So it sounds like dump+eval does not create a true copy of the data structure.

None that I am aware of will retain knowledge that one scalar in the structure is a reference to another scalar in that same structure.

Ie. If you have this structure (which is different to what you showed, but may be what you intended):

$a = 'one'; 
$b = 'two'; 
@d = ( $a, $b, \$b );;
[download]

Then neither Storable:

use Storable qw[ freeze thaw ];;

$x = freeze \@d;;
$y = thaw $x;;

pp $y;;
["one", "two", \"two"]

${ $y->[2] } = 'three';;

pp $y;;
["one", "two", \"three"]
[download]

Nor Clone successfully capture that scalar reference relationship:

use Clone qw[ clone ];;

$x = clone \@d;;

pp $x;;
["one", "two", \"two"]

${ $x->[2] } = 'three';;

pp $x;;
["one", "two", \"three"]
[download]

You could say that was a bug -- and I'd probably agree with you -- but it is a long standing one that it would appear that no one has till now seen as a problem. Which is another way of saying it is unlikely to be fixed -- or even accepted as a bug -- any time soon.

I think that Sereal might retain aliases properly. The spec at least has a type for a proper alias, and I think I remember hearing in a talk about Sereal that it does retain that. Also, Data::Dump::Stream, by demerphq might also properly retain aliases.

The spec at least has a type for a proper alias

Hm. Wouldn't REFP be more appropriate? (In that the described problem is a dereferencable, reference rather than an alias.)

That said, Storable & Clone (and even Dump) handle references to arrays and hashes, including ensuring that multiple embedded references to the same array or hash retain that property once thawed. It is a bit strange that they don't handle scalar references correctly.

---

With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'

Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.

"Science is about questioning the status quo. Questioning authority".

In the absence of evidence, opinion is indistinguishable from prejudice.