Module::Build::Database::PostgreSQL - PostgreSQL implementation for MBD
In Build.PL :
my $builder = Module::Build::Database->new( database_type => "PostgreSQL", database_options => { name => "my_database_name", schema => "my_schema_name", # Extra items for scratch databases : append_to_conf => "text to add to postgresql.conf", after_create => q[create schema audit;], }, database_extensions => { postgis => { schema => "public", }, # directory with postgis.sql and spatial_ref_sys.sql postgis_base => '/usr/local/share/postgresql/contrib' }, );
Postgres driver for Module::Build::Database.
All of the options above may be changed via the Module::Build option handling, e.g.
perl Build.PL --database_options name=my_name perl Build.PL --postgis_base=/usr/local/share/postgresql/contrib
The options are as follows ;
the name of the database (i.e. 'create database $name')
the name of the schema to be managed by MBD
extra options to append to postgresql.conf
before starting test instances of postgres
extra SQL to run after running a 'create database' statement. Note that this will be run in several different situations :
An example of using the after_create statement would be to create a second schema which will not be managed by MBD, but on which the MBD-managed schema depends.
To specify a server side procedural language you can use the database_extension
-> languages
option, like so:
my $builder = Module::Build::Database->new( database_extension => { languages => [ 'plperl', 'pltcl' ], }, );
Trying to create languages to a patch will not work because they not stored in the main schema and will not be included in base.sql
when you run Build dbdist
.
This is also similar to
after_create => 'create extension ...',
except it is executed on every dbinstall meaning you can use this to add extensions to existing database deployments.
Specify the directory containing postgis.sql and spatial_ref_sys.sql. If specified these SQL files will be loaded so that you can use PostGIS in your database.
If set to true, and if you are not using a persistent scratch database (see next option), then the scratch database will not be stopped and torn down after running Build dbtest
or Build dbfakeinstall
.
You can use this option to specify the connection settings for a persistent scratch or temporary database instance, used by the Build dbtest
and Build dbfakeinstall
to test schema. IMPORTANT: the Build dbtest
and Build dbfakeinstall
will drop and re-create databases on the scratch instance with the same name as the database on your production instance so it is very important that if you use a persistent scratch database that it be dedicated to that task.
my $builder = Module::Build::Database->new( scratch_database => { PGHOST => 'databasehost', PGPORT => '5555', PGUSER => 'dbuser', }, );
If you specify any one of these keys for this option (PGHOST
, PGPORT
, PGUSER
) then MBD will use a persistent scratch database. Any missing values will use the default.
You can also specify these settings using environment variables:
% export MBD_SCRATCH_PGHOST=databasehost % export MBD_SCRATCH_PGPORT=5555 % export MBD_SCRATCH_PGUSER=dbuser
By default this module will create its own scratch PostgreSQL instance that uses unix domain sockets for communication each time it needs one when you use the Build dbtest
or Build dbfakeinstall
commands. Situations where you might need to use a persistent scratch database:
The environment variables understood by psql
: PGUSER
, PGHOST
and PGPORT
will be used when connecting to a live database (for dbinstall and fakeinstall). PGDATABASE
will be ignored; the name of the database should be specified in Build.PL instead.