From sirrobert@gmail.com

(distilled from log)

11​:06 < sirrobert> r​: my @​arr; @​arr.push(%(a => 1, b => 2)); say @​arr.perl;
11​:06 <+p6eval> rakudo 097361​: OUTPUT«Array.new("a" => 1, "b" => 2)␤»
11​:06 < sirrobert> I expected that to give me an array with one element
that was a hash
11​:06 < sirrobert> what did I do wrong?
11​:07 < sirrobert> or maybe​: how should I think about that?
11​:07 < masak> push expects a list of things. if it finds a hash, it
listifies the hash.
11​:07 < sirrobert> ok, how do I get it not to listify things?
11​:08 < sirrobert> if I want an array of hashes
11​:08 < masak> r​: my @​a; @​a.push({ a => 1, b => 2}); say @​a.perl
11​:08 <+p6eval> rakudo 097361​: OUTPUT«Array.new({"a" => 1, "b" => 2})␤»
11​:08 < sirrobert> ok
11​:08 < sirrobert> so what I really have is​:
11​:09 < sirrobert> r​: my @​a; @​a.push(%((1,2,3) Z=> (<a b c>))); say @​a.perl;
11​:09 <+p6eval> rakudo 097361​: OUTPUT«Array.new("1" => "a", "2" => "b", "3"
=> "c")␤»
11​:09 < sirrobert> is the "right" solution here to create the hash
separately? but it
  still gets expanded...
11​:09 < sirrobert> I think
11​:09 < masak> yeah.
11​:09 < masak> r​: my @​a; my %h = (1,2,3) Z=> (<a b c>); @​a.push(%h); say
@​a.perl
11​:09 <+p6eval> rakudo 097361​: OUTPUT«Array.new("1" => "a", "2" => "b", "3"
=> "c")␤»
11​:09 < masak> indeed.
11​:09 < masak> r​: my @​a; my %h = (1,2,3) Z=> (<a b c>); @​a.push({%h}); say
@​a.perl
11​:10 <+p6eval> rakudo 097361​: OUTPUT«Array.new({"1" => "a", "2" => "b",
"3" => "c"})␤»
11​:10 < sirrobert> ohh.. weird
11​:10 < sirrobert> ok, I can live with that =)
11​:11 < pmichaud> r​: my @​a; @​a.push({ (1,2,3) Z <a b c> }); say @​a.perl
11​:11 <+p6eval> rakudo 097361​: OUTPUT«Array.new(Block.new())␤»
11​:11 < moritz> sirrobert​: what's weird about it?
11​:11 < sirrobert> the {%} seems redundant
11​:11 < moritz> well, the {} makes an hash-in-an-item
11​:11 < sirrobert> moritz​: a workaround for the listification magic of push
11​:11 < moritz> you can also say %h.item
11​:12 < sirrobert> ok, that's a good paradigm for me to adopt for it
11​:12 < sirrobert> I didn't previously have a concept of 'item'
11​:13 < sirrobert> huh... push({%...}) interprets it as a hash but push​:
{%...} interprets
  it as a block
11​:14 < moritz> that's weird
11​:14 < sirrobert> verifying now ...
11​:14 < moritz> r​: my @​h; @​h.push​: {%*ENV}; say @​h[0].WHAT
11​:14 <+p6eval> rakudo 097361​: OUTPUT«Block()␤»
11​:15 < moritz> r​: my @​h; @​h.push({%*ENV}); say @​h[0].WHAT
11​:15 <+p6eval> rakudo 097361​: OUTPUT«Block()␤»
11​:15 < moritz> sirrobert​: seems to be a block either way :(
11​:15 < moritz> r​: my %a; my @​h; @​h.push({%a}); say @​h[0].WHAT
11​:15 <+p6eval> rakudo 097361​: OUTPUT«Hash()␤»
11​:15 < moritz> r​: my %a; my @​h; @​h.push​: {%a}; say @​h[0].WHAT
11​:15 <+p6eval> rakudo 097361​: OUTPUT«Hash()␤»
11​:15 < moritz> erm, what?
11​:15 < sirrobert> moritz​: erm, indeed =)
11​:16 < moritz> %*ENV makes it a block, %a a hash?
11​:16 * moritz smeells a rakudobug
11​:18 < sirrobert> r​: my $h = {%(<a b>) Z=> (1,2)}; say $h.WHAT;
11​:18 <+p6eval> rakudo 097361​: OUTPUT«Block()␤»
11​:19 < masak> moritz​: that feels buggish, yes.
11​:19 < masak> moritz​: care to report it?
11​:19 < sirrobert> sure
11​:19 < sirrobert> if I copy/paste this discussion is that enough? I don't
quite know how
  to describe it.
11​:19 < moritz> yes, it's enough
11​:19 < sirrobert> ok, one sec

From @moritz

A bit more concise​:

19​:09 < moritz> r​: my %a; say {%a}.WHAT
19​:09 <+p6eval> rakudo 097361​: OUTPUT«Hash()␤»
19​:10 < moritz> r​: my %*a; say {%*a}.WHAT
19​:10 < p6eval> rakudo 097361​: OUTPUT«Block()␤»

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

From @moritz

On Wed Sep 19 10​:12​:02 2012, moritz wrote​:

A bit more concise​:

19​:09 < moritz> r​: my %a; say {%a}.WHAT
19​:09 <+p6eval> rakudo 097361​: OUTPUT«Hash()␤»
19​:10 < moritz> r​: my %*a; say {%*a}.WHAT
19​:10 < p6eval> rakudo 097361​: OUTPUT«Block()␤»

And both should say "Hash".

From @jnthn

On Wed Sep 19 10​:12​:02 2012, moritz wrote​:

A bit more concise​:

19​:09 < moritz> r​: my %a; say {%a}.WHAT
19​:09 <+p6eval> rakudo 097361​: OUTPUT«Hash()␤»
19​:10 < moritz> r​: my %*a; say {%*a}.WHAT
19​:10 < p6eval> rakudo 097361​: OUTPUT«Block()␤»

Implemented this case, and added a test in S06-other/anon-hashes-vs-block.t.

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