Learned brothers, I have a question:

I am working on an application that uses DBIx::Class to store it's data. There are around 100 tables, with plenty of joins and other relationships between the various tables. It is a complicated beast.

As with any large project, I have a number of tests. Most of them work by starting with an empty copy of the database, populating a small number of tables with a small amount of data, and then testing that the API calls retrieve the expected data. For example:

use strict;
use warnings;
use Test::More tests => 134;
use My::Company::Schema;

# Creates an empty SQLite DB in memory and deploys the schema to it.
my $schema = My::Company::Schema->connect;

# set-up some test data
my $tblFoo = $schema->resultset('Foo');

my $foo_item_1 = $tblFoo->create({
    name => 'item1',
    # Other fields
});

my $foo_item_2 = $tblFoo->create({
    name => 'item2',
    # More fields
});

my $bar_item = $foo_item_2->add_to_relation({'key'=>'value'});

# More code to populate the test DB.

# Tests:
my $found_entry_foo = $tblFoo_rs->big_complex_api_call();

is( $found_entry_foo->id, $bar_item->id, "The correct entry from tblFo
+o was found" );
[download]

The problem is, is that test above contains a bug. On the last line the primary key of the returned row is compared with the expected value from the WRONG table. The idiom is used in thousands of tests, and I have a nasty feeling that the wrong comparisons are done in quite a few places.

The reason the bug goes un-detected, is that the test tables only contain a few rows, and the numeric primary keys start from 1, and number incrementally, so there is a good chance that row 2 in one table could get confused with row 2 in another.

Is there an extension, or directive I can give to DBIx::Class, that will cause it to start the numbering of primary keys from a random large number, in order to flush out bugs like this? I did not find anything from reading the DBIx::Class docs, or from Google, though I could be looking in the wrong place.