/usr/share/perl5/DBD/Mock.pm is in libdbd-mock-perl 1.45-2.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 | package DBD::Mock;
# --------------------------------------------------------------------------- #
# Copyright (c) 2004-2007 Stevan Little, Chris Winters
# (spawned from original code Copyright (c) 1994 Tim Bunce)
# --------------------------------------------------------------------------- #
# You may distribute under the terms of either the GNU General Public
# License or the Artistic License, as specified in the Perl README file.
# --------------------------------------------------------------------------- #
use 5.008001;
use strict;
use warnings;
use DBI;
use DBD::Mock::dr;
use DBD::Mock::db;
use DBD::Mock::st;
use DBD::Mock::StatementTrack;
use DBD::Mock::StatementTrack::Iterator;
use DBD::Mock::Session;
use DBD::Mock::Pool;
use DBD::Mock::Pool::db;
sub import {
shift;
$DBI::connect_via = "DBD::Mock::Pool::connect"
if ( @_ && lc( $_[0] ) eq "pool" );
}
our $VERSION = '1.45';
our $drh = undef; # will hold driver handle
our $err = 0; # will hold any error codes
our $errstr = ''; # will hold any error messages
sub driver {
return $drh if defined $drh;
my ( $class, $attributes ) = @_;
$attributes = {}
unless ( defined($attributes) && ( ref($attributes) eq 'HASH' ) );
$drh = DBI::_new_drh(
"${class}::dr",
{
Name => 'Mock',
Version => $DBD::Mock::VERSION,
Attribution =>
'DBD Mock driver by Chris Winters & Stevan Little (orig. from Tim Bunce)',
Err => \$DBD::Mock::err,
Errstr => \$DBD::Mock::errstr,
# mock attributes
mock_connect_fail => 0,
# and pass in any extra attributes given
%{$attributes}
}
);
return $drh;
}
sub CLONE { undef $drh }
# NOTE:
# this feature is still quite experimental. It is defaulted to
# be off, but it can be turned on by doing this:
# $DBD::Mock::AttributeAliasing++;
# and then turned off by doing:
# $DBD::Mock::AttributeAliasing = 0;
# we shall see how this feature works out.
our $AttributeAliasing = 0;
my %AttributeAliases = (
mysql => {
db => {
# aliases can either be a string which is obvious
mysql_insertid => 'mock_last_insert_id'
},
st => {
# but they can also be a subroutine reference whose
# first argument will be either the $dbh or the $sth
# depending upon which context it is aliased in.
mysql_insertid =>
sub { (shift)->{Database}->{'mock_last_insert_id'} }
}
},
);
sub _get_mock_attribute_aliases {
my ($dbname) = @_;
( exists $AttributeAliases{ lc($dbname) } )
|| die "Attribute aliases not available for '$dbname'";
return $AttributeAliases{ lc($dbname) };
}
sub _set_mock_attribute_aliases {
my ( $dbname, $dbh_or_sth, $key, $value ) = @_;
return $AttributeAliases{ lc($dbname) }->{$dbh_or_sth}->{$key} = $value;
}
## Some useful constants
use constant NULL_RESULTSET => [ [] ];
1;
__END__
=head1 NAME
DBD::Mock - Mock database driver for testing
=head1 SYNOPSIS
use DBI;
# connect to your as normal, using 'Mock' as your driver name
my $dbh = DBI->connect( 'DBI:Mock:', '', '' )
|| die "Cannot create handle: $DBI::errstr\n";
# create a statement handle as normal and execute with parameters
my $sth = $dbh->prepare( 'SELECT this, that FROM foo WHERE id = ?' );
$sth->execute( 15 );
# Now query the statement handle as to what has been done with it
my $mock_params = $sth->{mock_params};
print "Used statement: ", $sth->{mock_statement}, "\n",
"Bound parameters: ", join( ', ', @{ $mock_params } ), "\n";
=head1 DESCRIPTION
Testing with databases can be tricky. If you are developing a system married to a single database then you can make some assumptions about your environment and ask the user to provide relevant connection information. But if you need to test a framework that uses DBI, particularly a framework that uses different types of persistence schemes, then it may be more useful to simply verify what the framework is trying to do -- ensure the right SQL is generated and that the correct parameters are bound. C<DBD::Mock> makes it easy to just modify your configuration (presumably held outside your code) and just use it instead of C<DBD::Foo> (like L<DBD::Pg> or L<DBD::mysql>) in your framework.
There is no distinct area where using this module makes sense. (Some people may successfully argue that this is a solution looking for a problem...) Indeed, if you can assume your users have something like L<DBD::AnyData> or L<DBD::SQLite> or if you do not mind creating a dependency on them then it makes far more sense to use these legitimate driver implementations and test your application in the real world -- at least as much of the real world as you can create in your tests...
And if your database handle exists as a package variable or something else easily replaced at test-time then it may make more sense to use L<Test::MockObject> to create a fully dynamic handle. There is an excellent article by chromatic about using L<Test::MockObject> in this and other ways, strongly recommended. (See L<SEE ALSO> for a link)
=head2 How does it work?
C<DBD::Mock> comprises a set of classes used by DBI to implement a database driver. But instead of connecting to a datasource and manipulating data found there it tracks all the calls made to the database handle and any created statement handles. You can then inspect them to ensure what you wanted to happen actually happened. For instance, say you have a configuration file with your database connection information:
[DBI]
dsn = DBI:Pg:dbname=myapp
user = foo
password = bar
And this file is read in at process startup and the handle stored for other procedures to use:
package ObjectDirectory;
my ( $DBH );
sub run_at_startup {
my ( $class, $config ) = @_;
$config ||= read_configuration( ... );
my $dsn = $config->{DBI}{dsn};
my $user = $config->{DBI}{user};
my $pass = $config->{DBI}{password};
$DBH = DBI->connect( $dsn, $user, $pass ) || die ...;
}
sub get_database_handle {
return $DBH;
}
A procedure might use it like this (ignoring any error handling for the moment):
package My::UserActions;
sub fetch_user {
my ( $class, $login ) = @_;
my $dbh = ObjectDirectory->get_database_handle;
my $sql = q{
SELECT login_name, first_name, last_name, creation_date, num_logins
FROM users
WHERE login_name = ?
};
my $sth = $dbh->prepare( $sql );
$sth->execute( $login );
my $row = $sth->fetchrow_arrayref;
return ( $row ) ? User->new( $row ) : undef;
}
So for the purposes of our tests we just want to ensure that:
=over 4
=item 1. The right SQL is being executed
=item 2. The right parameters are bound
=back
Assume whether the SQL actually B<works> or not is irrelevant for this test :-)
To do that our test might look like:
my $config = ObjectDirectory->read_configuration( ... );
$config->{DBI}{dsn} = 'DBI:Mock:';
ObjectDirectory->run_at_startup( $config );
my $login_name = 'foobar';
my $user = My::UserActions->fetch_user( $login_name );
# Get the handle from ObjectDirectory;
# this is the same handle used in the
# 'fetch_user()' procedure above
my $dbh = ObjectDirectory->get_database_handle();
# Ask the database handle for the history
# of all statements executed against it
my $history = $dbh->{mock_all_history};
# Now query that history record to
# see if our expectations match reality
is(scalar(@{$history}), 1, 'Correct number of statements executed' ;
my $login_st = $history->[0];
like($login_st->statement,
qr/SELECT login_name.*FROM users WHERE login_name = ?/sm,
'Correct statement generated' );
my $params = $login_st->bound_params;
is(scalar(@{$params}), 1, 'Correct number of parameters bound');
is($params->[0], $login_name, 'Correct value for parameter 1' );
# Reset the handle for future operations
$dbh->{mock_clear_history} = 1;
The list of properties and what they return is listed below. But in an overall view:
=over 4
=item *
A database handle contains the history of all statements created against it. Other properties set for the handle (e.g., 'PrintError', 'RaiseError') are left alone and can be queried as normal, but they do not affect anything. (A future feature may track the sequence/history of these assignments but if there is no demand it probably will not get implemented.)
=item *
A statement handle contains the statement it was prepared with plus all bound parameters or parameters passed via C<execute()>. It can also contain predefined results for the statement handle to 'fetch', track how many fetches were called and what its current record is.
=back
=head2 A Word of Warning
This may be an incredibly naive implementation of a DBD. But it works for me ...
=head1 DBD::Mock
Since this is a normal DBI statement handle we need to expose our tracking information as properties (accessed like a hash) rather than methods.
=head2 Database Driver Properties
=over 4
=item B<mock_connect_fail>
This is a boolean property which when set to true (C<1>) will not allow DBI to connect. This can be used to simulate a DSN error or authentication failure. This can then be set back to false (C<0>) to resume normal DBI operations. Here is an example of how this works:
# install the DBD::Mock driver
my $drh = DBI->install_driver('Mock');
$drh->{mock_connect_fail} = 1;
# this connection will fail
my $dbh = DBI->connect('dbi:Mock:', '', '') || die "Cannot connect";
# this connection will throw an exception
my $dbh = DBI->connect('dbi:Mock:', '', '', { RaiseError => 1 });
$drh->{mock_connect_fail} = 0;
# this will work now ...
my $dbh = DBI->connect(...);
This feature is conceptually different from the 'mock_can_connect' attribute of the C<$dbh> in that it has a driver-wide scope, where 'mock_can_connect' is handle-wide scope. It also only prevents the initial connection, any C<$dbh> handles created prior to setting 'mock_connect_fail' to true (C<1>) will still go on working just fine.
=item B<mock_data_sources>
This is an ARRAY reference which holds fake data sources which are returned by the Driver and Database Handle's C<data_source()> method.
=item B<mock_add_data_sources>
This takes a string and adds it to the 'mock_data_sources' attribute.
=back
=head2 Database Handle Properties
=over 4
=item B<mock_all_history>
Returns an array reference with all history (a.k.a. C<DBD::Mock::StatementTrack>) objects created against the database handle in the order they were created. Each history object can then report information about the SQL statement used to create it, the bound parameters, etc..
=item B<mock_all_history_iterator>
Returns a C<DBD::Mock::StatementTrack::Iterator> object which will iterate through the current set of C<DBD::Mock::StatementTrack> object in the history. See the B<DBD::Mock::StatementTrack::Iterator> documentation below for more information.
=item B<mock_clear_history>
If set to a true value all previous statement history operations will be erased. This B<includes> the history of currently open handles, so if you do something like:
my $dbh = get_handle( ... );
my $sth = $dbh->prepare( ... );
$dbh->{mock_clear_history} = 1;
$sth->execute( 'Foo' );
You will have no way to learn from the database handle that the statement parameter 'Foo' was bound.
This is useful mainly to ensure you can isolate the statement histories from each other. A typical sequence will look like:
set handle to framework
perform operations
analyze mock database handle
reset mock database handle history
perform more operations
analyze mock database handle
reset mock database handle history
...
=item B<mock_can_connect>
This statement allows you to simulate a downed database connection. This is useful in testing how your application/tests will perform in the face of some kind of catastrophic event such as a network outage or database server failure. It is a simple boolean value which defaults to on, and can be set like this:
# turn the database off
$dbh->{mock_can_connect} = 0;
# turn it back on again
$dbh->{mock_can_connect} = 1;
The statement handle checks this value as well, so something like this
will fail in the expected way:
$dbh = DBI->connect( 'DBI:Mock:', '', '' );
$dbh->{mock_can_connect} = 0;
# blows up!
my $sth = eval { $dbh->prepare( 'SELECT foo FROM bar' ) });
if ( $@ ) {
# Here, $DBI::errstr = 'No connection present'
}
Turning off the database after a statement prepare will fail on the statement C<execute()>, which is hopefully what you would expect:
$dbh = DBI->connect( 'DBI:Mock:', '', '' );
# ok!
my $sth = eval { $dbh->prepare( 'SELECT foo FROM bar' ) });
$dbh->{mock_can_connect} = 0;
# blows up!
$sth->execute;
Similarly:
$dbh = DBI->connect( 'DBI:Mock:', '', '' );
# ok!
my $sth = eval { $dbh->prepare( 'SELECT foo FROM bar' ) });
# ok!
$sth->execute;
$dbh->{mock_can_connect} = 0;
# blows up!
my $row = $sth->fetchrow_arrayref;
Note: The handle attribute C<Active> and the handle method C<ping> will behave according to the value of C<mock_can_connect>. So if C<mock_can_connect> were to be set to 0 (or off), then both C<Active> and C<ping> would return false values (or 0).
=item B<mock_add_resultset( \@resultset | \%sql_and_resultset )>
This stocks the database handle with a record set, allowing you to seed data for your application to see if it works properly.. Each recordset is a simple arrayref of arrays with the first arrayref being the fieldnames used. Every time a statement handle is created it asks the database handle if it has any resultsets available and if so uses it.
Here is a sample usage, partially from the test suite:
my @user_results = (
[ 'login', 'first_name', 'last_name' ],
[ 'cwinters', 'Chris', 'Winters' ],
[ 'bflay', 'Bobby', 'Flay' ],
[ 'alincoln', 'Abe', 'Lincoln' ],
);
my @generic_results = (
[ 'foo', 'bar' ],
[ 'this_one', 'that_one' ],
[ 'this_two', 'that_two' ],
);
my $dbh = DBI->connect( 'DBI:Mock:', '', '' );
$dbh->{mock_add_resultset} = \@user_results; # add first resultset
$dbh->{mock_add_resultset} = \@generic_results; # add second resultset
my ( $sth );
eval {
$sth = $dbh->prepare( 'SELECT login, first_name, last_name FROM foo' );
$sth->execute();
};
# this will fetch rows from the first resultset...
my $row1 = $sth->fetchrow_arrayref;
my $user1 = User->new( login => $row->[0],
first => $row->[1],
last => $row->[2] );
is( $user1->full_name, 'Chris Winters' );
my $row2 = $sth->fetchrow_arrayref;
my $user2 = User->new( login => $row->[0],
first => $row->[1],
last => $row->[2] );
is( $user2->full_name, 'Bobby Flay' );
...
my $sth_generic = $dbh->prepare( 'SELECT foo, bar FROM baz' );
$sth_generic->execute;
# this will fetch rows from the second resultset...
my $row = $sth->fetchrow_arrayref;
You can also associate a resultset with a particular SQL statement instead of adding them in the order they will be fetched:
$dbh->{mock_add_resultset} = {
sql => 'SELECT foo, bar FROM baz',
results => [
[ 'foo', 'bar' ],
[ 'this_one', 'that_one' ],
[ 'this_two', 'that_two' ],
],
};
This will return the given results when the statement 'SELECT foo, bar FROM baz' is prepared. Note that they will be returned B<every time> the statement is prepared, not just the first. It should also be noted that if you want, for some reason, to change the result set bound to a particular SQL statement, all you need to do is add the result set again with the same SQL statement and DBD::Mock will overwrite it.
It should also be noted that the C<rows> method will return the number of records stocked in the result set. So if your code/application makes use of the C<$sth-E<gt>rows> method for things like UPDATE and DELETE calls you should stock the result set like so:
$dbh->{mock_add_resultset} = {
sql => 'UPDATE foo SET baz = 1, bar = 2',
# this will appear to have updated 3 rows
results => [[ 'rows' ], [], [], []],
};
# or ...
$dbh->{mock_add_resultset} = {
sql => 'DELETE FROM foo WHERE bar = 2',
# this will appear to have deleted 1 row
results => [[ 'rows' ], []],
};
Now I admit this is not the most elegant way to go about this, but it works for me for now, and until I can come up with a better method, or someone sends me a patch ;) it will do for now.
If you want a given statement to fail, you will have to use the hashref method and add a 'failure' key. That key can be handed an arrayref with the error number and error string, in that order. It can also be handed a hashref with two keys - errornum and errorstring. If the 'failure' key has no useful value associated with it, the errornum will be '1' and the errorstring will be 'Unknown error'.
=item B<mock_get_info>
This attribute can be used to set up values for get_info(). It takes a hashref of attribute_name/value pairs. See L<DBI> for more information on the information types and their meaning.
=item B<mock_session>
This attribute can be used to set a current DBD::Mock::Session object. For more information on this, see the L<DBD::Mock::Session> docs below. This attribute can also be used to remove the current session from the C<$dbh> simply by setting it to C<undef>.
=item B<mock_last_insert_id>
This attribute is incremented each time an INSERT statement is passed to C<prepare> on a per-handle basis. It's starting value can be set with the 'mock_start_insert_id' attribute (see below).
$dbh->{mock_start_insert_id} = 10;
my $sth = $dbh->prepare('INSERT INTO Foo (foo, bar) VALUES(?, ?)');
$sth->execute(1, 2);
# $dbh->{mock_last_insert_id} == 10
$sth->execute(3, 4);
# $dbh->{mock_last_insert_id} == 11
For more examples, please refer to the test file F<t/025_mock_last_insert_id.t>.
=item B<mock_start_insert_id>
This attribute can be used to set a start value for the 'mock_last_insert_id' attribute. It can also be used to effectively reset the 'mock_last_insert_id' attribute as well.
This attribute also can be used with an ARRAY ref parameter, it's behavior is slightly different in that instead of incrementing the value for every C<prepare> it will only increment for each C<execute>. This allows it to be used over multiple C<execute> calls in a single C<$sth>. It's usage looks like this:
$dbh->{mock_start_insert_id} = [ 'Foo', 10 ];
$dbh->{mock_start_insert_id} = [ 'Baz', 20 ];
my $sth1 = $dbh->prepare('INSERT INTO Foo (foo, bar) VALUES(?, ?)');
my $sth2 = $dbh->prepare('INSERT INTO Baz (baz, buz) VALUES(?, ?)');
$sth1->execute(1, 2);
# $dbh->{mock_last_insert_id} == 10
$sth2->execute(3, 4);
# $dbh->{mock_last_insert_id} == 20
Note that DBD::Mock's matching of table names in 'INSERT' statements is fairly simple, so if your table names are quoted in the insert statement (C<INSERT INTO "Foo">) then you need to quote the name for C<mock_start_insert_id>:
$dbh->{mock_start_insert_id} = [ q{"Foo"}, 10 ];
=item B<mock_add_parser>
DBI provides some simple parsing capabilities for 'SELECT' statements to ensure that placeholders are bound properly. And typically you may simply want to check after the fact that a statement is syntactically correct, or at least what you expect.
But other times you may want to parse the statement as it is prepared rather than after the fact. There is a hook in this mock database driver for you to provide your own parsing routine or object.
The syntax is simple:
$dbh->{mock_add_parser} = sub {
my ( $sql ) = @_;
unless ( $sql =~ /some regex/ ) {
die "does not contain secret fieldname";
}
};
You can also add more than one for a handle. They will be called in order, and the first one to fail will halt the parsing process:
$dbh->{mock_add_parser} = \&parse_update_sql;
$dbh->{mock_add-parser} = \&parse_insert_sql;
Depending on the 'PrintError' and 'RaiseError' settings in the database handle any parsing errors encountered will issue a C<warn> or C<die>. No matter what the statement handle will be C<undef>.
Instead of providing a subroutine reference you can use an object. The only requirement is that it implements the method C<parse()> and takes a SQL statement as the only argument. So you should be able to do something like the following (untested):
my $parser = SQL::Parser->new( 'mysql', { RaiseError => 1 } );
$dbh->{mock_add_parser} = $parser;
=item B<mock_data_sources> & B<mock_add_data_sources>
These properties will dispatch to the Driver's properties of the same name.
=back
=head2 Database Driver Methods
=over 4
=item B<last_insert_id>
This returns the value of C<mock_last_insert_id>.
=back
In order to capture begin_work(), commit(), and rollback(), DBD::Mock will create statements for them, as if you had issued them in the appropriate SQL command line program. They will go through the standard prepare()-execute() cycle, meaning that any custom SQL parsers will be triggered and DBD::Mock::Session will need to know about these statements.
=over 4
=item B<begin_work>
This will create a statement with SQL of "BEGIN WORK" and no parameters.
=item B<commit>
This will create a statement with SQL of "COMMIT" and no parameters.
=item B<rollback>
This will create a statement with SQL of "ROLLBACK" and no parameters.
=back
=head2 Statement Handle Properties
=over 4
=item B<Active>
Returns true if the handle is a 'SELECT' and has more records to fetch, false otherwise. (From the DBI.)
=item B<mock_statement>
The SQL statement this statement handle was C<prepare>d with. So if the handle were created with:
my $sth = $dbh->prepare( 'SELECT * FROM foo' );
This would return:
SELECT * FROM foo
The original statement is unmodified so if you are checking against it in tests you may want to use a regex rather than a straight equality check. (However if you use a phrasebook to store your SQL externally you are a step ahead...)
=item B<mock_fields>
Fields used by the statement. As said elsewhere we do no analysis or parsing to find these, you need to define them beforehand. That said, you do not actually need this very often.
Note that this returns the same thing as the normal statement property 'FIELD'.
=item B<mock_params>
Returns an arrayref of parameters bound to this statement in the order specified by the bind type. For instance, if you created and stocked a handle with:
my $sth = $dbh->prepare( 'SELECT * FROM foo WHERE id = ? AND is_active = ?' );
$sth->bind_param( 2, 'yes' );
$sth->bind_param( 1, 7783 );
This would return:
[ 7738, 'yes' ]
The same result will occur if you pass the parameters via C<execute()> instead:
my $sth = $dbh->prepare( 'SELECT * FROM foo WHERE id = ? AND is_active = ?' );
$sth->execute( 7783, 'yes' );
The same using named parameters
my $sth = $dbh->prepare( 'SELECT * FROM foo WHERE id = :id AND is_active = :active' );
$sth->bind_param( ':id' => 7783 );
$sth->bind_param( ':active' => 'yes' );
=item B<mock_records>
An arrayref of arrayrefs representing the records the mock statement was stocked with.
=item B<mock_num_records>
Number of records the mock statement was stocked with; if never stocked it is still 0. (Some weirdos might expect undef...)
=item B<mock_num_rows>
This returns the same value as I<mock_num_records>. And is what is returned by the C<rows> method of the statement handle.
=item B<mock_current_record_num>
Current record the statement is on; returns 0 in the instances when you have not yet called C<execute()> and if you have not yet called a C<fetch> method after the execute.
=item B<mock_is_executed>
Whether C<execute()> has been called against the statement handle. Returns 'yes' if so, 'no' if not.
=item B<mock_is_finished>
Whether C<finish()> has been called against the statement handle. Returns 'yes' if so, 'no' if not.
=item B<mock_is_depleted>
Returns 'yes' if all the records in the recordset have been returned. If no C<fetch()> was executed against the statement, or If no return data was set this will return 'no'.
=item B<mock_my_history>
Returns a C<DBD::Mock::StatementTrack> object which tracks the actions performed by this statement handle. Most of the actions are separately available from the properties listed above, so you should never need this.
=back
=head1 DBD::Mock::Pool
This module can be used to emulate Apache::DBI style DBI connection pooling. Just as with Apache::DBI, you must enable DBD::Mock::Pool before loading DBI.
use DBD::Mock qw(Pool);
# followed by ...
use DBI;
While this may not seem to make a lot of sense in a single-process testing scenario, it can be useful when testing code which assumes a multi-process Apache::DBI pooled environment.
=head1 DBD::Mock::StatementTrack
Under the hood this module does most of the work with a C<DBD::Mock::StatementTrack> object. This is most useful when you are reviewing multiple statements at a time, otherwise you might want to use the C<mock_*> statement handle attributes instead.
=over 4
=item B<new( %params )>
Takes the following parameters:
=over 4
=item *
B<return_data>: Arrayref of return data records
=item *
B<fields>: Arrayref of field names
=item *
B<bound_params>: Arrayref of bound parameters
=back
=item B<statement> (Statement attribute 'mock_statement')
Gets/sets the SQL statement used.
=item B<fields> (Statement attribute 'mock_fields')
Gets/sets the fields to use for this statement.
=item B<bound_params> (Statement attribute 'mock_params')
Gets/set the bound parameters to use for this statement.
=item B<return_data> (Statement attribute 'mock_records')
Gets/sets the data to return when asked (that is, when someone calls 'fetch' on the statement handle).
=item B<current_record_num> (Statement attribute 'mock_current_record_num')
Gets/sets the current record number.
=item B<is_active()> (Statement attribute 'Active')
Returns true if the statement is a SELECT and has more records to fetch, false otherwise. (This is from the DBI, see the 'Active' docs under 'ATTRIBUTES COMMON TO ALL HANDLES'.)
=item B<is_executed( $yes_or_no )> (Statement attribute 'mock_is_executed')
Sets the state of the tracker 'executed' flag.
=item B<is_finished( $yes_or_no )> (Statement attribute 'mock_is_finished')
If set to 'yes' tells the tracker that the statement is finished. This resets the current record number to '0' and clears out the array ref of returned records.
=item B<is_depleted()> (Statement attribute 'mock_is_depleted')
Returns true if the current record number is greater than the number of records set to return.
=item B<num_fields>
Returns the number of fields set in the 'fields' parameter.
=item B<num_rows>
Returns the number of records in the current result set.
=item B<num_params>
Returns the number of parameters set in the 'bound_params' parameter.
=item B<bound_param( $param_num, $value )>
Sets bound parameter C<$param_num> to C<$value>. Returns the arrayref of currently-set bound parameters. This corresponds to the 'bind_param' statement handle call.
=item B<bound_param_trailing( @params )>
Pushes C<@params> onto the list of already-set bound parameters.
=item B<mark_executed()>
Tells the tracker that the statement has been executed and resets the current record number to '0'.
=item B<next_record()>
If the statement has been depleted (all records returned) returns undef; otherwise it gets the current recordfor returning, increments the current record number and returns the current record.
=item B<to_string()>
Tries to give an decent depiction of the object state for use in debugging.
=back
=head1 DBD::Mock::StatementTrack::Iterator
This object can be used to iterate through the current set of C<DBD::Mock::StatementTrack> objects in the history by fetching the 'mock_all_history_iterator' attribute from a database handle. This object is very simple and is meant to be a convience to make writing long test script easier. Aside from the constructor (C<new>) this object has only one method.
=over 4
B<next>
Calling C<next> will return the next C<DBD::Mock::StatementTrack> object in the history. If there are no more C<DBD::Mock::StatementTrack> objects available, then this method will return false.
B<reset>
This will reset the internal pointer to the beginning of the statement history.
=back
=head1 DBD::Mock::Session
The DBD::Mock::Session object is an alternate means of specifying the SQL statements and result sets for DBD::Mock. The idea is that you can specify a complete 'session' of usage, which will be verified through DBD::Mock. Here is an example:
my $session = DBD::Mock::Session->new('my_session' => (
{
statement => "SELECT foo FROM bar", # as a string
results => [[ 'foo' ], [ 'baz' ]]
},
{
statement => qr/UPDATE bar SET foo \= \'bar\'/, # as a reg-exp
results => [[]]
},
{
statement => sub { # as a CODE ref
my ($SQL, $state) = @_;
return $SQL eq "SELECT foo FROM bar";
},
results => [[ 'foo' ], [ 'bar' ]]
},
{
# with bound parameters
statement => "SELECT foo FROM bar WHERE baz = ? AND borg = ?",
# check exact bound param value,
# then check it against regexp
bound_params => [ 10, qr/\d+/ ],
results => [[ 'foo' ], [ 'baz' ]]
}
));
As you can see, a session is essentially made up a list of HASH references we call 'states'. Each state has a 'statement' and a set of 'results'. If DBD::Mock finds a session in the 'mock_session' attribute, then it will pass the current C<$dbh> and SQL statement to that DBD::Mock::Session. The SQL statement will be checked against the 'statement' field in the current state. If it passes, then the 'results' of the current state will get feed to DBD::Mock through the 'mock_add_resultset' attribute. We then advance to the next state in the session, and wait for the next call through DBD::Mock. If at any time the SQL statement does not match the current state's 'statement', or the session runs out of available states, an error will be raised (and propagated through the normal DBI error handling based on your values for RaiseError and PrintError).
Also, as can be seen in the the session element, bound parameters can also be supplied and tested. In this statement, the SQL is compared, then when the statement is executed, the bound parameters are also checked. The bound parameters much match in both number of parameters and the parameters themselves, or an error will be raised.
As can also be seen in the example above, 'statement' fields can come in many forms. The simplest is a string, which will be compared using C<eq> against the currently running statement. The next is a reg-exp reference, this too will get compared against the currently running statement. The last option is a CODE ref, this is sort of a catch-all to allow for a wide range of SQL comparison approaches (including using modules like SQL::Statement or SQL::Parser for detailed functional comparisons). The first argument to the CODE ref will be the currently active SQL statement to compare against, the second argument is a reference to the current state HASH (in case you need to alter the results, or store extra information). The CODE is evaluated in boolean context and throws and exception if it is false.
=over 4
B<new ($session_name, @session_states)>
A C<$session_name> can be optionally be specified, along with at least one C<@session_states>. If you don't specify a C<$session_name>, then a default one will be created for you. The C<@session_states> must all be HASH references as well, if this conditions fail, an exception will be thrown.
B<verify_statement ($dbh, $SQL)>
This will check the C<$SQL> against the current state's 'statement' value, and if it passes will add the current state's 'results' to the C<$dbh>. If for some reason the 'statement' value is bad, not of the prescribed type, an exception is thrown. See above for more details.
B<verify_bound_params ($dbh, $params)>
If the 'bound_params' slot is available in the current state, this will check the C<$params> against the current state's 'bound_params' value. Both number of parameters and the parameters themselves must match, or an error will be raised.
B<reset>
Calling this method will reset the state of the session object so that it can be reused.
=back
=head1 EXPERIMENTAL FUNCTIONALITY
All functionality listed here is highly experimental and should be used with great caution (if at all).
=over 4
=item Error handling in I<mock_add_resultset>
We have added experimental erro handling in I<mock_add_resultset> the best example is the test file F<t/023_statement_failure.t>, but it looks something like this:
$dbh->{mock_add_resultset} = {
sql => 'SELECT foo FROM bar',
results => DBD::Mock->NULL_RESULTSET,
failure => [ 5, 'Ooops!' ],
};
The C<5> is the DBI error number, and C<'Ooops!'> is the error string passed to DBI. This basically allows you to force an error condition to occur when a given SQL statement is execute. We are currently working on allowing more control on the 'when' and 'where' the error happens, look for it in future releases.
=item Attribute Aliasing
Basically this feature allows you to alias attributes to other attributes. So for instance, you can alias a commonly expected attribute like 'mysql_insertid' to something DBD::Mock already has like 'mock_last_insert_id'. While you can also just set 'mysql_insertid' yourself, this functionality allows it to take advantage of things like the autoincrementing of the 'mock_last_insert_id' attribute.
Right now this feature is highly experimental, and has been added as a first attempt to automatically handle some of the DBD specific attributes which are commonly used/accessed in DBI programming. The functionality is off by default so as to not cause any issues with backwards compatibility, but can easily be turned on and off like this:
# turn it on
$DBD::Mock::AttributeAliasing++;
# turn it off
$DBD::Mock::AttributeAliasing = 0;
Once this is turned on, you will need to choose a database specific attribute aliasing table like so:
DBI->connect('dbi:Mock:MySQL', '', '');
The 'MySQL' in the DSN will be picked up and the MySQL specific attribute aliasing will be used.
Right now only MySQL is supported by this feature, and even that support is very minimal. Currently the MySQL C<$dbh> and C<$sth> attributes 'mysql_insertid' are aliased to the C<$dbh> attribute 'mock_last_insert_id'. It is possible to add more aliases though, using the C<DBD::Mock:_set_mock_attribute_aliases> function (see the source code for details).
=back
=head1 BUGS
=over
=item Odd $dbh attribute behavior
When writing the test suite I encountered some odd behavior with some C<$dbh> attributes. I still need to get deeper into how DBD's work to understand what it is that is actually doing wrong.
=back
=head1 TO DO
=over
=item Make DBD specific handlers
Each DBD has its own quirks and issues, it would be nice to be able to handle those issues with DBD::Mock in some way. I have an number of ideas already, but little time to sit down and really flesh them out. If you have any suggestions or thoughts, feel free to email me with them.
=item Enhance the DBD::Mock::StatementTrack object
I would like to have the DBD::Mock::StatementTrack object handle more of the mock_* attributes. This would encapsulate much of the mock_* behavior in one place, which would be a good thing.
I would also like to add the ability to bind a subroutine (or possibly an object) to the result set, so that the results can be somewhat more dynamic and allow for a more realistic interaction.
=back
=head1 SEE ALSO
L<DBI>
L<DBD::NullP>, which provided a good starting point
L<Test::MockObject>, which provided the approach
Test::MockObject article - L<http://www.perl.com/pub/a/2002/07/10/tmo.html>
Perl Code Kata: Testing Databases - L<http://www.perl.com/pub/a/2005/02/10/database_kata.html>
=head1 DISCUSSION GROUP
We have created a B<DBD::Mock> google group for discussion/questions about this module.
L<http://groups.google.com/group/DBDMock>
=head1 ACKNOWLEDGEMENTS
=over 4
=item Thanks to Ryan Gerry for his patch in RT #26604
=item Thanks to Marc Beyer for his patch in RT #16951
=item Thanks to Justin DeVuyst for the mock_connect_fail idea
=item Thanks to Thilo Planz for the code for C<bind_param_inout>
=item Thanks to Shlomi Fish for help tracking down RT Bug #11515
=item Thanks to Collin Winter for the patch to fix the C<begin_work()>, C<commit()> and C<rollback()> methods.
=item Thanks to Andrew McHarg E<lt>amcharg@acm.orgE<gt> for C<fetchall_hashref()>, C<fetchrow_hashref()> and C<selectcol_arrayref()> methods and tests.
=item Thanks to Andrew W. Gibbs for the C<mock_last_insert_ids> patch and test
=item Thanks to Chas Owens for patch and test for the C<mock_can_prepare>, C<mock_can_execute>, and C<mock_can_fetch> features.
=back
=head1 COPYRIGHT
Copyright (C) 2004 Chris Winters E<lt>chris@cwinters.comE<gt>
Copyright (C) 2004-2007 Stevan Little E<lt>stevan@iinteractive.comE<gt>
Copyright (C) 2007 Rob Kinyon E<lt>rob.kinyon@gmail.comE<gt>
Copyright (C) 2011 Mariano Wahlmann E<lt>dichoso _at_ gmail.comE<gt>
This library is free software; you can redistribute it and/or modify
it under the same terms as Perl itself.
=head1 AUTHORS
Chris Winters E<lt>chris@cwinters.comE<gt>
Stevan Little E<lt>stevan@iinteractive.comE<gt>
Rob Kinyon E<lt>rob.kinyon@gmail.comE<gt>
Mariano Wahlmann E<lt>dichoso _at_ gmail.com <gt>
=cut
|