I don't want to protect $x. I just want to touch $x (store the value into it)
when Scope::Guard object gets destroyed, i.e. when all references to it go out
of scope.
In simple and obvious example below, the $g gets destroyed and inner sub
called immediately after it goes out of scope.
use strict;
use Scope::Guard;
my $x;
{
my $g = Scope::Guard->new(sub {$x; warn "destroyed"});
}
warn "end";
But I checked, and putting braces around the dcl of $x, doesn't
change the output. So I would bet scopeguard has stored a reference to the
interior sub so it can be called when "$x" goes out of scope.
Your bet is wrong. The sub is called the guard gets destroyed, not the object
it references. Look at the Scope::Guard code. Is very short and self
explanatory.
All you did by assigning undef to $code is lose the reference to the outside
sub.. SGnew still has a reference to the inner sub and $x...
Yes, but when I lose reference to the outside sub, I should also lose the last reference to $g,
because the sub is the only object which sees $g. But it doesnot happen.
Does that make more sense?
I need to read comments above, but it still doesnot make more sense to me.
However constructed is the inner sub (stored in $g) it is
just another object referenced by $g, not referencing it.
So I don't understand how it (the inner sub) can affect the reference count to
$g.
The example I gave doesnot make sense itself, it is just the simplest
demonstration of the behaviour I was able to make up.
My original was more like the test below: ($x is \@res, I just try to log into res,
when the object gets destroyed).
use strict;
use Test::More tests=>2;
use Scope::Guard;
sub try {
my @res;
my @subs = do {
my $g;
( sub {
push @res, 'a';
$g = Scope::Guard->new(
sub {
warn "destroyed";
push @res, 'destroyed'
} );
},
sub {
push @res, 'b';
undef $g;
}
);
};
( shift @subs )->();
is_deeply( \@res, ['a'] );
undef @subs;
is_deeply( \@res, [ 'a', 'destroyed' ] );
}
try();
|