From @masak

<masak> rakudo​: class A { method foo(​::T $ :) { say T.perl } }; A.new.foo
<p6eval> rakudo 2953ac​: OUTPUT«A␤»
<masak> std​: class A { method foo(​::T $​:) {} }
<p6eval> std 28279​: OUTPUT«ok 00​:03 40m␤»
<masak> rakudo​: class A { method foo(​::T $​:) {} }
<p6eval> rakudo 2953ac​: OUTPUT«Invalid twigil used in signature parameter. [...]
* masak cackles
<masak> pmichaud​: is it reasonable to expect Rakudo to parse the above
'$​:' as '$ :'
<pmichaud> is std parsing it that way?
<masak> good question.
<masak> it's not throwing an error, at least.
<TimToady> it has a param sep of '​:'
<TimToady> so I think it's making it the invocant
<pmichaud> okay
<pmichaud> that's fine -- that's what masak++ is asking for
<pmichaud> let's see what Rakudo is doing
<TimToady> $​: requires a name at STD​:1783
<pmichaud> yes, I see that
<TimToady> I don't know why param_var isn't slurping it as a twigil though
<pmichaud> right, that's my point :)
<pmichaud> looks like param_var wants a \w after the <twigil>
<pmichaud> [<?before <.twigil>\w> <twigil>]?
<TimToady> line 3062, yes
<TimToady> that would be it
<pmichaud> Rakudo doesn't have that check.
<pmichaud> We can add it.
<masak> \o/
<masak> pmichaud​: should I submit this to RT?
<pmichaud> sure.
* masak makes it so
<TimToady> why do I get the feeling that colon is overloaded to within
an inch of its life already?

From @pmichaud

Now fixed in a7ac444 -- just needs tests.

Thanks!

Pm

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

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