c>; say %e.perl # scoping issue?
[23​:44​:43] <+camelia> rakudo 78ccd9​: OUTPUT«("Str|a" => Any => 1, "Str|b" => Any => 2, "Str|c" => Any => 3).hash␤»
[23​:44​:53] <lizmat> why are the pairs all Any?
[23​:46​:00] <lizmat> r​: my %e; for <a b b c c c> -> $k { (%e{$k.WHICH} //= ($k => 0)).value++}; say %e.perl # this seems to work
[23​:46​:02] <+camelia> rakudo 78ccd9​: OUTPUT«("Str|a" => "a" => 1, "Str|b" => "b" => 2, "Str|c" => "c" => 3).hash␤»
[23​:46​:19] <lizmat> feels like a rakudobug to me
[23​:46​:59] <FROGGS> r​: my %e; -> $_ { (%e{$_.WHICH} //= ($_ => 0)).value++ } for <a b b c c c>; say %e.perl # scoping issue?
[23​:47​:00] <+camelia> rakudo 78ccd9​: OUTPUT«("Str|a" => "a" => 1, "Str|b" => "b" => 2, "Str|c" => "c" => 3).hash␤»
[23​:47​:21] <jnthn> lizmat​: I'm tired, but yes, looks like...
[23​:47​:45] <lizmat> will use FROGGS++ quick fix
[23​:48​:01] <lizmat> and report as rakudobug

From @diakopter

new behavior​:

15​:07 <diakopter> m​: my %e; (%e{$_.WHICH} //= ($_ => 0)).value++ for <a b b c c c>; say %e.perl │·······························
15​:07 <camelia> rakudo-moar 61d231​: OUTPUT«Cannot call postfix​:<++>(Int); none of these signatures match​:␤ (Mu​:D │·······························
  $a is rw)␤ (Mu​:U $a is rw)␤ (Int​:D $a is rw)␤ (int $a is rw)␤ (Bool​:U $a is rw)␤ │·······························
  (Bool​:D $a is rw)␤ (Num​:D $a is rw)␤ (Num​:U $a is rw)␤ (num …» │·······························

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

From @smls

@​lizmat, @​FROGGS​:

Do you know how the original scoping issue reported in this RT, was fixed?

#​122709 and #​126569 suffer from similar scoping bugs involving $_ inside parens under a statement-modifier, and *haven't* been fixed yet.

@​matthew​:

That new behavior is to be expected.

It's a result of the fact that unlike hashes, pairs no longer automatically containerize their values.

And the ++ operator does not work on objects of type Int (those are immutable!), it only works on a Scalar container which happens to *contain* an Int.

For example​:

  say 0.VAR.^name; # Int
  say (a => 0).value.VAR.^name; # Int
 
  say (my $x = 0).VAR.^name; # Scalar
  say (a => (my $ = 0)).value.VAR.^name; # Scalar

In order to make your example work, you'd have to manually containerize the values of the keys you construct​:

  my %e; (%e{$_.WHICH} //= ($_ => (my $ = 0))).value++ for <a b b c c c>; say %e.perl
  # Output​:
  # {"Str|a" => (Any) => 3, "Str|b" => (Any) => 3, "Str|c" => (Any) => 3}

Or use a Hash of Arrays, instead of a Hash of Pairs.

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