Once I know all the "gotchas" I can make a decision.

Many thanks,

I don't think I was as clear as I needed to be. Try the following code with a hash tied against the various DBMs.

use Data::Dumper;

# %x is the tied hash.
$x{foo} = [ 1 .. 3, { a => { b => 'c' } } ];
print Dumper \%x;
[download]

Using DBM::Deep, the following will be printed:

$VAR1 = {
          'foo' => bless( [
                            '1',
                            '2',
                            '3',
                            bless( {
                                     'a' => bless( {
                                                     'b' => 'c'
                                                   }, 'DBM::Deep::Hash
+' )
                                   }, 'DBM::Deep::Hash' )
                          ], 'DBM::Deep::Array' )
        };
[download]

Ignoring the random blessings, the entire data structure as Perl would see it is encoded. Using BerkeleyDB, you see

$VAR1 = {
          'foo' => 'ARRAY(0x18014f4)'
        };
[download]

Note: this is as designed for BDB - it isn't meant to handle anything as the value other than a simple binary string. DBM::Deep is.

---

My criteria for good software:

1. Does it work?
2. Can someone else come in, make a change, and be reasonably certain no bugs were introduced?