[perl #126477] [GLR] .assuming on sub with is copy parameter fails

[Brian S. Julin via RT]

On Thu, 11 Aug 2016 09:24:31 -0700, coke wrote:
> On Wed Oct 28 13:40:45 2015, gfldex wrote:
> > sub foo(@a is copy, :$flag){ say @a };
> >  my @a = 1,2;
> > (&foo.assuming(:flag))(@a);
> > 
> > # OUTPUT«replace this Array is copy logic␤  in sub __PRIMED_ANON at
> > EVAL_2:2␤  in sub __PRIMED_ANON at
> > EVAL_2:1␤  in block <unit> at /tmp/TlcJQPUFbc:1␤
> 
> Coming from:
> 
> src/Perl6/Metamodel/BOOTSTRAP.nqp
> 435:                        nqp::die('replace this Array is copy logic');
> 
> Which is right below a
> 
> # XXX GLR
> 
> which appears 8 times in rakudo's src/ dir.
> 

Yeah this has nothing to do with .assuming.  Any '@ is copy' that hits the
slowpath binder will hit this.  Currently a way to force hitting the slowpath 
binder
is to provide a named parameter with a lot of alternate names:

$ perl6 -e 'sub foo(@a is copy, :f(:g(:h(:$i)))){ @a }; my @a = 1,2; 
&foo(@a).say;'
replace this Array is copy logic
  in sub foo at -e line 1
  in block <unit> at -e line 1

It's just that many (all?) things that get munged through .assuming hit that 
binder.