Oh, that's annoying. In fairness, a documented feature of mine is that it explicitly avoids autovivifying keys. So with Tie::Hash::Vivify, this is true:

tie my %hash, 'Tie::Hash::Vivify', sub { 10 };
say $hash{foo};  # says 10
my $is_this_true = exists $hash{foo};
[download]

Whereas this is false:

tie my %hash, 'Hash::DefaultValue', sub { 10 };
say $hash{foo};
my $is_this_true = exists $hash{foo};
[download]

Another feature of Hash::DefaultValue is that the coderef gets passed a copy of the key, so can use that information when generating the default value.

tie my %hash, 'Hash::DefaultValue', sub { uc };
say $hash{foo}; # says "FOO"
[download]

I mostly use Tie::Hash::Vivify to prevent autovivification thus:

  my $hashref = Tie::Hash::Vivify->new(
    sub { confess("No auto-vivifying!\n".Dumper(\@_)) }
  )
[download]

to avoid those annoyingly hard-to-track-down bugs where you've typoed a hash key.

The autovivification module is pretty good for this kind of thing. Offers you the ability to simply switch off autovivification, warn about it, or croak. Rather than being tied to a particular hash, it's lexically scoped. And it's context-sensitive, so you could decide that 'store' operations are allowed to autovivify, like:

{
  no autovivification qw(strict);
  use autovivification qw(store);
  my %hash;
  $hash{foo}{bar}{baz} = 1;     # autovivifies OK
  delete $hash{fool}{bar}{baz}; # dies
}
[download]

If only autovivification weren't so tricky to type.

perl -E'sub Monkey::do{say$_,for@_,do{($monkey=[caller(0)]->[3])=~s{::}{ }and$monkey}}"Monkey say"->Monkey::do'