Commit | Line | Data |
ffed8b01 |
1 | package DBM::Deep; |
2 | |
3 | ## |
4 | # DBM::Deep |
5 | # |
6 | # Description: |
d0b74c17 |
7 | # Multi-level database module for storing hash trees, arrays and simple |
8 | # key/value pairs into FTP-able, cross-platform binary database files. |
ffed8b01 |
9 | # |
d0b74c17 |
10 | # Type `perldoc DBM::Deep` for complete documentation. |
ffed8b01 |
11 | # |
12 | # Usage Examples: |
d0b74c17 |
13 | # my %db; |
14 | # tie %db, 'DBM::Deep', 'my_database.db'; # standard tie() method |
ffed8b01 |
15 | # |
d0b74c17 |
16 | # my $db = new DBM::Deep( 'my_database.db' ); # preferred OO method |
17 | # |
18 | # $db->{my_scalar} = 'hello world'; |
19 | # $db->{my_hash} = { larry => 'genius', hashes => 'fast' }; |
20 | # $db->{my_array} = [ 1, 2, 3, time() ]; |
21 | # $db->{my_complex} = [ 'hello', { perl => 'rules' }, 42, 99 ]; |
22 | # push @{$db->{my_array}}, 'another value'; |
23 | # my @key_list = keys %{$db->{my_hash}}; |
24 | # print "This module " . $db->{my_complex}->[1]->{perl} . "!\n"; |
ffed8b01 |
25 | # |
26 | # Copyright: |
d0b74c17 |
27 | # (c) 2002-2006 Joseph Huckaby. All Rights Reserved. |
28 | # This program is free software; you can redistribute it and/or |
29 | # modify it under the same terms as Perl itself. |
ffed8b01 |
30 | ## |
31 | |
460b1067 |
32 | use 5.6.0; |
33 | |
ffed8b01 |
34 | use strict; |
460b1067 |
35 | use warnings; |
8b957036 |
36 | |
d8db2929 |
37 | our $VERSION = q(0.99_03); |
86867f3a |
38 | |
596e9574 |
39 | use Fcntl qw( :DEFAULT :flock :seek ); |
12b96196 |
40 | |
41 | use Clone::Any '_clone_data'; |
ffed8b01 |
42 | use Digest::MD5 (); |
a8fdabda |
43 | use FileHandle::Fmode (); |
ffed8b01 |
44 | use Scalar::Util (); |
ffed8b01 |
45 | |
129ea236 |
46 | use DBM::Deep::Engine2; |
460b1067 |
47 | use DBM::Deep::File; |
95967a5e |
48 | |
ffed8b01 |
49 | ## |
50 | # Setup constants for users to pass to new() |
51 | ## |
129ea236 |
52 | sub TYPE_HASH () { DBM::Deep::Engine2->SIG_HASH } |
53 | sub TYPE_ARRAY () { DBM::Deep::Engine2->SIG_ARRAY } |
ffed8b01 |
54 | |
0ca7ea98 |
55 | sub _get_args { |
56 | my $proto = shift; |
57 | |
58 | my $args; |
59 | if (scalar(@_) > 1) { |
60 | if ( @_ % 2 ) { |
61 | $proto->_throw_error( "Odd number of parameters to " . (caller(1))[2] ); |
62 | } |
63 | $args = {@_}; |
64 | } |
d0b74c17 |
65 | elsif ( ref $_[0] ) { |
4d35d856 |
66 | unless ( eval { local $SIG{'__DIE__'}; %{$_[0]} || 1 } ) { |
0ca7ea98 |
67 | $proto->_throw_error( "Not a hashref in args to " . (caller(1))[2] ); |
68 | } |
69 | $args = $_[0]; |
70 | } |
d0b74c17 |
71 | else { |
0ca7ea98 |
72 | $args = { file => shift }; |
73 | } |
74 | |
75 | return $args; |
76 | } |
77 | |
ffed8b01 |
78 | sub new { |
d0b74c17 |
79 | ## |
80 | # Class constructor method for Perl OO interface. |
81 | # Calls tie() and returns blessed reference to tied hash or array, |
82 | # providing a hybrid OO/tie interface. |
83 | ## |
84 | my $class = shift; |
85 | my $args = $class->_get_args( @_ ); |
86 | |
87 | ## |
88 | # Check if we want a tied hash or array. |
89 | ## |
90 | my $self; |
91 | if (defined($args->{type}) && $args->{type} eq TYPE_ARRAY) { |
6fe26b29 |
92 | $class = 'DBM::Deep::Array'; |
93 | require DBM::Deep::Array; |
d0b74c17 |
94 | tie @$self, $class, %$args; |
95 | } |
96 | else { |
6fe26b29 |
97 | $class = 'DBM::Deep::Hash'; |
98 | require DBM::Deep::Hash; |
d0b74c17 |
99 | tie %$self, $class, %$args; |
100 | } |
ffed8b01 |
101 | |
d0b74c17 |
102 | return bless $self, $class; |
ffed8b01 |
103 | } |
104 | |
96041a25 |
105 | # This initializer is called from the various TIE* methods. new() calls tie(), |
106 | # which allows for a single point of entry. |
0795f290 |
107 | sub _init { |
0795f290 |
108 | my $class = shift; |
994ccd8e |
109 | my ($args) = @_; |
0795f290 |
110 | |
83371fe3 |
111 | $args->{storage} = DBM::Deep::File->new( $args ) |
112 | unless exists $args->{storage}; |
460b1067 |
113 | |
114 | # locking implicitly enables autoflush |
115 | if ($args->{locking}) { $args->{autoflush} = 1; } |
116 | |
0795f290 |
117 | # These are the defaults to be optionally overridden below |
118 | my $self = bless { |
95967a5e |
119 | type => TYPE_HASH, |
e06824f8 |
120 | base_offset => undef, |
359a01ac |
121 | |
122 | parent => undef, |
123 | parent_key => undef, |
124 | |
83371fe3 |
125 | storage => undef, |
0795f290 |
126 | }, $class; |
129ea236 |
127 | $self->{engine} = DBM::Deep::Engine2->new( { %{$args}, obj => $self } ); |
8db25060 |
128 | |
fde3db1a |
129 | # Grab the parameters we want to use |
0795f290 |
130 | foreach my $param ( keys %$self ) { |
131 | next unless exists $args->{$param}; |
3e9498a1 |
132 | $self->{$param} = $args->{$param}; |
ffed8b01 |
133 | } |
d0b74c17 |
134 | |
72e315ac |
135 | $self->_engine->setup_fh( $self ); |
0795f290 |
136 | |
83371fe3 |
137 | $self->_storage->set_db( $self ); |
359a01ac |
138 | |
0795f290 |
139 | return $self; |
ffed8b01 |
140 | } |
141 | |
ffed8b01 |
142 | sub TIEHASH { |
6fe26b29 |
143 | shift; |
144 | require DBM::Deep::Hash; |
145 | return DBM::Deep::Hash->TIEHASH( @_ ); |
ffed8b01 |
146 | } |
147 | |
148 | sub TIEARRAY { |
6fe26b29 |
149 | shift; |
150 | require DBM::Deep::Array; |
151 | return DBM::Deep::Array->TIEARRAY( @_ ); |
ffed8b01 |
152 | } |
153 | |
ffed8b01 |
154 | sub lock { |
994ccd8e |
155 | my $self = shift->_get_self; |
83371fe3 |
156 | return $self->_storage->lock( $self, @_ ); |
ffed8b01 |
157 | } |
158 | |
159 | sub unlock { |
994ccd8e |
160 | my $self = shift->_get_self; |
83371fe3 |
161 | return $self->_storage->unlock( $self, @_ ); |
ffed8b01 |
162 | } |
163 | |
906c8e01 |
164 | sub _copy_value { |
165 | my $self = shift->_get_self; |
166 | my ($spot, $value) = @_; |
167 | |
168 | if ( !ref $value ) { |
169 | ${$spot} = $value; |
170 | } |
171 | elsif ( eval { local $SIG{__DIE__}; $value->isa( 'DBM::Deep' ) } ) { |
f9c33187 |
172 | ${$spot} = $value->_repr; |
906c8e01 |
173 | $value->_copy_node( ${$spot} ); |
174 | } |
175 | else { |
176 | my $r = Scalar::Util::reftype( $value ); |
177 | my $c = Scalar::Util::blessed( $value ); |
178 | if ( $r eq 'ARRAY' ) { |
179 | ${$spot} = [ @{$value} ]; |
180 | } |
181 | else { |
182 | ${$spot} = { %{$value} }; |
183 | } |
95bbd935 |
184 | ${$spot} = bless ${$spot}, $c |
906c8e01 |
185 | if defined $c; |
186 | } |
187 | |
188 | return 1; |
189 | } |
190 | |
261d1296 |
191 | sub _copy_node { |
f9c33187 |
192 | die "Must be implemented in a child class\n"; |
193 | } |
906c8e01 |
194 | |
f9c33187 |
195 | sub _repr { |
196 | die "Must be implemented in a child class\n"; |
ffed8b01 |
197 | } |
198 | |
199 | sub export { |
d0b74c17 |
200 | ## |
201 | # Recursively export into standard Perl hashes and arrays. |
202 | ## |
994ccd8e |
203 | my $self = shift->_get_self; |
d0b74c17 |
204 | |
f9c33187 |
205 | my $temp = $self->_repr; |
d0b74c17 |
206 | |
207 | $self->lock(); |
208 | $self->_copy_node( $temp ); |
209 | $self->unlock(); |
210 | |
68f943b3 |
211 | # This will always work because $self, after _get_self() is a HASH |
212 | if ( $self->{parent} ) { |
213 | my $c = Scalar::Util::blessed( |
214 | $self->{parent}->get($self->{parent_key}) |
215 | ); |
9a772062 |
216 | if ( $c && !$c->isa( 'DBM::Deep' ) ) { |
68f943b3 |
217 | bless $temp, $c; |
218 | } |
219 | } |
220 | |
d0b74c17 |
221 | return $temp; |
ffed8b01 |
222 | } |
223 | |
224 | sub import { |
d0b74c17 |
225 | ## |
226 | # Recursively import Perl hash/array structure |
227 | ## |
d0b74c17 |
228 | if (!ref($_[0])) { return; } # Perl calls import() on use -- ignore |
229 | |
994ccd8e |
230 | my $self = shift->_get_self; |
231 | my ($struct) = @_; |
d0b74c17 |
232 | |
c9cec40e |
233 | # struct is not a reference, so just import based on our type |
d0b74c17 |
234 | if (!ref($struct)) { |
f9c33187 |
235 | $struct = $self->_repr( @_ ); |
d0b74c17 |
236 | } |
237 | |
12b96196 |
238 | #XXX This isn't the best solution. Better would be to use Data::Walker, |
239 | #XXX but that's a lot more thinking than I want to do right now. |
7a960a12 |
240 | eval { |
12b96196 |
241 | $self->begin_work; |
242 | $self->_import( _clone_data( $struct ) ); |
243 | $self->commit; |
7a960a12 |
244 | }; if ( $@ ) { |
245 | $self->rollback; |
246 | die $@; |
247 | } |
248 | |
249 | return 1; |
ffed8b01 |
250 | } |
251 | |
13ff93d5 |
252 | #XXX Need to keep track of who has a fh to this file in order to |
253 | #XXX close them all prior to optimize on Win32/cygwin |
ffed8b01 |
254 | sub optimize { |
d0b74c17 |
255 | ## |
256 | # Rebuild entire database into new file, then move |
257 | # it back on top of original. |
258 | ## |
994ccd8e |
259 | my $self = shift->_get_self; |
cc4bef86 |
260 | |
261 | #XXX Need to create a new test for this |
83371fe3 |
262 | # if ($self->_storage->{links} > 1) { |
1400a48e |
263 | # $self->_throw_error("Cannot optimize: reference count is greater than 1"); |
d0b74c17 |
264 | # } |
265 | |
7a960a12 |
266 | #XXX Do we have to lock the tempfile? |
267 | |
d0b74c17 |
268 | my $db_temp = DBM::Deep->new( |
83371fe3 |
269 | file => $self->_storage->{file} . '.tmp', |
d0b74c17 |
270 | type => $self->_type |
271 | ); |
d0b74c17 |
272 | |
273 | $self->lock(); |
274 | $self->_copy_node( $db_temp ); |
275 | undef $db_temp; |
276 | |
277 | ## |
278 | # Attempt to copy user, group and permissions over to new file |
279 | ## |
280 | my @stats = stat($self->_fh); |
281 | my $perms = $stats[2] & 07777; |
282 | my $uid = $stats[4]; |
283 | my $gid = $stats[5]; |
83371fe3 |
284 | chown( $uid, $gid, $self->_storage->{file} . '.tmp' ); |
285 | chmod( $perms, $self->_storage->{file} . '.tmp' ); |
d0b74c17 |
286 | |
ffed8b01 |
287 | # q.v. perlport for more information on this variable |
90f93b43 |
288 | if ( $^O eq 'MSWin32' || $^O eq 'cygwin' ) { |
d0b74c17 |
289 | ## |
290 | # Potential race condition when optmizing on Win32 with locking. |
291 | # The Windows filesystem requires that the filehandle be closed |
292 | # before it is overwritten with rename(). This could be redone |
293 | # with a soft copy. |
294 | ## |
295 | $self->unlock(); |
83371fe3 |
296 | $self->_storage->close; |
d0b74c17 |
297 | } |
298 | |
83371fe3 |
299 | if (!rename $self->_storage->{file} . '.tmp', $self->_storage->{file}) { |
300 | unlink $self->_storage->{file} . '.tmp'; |
d0b74c17 |
301 | $self->unlock(); |
1400a48e |
302 | $self->_throw_error("Optimize failed: Cannot copy temp file over original: $!"); |
d0b74c17 |
303 | } |
304 | |
305 | $self->unlock(); |
83371fe3 |
306 | $self->_storage->close; |
307 | $self->_storage->open; |
72e315ac |
308 | $self->_engine->setup_fh( $self ); |
70b55428 |
309 | |
d0b74c17 |
310 | return 1; |
ffed8b01 |
311 | } |
312 | |
313 | sub clone { |
d0b74c17 |
314 | ## |
315 | # Make copy of object and return |
316 | ## |
994ccd8e |
317 | my $self = shift->_get_self; |
d0b74c17 |
318 | |
319 | return DBM::Deep->new( |
c3aafc14 |
320 | type => $self->_type, |
d0b74c17 |
321 | base_offset => $self->_base_offset, |
83371fe3 |
322 | storage => $self->_storage, |
c3aafc14 |
323 | parent => $self->{parent}, |
324 | parent_key => $self->{parent_key}, |
d0b74c17 |
325 | ); |
ffed8b01 |
326 | } |
327 | |
328 | { |
329 | my %is_legal_filter = map { |
330 | $_ => ~~1, |
331 | } qw( |
332 | store_key store_value |
333 | fetch_key fetch_value |
334 | ); |
335 | |
336 | sub set_filter { |
337 | ## |
338 | # Setup filter function for storing or fetching the key or value |
339 | ## |
994ccd8e |
340 | my $self = shift->_get_self; |
341 | my $type = lc shift; |
342 | my $func = shift; |
d0b74c17 |
343 | |
ffed8b01 |
344 | if ( $is_legal_filter{$type} ) { |
83371fe3 |
345 | $self->_storage->{"filter_$type"} = $func; |
ffed8b01 |
346 | return 1; |
347 | } |
348 | |
349 | return; |
350 | } |
351 | } |
352 | |
fee0243f |
353 | sub begin_work { |
354 | my $self = shift->_get_self; |
83371fe3 |
355 | return $self->_storage->begin_transaction; |
fee0243f |
356 | } |
357 | |
358 | sub rollback { |
359 | my $self = shift->_get_self; |
83371fe3 |
360 | return $self->_storage->end_transaction; |
fee0243f |
361 | } |
362 | |
359a01ac |
363 | sub commit { |
364 | my $self = shift->_get_self; |
83371fe3 |
365 | return $self->_storage->commit_transaction; |
359a01ac |
366 | } |
fee0243f |
367 | |
ffed8b01 |
368 | ## |
369 | # Accessor methods |
370 | ## |
371 | |
72e315ac |
372 | sub _engine { |
373 | my $self = $_[0]->_get_self; |
374 | return $self->{engine}; |
375 | } |
376 | |
83371fe3 |
377 | sub _storage { |
2ac02042 |
378 | my $self = $_[0]->_get_self; |
83371fe3 |
379 | return $self->{storage}; |
ffed8b01 |
380 | } |
381 | |
4d35d856 |
382 | sub _type { |
2ac02042 |
383 | my $self = $_[0]->_get_self; |
d0b74c17 |
384 | return $self->{type}; |
ffed8b01 |
385 | } |
386 | |
4d35d856 |
387 | sub _base_offset { |
2ac02042 |
388 | my $self = $_[0]->_get_self; |
d0b74c17 |
389 | return $self->{base_offset}; |
ffed8b01 |
390 | } |
391 | |
994ccd8e |
392 | sub _fh { |
994ccd8e |
393 | my $self = $_[0]->_get_self; |
83371fe3 |
394 | return $self->_storage->{fh}; |
994ccd8e |
395 | } |
396 | |
ffed8b01 |
397 | ## |
398 | # Utility methods |
399 | ## |
400 | |
261d1296 |
401 | sub _throw_error { |
95967a5e |
402 | die "DBM::Deep: $_[1]\n"; |
ffed8b01 |
403 | } |
404 | |
359a01ac |
405 | sub _find_parent { |
406 | my $self = shift; |
cfd97a7f |
407 | |
408 | my $base = ''; |
633df1fd |
409 | #XXX This if() is redundant |
cfd97a7f |
410 | if ( my $parent = $self->{parent} ) { |
411 | my $child = $self; |
25c7c8d6 |
412 | while ( $parent->{parent} ) { |
cfd97a7f |
413 | $base = ( |
414 | $parent->_type eq TYPE_HASH |
415dcbb7 |
415 | ? "\{q{$child->{parent_key}}\}" |
cfd97a7f |
416 | : "\[$child->{parent_key}\]" |
417 | ) . $base; |
418 | |
419 | $child = $parent; |
420 | $parent = $parent->{parent}; |
25c7c8d6 |
421 | } |
422 | if ( $base ) { |
415dcbb7 |
423 | $base = "\$db->get( q{$child->{parent_key}} )->" . $base; |
25c7c8d6 |
424 | } |
425 | else { |
415dcbb7 |
426 | $base = "\$db->get( q{$child->{parent_key}} )"; |
359a01ac |
427 | } |
359a01ac |
428 | } |
25c7c8d6 |
429 | return $base; |
359a01ac |
430 | } |
431 | |
ffed8b01 |
432 | sub STORE { |
d0b74c17 |
433 | ## |
434 | # Store single hash key/value or array element in database. |
435 | ## |
436 | my $self = shift->_get_self; |
359a01ac |
437 | my ($key, $value, $orig_key) = @_; |
c3aafc14 |
438 | $orig_key = $key unless defined $orig_key; |
81d3d316 |
439 | |
a8fdabda |
440 | if ( !FileHandle::Fmode::is_W( $self->_fh ) ) { |
acd4faf2 |
441 | $self->_throw_error( 'Cannot write to a readonly filehandle' ); |
442 | } |
d0b74c17 |
443 | |
504185fb |
444 | #XXX The second condition needs to disappear |
c3aafc14 |
445 | if ( !( $self->_type eq TYPE_ARRAY && $orig_key eq 'length') ) { |
4768a580 |
446 | my $rhs; |
447 | |
448 | my $r = Scalar::Util::reftype( $value ) || ''; |
449 | if ( $r eq 'HASH' ) { |
450 | $rhs = '{}'; |
451 | } |
452 | elsif ( $r eq 'ARRAY' ) { |
453 | $rhs = '[]'; |
454 | } |
455 | elsif ( defined $value ) { |
456 | $rhs = "'$value'"; |
457 | } |
458 | else { |
459 | $rhs = "undef"; |
460 | } |
461 | |
462 | if ( my $c = Scalar::Util::blessed( $value ) ) { |
463 | $rhs = "bless $rhs, '$c'"; |
464 | } |
465 | |
25c7c8d6 |
466 | my $lhs = $self->_find_parent; |
467 | if ( $lhs ) { |
468 | if ( $self->_type eq TYPE_HASH ) { |
415dcbb7 |
469 | $lhs .= "->\{q{$orig_key}\}"; |
25c7c8d6 |
470 | } |
471 | else { |
472 | $lhs .= "->\[$orig_key\]"; |
473 | } |
474 | |
475 | $lhs .= "=$rhs;"; |
476 | } |
477 | else { |
415dcbb7 |
478 | $lhs = "\$db->put(q{$orig_key},$rhs);"; |
25c7c8d6 |
479 | } |
480 | |
83371fe3 |
481 | $self->_storage->audit($lhs); |
4768a580 |
482 | } |
359a01ac |
483 | |
d0b74c17 |
484 | ## |
485 | # Request exclusive lock for writing |
486 | ## |
487 | $self->lock( LOCK_EX ); |
488 | |
0cb639bd |
489 | # User may be storing a complex value, in which case we do not want it run |
490 | # through the filtering system. |
83371fe3 |
491 | if ( !ref($value) && $self->_storage->{filter_store_value} ) { |
492 | $value = $self->_storage->{filter_store_value}->( $value ); |
d0b74c17 |
493 | } |
494 | |
40956c06 |
495 | $self->_engine->write_value( $self->_storage->transaction_id, $self->_base_offset, $key, $value, $orig_key ); |
d0b74c17 |
496 | |
497 | $self->unlock(); |
498 | |
86867f3a |
499 | return 1; |
ffed8b01 |
500 | } |
501 | |
502 | sub FETCH { |
d0b74c17 |
503 | ## |
504 | # Fetch single value or element given plain key or array index |
505 | ## |
cb79ec85 |
506 | my $self = shift->_get_self; |
a97c8f67 |
507 | my ($key, $orig_key) = @_; |
0cb639bd |
508 | $orig_key = $key unless defined $orig_key; |
ffed8b01 |
509 | |
d0b74c17 |
510 | ## |
511 | # Request shared lock for reading |
512 | ## |
513 | $self->lock( LOCK_SH ); |
514 | |
40956c06 |
515 | my $result = $self->_engine->read_value( $self->_storage->transaction_id, $self->_base_offset, $key, $orig_key ); |
d0b74c17 |
516 | |
517 | $self->unlock(); |
518 | |
a86430bd |
519 | # Filters only apply to scalar values, so the ref check is making |
520 | # sure the fetched bucket is a scalar, not a child hash or array. |
83371fe3 |
521 | return ($result && !ref($result) && $self->_storage->{filter_fetch_value}) |
522 | ? $self->_storage->{filter_fetch_value}->($result) |
cb79ec85 |
523 | : $result; |
ffed8b01 |
524 | } |
525 | |
526 | sub DELETE { |
d0b74c17 |
527 | ## |
528 | # Delete single key/value pair or element given plain key or array index |
529 | ## |
a97c8f67 |
530 | my $self = shift->_get_self; |
531 | my ($key, $orig_key) = @_; |
c3aafc14 |
532 | $orig_key = $key unless defined $orig_key; |
d0b74c17 |
533 | |
a8fdabda |
534 | if ( !FileHandle::Fmode::is_W( $self->_fh ) ) { |
a86430bd |
535 | $self->_throw_error( 'Cannot write to a readonly filehandle' ); |
536 | } |
d0b74c17 |
537 | |
4768a580 |
538 | if ( defined $orig_key ) { |
539 | my $lhs = $self->_find_parent; |
25c7c8d6 |
540 | if ( $lhs ) { |
83371fe3 |
541 | $self->_storage->audit( "delete $lhs;" ); |
a97c8f67 |
542 | } |
4768a580 |
543 | else { |
83371fe3 |
544 | $self->_storage->audit( "\$db->delete('$orig_key');" ); |
4768a580 |
545 | } |
a97c8f67 |
546 | } |
547 | |
d0b74c17 |
548 | ## |
549 | # Request exclusive lock for writing |
550 | ## |
551 | $self->lock( LOCK_EX ); |
552 | |
d0b74c17 |
553 | ## |
554 | # Delete bucket |
555 | ## |
40956c06 |
556 | my $value = $self->_engine->delete_key( $self->_storage->transaction_id, $self->_base_offset, $key, $orig_key ); |
a86430bd |
557 | |
83371fe3 |
558 | if (defined $value && !ref($value) && $self->_storage->{filter_fetch_value}) { |
559 | $value = $self->_storage->{filter_fetch_value}->($value); |
3b6a5056 |
560 | } |
561 | |
d0b74c17 |
562 | $self->unlock(); |
563 | |
564 | return $value; |
ffed8b01 |
565 | } |
566 | |
567 | sub EXISTS { |
d0b74c17 |
568 | ## |
569 | # Check if a single key or element exists given plain key or array index |
570 | ## |
a97c8f67 |
571 | my $self = shift->_get_self; |
572 | my ($key) = @_; |
d0b74c17 |
573 | |
d0b74c17 |
574 | ## |
575 | # Request shared lock for reading |
576 | ## |
577 | $self->lock( LOCK_SH ); |
578 | |
40956c06 |
579 | my $result = $self->_engine->key_exists( $self->_storage->transaction_id, $self->_base_offset, $key ); |
d0b74c17 |
580 | |
581 | $self->unlock(); |
582 | |
583 | return $result; |
ffed8b01 |
584 | } |
585 | |
586 | sub CLEAR { |
d0b74c17 |
587 | ## |
588 | # Clear all keys from hash, or all elements from array. |
589 | ## |
a97c8f67 |
590 | my $self = shift->_get_self; |
ffed8b01 |
591 | |
a8fdabda |
592 | if ( !FileHandle::Fmode::is_W( $self->_fh ) ) { |
a86430bd |
593 | $self->_throw_error( 'Cannot write to a readonly filehandle' ); |
594 | } |
595 | |
4768a580 |
596 | { |
a97c8f67 |
597 | my $lhs = $self->_find_parent; |
598 | |
a97c8f67 |
599 | if ( $self->_type eq TYPE_HASH ) { |
e82621dd |
600 | $lhs = '%{' . $lhs . '}'; |
a97c8f67 |
601 | } |
602 | else { |
e82621dd |
603 | $lhs = '@{' . $lhs . '}'; |
a97c8f67 |
604 | } |
605 | |
83371fe3 |
606 | $self->_storage->audit( "$lhs = ();" ); |
a97c8f67 |
607 | } |
608 | |
d0b74c17 |
609 | ## |
610 | # Request exclusive lock for writing |
611 | ## |
612 | $self->lock( LOCK_EX ); |
613 | |
f9a320bb |
614 | if ( $self->_type eq TYPE_HASH ) { |
615 | my $key = $self->first_key; |
616 | while ( $key ) { |
83c43bb5 |
617 | # Retrieve the key before deleting because we depend on next_key |
f9a320bb |
618 | my $next_key = $self->next_key( $key ); |
40956c06 |
619 | $self->_engine->delete_key( $self->_storage->transaction_id, $self->_base_offset, $key, $key ); |
f9a320bb |
620 | $key = $next_key; |
621 | } |
622 | } |
623 | else { |
624 | my $size = $self->FETCHSIZE; |
c3aafc14 |
625 | for my $key ( 0 .. $size - 1 ) { |
40956c06 |
626 | $self->_engine->delete_key( $self->_storage->transaction_id, $self->_base_offset, $key, $key ); |
f9a320bb |
627 | } |
628 | $self->STORESIZE( 0 ); |
629 | } |
f9c33187 |
630 | #XXX This needs updating to use _release_space |
f9a320bb |
631 | # $self->_engine->write_tag( |
632 | # $self->_base_offset, $self->_type, |
633 | # chr(0)x$self->_engine->{index_size}, |
634 | # ); |
d0b74c17 |
635 | |
636 | $self->unlock(); |
637 | |
638 | return 1; |
ffed8b01 |
639 | } |
640 | |
ffed8b01 |
641 | ## |
642 | # Public method aliases |
643 | ## |
7f441181 |
644 | sub put { (shift)->STORE( @_ ) } |
645 | sub store { (shift)->STORE( @_ ) } |
646 | sub get { (shift)->FETCH( @_ ) } |
647 | sub fetch { (shift)->FETCH( @_ ) } |
baa27ab6 |
648 | sub delete { (shift)->DELETE( @_ ) } |
649 | sub exists { (shift)->EXISTS( @_ ) } |
650 | sub clear { (shift)->CLEAR( @_ ) } |
ffed8b01 |
651 | |
652 | 1; |
ffed8b01 |
653 | __END__ |
654 | |
655 | =head1 NAME |
656 | |
657 | DBM::Deep - A pure perl multi-level hash/array DBM |
658 | |
659 | =head1 SYNOPSIS |
660 | |
661 | use DBM::Deep; |
662 | my $db = DBM::Deep->new( "foo.db" ); |
d0b74c17 |
663 | |
eff6a245 |
664 | $db->{key} = 'value'; |
ffed8b01 |
665 | print $db->{key}; |
d0b74c17 |
666 | |
eff6a245 |
667 | $db->put('key' => 'value'); |
ffed8b01 |
668 | print $db->get('key'); |
d0b74c17 |
669 | |
ffed8b01 |
670 | # true multi-level support |
671 | $db->{my_complex} = [ |
d0b74c17 |
672 | 'hello', { perl => 'rules' }, |
673 | 42, 99, |
90f93b43 |
674 | ]; |
ffed8b01 |
675 | |
eff6a245 |
676 | tie my %db, 'DBM::Deep', 'foo.db'; |
677 | $db{key} = 'value'; |
678 | print $db{key}; |
ffed8b01 |
679 | |
eff6a245 |
680 | tied(%db)->put('key' => 'value'); |
681 | print tied(%db)->get('key'); |
8db25060 |
682 | |
eff6a245 |
683 | =head1 DESCRIPTION |
8db25060 |
684 | |
eff6a245 |
685 | A unique flat-file database module, written in pure perl. True multi-level |
686 | hash/array support (unlike MLDBM, which is faked), hybrid OO / tie() |
687 | interface, cross-platform FTPable files, ACID transactions, and is quite fast. |
688 | Can handle millions of keys and unlimited levels without significant |
689 | slow-down. Written from the ground-up in pure perl -- this is NOT a wrapper |
690 | around a C-based DBM. Out-of-the-box compatibility with Unix, Mac OS X and |
691 | Windows. |
ffed8b01 |
692 | |
eff6a245 |
693 | =head1 VERSION DIFFERENCES |
ffed8b01 |
694 | |
eff6a245 |
695 | B<NOTE>: 0.99_01 and above have significant file format differences from 0.983 and |
696 | before. There will be a backwards-compatibility layer in 1.00, but that is |
697 | slated for a later 0.99_x release. This version is B<NOT> backwards compatible |
698 | with 0.983 and before. |
ffed8b01 |
699 | |
700 | =head1 SETUP |
701 | |
d0b74c17 |
702 | Construction can be done OO-style (which is the recommended way), or using |
ffed8b01 |
703 | Perl's tie() function. Both are examined here. |
704 | |
705 | =head2 OO CONSTRUCTION |
706 | |
707 | The recommended way to construct a DBM::Deep object is to use the new() |
eff6a245 |
708 | method, which gets you a blessed I<and> tied hash (or array) reference. |
ffed8b01 |
709 | |
a8fdabda |
710 | my $db = DBM::Deep->new( "foo.db" ); |
ffed8b01 |
711 | |
712 | This opens a new database handle, mapped to the file "foo.db". If this |
d0b74c17 |
713 | file does not exist, it will automatically be created. DB files are |
ffed8b01 |
714 | opened in "r+" (read/write) mode, and the type of object returned is a |
715 | hash, unless otherwise specified (see L<OPTIONS> below). |
716 | |
ffed8b01 |
717 | You can pass a number of options to the constructor to specify things like |
eff6a245 |
718 | locking, autoflush, etc. This is done by passing an inline hash (or hashref): |
ffed8b01 |
719 | |
a8fdabda |
720 | my $db = DBM::Deep->new( |
721 | file => "foo.db", |
722 | locking => 1, |
723 | autoflush => 1 |
724 | ); |
ffed8b01 |
725 | |
726 | Notice that the filename is now specified I<inside> the hash with |
d0b74c17 |
727 | the "file" parameter, as opposed to being the sole argument to the |
ffed8b01 |
728 | constructor. This is required if any options are specified. |
729 | See L<OPTIONS> below for the complete list. |
730 | |
ffed8b01 |
731 | You can also start with an array instead of a hash. For this, you must |
732 | specify the C<type> parameter: |
733 | |
a8fdabda |
734 | my $db = DBM::Deep->new( |
735 | file => "foo.db", |
736 | type => DBM::Deep->TYPE_ARRAY |
737 | ); |
ffed8b01 |
738 | |
739 | B<Note:> Specifing the C<type> parameter only takes effect when beginning |
740 | a new DB file. If you create a DBM::Deep object with an existing file, the |
90f93b43 |
741 | C<type> will be loaded from the file header, and an error will be thrown if |
742 | the wrong type is passed in. |
ffed8b01 |
743 | |
744 | =head2 TIE CONSTRUCTION |
745 | |
90f93b43 |
746 | Alternately, you can create a DBM::Deep handle by using Perl's built-in |
747 | tie() function. The object returned from tie() can be used to call methods, |
eff6a245 |
748 | such as lock() and unlock(). (That object can be retrieved from the tied |
749 | variable at any time using tied() - please see L<perltie/> for more info. |
ffed8b01 |
750 | |
a8fdabda |
751 | my %hash; |
752 | my $db = tie %hash, "DBM::Deep", "foo.db"; |
d0b74c17 |
753 | |
a8fdabda |
754 | my @array; |
755 | my $db = tie @array, "DBM::Deep", "bar.db"; |
ffed8b01 |
756 | |
757 | As with the OO constructor, you can replace the DB filename parameter with |
758 | a hash containing one or more options (see L<OPTIONS> just below for the |
759 | complete list). |
760 | |
a8fdabda |
761 | tie %hash, "DBM::Deep", { |
762 | file => "foo.db", |
763 | locking => 1, |
764 | autoflush => 1 |
765 | }; |
ffed8b01 |
766 | |
767 | =head2 OPTIONS |
768 | |
769 | There are a number of options that can be passed in when constructing your |
770 | DBM::Deep objects. These apply to both the OO- and tie- based approaches. |
771 | |
772 | =over |
773 | |
774 | =item * file |
775 | |
776 | Filename of the DB file to link the handle to. You can pass a full absolute |
d0b74c17 |
777 | filesystem path, partial path, or a plain filename if the file is in the |
714618f0 |
778 | current working directory. This is a required parameter (though q.v. fh). |
779 | |
780 | =item * fh |
781 | |
782 | If you want, you can pass in the fh instead of the file. This is most useful for doing |
783 | something like: |
784 | |
785 | my $db = DBM::Deep->new( { fh => \*DATA } ); |
786 | |
787 | You are responsible for making sure that the fh has been opened appropriately for your |
788 | needs. If you open it read-only and attempt to write, an exception will be thrown. If you |
789 | open it write-only or append-only, an exception will be thrown immediately as DBM::Deep |
790 | needs to read from the fh. |
791 | |
eff6a245 |
792 | =item * audit_file / audit_fh |
793 | |
794 | These are just like file/fh, except for auditing. Please see L</AUDITING> for |
795 | more information. |
796 | |
714618f0 |
797 | =item * file_offset |
798 | |
799 | This is the offset within the file that the DBM::Deep db starts. Most of the time, you will |
800 | not need to set this. However, it's there if you want it. |
801 | |
802 | If you pass in fh and do not set this, it will be set appropriately. |
ffed8b01 |
803 | |
ffed8b01 |
804 | =item * type |
805 | |
806 | This parameter specifies what type of object to create, a hash or array. Use |
359a01ac |
807 | one of these two constants: |
808 | |
809 | =over 4 |
810 | |
811 | =item * C<DBM::Deep-E<gt>TYPE_HASH> |
812 | |
813 | =item * C<DBM::Deep-E<gt>TYPE_ARRAY>. |
814 | |
815 | =back |
816 | |
d0b74c17 |
817 | This only takes effect when beginning a new file. This is an optional |
ffed8b01 |
818 | parameter, and defaults to C<DBM::Deep-E<gt>TYPE_HASH>. |
819 | |
820 | =item * locking |
821 | |
eff6a245 |
822 | Specifies whether locking is to be enabled. DBM::Deep uses Perl's flock() |
823 | function to lock the database in exclusive mode for writes, and shared mode |
824 | for reads. Pass any true value to enable. This affects the base DB handle |
825 | I<and any child hashes or arrays> that use the same DB file. This is an |
826 | optional parameter, and defaults to 0 (disabled). See L<LOCKING> below for |
827 | more. |
ffed8b01 |
828 | |
829 | =item * autoflush |
830 | |
d0b74c17 |
831 | Specifies whether autoflush is to be enabled on the underlying filehandle. |
832 | This obviously slows down write operations, but is required if you may have |
833 | multiple processes accessing the same DB file (also consider enable I<locking>). |
834 | Pass any true value to enable. This is an optional parameter, and defaults to 0 |
ffed8b01 |
835 | (disabled). |
836 | |
837 | =item * autobless |
838 | |
359a01ac |
839 | If I<autobless> mode is enabled, DBM::Deep will preserve the class something |
840 | is blessed into, and restores it when fetched. This is an optional parameter, and defaults to 1 (enabled). |
841 | |
842 | B<Note:> If you use the OO-interface, you will not be able to call any methods |
843 | of DBM::Deep on the blessed item. This is considered to be a feature. |
ffed8b01 |
844 | |
845 | =item * filter_* |
846 | |
359a01ac |
847 | See L</FILTERS> below. |
ffed8b01 |
848 | |
ffed8b01 |
849 | =back |
850 | |
851 | =head1 TIE INTERFACE |
852 | |
853 | With DBM::Deep you can access your databases using Perl's standard hash/array |
90f93b43 |
854 | syntax. Because all DBM::Deep objects are I<tied> to hashes or arrays, you can |
855 | treat them as such. DBM::Deep will intercept all reads/writes and direct them |
856 | to the right place -- the DB file. This has nothing to do with the |
857 | L<TIE CONSTRUCTION> section above. This simply tells you how to use DBM::Deep |
858 | using regular hashes and arrays, rather than calling functions like C<get()> |
859 | and C<put()> (although those work too). It is entirely up to you how to want |
860 | to access your databases. |
ffed8b01 |
861 | |
862 | =head2 HASHES |
863 | |
864 | You can treat any DBM::Deep object like a normal Perl hash reference. Add keys, |
865 | or even nested hashes (or arrays) using standard Perl syntax: |
866 | |
a8fdabda |
867 | my $db = DBM::Deep->new( "foo.db" ); |
d0b74c17 |
868 | |
a8fdabda |
869 | $db->{mykey} = "myvalue"; |
870 | $db->{myhash} = {}; |
871 | $db->{myhash}->{subkey} = "subvalue"; |
ffed8b01 |
872 | |
a8fdabda |
873 | print $db->{myhash}->{subkey} . "\n"; |
ffed8b01 |
874 | |
875 | You can even step through hash keys using the normal Perl C<keys()> function: |
876 | |
a8fdabda |
877 | foreach my $key (keys %$db) { |
878 | print "$key: " . $db->{$key} . "\n"; |
879 | } |
ffed8b01 |
880 | |
881 | Remember that Perl's C<keys()> function extracts I<every> key from the hash and |
d0b74c17 |
882 | pushes them onto an array, all before the loop even begins. If you have an |
eff6a245 |
883 | extremely large hash, this may exhaust Perl's memory. Instead, consider using |
d0b74c17 |
884 | Perl's C<each()> function, which pulls keys/values one at a time, using very |
ffed8b01 |
885 | little memory: |
886 | |
a8fdabda |
887 | while (my ($key, $value) = each %$db) { |
888 | print "$key: $value\n"; |
889 | } |
ffed8b01 |
890 | |
891 | Please note that when using C<each()>, you should always pass a direct |
892 | hash reference, not a lookup. Meaning, you should B<never> do this: |
893 | |
a8fdabda |
894 | # NEVER DO THIS |
895 | while (my ($key, $value) = each %{$db->{foo}}) { # BAD |
ffed8b01 |
896 | |
897 | This causes an infinite loop, because for each iteration, Perl is calling |
898 | FETCH() on the $db handle, resulting in a "new" hash for foo every time, so |
d0b74c17 |
899 | it effectively keeps returning the first key over and over again. Instead, |
ffed8b01 |
900 | assign a temporary variable to C<$db->{foo}>, then pass that to each(). |
901 | |
902 | =head2 ARRAYS |
903 | |
904 | As with hashes, you can treat any DBM::Deep object like a normal Perl array |
d0b74c17 |
905 | reference. This includes inserting, removing and manipulating elements, |
ffed8b01 |
906 | and the C<push()>, C<pop()>, C<shift()>, C<unshift()> and C<splice()> functions. |
d0b74c17 |
907 | The object must have first been created using type C<DBM::Deep-E<gt>TYPE_ARRAY>, |
ffed8b01 |
908 | or simply be a nested array reference inside a hash. Example: |
909 | |
a8fdabda |
910 | my $db = DBM::Deep->new( |
911 | file => "foo-array.db", |
912 | type => DBM::Deep->TYPE_ARRAY |
913 | ); |
d0b74c17 |
914 | |
a8fdabda |
915 | $db->[0] = "foo"; |
916 | push @$db, "bar", "baz"; |
917 | unshift @$db, "bah"; |
d0b74c17 |
918 | |
a8fdabda |
919 | my $last_elem = pop @$db; # baz |
920 | my $first_elem = shift @$db; # bah |
921 | my $second_elem = $db->[1]; # bar |
d0b74c17 |
922 | |
a8fdabda |
923 | my $num_elements = scalar @$db; |
ffed8b01 |
924 | |
925 | =head1 OO INTERFACE |
926 | |
927 | In addition to the I<tie()> interface, you can also use a standard OO interface |
928 | to manipulate all aspects of DBM::Deep databases. Each type of object (hash or |
d0b74c17 |
929 | array) has its own methods, but both types share the following common methods: |
eff6a245 |
930 | C<put()>, C<get()>, C<exists()>, C<delete()> and C<clear()>. C<fetch()> and |
931 | C<store(> are aliases to C<put()> and C<get()>, respectively. |
ffed8b01 |
932 | |
933 | =over |
934 | |
4d35d856 |
935 | =item * new() / clone() |
936 | |
937 | These are the constructor and copy-functions. |
938 | |
90f93b43 |
939 | =item * put() / store() |
ffed8b01 |
940 | |
941 | Stores a new hash key/value pair, or sets an array element value. Takes two |
942 | arguments, the hash key or array index, and the new value. The value can be |
943 | a scalar, hash ref or array ref. Returns true on success, false on failure. |
944 | |
a8fdabda |
945 | $db->put("foo", "bar"); # for hashes |
946 | $db->put(1, "bar"); # for arrays |
ffed8b01 |
947 | |
90f93b43 |
948 | =item * get() / fetch() |
ffed8b01 |
949 | |
950 | Fetches the value of a hash key or array element. Takes one argument: the hash |
d0b74c17 |
951 | key or array index. Returns a scalar, hash ref or array ref, depending on the |
ffed8b01 |
952 | data type stored. |
953 | |
a8fdabda |
954 | my $value = $db->get("foo"); # for hashes |
955 | my $value = $db->get(1); # for arrays |
ffed8b01 |
956 | |
957 | =item * exists() |
958 | |
d0b74c17 |
959 | Checks if a hash key or array index exists. Takes one argument: the hash key |
ffed8b01 |
960 | or array index. Returns true if it exists, false if not. |
961 | |
a8fdabda |
962 | if ($db->exists("foo")) { print "yay!\n"; } # for hashes |
963 | if ($db->exists(1)) { print "yay!\n"; } # for arrays |
ffed8b01 |
964 | |
965 | =item * delete() |
966 | |
967 | Deletes one hash key/value pair or array element. Takes one argument: the hash |
968 | key or array index. Returns true on success, false if not found. For arrays, |
969 | the remaining elements located after the deleted element are NOT moved over. |
970 | The deleted element is essentially just undefined, which is exactly how Perl's |
d0b74c17 |
971 | internal arrays work. Please note that the space occupied by the deleted |
972 | key/value or element is B<not> reused again -- see L<UNUSED SPACE RECOVERY> |
ffed8b01 |
973 | below for details and workarounds. |
974 | |
a8fdabda |
975 | $db->delete("foo"); # for hashes |
976 | $db->delete(1); # for arrays |
ffed8b01 |
977 | |
978 | =item * clear() |
979 | |
d0b74c17 |
980 | Deletes B<all> hash keys or array elements. Takes no arguments. No return |
981 | value. Please note that the space occupied by the deleted keys/values or |
982 | elements is B<not> reused again -- see L<UNUSED SPACE RECOVERY> below for |
ffed8b01 |
983 | details and workarounds. |
984 | |
a8fdabda |
985 | $db->clear(); # hashes or arrays |
ffed8b01 |
986 | |
4d35d856 |
987 | =item * lock() / unlock() |
988 | |
989 | q.v. Locking. |
990 | |
991 | =item * optimize() |
992 | |
eff6a245 |
993 | Recover lost disk space. This is important to do, especially if you use |
994 | transactions. |
4d35d856 |
995 | |
996 | =item * import() / export() |
997 | |
998 | Data going in and out. |
999 | |
ffed8b01 |
1000 | =back |
1001 | |
1002 | =head2 HASHES |
1003 | |
d0b74c17 |
1004 | For hashes, DBM::Deep supports all the common methods described above, and the |
ffed8b01 |
1005 | following additional methods: C<first_key()> and C<next_key()>. |
1006 | |
1007 | =over |
1008 | |
1009 | =item * first_key() |
1010 | |
d0b74c17 |
1011 | Returns the "first" key in the hash. As with built-in Perl hashes, keys are |
1012 | fetched in an undefined order (which appears random). Takes no arguments, |
ffed8b01 |
1013 | returns the key as a scalar value. |
1014 | |
a8fdabda |
1015 | my $key = $db->first_key(); |
ffed8b01 |
1016 | |
1017 | =item * next_key() |
1018 | |
1019 | Returns the "next" key in the hash, given the previous one as the sole argument. |
1020 | Returns undef if there are no more keys to be fetched. |
1021 | |
a8fdabda |
1022 | $key = $db->next_key($key); |
ffed8b01 |
1023 | |
1024 | =back |
1025 | |
1026 | Here are some examples of using hashes: |
1027 | |
a8fdabda |
1028 | my $db = DBM::Deep->new( "foo.db" ); |
d0b74c17 |
1029 | |
a8fdabda |
1030 | $db->put("foo", "bar"); |
1031 | print "foo: " . $db->get("foo") . "\n"; |
d0b74c17 |
1032 | |
a8fdabda |
1033 | $db->put("baz", {}); # new child hash ref |
1034 | $db->get("baz")->put("buz", "biz"); |
1035 | print "buz: " . $db->get("baz")->get("buz") . "\n"; |
d0b74c17 |
1036 | |
a8fdabda |
1037 | my $key = $db->first_key(); |
1038 | while ($key) { |
1039 | print "$key: " . $db->get($key) . "\n"; |
1040 | $key = $db->next_key($key); |
1041 | } |
d0b74c17 |
1042 | |
a8fdabda |
1043 | if ($db->exists("foo")) { $db->delete("foo"); } |
ffed8b01 |
1044 | |
1045 | =head2 ARRAYS |
1046 | |
d0b74c17 |
1047 | For arrays, DBM::Deep supports all the common methods described above, and the |
1048 | following additional methods: C<length()>, C<push()>, C<pop()>, C<shift()>, |
ffed8b01 |
1049 | C<unshift()> and C<splice()>. |
1050 | |
1051 | =over |
1052 | |
1053 | =item * length() |
1054 | |
1055 | Returns the number of elements in the array. Takes no arguments. |
1056 | |
a8fdabda |
1057 | my $len = $db->length(); |
ffed8b01 |
1058 | |
1059 | =item * push() |
1060 | |
d0b74c17 |
1061 | Adds one or more elements onto the end of the array. Accepts scalars, hash |
ffed8b01 |
1062 | refs or array refs. No return value. |
1063 | |
a8fdabda |
1064 | $db->push("foo", "bar", {}); |
ffed8b01 |
1065 | |
1066 | =item * pop() |
1067 | |
1068 | Fetches the last element in the array, and deletes it. Takes no arguments. |
1069 | Returns undef if array is empty. Returns the element value. |
1070 | |
a8fdabda |
1071 | my $elem = $db->pop(); |
ffed8b01 |
1072 | |
1073 | =item * shift() |
1074 | |
d0b74c17 |
1075 | Fetches the first element in the array, deletes it, then shifts all the |
1076 | remaining elements over to take up the space. Returns the element value. This |
1077 | method is not recommended with large arrays -- see L<LARGE ARRAYS> below for |
ffed8b01 |
1078 | details. |
1079 | |
a8fdabda |
1080 | my $elem = $db->shift(); |
ffed8b01 |
1081 | |
1082 | =item * unshift() |
1083 | |
d0b74c17 |
1084 | Inserts one or more elements onto the beginning of the array, shifting all |
1085 | existing elements over to make room. Accepts scalars, hash refs or array refs. |
1086 | No return value. This method is not recommended with large arrays -- see |
ffed8b01 |
1087 | <LARGE ARRAYS> below for details. |
1088 | |
a8fdabda |
1089 | $db->unshift("foo", "bar", {}); |
ffed8b01 |
1090 | |
1091 | =item * splice() |
1092 | |
d0b74c17 |
1093 | Performs exactly like Perl's built-in function of the same name. See L<perldoc |
ffed8b01 |
1094 | -f splice> for usage -- it is too complicated to document here. This method is |
1095 | not recommended with large arrays -- see L<LARGE ARRAYS> below for details. |
1096 | |
1097 | =back |
1098 | |
1099 | Here are some examples of using arrays: |
1100 | |
a8fdabda |
1101 | my $db = DBM::Deep->new( |
1102 | file => "foo.db", |
1103 | type => DBM::Deep->TYPE_ARRAY |
1104 | ); |
d0b74c17 |
1105 | |
a8fdabda |
1106 | $db->push("bar", "baz"); |
1107 | $db->unshift("foo"); |
1108 | $db->put(3, "buz"); |
d0b74c17 |
1109 | |
a8fdabda |
1110 | my $len = $db->length(); |
1111 | print "length: $len\n"; # 4 |
d0b74c17 |
1112 | |
a8fdabda |
1113 | for (my $k=0; $k<$len; $k++) { |
1114 | print "$k: " . $db->get($k) . "\n"; |
1115 | } |
d0b74c17 |
1116 | |
a8fdabda |
1117 | $db->splice(1, 2, "biz", "baf"); |
d0b74c17 |
1118 | |
a8fdabda |
1119 | while (my $elem = shift @$db) { |
1120 | print "shifted: $elem\n"; |
1121 | } |
ffed8b01 |
1122 | |
1123 | =head1 LOCKING |
1124 | |
d0b74c17 |
1125 | Enable automatic file locking by passing a true value to the C<locking> |
ffed8b01 |
1126 | parameter when constructing your DBM::Deep object (see L<SETUP> above). |
1127 | |
a8fdabda |
1128 | my $db = DBM::Deep->new( |
1129 | file => "foo.db", |
1130 | locking => 1 |
1131 | ); |
ffed8b01 |
1132 | |
d0b74c17 |
1133 | This causes DBM::Deep to C<flock()> the underlying filehandle with exclusive |
1134 | mode for writes, and shared mode for reads. This is required if you have |
1135 | multiple processes accessing the same database file, to avoid file corruption. |
1136 | Please note that C<flock()> does NOT work for files over NFS. See L<DB OVER |
ffed8b01 |
1137 | NFS> below for more. |
1138 | |
1139 | =head2 EXPLICIT LOCKING |
1140 | |
d0b74c17 |
1141 | You can explicitly lock a database, so it remains locked for multiple |
1142 | transactions. This is done by calling the C<lock()> method, and passing an |
90f93b43 |
1143 | optional lock mode argument (defaults to exclusive mode). This is particularly |
d0b74c17 |
1144 | useful for things like counters, where the current value needs to be fetched, |
ffed8b01 |
1145 | then incremented, then stored again. |
1146 | |
a8fdabda |
1147 | $db->lock(); |
1148 | my $counter = $db->get("counter"); |
1149 | $counter++; |
1150 | $db->put("counter", $counter); |
1151 | $db->unlock(); |
d0b74c17 |
1152 | |
a8fdabda |
1153 | # or... |
ffed8b01 |
1154 | |
a8fdabda |
1155 | $db->lock(); |
1156 | $db->{counter}++; |
1157 | $db->unlock(); |
ffed8b01 |
1158 | |
1159 | You can pass C<lock()> an optional argument, which specifies which mode to use |
68f943b3 |
1160 | (exclusive or shared). Use one of these two constants: |
1161 | C<DBM::Deep-E<gt>LOCK_EX> or C<DBM::Deep-E<gt>LOCK_SH>. These are passed |
1162 | directly to C<flock()>, and are the same as the constants defined in Perl's |
1163 | L<Fcntl/> module. |
ffed8b01 |
1164 | |
a8fdabda |
1165 | $db->lock( $db->LOCK_SH ); |
1166 | # something here |
1167 | $db->unlock(); |
ffed8b01 |
1168 | |
ffed8b01 |
1169 | =head1 IMPORTING/EXPORTING |
1170 | |
1171 | You can import existing complex structures by calling the C<import()> method, |
1172 | and export an entire database into an in-memory structure using the C<export()> |
1173 | method. Both are examined here. |
1174 | |
1175 | =head2 IMPORTING |
1176 | |
1177 | Say you have an existing hash with nested hashes/arrays inside it. Instead of |
d0b74c17 |
1178 | walking the structure and adding keys/elements to the database as you go, |
1179 | simply pass a reference to the C<import()> method. This recursively adds |
ffed8b01 |
1180 | everything to an existing DBM::Deep object for you. Here is an example: |
1181 | |
a8fdabda |
1182 | my $struct = { |
1183 | key1 => "value1", |
1184 | key2 => "value2", |
1185 | array1 => [ "elem0", "elem1", "elem2" ], |
1186 | hash1 => { |
1187 | subkey1 => "subvalue1", |
1188 | subkey2 => "subvalue2" |
1189 | } |
1190 | }; |
d0b74c17 |
1191 | |
a8fdabda |
1192 | my $db = DBM::Deep->new( "foo.db" ); |
1193 | $db->import( $struct ); |
d0b74c17 |
1194 | |
a8fdabda |
1195 | print $db->{key1} . "\n"; # prints "value1" |
d0b74c17 |
1196 | |
1197 | This recursively imports the entire C<$struct> object into C<$db>, including |
ffed8b01 |
1198 | all nested hashes and arrays. If the DBM::Deep object contains exsiting data, |
d0b74c17 |
1199 | keys are merged with the existing ones, replacing if they already exist. |
1200 | The C<import()> method can be called on any database level (not just the base |
ffed8b01 |
1201 | level), and works with both hash and array DB types. |
1202 | |
ffed8b01 |
1203 | B<Note:> Make sure your existing structure has no circular references in it. |
eff6a245 |
1204 | These will cause an infinite loop when importing. There are plans to fix this |
1205 | in a later release. |
ffed8b01 |
1206 | |
1207 | =head2 EXPORTING |
1208 | |
d0b74c17 |
1209 | Calling the C<export()> method on an existing DBM::Deep object will return |
1210 | a reference to a new in-memory copy of the database. The export is done |
ffed8b01 |
1211 | recursively, so all nested hashes/arrays are all exported to standard Perl |
1212 | objects. Here is an example: |
1213 | |
a8fdabda |
1214 | my $db = DBM::Deep->new( "foo.db" ); |
d0b74c17 |
1215 | |
a8fdabda |
1216 | $db->{key1} = "value1"; |
1217 | $db->{key2} = "value2"; |
1218 | $db->{hash1} = {}; |
1219 | $db->{hash1}->{subkey1} = "subvalue1"; |
1220 | $db->{hash1}->{subkey2} = "subvalue2"; |
d0b74c17 |
1221 | |
a8fdabda |
1222 | my $struct = $db->export(); |
d0b74c17 |
1223 | |
a8fdabda |
1224 | print $struct->{key1} . "\n"; # prints "value1" |
ffed8b01 |
1225 | |
1226 | This makes a complete copy of the database in memory, and returns a reference |
d0b74c17 |
1227 | to it. The C<export()> method can be called on any database level (not just |
1228 | the base level), and works with both hash and array DB types. Be careful of |
1229 | large databases -- you can store a lot more data in a DBM::Deep object than an |
ffed8b01 |
1230 | in-memory Perl structure. |
1231 | |
ffed8b01 |
1232 | B<Note:> Make sure your database has no circular references in it. |
eff6a245 |
1233 | These will cause an infinite loop when exporting. There are plans to fix this |
1234 | in a later release. |
ffed8b01 |
1235 | |
1236 | =head1 FILTERS |
1237 | |
1238 | DBM::Deep has a number of hooks where you can specify your own Perl function |
1239 | to perform filtering on incoming or outgoing data. This is a perfect |
1240 | way to extend the engine, and implement things like real-time compression or |
d0b74c17 |
1241 | encryption. Filtering applies to the base DB level, and all child hashes / |
1242 | arrays. Filter hooks can be specified when your DBM::Deep object is first |
1243 | constructed, or by calling the C<set_filter()> method at any time. There are |
ffed8b01 |
1244 | four available filter hooks, described below: |
1245 | |
1246 | =over |
1247 | |
1248 | =item * filter_store_key |
1249 | |
d0b74c17 |
1250 | This filter is called whenever a hash key is stored. It |
ffed8b01 |
1251 | is passed the incoming key, and expected to return a transformed key. |
1252 | |
1253 | =item * filter_store_value |
1254 | |
d0b74c17 |
1255 | This filter is called whenever a hash key or array element is stored. It |
ffed8b01 |
1256 | is passed the incoming value, and expected to return a transformed value. |
1257 | |
1258 | =item * filter_fetch_key |
1259 | |
d0b74c17 |
1260 | This filter is called whenever a hash key is fetched (i.e. via |
ffed8b01 |
1261 | C<first_key()> or C<next_key()>). It is passed the transformed key, |
1262 | and expected to return the plain key. |
1263 | |
1264 | =item * filter_fetch_value |
1265 | |
d0b74c17 |
1266 | This filter is called whenever a hash key or array element is fetched. |
ffed8b01 |
1267 | It is passed the transformed value, and expected to return the plain value. |
1268 | |
1269 | =back |
1270 | |
1271 | Here are the two ways to setup a filter hook: |
1272 | |
a8fdabda |
1273 | my $db = DBM::Deep->new( |
1274 | file => "foo.db", |
1275 | filter_store_value => \&my_filter_store, |
1276 | filter_fetch_value => \&my_filter_fetch |
1277 | ); |
d0b74c17 |
1278 | |
a8fdabda |
1279 | # or... |
d0b74c17 |
1280 | |
a8fdabda |
1281 | $db->set_filter( "filter_store_value", \&my_filter_store ); |
1282 | $db->set_filter( "filter_fetch_value", \&my_filter_fetch ); |
ffed8b01 |
1283 | |
1284 | Your filter function will be called only when dealing with SCALAR keys or |
1285 | values. When nested hashes and arrays are being stored/fetched, filtering |
d0b74c17 |
1286 | is bypassed. Filters are called as static functions, passed a single SCALAR |
ffed8b01 |
1287 | argument, and expected to return a single SCALAR value. If you want to |
1288 | remove a filter, set the function reference to C<undef>: |
1289 | |
a8fdabda |
1290 | $db->set_filter( "filter_store_value", undef ); |
ffed8b01 |
1291 | |
1292 | =head2 REAL-TIME ENCRYPTION EXAMPLE |
1293 | |
d0b74c17 |
1294 | Here is a working example that uses the I<Crypt::Blowfish> module to |
ffed8b01 |
1295 | do real-time encryption / decryption of keys & values with DBM::Deep Filters. |
d0b74c17 |
1296 | Please visit L<http://search.cpan.org/search?module=Crypt::Blowfish> for more |
ffed8b01 |
1297 | on I<Crypt::Blowfish>. You'll also need the I<Crypt::CBC> module. |
1298 | |
a8fdabda |
1299 | use DBM::Deep; |
1300 | use Crypt::Blowfish; |
1301 | use Crypt::CBC; |
1302 | |
1303 | my $cipher = Crypt::CBC->new({ |
1304 | 'key' => 'my secret key', |
1305 | 'cipher' => 'Blowfish', |
1306 | 'iv' => '$KJh#(}q', |
1307 | 'regenerate_key' => 0, |
1308 | 'padding' => 'space', |
1309 | 'prepend_iv' => 0 |
1310 | }); |
1311 | |
1312 | my $db = DBM::Deep->new( |
1313 | file => "foo-encrypt.db", |
1314 | filter_store_key => \&my_encrypt, |
1315 | filter_store_value => \&my_encrypt, |
1316 | filter_fetch_key => \&my_decrypt, |
1317 | filter_fetch_value => \&my_decrypt, |
1318 | ); |
1319 | |
1320 | $db->{key1} = "value1"; |
1321 | $db->{key2} = "value2"; |
1322 | print "key1: " . $db->{key1} . "\n"; |
1323 | print "key2: " . $db->{key2} . "\n"; |
1324 | |
1325 | undef $db; |
1326 | exit; |
1327 | |
1328 | sub my_encrypt { |
1329 | return $cipher->encrypt( $_[0] ); |
1330 | } |
1331 | sub my_decrypt { |
1332 | return $cipher->decrypt( $_[0] ); |
1333 | } |
ffed8b01 |
1334 | |
1335 | =head2 REAL-TIME COMPRESSION EXAMPLE |
1336 | |
1337 | Here is a working example that uses the I<Compress::Zlib> module to do real-time |
1338 | compression / decompression of keys & values with DBM::Deep Filters. |
d0b74c17 |
1339 | Please visit L<http://search.cpan.org/search?module=Compress::Zlib> for |
ffed8b01 |
1340 | more on I<Compress::Zlib>. |
1341 | |
a8fdabda |
1342 | use DBM::Deep; |
1343 | use Compress::Zlib; |
1344 | |
1345 | my $db = DBM::Deep->new( |
1346 | file => "foo-compress.db", |
1347 | filter_store_key => \&my_compress, |
1348 | filter_store_value => \&my_compress, |
1349 | filter_fetch_key => \&my_decompress, |
1350 | filter_fetch_value => \&my_decompress, |
1351 | ); |
1352 | |
1353 | $db->{key1} = "value1"; |
1354 | $db->{key2} = "value2"; |
1355 | print "key1: " . $db->{key1} . "\n"; |
1356 | print "key2: " . $db->{key2} . "\n"; |
1357 | |
1358 | undef $db; |
1359 | exit; |
1360 | |
1361 | sub my_compress { |
1362 | return Compress::Zlib::memGzip( $_[0] ) ; |
1363 | } |
1364 | sub my_decompress { |
1365 | return Compress::Zlib::memGunzip( $_[0] ) ; |
1366 | } |
ffed8b01 |
1367 | |
1368 | B<Note:> Filtering of keys only applies to hashes. Array "keys" are |
1369 | actually numerical index numbers, and are not filtered. |
1370 | |
1371 | =head1 ERROR HANDLING |
1372 | |
1373 | Most DBM::Deep methods return a true value for success, and call die() on |
95967a5e |
1374 | failure. You can wrap calls in an eval block to catch the die. |
ffed8b01 |
1375 | |
a8fdabda |
1376 | my $db = DBM::Deep->new( "foo.db" ); # create hash |
1377 | eval { $db->push("foo"); }; # ILLEGAL -- push is array-only call |
d0b74c17 |
1378 | |
a8fdabda |
1379 | print $@; # prints error message |
429e4192 |
1380 | |
ffed8b01 |
1381 | =head1 LARGEFILE SUPPORT |
1382 | |
1383 | If you have a 64-bit system, and your Perl is compiled with both LARGEFILE |
1384 | and 64-bit support, you I<may> be able to create databases larger than 2 GB. |
1385 | DBM::Deep by default uses 32-bit file offset tags, but these can be changed |
044e6288 |
1386 | by specifying the 'pack_size' parameter when constructing the file. |
ffed8b01 |
1387 | |
a8fdabda |
1388 | DBM::Deep->new( |
1389 | filename => $filename, |
1390 | pack_size => 'large', |
1391 | ); |
ffed8b01 |
1392 | |
d0b74c17 |
1393 | This tells DBM::Deep to pack all file offsets with 8-byte (64-bit) quad words |
1394 | instead of 32-bit longs. After setting these values your DB files have a |
ffed8b01 |
1395 | theoretical maximum size of 16 XB (exabytes). |
1396 | |
044e6288 |
1397 | You can also use C<pack_size =E<gt> 'small'> in order to use 16-bit file |
1398 | offsets. |
1399 | |
ffed8b01 |
1400 | B<Note:> Changing these values will B<NOT> work for existing database files. |
044e6288 |
1401 | Only change this for new files. Once the value has been set, it is stored in |
1402 | the file's header and cannot be changed for the life of the file. These |
1403 | parameters are per-file, meaning you can access 32-bit and 64-bit files, as |
1404 | you chose. |
ffed8b01 |
1405 | |
044e6288 |
1406 | B<Note:> We have not personally tested files larger than 2 GB -- all my |
1407 | systems have only a 32-bit Perl. However, I have received user reports that |
1408 | this does indeed work! |
ffed8b01 |
1409 | |
1410 | =head1 LOW-LEVEL ACCESS |
1411 | |
90f93b43 |
1412 | If you require low-level access to the underlying filehandle that DBM::Deep uses, |
4d35d856 |
1413 | you can call the C<_fh()> method, which returns the handle: |
ffed8b01 |
1414 | |
a8fdabda |
1415 | my $fh = $db->_fh(); |
ffed8b01 |
1416 | |
1417 | This method can be called on the root level of the datbase, or any child |
1418 | hashes or arrays. All levels share a I<root> structure, which contains things |
90f93b43 |
1419 | like the filehandle, a reference counter, and all the options specified |
460b1067 |
1420 | when you created the object. You can get access to this file object by |
83371fe3 |
1421 | calling the C<_storage()> method. |
ffed8b01 |
1422 | |
83371fe3 |
1423 | my $file_obj = $db->_storage(); |
ffed8b01 |
1424 | |
1425 | This is useful for changing options after the object has already been created, |
f5be9b03 |
1426 | such as enabling/disabling locking. You can also store your own temporary user |
1427 | data in this structure (be wary of name collision), which is then accessible from |
1428 | any child hash or array. |
ffed8b01 |
1429 | |
1430 | =head1 CUSTOM DIGEST ALGORITHM |
1431 | |
1432 | DBM::Deep by default uses the I<Message Digest 5> (MD5) algorithm for hashing |
1433 | keys. However you can override this, and use another algorithm (such as SHA-256) |
d0b74c17 |
1434 | or even write your own. But please note that DBM::Deep currently expects zero |
044e6288 |
1435 | collisions, so your algorithm has to be I<perfect>, so to speak. Collision |
1436 | detection may be introduced in a later version. |
ffed8b01 |
1437 | |
044e6288 |
1438 | You can specify a custom digest algorithm by passing it into the parameter |
1439 | list for new(), passing a reference to a subroutine as the 'digest' parameter, |
1440 | and the length of the algorithm's hashes (in bytes) as the 'hash_size' |
1441 | parameter. Here is a working example that uses a 256-bit hash from the |
d0b74c17 |
1442 | I<Digest::SHA256> module. Please see |
044e6288 |
1443 | L<http://search.cpan.org/search?module=Digest::SHA256> for more information. |
ffed8b01 |
1444 | |
a8fdabda |
1445 | use DBM::Deep; |
1446 | use Digest::SHA256; |
d0b74c17 |
1447 | |
a8fdabda |
1448 | my $context = Digest::SHA256::new(256); |
d0b74c17 |
1449 | |
a8fdabda |
1450 | my $db = DBM::Deep->new( |
1451 | filename => "foo-sha.db", |
1452 | digest => \&my_digest, |
1453 | hash_size => 32, |
1454 | ); |
d0b74c17 |
1455 | |
a8fdabda |
1456 | $db->{key1} = "value1"; |
1457 | $db->{key2} = "value2"; |
1458 | print "key1: " . $db->{key1} . "\n"; |
1459 | print "key2: " . $db->{key2} . "\n"; |
d0b74c17 |
1460 | |
a8fdabda |
1461 | undef $db; |
1462 | exit; |
d0b74c17 |
1463 | |
a8fdabda |
1464 | sub my_digest { |
1465 | return substr( $context->hash($_[0]), 0, 32 ); |
1466 | } |
ffed8b01 |
1467 | |
1468 | B<Note:> Your returned digest strings must be B<EXACTLY> the number |
044e6288 |
1469 | of bytes you specify in the hash_size parameter (in this case 32). |
ffed8b01 |
1470 | |
260a80b4 |
1471 | B<Note:> If you do choose to use a custom digest algorithm, you must set it |
1472 | every time you access this file. Otherwise, the default (MD5) will be used. |
1473 | |
ffed8b01 |
1474 | =head1 CIRCULAR REFERENCES |
1475 | |
1476 | DBM::Deep has B<experimental> support for circular references. Meaning you |
1477 | can have a nested hash key or array element that points to a parent object. |
1478 | This relationship is stored in the DB file, and is preserved between sessions. |
1479 | Here is an example: |
1480 | |
a8fdabda |
1481 | my $db = DBM::Deep->new( "foo.db" ); |
d0b74c17 |
1482 | |
a8fdabda |
1483 | $db->{foo} = "bar"; |
1484 | $db->{circle} = $db; # ref to self |
d0b74c17 |
1485 | |
a8fdabda |
1486 | print $db->{foo} . "\n"; # prints "bar" |
1487 | print $db->{circle}->{foo} . "\n"; # prints "bar" again |
ffed8b01 |
1488 | |
69c94980 |
1489 | B<Note>: Passing the object to a function that recursively walks the |
ffed8b01 |
1490 | object tree (such as I<Data::Dumper> or even the built-in C<optimize()> or |
69c94980 |
1491 | C<export()> methods) will result in an infinite loop. This will be fixed in |
1492 | a future release. |
ffed8b01 |
1493 | |
eff6a245 |
1494 | =head1 AUDITING |
1495 | |
1496 | New in 0.99_01 is the ability to audit your databases actions. By passing in |
1497 | audit_file (or audit_fh) to the constructor, all actions will be logged to |
1498 | that file. The format is one that is suitable for eval'ing against the |
1499 | database to replay the actions. Please see t/33_audit_trail.t for an example |
1500 | of how to do this. |
1501 | |
1502 | =head1 TRANSACTIONS |
1503 | |
1504 | New in 0.99_01 is ACID transactions. Every DBM::Deep object is completely |
1505 | transaction-ready - it is not an option you have to turn on. Three new methods |
1506 | have been added to support them. They are: |
1507 | |
1508 | =over 4 |
1509 | |
1510 | =item * begin_work() |
1511 | |
1512 | This starts a transaction. |
1513 | |
1514 | =item * commit() |
1515 | |
1516 | This applies the changes done within the transaction to the mainline and ends |
1517 | the transaction. |
1518 | |
1519 | =item * rollback() |
1520 | |
1521 | This discards the changes done within the transaction to the mainline and ends |
1522 | the transaction. |
1523 | |
1524 | =back |
1525 | |
1526 | Transactions in DBM::Deep are done using the MVCC method, the same method used |
1527 | by the InnoDB MySQL table type. |
1528 | |
ffed8b01 |
1529 | =head1 CAVEATS / ISSUES / BUGS |
1530 | |
1531 | This section describes all the known issues with DBM::Deep. It you have found |
1532 | something that is not listed here, please send e-mail to L<jhuckaby@cpan.org>. |
1533 | |
1534 | =head2 UNUSED SPACE RECOVERY |
1535 | |
14a3acb6 |
1536 | One major caveat with DBM::Deep is that space occupied by existing keys and |
ffed8b01 |
1537 | values is not recovered when they are deleted. Meaning if you keep deleting |
1538 | and adding new keys, your file will continuously grow. I am working on this, |
d0b74c17 |
1539 | but in the meantime you can call the built-in C<optimize()> method from time to |
ffed8b01 |
1540 | time (perhaps in a crontab or something) to recover all your unused space. |
1541 | |
a8fdabda |
1542 | $db->optimize(); # returns true on success |
ffed8b01 |
1543 | |
1544 | This rebuilds the ENTIRE database into a new file, then moves it on top of |
1545 | the original. The new file will have no unused space, thus it will take up as |
d0b74c17 |
1546 | little disk space as possible. Please note that this operation can take |
1547 | a long time for large files, and you need enough disk space to temporarily hold |
1548 | 2 copies of your DB file. The temporary file is created in the same directory |
1549 | as the original, named with a ".tmp" extension, and is deleted when the |
1550 | operation completes. Oh, and if locking is enabled, the DB is automatically |
ffed8b01 |
1551 | locked for the entire duration of the copy. |
1552 | |
d0b74c17 |
1553 | B<WARNING:> Only call optimize() on the top-level node of the database, and |
1554 | make sure there are no child references lying around. DBM::Deep keeps a reference |
ffed8b01 |
1555 | counter, and if it is greater than 1, optimize() will abort and return undef. |
1556 | |
eea0d863 |
1557 | =head2 REFERENCES |
1558 | |
1559 | (The reasons given assume a high level of Perl understanding, specifically of |
1560 | references. You can safely skip this section.) |
1561 | |
1562 | Currently, the only references supported are HASH and ARRAY. The other reference |
1563 | types (SCALAR, CODE, GLOB, and REF) cannot be supported for various reasons. |
1564 | |
1565 | =over 4 |
1566 | |
1567 | =item * GLOB |
1568 | |
1569 | These are things like filehandles and other sockets. They can't be supported |
1570 | because it's completely unclear how DBM::Deep should serialize them. |
1571 | |
1572 | =item * SCALAR / REF |
1573 | |
1574 | The discussion here refers to the following type of example: |
1575 | |
1576 | my $x = 25; |
1577 | $db->{key1} = \$x; |
1578 | |
1579 | $x = 50; |
1580 | |
1581 | # In some other process ... |
1582 | |
1583 | my $val = ${ $db->{key1} }; |
1584 | |
1585 | is( $val, 50, "What actually gets stored in the DB file?" ); |
1586 | |
1587 | The problem is one of synchronization. When the variable being referred to |
1588 | changes value, the reference isn't notified. This means that the new value won't |
1589 | be stored in the datafile for other processes to read. There is no TIEREF. |
1590 | |
1591 | It is theoretically possible to store references to values already within a |
1592 | DBM::Deep object because everything already is synchronized, but the change to |
1593 | the internals would be quite large. Specifically, DBM::Deep would have to tie |
1594 | every single value that is stored. This would bloat the RAM footprint of |
1595 | DBM::Deep at least twofold (if not more) and be a significant performance drain, |
1596 | all to support a feature that has never been requested. |
1597 | |
1598 | =item * CODE |
1599 | |
1990c72d |
1600 | L<Data::Dump::Streamer/> provides a mechanism for serializing coderefs, |
1601 | including saving off all closure state. However, just as for SCALAR and REF, |
1602 | that closure state may change without notifying the DBM::Deep object storing |
1603 | the reference. |
eea0d863 |
1604 | |
1605 | =back |
1606 | |
ffed8b01 |
1607 | =head2 FILE CORRUPTION |
1608 | |
14a3acb6 |
1609 | The current level of error handling in DBM::Deep is minimal. Files I<are> checked |
1610 | for a 32-bit signature when opened, but other corruption in files can cause |
1611 | segmentation faults. DBM::Deep may try to seek() past the end of a file, or get |
ffed8b01 |
1612 | stuck in an infinite loop depending on the level of corruption. File write |
1613 | operations are not checked for failure (for speed), so if you happen to run |
d0b74c17 |
1614 | out of disk space, DBM::Deep will probably fail in a bad way. These things will |
ffed8b01 |
1615 | be addressed in a later version of DBM::Deep. |
1616 | |
1617 | =head2 DB OVER NFS |
1618 | |
d8db2929 |
1619 | Beware of using DBM::Deep files over NFS. DBM::Deep uses flock(), which works |
1620 | well on local filesystems, but will NOT protect you from file corruption over |
1621 | NFS. I've heard about setting up your NFS server with a locking daemon, then |
1622 | using lockf() to lock your files, but your mileage may vary there as well. |
1623 | From what I understand, there is no real way to do it. However, if you need |
1624 | access to the underlying filehandle in DBM::Deep for using some other kind of |
1625 | locking scheme like lockf(), see the L<LOW-LEVEL ACCESS> section above. |
ffed8b01 |
1626 | |
1627 | =head2 COPYING OBJECTS |
1628 | |
d0b74c17 |
1629 | Beware of copying tied objects in Perl. Very strange things can happen. |
1630 | Instead, use DBM::Deep's C<clone()> method which safely copies the object and |
ffed8b01 |
1631 | returns a new, blessed, tied hash or array to the same level in the DB. |
1632 | |
a8fdabda |
1633 | my $copy = $db->clone(); |
ffed8b01 |
1634 | |
90f93b43 |
1635 | B<Note>: Since clone() here is cloning the object, not the database location, any |
d8db2929 |
1636 | modifications to either $db or $copy will be visible to both. |
90f93b43 |
1637 | |
ffed8b01 |
1638 | =head2 LARGE ARRAYS |
1639 | |
1640 | Beware of using C<shift()>, C<unshift()> or C<splice()> with large arrays. |
1641 | These functions cause every element in the array to move, which can be murder |
1642 | on DBM::Deep, as every element has to be fetched from disk, then stored again in |
90f93b43 |
1643 | a different location. This will be addressed in the forthcoming version 1.00. |
ffed8b01 |
1644 | |
9be51a89 |
1645 | =head2 WRITEONLY FILES |
1646 | |
1647 | If you pass in a filehandle to new(), you may have opened it in either a readonly or |
1648 | writeonly mode. STORE will verify that the filehandle is writable. However, there |
1649 | doesn't seem to be a good way to determine if a filehandle is readable. And, if the |
1650 | filehandle isn't readable, it's not clear what will happen. So, don't do that. |
1651 | |
261d1296 |
1652 | =head1 CODE COVERAGE |
1653 | |
eff6a245 |
1654 | B<Devel::Cover> is used to test the code coverage of the tests. Below is the |
1655 | B<Devel::Cover> report on this distribution's test suite. |
7910cf68 |
1656 | |
eff6a245 |
1657 | ---------------------------- ------ ------ ------ ------ ------ ------ ------ |
1658 | File stmt bran cond sub pod time total |
1659 | ---------------------------- ------ ------ ------ ------ ------ ------ ------ |
1660 | blib/lib/DBM/Deep.pm 96.2 89.0 75.0 95.8 89.5 36.0 92.9 |
1661 | blib/lib/DBM/Deep/Array.pm 96.1 88.3 100.0 96.4 100.0 15.9 94.7 |
1662 | blib/lib/DBM/Deep/Engine.pm 96.6 86.6 89.5 100.0 0.0 20.0 91.0 |
1663 | blib/lib/DBM/Deep/File.pm 99.4 88.3 55.6 100.0 0.0 19.6 89.5 |
1664 | blib/lib/DBM/Deep/Hash.pm 98.5 83.3 100.0 100.0 100.0 8.5 96.3 |
1665 | Total 96.9 87.4 81.2 98.0 38.5 100.0 92.1 |
1666 | ---------------------------- ------ ------ ------ ------ ------ ------ ------ |
37c5bcf0 |
1667 | |
1668 | =head1 MORE INFORMATION |
1669 | |
1670 | Check out the DBM::Deep Google Group at L<http://groups.google.com/group/DBM-Deep> |
eff6a245 |
1671 | or send email to L<DBM-Deep@googlegroups.com>. You can also visit #dbm-deep on |
1672 | irc.perl.org |
ffed8b01 |
1673 | |
d8db2929 |
1674 | The source code repository is at L<http://svn.perl.org/modules/DBM-Deep> |
1675 | |
eff6a245 |
1676 | =head1 MAINTAINERS |
37c5bcf0 |
1677 | |
aeeb5497 |
1678 | Rob Kinyon, L<rkinyon@cpan.org> |
ffed8b01 |
1679 | |
eff6a245 |
1680 | Originally written by Joseph Huckaby, L<jhuckaby@cpan.org> |
1681 | |
ffed8b01 |
1682 | Special thanks to Adam Sah and Rich Gaushell! You know why :-) |
1683 | |
1684 | =head1 SEE ALSO |
1685 | |
1686 | perltie(1), Tie::Hash(3), Digest::MD5(3), Fcntl(3), flock(2), lockf(3), nfs(5), |
1687 | Digest::SHA256(3), Crypt::Blowfish(3), Compress::Zlib(3) |
1688 | |
1689 | =head1 LICENSE |
1690 | |
aeeb5497 |
1691 | Copyright (c) 2002-2006 Joseph Huckaby. All Rights Reserved. |
ffed8b01 |
1692 | This is free software, you may use it and distribute it under the |
1693 | same terms as Perl itself. |
1694 | |
1695 | =cut |