doc >>. on lists of Callables

[gfldex]

say (&sin, &cos, &sqrt)».(0.5);
rakudo-moar dcf3e28c5: OUTPUT: «(0.479425538604203 0.877582561890373 0.707106781186548)␤»

[smls]

That's not a special construct, but a combination of these features:

1. ( ) is a postcircumfix operator (documented here)
2. postcircumfixes may be written in dot notation (documented here)
3. postcircumfixes can be hypered (not yet documented):

@foo»[0]     # works
@foo».[0]    # works
@foo»{"a"}   # works
@foo».{"a"}  # works
@foo»<a>     # works
@foo».<a>    # works
@foo»()      # does NOT work, for some reason - bug?
@foo».()     # works

No.3 should be documented in operators#Hyper_Operators, which already documents hypered postfix operators and method calls.

(There's also this section, which is a bit dubious because that's not actually an operator as such.)

[tisonkun]

@smls

@foo»(0.5) # does NOT work, for some reason - bug?
turns to, if @foo = [&sin, &cos, &sqrt]
(0.5).CALL-ME(&sin) and so on
so it complains "No such method 'CALL-ME' for invocant of type 'Rat'"

not sure about above, it may be wrong.
a common use of >> is:

sub postfix:<!> ($a) { -$a };

say [1,2,3,4]>>!; # [-1 -2 -3 -4]

However, @foo».(0.5) # works
turns to
&sin.(0.5) and so on

Not a bug I think

[smls]

turns to [...] (0.5).CALL-ME(&sin) and so on

Shouldn't it turn into &sin.CALL-ME(0.5)?

Also, it looks like what's currently happening is even weirder:

➜  [&sin, &cos, &sqrt]»( &say );
sub sin ($) { #`(Sub+{is-pure}|93829902580832) ... }sub say (| is raw) { #`(Sub|93829903363016) ... }
sub cos ($) { #`(Sub+{is-pure}|93829902582048) ... }sub say (| is raw) { #`(Sub|93829903363016) ... }
sub sqrt ($) { #`(Sub+{is-pure}|93829902595424) ... }sub say (| is raw) { #`(Sub|93829903363016) ... }

[tisonkun]

@smls
interesting,

> dd [&sin, &cos, &sqrt]»( &say );
sub sin ($) { #`(Sub+{is-pure}|140321546561448) ... }sub say (| is raw) { #`(Sub|140321547176936) ... }
sub cos ($) { #`(Sub+{is-pure}|140321546562664) ... }sub say (| is raw) { #`(Sub|140321547176936) ... }
sub sqrt ($) { #`(Sub+{is-pure}|140321546576040) ... }sub say (| is raw) { #`(Sub|140321547176936) ... }
[Bool::True, Bool::True, Bool::True]

I am going to rakudo to find out how >> works

above seems &sin -> &sin.&say, but what it does is more than that.

[tisonkun]

@smls

https://github.com/rakudo/rakudo/blob/23776247298f24fe374903e5239cdd61f173078a/src/Perl6/Actions.nqp#L7511

take a look at examples below

13:35 | wander | m: sub see-what-pass { dd @; }; [1,2,3]>>( &see-what-pass );
13:35 | camelia | rakudo-moar b386963a3: OUTPUT: «[1, sub see-what-pass (*@) { #(Sub\|60245232) ... }]␤[2, sub see-what-pass (*@_) { #(Sub|60245232) ... }]␤[3, sub see-what-pass (*@) { #`(Sub|60245232) ... }]␤»
13:37 | wander | m: sub postfix:<!> { dd @; }; [1,2,3]>>!;
13:37 | camelia | rakudo-moar b386963a3: OUTPUT: «[1]␤[2]␤[3]␤»

so, normally use >> with postfix operator

[tisonkun]

@gfldex I think "doc >>. on lists of Callables" is not so different from examples existing.

see

say (<a b>, <c d e>)».elems;        # OUTPUT: «(2 3)␤» 
say (<a b>, <c d e>)».&{ .elems };  # OUTPUT: «((1 1) (1 1 1))␤» 

but if you find it worth to add new examples, you can do it by yourself and send a PR.

[tisonkun]

@AlexDaniel , I don't think we have to document ">>. on lists of Callables".

Because it's no more than [obj1, obj2, ...]>>.method

And we may be going to rearrange the doc in a few days, so keep the doc from redundant.

To sum up, this issue can be closed.