Definitely a nice hack, robin++ ... but it leaves a question open. Consider:

local *foo:

*foo = [qw (a b c)];
{
  local *foo = [qw(1 2 3)];
  print "local \@{\*foo{ARRAY}} = qw(@{*foo{ARRAY}})\n";
  print "local \@foo = qw(@foo)\n";
}
print "global \@foo = qw(@foo)\n";
__END__
local @{*foo{ARRAY}} = qw(1 2 3)
local @foo = qw(1 2 3)
global @foo = qw(a b c)
[download]

my *foo:

*foo = [qw (a b c)];
{
  my *foo = [qw(1 2 3)];
  print "my \@{\*foo{ARRAY}} = qw(@{*foo{ARRAY}})\n";
  print "my \@foo = qw(@foo)\n";
}
print "global \@foo = qw(@foo)\n";
__END__
my @{*foo{ARRAY}} = qw(1 2 3)
my @foo = qw(a b c)
global @foo = qw(a b c)
[download]

Hmm. The lexical glob is created ok, but other than with local, glob corresponding variables aren't treated as aliases to the typeglob entries in the current scope.

Now, should that happen? I think, it should; but such lookup should also warn about "masking an earlier @foo in the same scope".

What do you think?

--shmem

_($_=" "x(1<<5)."?\n".q·/)Oo.  G°\        /
                              /\_¯/(q    /
----------------------------  \__(m.====·.(_("always off the crowd"))."·
");sub _{s./.($e="'Itrs `mnsgdq Gdbj O`qkdq")=~y/"-y/#-z/;$e.e && print}