[dbsrgits/DBM-Deep.git] / lib / DBM / Deep.pod

=head1 NAME

DBM::Deep - A pure perl multi-level hash/array DBM that supports transactions

=head1 SYNOPSIS

  use DBM::Deep;
  my $db = DBM::Deep->new( "foo.db" );

  $db->{key} = 'value';
  print $db->{key};

  $db->put('key' => 'value');
  print $db->get('key');

  # true multi-level support
  $db->{my_complex} = [
      'hello', { perl => 'rules' },
      42, 99,
  ];

  $db->begin_work;

  # Do stuff here

  $db->rollback;
  $db->commit;

  tie my %db, 'DBM::Deep', 'foo.db';
  $db{key} = 'value';
  print $db{key};

  tied(%db)->put('key' => 'value');
  print tied(%db)->get('key');

=head1 DESCRIPTION

A unique flat-file database module, written in pure perl. True multi-level
hash/array support (unlike MLDBM, which is faked), hybrid OO / tie()
interface, cross-platform FTPable files, ACID transactions, and is quite fast.
Can handle millions of keys and unlimited levels without significant
slow-down. Written from the ground-up in pure perl -- this is NOT a wrapper
around a C-based DBM. Out-of-the-box compatibility with Unix, Mac OS X and
Windows.

=head1 VERSION DIFFERENCES

B<NOTE>: 1.0000 has significant file format differences from prior versions.
THere is a backwards-compatibility layer at C<utils/upgrade_db.pl>. Files
created by 1.0000 or higher are B<NOT> compatible with scripts using prior
versions.

=head1 SETUP

Construction can be done OO-style (which is the recommended way), or using
Perl's tie() function. Both are examined here.

=head2 OO Construction

The recommended way to construct a DBM::Deep object is to use the new()
method, which gets you a blessed I<and> tied hash (or array) reference.

  my $db = DBM::Deep->new( "foo.db" );

This opens a new database handle, mapped to the file "foo.db". If this
file does not exist, it will automatically be created. DB files are
opened in "r+" (read/write) mode, and the type of object returned is a
hash, unless otherwise specified (see L<OPTIONS> below).

You can pass a number of options to the constructor to specify things like
locking, autoflush, etc. This is done by passing an inline hash (or hashref):

  my $db = DBM::Deep->new(
      file      => "foo.db",
      locking   => 1,
      autoflush => 1
  );

Notice that the filename is now specified I<inside> the hash with
the "file" parameter, as opposed to being the sole argument to the
constructor. This is required if any options are specified.
See L<OPTIONS> below for the complete list.

You can also start with an array instead of a hash. For this, you must
specify the C<type> parameter:

  my $db = DBM::Deep->new(
      file => "foo.db",
      type => DBM::Deep->TYPE_ARRAY
  );

B<Note:> Specifing the C<type> parameter only takes effect when beginning
a new DB file. If you create a DBM::Deep object with an existing file, the
C<type> will be loaded from the file header, and an error will be thrown if
the wrong type is passed in.

=head2 Tie Construction

Alternately, you can create a DBM::Deep handle by using Perl's built-in
tie() function. The object returned from tie() can be used to call methods,
such as lock() and unlock(). (That object can be retrieved from the tied
variable at any time using tied() - please see L<perltie/> for more info.

  my %hash;
  my $db = tie %hash, "DBM::Deep", "foo.db";

  my @array;
  my $db = tie @array, "DBM::Deep", "bar.db";

As with the OO constructor, you can replace the DB filename parameter with
a hash containing one or more options (see L<OPTIONS> just below for the
complete list).

  tie %hash, "DBM::Deep", {
      file => "foo.db",
      locking => 1,
      autoflush => 1
  };

=head2 Options

There are a number of options that can be passed in when constructing your
DBM::Deep objects. These apply to both the OO- and tie- based approaches.

=over

=item * file

Filename of the DB file to link the handle to. You can pass a full absolute
filesystem path, partial path, or a plain filename if the file is in the
current working directory. This is a required parameter (though q.v. fh).

=item * fh

If you want, you can pass in the fh instead of the file. This is most useful for doing
something like:

  my $db = DBM::Deep->new( { fh => \*DATA } );

You are responsible for making sure that the fh has been opened appropriately for your
needs. If you open it read-only and attempt to write, an exception will be thrown. If you
open it write-only or append-only, an exception will be thrown immediately as DBM::Deep
needs to read from the fh.

=item * file_offset

This is the offset within the file that the DBM::Deep db starts. Most of the time, you will
not need to set this. However, it's there if you want it.

If you pass in fh and do not set this, it will be set appropriately.

=item * type

This parameter specifies what type of object to create, a hash or array. Use
one of these two constants:

=over 4

=item * C<DBM::Deep-E<gt>TYPE_HASH>

=item * C<DBM::Deep-E<gt>TYPE_ARRAY>.

=back

This only takes effect when beginning a new file. This is an optional
parameter, and defaults to C<DBM::Deep-E<gt>TYPE_HASH>.

=item * locking

Specifies whether locking is to be enabled. DBM::Deep uses Perl's flock()
function to lock the database in exclusive mode for writes, and shared mode
for reads. Pass any true value to enable. This affects the base DB handle
I<and any child hashes or arrays> that use the same DB file. This is an
optional parameter, and defaults to 1 (enabled). See L<LOCKING> below for
more.

=item * autoflush

Specifies whether autoflush is to be enabled on the underlying filehandle.
This obviously slows down write operations, but is required if you may have
multiple processes accessing the same DB file (also consider enable I<locking>).
Pass any true value to enable. This is an optional parameter, and defaults to 1
(enabled).

=item * filter_*

See L</FILTERS> below.

=back

The following parameters may be specified in the constructor the first time the
datafile is created. However, they will be stored in the header of the file and
cannot be overridden by subsequent openings of the file - the values will be set
from the values stored in the datafile's header.

=over 4

=item * num_txns

This is the number of transactions that can be running at one time. The
default is one - the HEAD. The minimum is one and the maximum is 255. The more
transactions, the larger and quicker the datafile grows.

See L</TRANSACTIONS> below.

=item * max_buckets

This is the number of entries that can be added before a reindexing. The larger
this number is made, the larger a file gets, but the better performance you will
have. The default and minimum number this can be is 16. The maximum is 256, but
more than 64 isn't recommended.

=item * data_sector_size

This is the size in bytes of a given data sector. Data sectors will chain, so
a value of any size can be stored. However, chaining is expensive in terms of
time. Setting this value to something close to the expected common length of
your scalars will improve your performance. If it is too small, your file will
have a lot of chaining. If it is too large, your file will have a lot of dead
space in it.

The default for this is 64 bytes. The minimum value is 32 and the maximum is
256 bytes.

B<Note:> There are between 6 and 10 bytes taken up in each data sector for
bookkeeping. (It's 4 + the number of bytes in your L</pack_size>.) This is
included within the data_sector_size, thus the effective value is 6-10 bytes
less than what you specified.

=item * pack_size

This is the size of the file pointer used throughout the file. The valid values
are:

=over 4

=item * small

This uses 2-byte offsets, allowing for a maximum file size of 65 KB.

=item * medium (default)

This uses 4-byte offsets, allowing for a maximum file size of 4 GB.

=item * large

This uses 8-byte offsets, allowing for a maximum file size of 16 XB
(exabytes). This can only be enabled if your Perl is compiled for 64-bit.

=back

See L</LARGEFILE SUPPORT> for more information.

=back

=head1 TIE INTERFACE

With DBM::Deep you can access your databases using Perl's standard hash/array
syntax. Because all DBM::Deep objects are I<tied> to hashes or arrays, you can
treat them as such. DBM::Deep will intercept all reads/writes and direct them
to the right place -- the DB file. This has nothing to do with the
L<TIE CONSTRUCTION> section above. This simply tells you how to use DBM::Deep
using regular hashes and arrays, rather than calling functions like C<get()>
and C<put()> (although those work too). It is entirely up to you how to want
to access your databases.

=head2 Hashes

You can treat any DBM::Deep object like a normal Perl hash reference. Add keys,
or even nested hashes (or arrays) using standard Perl syntax:

  my $db = DBM::Deep->new( "foo.db" );

  $db->{mykey} = "myvalue";
  $db->{myhash} = {};
  $db->{myhash}->{subkey} = "subvalue";

  print $db->{myhash}->{subkey} . "\n";

You can even step through hash keys using the normal Perl C<keys()> function:

  foreach my $key (keys %$db) {
      print "$key: " . $db->{$key} . "\n";
  }

Remember that Perl's C<keys()> function extracts I<every> key from the hash and
pushes them onto an array, all before the loop even begins. If you have an
extremely large hash, this may exhaust Perl's memory. Instead, consider using
Perl's C<each()> function, which pulls keys/values one at a time, using very
little memory:

  while (my ($key, $value) = each %$db) {
      print "$key: $value\n";
  }

Please note that when using C<each()>, you should always pass a direct
hash reference, not a lookup. Meaning, you should B<never> do this:

  # NEVER DO THIS
  while (my ($key, $value) = each %{$db->{foo}}) { # BAD

This causes an infinite loop, because for each iteration, Perl is calling
FETCH() on the $db handle, resulting in a "new" hash for foo every time, so
it effectively keeps returning the first key over and over again. Instead,
assign a temporary variable to C<$db->{foo}>, then pass that to each().

=head2 Arrays

As with hashes, you can treat any DBM::Deep object like a normal Perl array
reference. This includes inserting, removing and manipulating elements,
and the C<push()>, C<pop()>, C<shift()>, C<unshift()> and C<splice()> functions.
The object must have first been created using type C<DBM::Deep-E<gt>TYPE_ARRAY>,
or simply be a nested array reference inside a hash. Example:

  my $db = DBM::Deep->new(
      file => "foo-array.db",
      type => DBM::Deep->TYPE_ARRAY
  );

  $db->[0] = "foo";
  push @$db, "bar", "baz";
  unshift @$db, "bah";

  my $last_elem = pop @$db; # baz
  my $first_elem = shift @$db; # bah
  my $second_elem = $db->[1]; # bar

  my $num_elements = scalar @$db;

=head1 OO INTERFACE

In addition to the I<tie()> interface, you can also use a standard OO interface
to manipulate all aspects of DBM::Deep databases. Each type of object (hash or
array) has its own methods, but both types share the following common methods:
C<put()>, C<get()>, C<exists()>, C<delete()> and C<clear()>. C<fetch()> and
C<store(> are aliases to C<put()> and C<get()>, respectively.

=over

=item * new() / clone()

These are the constructor and copy-functions.

=item * put() / store()

Stores a new hash key/value pair, or sets an array element value. Takes two
arguments, the hash key or array index, and the new value. The value can be
a scalar, hash ref or array ref. Returns true on success, false on failure.

  $db->put("foo", "bar"); # for hashes
  $db->put(1, "bar"); # for arrays

=item * get() / fetch()

Fetches the value of a hash key or array element. Takes one argument: the hash
key or array index. Returns a scalar, hash ref or array ref, depending on the
data type stored.

  my $value = $db->get("foo"); # for hashes
  my $value = $db->get(1); # for arrays

=item * exists()

Checks if a hash key or array index exists. Takes one argument: the hash key
or array index. Returns true if it exists, false if not.

  if ($db->exists("foo")) { print "yay!\n"; } # for hashes
  if ($db->exists(1)) { print "yay!\n"; } # for arrays

=item * delete()

Deletes one hash key/value pair or array element. Takes one argument: the hash
key or array index. Returns true on success, false if not found. For arrays,
the remaining elements located after the deleted element are NOT moved over.
The deleted element is essentially just undefined, which is exactly how Perl's
internal arrays work.

  $db->delete("foo"); # for hashes
  $db->delete(1); # for arrays

=item * clear()

Deletes B<all> hash keys or array elements. Takes no arguments. No return
value.

  $db->clear(); # hashes or arrays

=item * lock() / unlock()

q.v. Locking.

=item * optimize()

Recover lost disk space. This is important to do, especially if you use
transactions.

=item * import() / export()

Data going in and out.

=item * begin_work() / commit() / rollback()

These are the transactional functions. L</TRANSACTIONS> for more information.

=back

=head2 Hashes

For hashes, DBM::Deep supports all the common methods described above, and the
following additional methods: C<first_key()> and C<next_key()>.

=over

=item * first_key()

Returns the "first" key in the hash. As with built-in Perl hashes, keys are
fetched in an undefined order (which appears random). Takes no arguments,
returns the key as a scalar value.

  my $key = $db->first_key();

=item * next_key()

Returns the "next" key in the hash, given the previous one as the sole argument.
Returns undef if there are no more keys to be fetched.

  $key = $db->next_key($key);

=back

Here are some examples of using hashes:

  my $db = DBM::Deep->new( "foo.db" );

  $db->put("foo", "bar");
  print "foo: " . $db->get("foo") . "\n";

  $db->put("baz", {}); # new child hash ref
  $db->get("baz")->put("buz", "biz");
  print "buz: " . $db->get("baz")->get("buz") . "\n";

  my $key = $db->first_key();
  while ($key) {
      print "$key: " . $db->get($key) . "\n";
      $key = $db->next_key($key);
  }

  if ($db->exists("foo")) { $db->delete("foo"); }

=head2 Arrays

For arrays, DBM::Deep supports all the common methods described above, and the
following additional methods: C<length()>, C<push()>, C<pop()>, C<shift()>,
C<unshift()> and C<splice()>.

=over

=item * length()

Returns the number of elements in the array. Takes no arguments.

  my $len = $db->length();

=item * push()

Adds one or more elements onto the end of the array. Accepts scalars, hash
refs or array refs. No return value.

  $db->push("foo", "bar", {});

=item * pop()

Fetches the last element in the array, and deletes it. Takes no arguments.
Returns undef if array is empty. Returns the element value.

  my $elem = $db->pop();

=item * shift()

Fetches the first element in the array, deletes it, then shifts all the
remaining elements over to take up the space. Returns the element value. This
method is not recommended with large arrays -- see L<LARGE ARRAYS> below for
details.

  my $elem = $db->shift();

=item * unshift()

Inserts one or more elements onto the beginning of the array, shifting all
existing elements over to make room. Accepts scalars, hash refs or array refs.
No return value. This method is not recommended with large arrays -- see
<LARGE ARRAYS> below for details.

  $db->unshift("foo", "bar", {});

=item * splice()

Performs exactly like Perl's built-in function of the same name. See L<perldoc
-f splice> for usage -- it is too complicated to document here. This method is
not recommended with large arrays -- see L<LARGE ARRAYS> below for details.

=back

Here are some examples of using arrays:

  my $db = DBM::Deep->new(
      file => "foo.db",
      type => DBM::Deep->TYPE_ARRAY
  );

  $db->push("bar", "baz");
  $db->unshift("foo");
  $db->put(3, "buz");

  my $len = $db->length();
  print "length: $len\n"; # 4

  for (my $k=0; $k<$len; $k++) {
      print "$k: " . $db->get($k) . "\n";
  }

  $db->splice(1, 2, "biz", "baf");

  while (my $elem = shift @$db) {
      print "shifted: $elem\n";
  }

=head1 LOCKING

Enable or disable automatic file locking by passing a boolean value to the
C<locking> parameter when constructing your DBM::Deep object (see L<SETUP>
above).

  my $db = DBM::Deep->new(
      file => "foo.db",
      locking => 1
  );

This causes DBM::Deep to C<flock()> the underlying filehandle with exclusive
mode for writes, and shared mode for reads. This is required if you have
multiple processes accessing the same database file, to avoid file corruption.
Please note that C<flock()> does NOT work for files over NFS. See L<DB OVER
NFS> below for more.

=head2 Explicit Locking

You can explicitly lock a database, so it remains locked for multiple
actions. This is done by calling the C<lock()> method, and passing an
optional lock mode argument (defaults to exclusive mode). This is particularly
useful for things like counters, where the current value needs to be fetched,
then incremented, then stored again.

  $db->lock();
  my $counter = $db->get("counter");
  $counter++;
  $db->put("counter", $counter);
  $db->unlock();

  # or...

  $db->lock();
  $db->{counter}++;
  $db->unlock();

You can pass C<lock()> an optional argument, which specifies which mode to use
(exclusive or shared). Use one of these two constants:
C<DBM::Deep-E<gt>LOCK_EX> or C<DBM::Deep-E<gt>LOCK_SH>. These are passed
directly to C<flock()>, and are the same as the constants defined in Perl's
L<Fcntl/> module.

  $db->lock( $db->LOCK_SH );
  # something here
  $db->unlock();

=head1 IMPORTING/EXPORTING

You can import existing complex structures by calling the C<import()> method,
and export an entire database into an in-memory structure using the C<export()>
method. Both are examined here.

=head2 Importing

Say you have an existing hash with nested hashes/arrays inside it. Instead of
walking the structure and adding keys/elements to the database as you go,
simply pass a reference to the C<import()> method. This recursively adds
everything to an existing DBM::Deep object for you. Here is an example:

  my $struct = {
      key1 => "value1",
      key2 => "value2",
      array1 => [ "elem0", "elem1", "elem2" ],
      hash1 => {
          subkey1 => "subvalue1",
          subkey2 => "subvalue2"
      }
  };

  my $db = DBM::Deep->new( "foo.db" );
  $db->import( $struct );

  print $db->{key1} . "\n"; # prints "value1"

This recursively imports the entire C<$struct> object into C<$db>, including
all nested hashes and arrays. If the DBM::Deep object contains exsiting data,
keys are merged with the existing ones, replacing if they already exist.
The C<import()> method can be called on any database level (not just the base
level), and works with both hash and array DB types.

B<Note:> Make sure your existing structure has no circular references in it.
These will cause an infinite loop when importing. There are plans to fix this
in a later release.

=head2 Exporting

Calling the C<export()> method on an existing DBM::Deep object will return
a reference to a new in-memory copy of the database. The export is done
recursively, so all nested hashes/arrays are all exported to standard Perl
objects. Here is an example:

  my $db = DBM::Deep->new( "foo.db" );

  $db->{key1} = "value1";
  $db->{key2} = "value2";
  $db->{hash1} = {};
  $db->{hash1}->{subkey1} = "subvalue1";
  $db->{hash1}->{subkey2} = "subvalue2";

  my $struct = $db->export();

  print $struct->{key1} . "\n"; # prints "value1"

This makes a complete copy of the database in memory, and returns a reference
to it. The C<export()> method can be called on any database level (not just
the base level), and works with both hash and array DB types. Be careful of
large databases -- you can store a lot more data in a DBM::Deep object than an
in-memory Perl structure.

B<Note:> Make sure your database has no circular references in it.
These will cause an infinite loop when exporting. There are plans to fix this
in a later release.

=head1 FILTERS

DBM::Deep has a number of hooks where you can specify your own Perl function
to perform filtering on incoming or outgoing data. This is a perfect
way to extend the engine, and implement things like real-time compression or
encryption. Filtering applies to the base DB level, and all child hashes /
arrays. Filter hooks can be specified when your DBM::Deep object is first
constructed, or by calling the C<set_filter()> method at any time. There are
four available filter hooks.

=head2 set_filter()

This method takes two paramters - the filter type and the filter subreference.
The four types are:

=over

=item * filter_store_key

This filter is called whenever a hash key is stored. It
is passed the incoming key, and expected to return a transformed key.

=item * filter_store_value

This filter is called whenever a hash key or array element is stored. It
is passed the incoming value, and expected to return a transformed value.

=item * filter_fetch_key

This filter is called whenever a hash key is fetched (i.e. via
C<first_key()> or C<next_key()>). It is passed the transformed key,
and expected to return the plain key.

=item * filter_fetch_value

This filter is called whenever a hash key or array element is fetched.
It is passed the transformed value, and expected to return the plain value.

=back

Here are the two ways to setup a filter hook:

  my $db = DBM::Deep->new(
      file => "foo.db",
      filter_store_value => \&my_filter_store,
      filter_fetch_value => \&my_filter_fetch
  );

  # or...

  $db->set_filter( "filter_store_value", \&my_filter_store );
  $db->set_filter( "filter_fetch_value", \&my_filter_fetch );

Your filter function will be called only when dealing with SCALAR keys or
values. When nested hashes and arrays are being stored/fetched, filtering
is bypassed. Filters are called as static functions, passed a single SCALAR
argument, and expected to return a single SCALAR value. If you want to
remove a filter, set the function reference to C<undef>:

  $db->set_filter( "filter_store_value", undef );

=head2 Examples

Please read L<DBM::Deep::Manual/> for examples of filters.

=head1 ERROR HANDLING

Most DBM::Deep methods return a true value for success, and call die() on
failure. You can wrap calls in an eval block to catch the die.

  my $db = DBM::Deep->new( "foo.db" ); # create hash
  eval { $db->push("foo"); }; # ILLEGAL -- push is array-only call

  print $@;           # prints error message

=head1 LARGEFILE SUPPORT

If you have a 64-bit system, and your Perl is compiled with both LARGEFILE
and 64-bit support, you I<may> be able to create databases larger than 4 GB.
DBM::Deep by default uses 32-bit file offset tags, but these can be changed
by specifying the 'pack_size' parameter when constructing the file.

  DBM::Deep->new(
      filename  => $filename,
      pack_size => 'large',
  );

This tells DBM::Deep to pack all file offsets with 8-byte (64-bit) quad words
instead of 32-bit longs. After setting these values your DB files have a
theoretical maximum size of 16 XB (exabytes).

You can also use C<pack_size =E<gt> 'small'> in order to use 16-bit file
offsets.

B<Note:> Changing these values will B<NOT> work for existing database files.
Only change this for new files. Once the value has been set, it is stored in
the file's header and cannot be changed for the life of the file. These
parameters are per-file, meaning you can access 32-bit and 64-bit files, as
you choose.

B<Note:> We have not personally tested files larger than 4 GB -- all our
systems have only a 32-bit Perl. However, we have received user reports that
this does indeed work.

=head1 LOW-LEVEL ACCESS

If you require low-level access to the underlying filehandle that DBM::Deep uses,
you can call the C<_fh()> method, which returns the handle:

  my $fh = $db->_fh();

This method can be called on the root level of the datbase, or any child
hashes or arrays. All levels share a I<root> structure, which contains things
like the filehandle, a reference counter, and all the options specified
when you created the object. You can get access to this file object by
calling the C<_storage()> method.

  my $file_obj = $db->_storage();

This is useful for changing options after the object has already been created,
such as enabling/disabling locking. You can also store your own temporary user
data in this structure (be wary of name collision), which is then accessible from
any child hash or array.

=head1 CIRCULAR REFERENCES

DBM::Deep has full support for circular references. Meaning you
can have a nested hash key or array element that points to a parent object.
This relationship is stored in the DB file, and is preserved between sessions.
Here is an example:

  my $db = DBM::Deep->new( "foo.db" );

  $db->{foo} = "bar";
  $db->{circle} = $db; # ref to self

  print $db->{foo} . "\n"; # prints "bar"
  print $db->{circle}->{foo} . "\n"; # prints "bar" again

This also works as expected with array and hash references. So, the following
works as expected:

  $db->{foo} = [ 1 .. 3 ];
  $db->{bar} = $db->{foo};

  push @{$db->{foo}}, 42;
  is( $db->{bar}[-1], 42 ); # Passes

This, however, does I<not> extend to assignments from one DB file to another.
So, the following will throw an error:

  my $db1 = DBM::Deep->new( "foo.db" );
  my $db2 = DBM::Deep->new( "bar.db" );

  $db1->{foo} = [];
  $db2->{foo} = $db1->{foo}; # dies

B<Note>: Passing the object to a function that recursively walks the
object tree (such as I<Data::Dumper> or even the built-in C<optimize()> or
C<export()> methods) will result in an infinite loop. This will be fixed in
a future release by adding singleton support.

=head1 TRANSACTIONS

As of 1.0000, DBM::Deep hass ACID transactions. Every DBM::Deep object is completely
transaction-ready - it is not an option you have to turn on. You do have to
specify how many transactions may run simultaneously (q.v. L</num_txns>).

Three new methods have been added to support them. They are:

=over 4

=item * begin_work()

This starts a transaction.

=item * commit()

This applies the changes done within the transaction to the mainline and ends
the transaction.

=item * rollback()

This discards the changes done within the transaction to the mainline and ends
the transaction.

=back

Transactions in DBM::Deep are done using a variant of the MVCC method, the
same method used by the InnoDB MySQL engine.

=head1 MIGRATION

As of 1.0000, the file format has changed. Furthermore, DBM::Deep is now
designed to potentially change file format between point-releases, if needed to
support a requested feature. To aid in this, a migration script is provided
within the CPAN distribution called C<utils/upgrade_db.pl>.

B<NOTE:> This script is not installed onto your system because it carries a copy
of every version prior to the current version.

=head1 TODO

The following are items that are planned to be added in future releases. These
are separate from the L<CAVEATS, ISSUES & BUGS> below.

=head2 Sub-Transactions

Right now, you cannot run a transaction within a transaction. Removing this
restriction is technically straightforward, but the combinatorial explosion of
possible usecases hurts my head. If this is something you want to see
immediately, please submit many testcases.

=head2 Caching

If a user is willing to assert upon opening the file that this process will be
the only consumer of that datafile, then there are a number of caching
possibilities that can be taken advantage of. This does, however, mean that
DBM::Deep is more vulnerable to losing data due to unflushed changes. It also
means a much larger in-memory footprint. As such, it's not clear exactly how
this should be done. Suggestions are welcome.

=head2 Ram-only

The techniques used in DBM::Deep simply require a seekable contiguous
datastore. This could just as easily be a large string as a file. By using
substr, the STM capabilities of DBM::Deep could be used within a
single-process. I have no idea how I'd specify this, though. Suggestions are
welcome.

=head2 Importing using Data::Walker

Right now, importing is done using C<Clone::clone()> to make a complete copy
in memory, then tying that copy. It would be much better to use
L<Data::Walker/> to walk the data structure instead, particularly in the case
of large datastructures.

=head2 Different contention resolution mechanisms

Currently, the only contention resolution mechanism is last-write-wins. This
is the mechanism used by most RDBMSes and should be good enough for most uses.
For advanced uses of STM, other contention mechanisms will be needed. If you
have an idea of how you'd like to see contention resolution in DBM::Deep,
please let me know.

=head1 CAVEATS, ISSUES & BUGS

This section describes all the known issues with DBM::Deep. These are issues
that are either intractable or depend on some feature within Perl working
exactly right. It you have found something that is not listed below, please
send an e-mail to L<rkinyon@cpan.org>. Likewise, if you think you know of a
way around one of these issues, please let me know.

=head2 References

(The following assumes a high level of Perl understanding, specifically of
references. Most users can safely skip this section.)

Currently, the only references supported are HASH and ARRAY. The other reference
types (SCALAR, CODE, GLOB, and REF) cannot be supported for various reasons.

=over 4

=item * GLOB

These are things like filehandles and other sockets. They can't be supported
because it's completely unclear how DBM::Deep should serialize them.

=item * SCALAR / REF

The discussion here refers to the following type of example:

  my $x = 25;
  $db->{key1} = \$x;

  $x = 50;

  # In some other process ...

  my $val = ${ $db->{key1} };

  is( $val, 50, "What actually gets stored in the DB file?" );

The problem is one of synchronization. When the variable being referred to
changes value, the reference isn't notified, which is kind of the point of
references. This means that the new value won't be stored in the datafile for
other processes to read. There is no TIEREF.

It is theoretically possible to store references to values already within a
DBM::Deep object because everything already is synchronized, but the change to
the internals would be quite large. Specifically, DBM::Deep would have to tie
every single value that is stored. This would bloat the RAM footprint of
DBM::Deep at least twofold (if not more) and be a significant performance drain,
all to support a feature that has never been requested.

=item * CODE

L<Data::Dump::Streamer/> provides a mechanism for serializing coderefs,
including saving off all closure state. This would allow for DBM::Deep to
store the code for a subroutine. Then, whenever the subroutine is read, the
code could be C<eval()>'ed into being. However, just as for SCALAR and REF,
that closure state may change without notifying the DBM::Deep object storing
the reference. Again, this would generally be considered a feature.

=back

=head2 External references and transactions

If you do C<my $x = $db-E<gt>{foo};>, then start a transaction, $x will be
referencing the database from outside the transaction. A fix for this (and other
issues with how external references into the database) is being looked into. This
is the skipped set of tests in t/39_singletons.t and a related issue is the focus
of t/37_delete_edge_cases.t

=head2 File corruption

The current level of error handling in DBM::Deep is minimal. Files I<are> checked
for a 32-bit signature when opened, but any other form of corruption in the
datafile can cause segmentation faults. DBM::Deep may try to C<seek()> past
the end of a file, or get stuck in an infinite loop depending on the level and
type of corruption. File write operations are not checked for failure (for
speed), so if you happen to run out of disk space, DBM::Deep will probably fail in
a bad way. These things will be addressed in a later version of DBM::Deep.

=head2 DB over NFS

Beware of using DBM::Deep files over NFS. DBM::Deep uses flock(), which works
well on local filesystems, but will NOT protect you from file corruption over
NFS. I've heard about setting up your NFS server with a locking daemon, then
using C<lockf()> to lock your files, but your mileage may vary there as well.
From what I understand, there is no real way to do it. However, if you need
access to the underlying filehandle in DBM::Deep for using some other kind of
locking scheme like C<lockf()>, see the L<LOW-LEVEL ACCESS> section above.

=head2 Copying Objects

Beware of copying tied objects in Perl. Very strange things can happen.
Instead, use DBM::Deep's C<clone()> method which safely copies the object and
returns a new, blessed and tied hash or array to the same level in the DB.

  my $copy = $db->clone();

B<Note>: Since clone() here is cloning the object, not the database location, any
modifications to either $db or $copy will be visible to both.

=head2 Large Arrays

Beware of using C<shift()>, C<unshift()> or C<splice()> with large arrays.
These functions cause every element in the array to move, which can be murder
on DBM::Deep, as every element has to be fetched from disk, then stored again in
a different location. This will be addressed in a future version.

=head2 Writeonly Files

If you pass in a filehandle to new(), you may have opened it in either a readonly or
writeonly mode. STORE will verify that the filehandle is writable. However, there
doesn't seem to be a good way to determine if a filehandle is readable. And, if the
filehandle isn't readable, it's not clear what will happen. So, don't do that.

=head2 Assignments Within Transactions

The following will I<not> work as one might expect:

  my $x = { a => 1 };

  $db->begin_work;
  $db->{foo} = $x;
  $db->rollback;

  is( $x->{a}, 1 ); # This will fail!

The problem is that the moment a reference used as the rvalue to a DBM::Deep
object's lvalue, it becomes tied itself. This is so that future changes to
C<$x> can be tracked within the DBM::Deep file and is considered to be a
feature. By the time the rollback occurs, there is no knowledge that there had
been an C<$x> or what memory location to assign an C<export()> to.

B<NOTE:> This does not affect importing because imports do a walk over the
reference to be imported in order to explicitly leave it untied.

=head1 CODE COVERAGE

B<Devel::Cover> is used to test the code coverage of the tests. Below is the
B<Devel::Cover> report on this distribution's test suite.

  ------------------------------------------ ------ ------ ------ ------ ------
  File                                         stmt   bran   cond    sub  total
  ------------------------------------------ ------ ------ ------ ------ ------
  blib/lib/DBM/Deep.pm                         96.9   88.3   90.5  100.0   95.7
  blib/lib/DBM/Deep/Array.pm                  100.0   95.7  100.0  100.0   99.0
  blib/lib/DBM/Deep/Engine.pm                  95.5   84.7   81.6   98.4   92.4
  blib/lib/DBM/Deep/File.pm                    97.2   81.6   66.7  100.0   91.9
  blib/lib/DBM/Deep/Hash.pm                   100.0  100.0  100.0  100.0  100.0
  Total                                        96.7   87.0   83.3   99.2   94.1
  ------------------------------------------ ------ ------ ------ ------ ------

=head1 MORE INFORMATION

Check out the DBM::Deep Google Group at L<http://groups.google.com/group/DBM-Deep>
or send email to L<DBM-Deep@googlegroups.com>. You can also visit #dbm-deep on
irc.perl.org

The source code repository is at L<http://svn.perl.org/modules/DBM-Deep>

=head1 MAINTAINERS

Rob Kinyon, L<rkinyon@cpan.org>

Originally written by Joseph Huckaby, L<jhuckaby@cpan.org>

=head1 SPONSORS

Stonehenge Consulting (L<http://www.stonehenge.com/>) sponsored the
developement of transactions and freespace management, leading to the 1.0000
release. A great debt of gratitude goes out to them for their continuing
leadership in and support of the Perl community.

=head1 CONTRIBUTORS

The following have contributed greatly to make DBM::Deep what it is today:

=over 4

=item * Adam Sah and Rich Gaushell for innumerable contributions early on.

=item * Dan Golden and others at YAPC::NA 2006 for helping me design through transactions.

=back

=head1 SEE ALSO

perltie(1), Tie::Hash(3), Digest::MD5(3), Fcntl(3), flock(2), lockf(3), nfs(5),
Digest::SHA256(3), Crypt::Blowfish(3), Compress::Zlib(3)

=head1 LICENSE

Copyright (c) 2007 Rob Kinyon. All Rights Reserved.
This is free software, you may use it and distribute it under the same terms
as Perl itself.

=cut