Re: Quantum Weirdness and the Increment Operator

[ysth]

in reply to Quantum Weirdness and the Increment Operator

This may prove useful to study:

my $x;
# unless overloaded, numeric values of two refs to the same thing are 
+equal
{package X; use Carp; sub is_x { carp \$x == \$_[0] ? "ok" : "not ok" 
+}}
X::is_x($x);
X::is_x(++$x);
X::is_x($x++);
X::is_x($x+0);

[barrachois]

Yes.

Part of what surprised me was that "return" gave me back a copy, so my "noop" subroutine should actually be called "copy".

 sub noop { 
   X::is_x( $_[0] );     # ok
   return $_[0];
 }
 X::is_x( noop($x) );    # not_ok

Update

The plot thickens.

I see over on Aliasing and return, how does return work? a discussion of just this issue, with a suggestion by japhy to use lvalue subs.

A quick test shows that one can indeed return $x itself this way - but the increment expression is still altered by a call to this kind of subroutine, too.

 sub noop_lval :lvalue { $_[0] }

 X:is_x( noop_lval($x) );                # ok !

 $m=20; print noop_lval(++$m) + $m++;    # 42 (still)

 $m=20; (++$m) + $m++;                   # 43

So I guess the copy that "return" makes isn't the whole issue.