Provided by: libtest-postgresql-perl_1.27-1_all bug


       Test::PostgreSQL - PostgreSQL runner for tests


         use DBI;
         use Test::PostgreSQL;
         use Test::More;

         # optionally
         # (if not already set at shell):
         # $ENV{POSTGRES_HOME} = '/path/to/my/pgsql/installation';

         my $pgsql = eval { Test::PostgreSQL->new() }
             or plan skip_all => $@;

         plan tests => XXX;

         my $dbh = DBI->connect($pgsql->dsn);


       "Test::PostgreSQL" automatically setups a PostgreSQL instance in a temporary directory,
       and destroys it when the perl script exits.

       This module is a fork of Test::postgresql, which was abandoned by its author several years


       "Test::PostgreSQL" object has the following attributes, overridable by passing
       corresponding argument to constructor:

       Database name to use in this "Test::PostgreSQL" instance. Default is "test".

       Database owner user name. Default is "postgres".

       Host name or IP address to use for PostgreSQL instance connections. Default is

       Base directory under which the PostgreSQL instance is being created. The property can be
       passed as a parameter to the constructor, in which case the directory will not be removed
       at exit.

       Connection port number to start with. If the port is already used we will increment the
       value and try again.

       Default: 15432.

       Whether to only connect via UNIX sockets; if false (the default), connections can occur
       via localhost. [This changes the "dsn" returned to only give the UNIX socket directory,
       and avoids any issues with conflicting TCP ports on localhost.]

       Unix socket directory to use if "unix_socket" is true. Default is "$basedir/tmp".

       Path to "pg_ctl" program which is part of the PostgreSQL distribution.

       Starting with PostgreSQL version 9.0 "pg_ctl" can be used to start/stop postgres without
       having to use fork/pipe and will be chosen automatically if "pg_ctl" is not set but the
       program is found and the version is recent enough.

       NOTE: do NOT use this with PostgreSQL versions prior to version 9.0.

       By default we will try to find "pg_ctl" in PostgresSQL directory.

       Path to "initdb" program which is part of the PostreSQL distribution. Default is to try
       and find it in PostgreSQL directory.

       Arguments to pass to "initdb" program when creating a new PostgreSQL database cluster for
       Test::PostgreSQL session.

       Defaults to "-U postgres -A trust". See "db_owner".

       Extra args to be appended to "initdb_args". Default is empty.

       Configuration to place in "$basedir/data/postgresql.conf". Use this to override PostgreSQL
       configuration defaults, e.g. to speed up PostgreSQL database init and seeding one might
       use something like this:

           my $pgsql = Test::PostgreSQL->new(
               pg_config => q|
               # foo baroo mymse throbbozongo
               fsync = off
               synchronous_commit = off
               full_page_writes = off
               bgwriter_lru_maxpages = 0
               shared_buffers = 512MB
               effective_cache_size = 512MB
               work_mem = 100MB

       Path to "postmaster" which is part of the PostgreSQL distribution. If not set, the
       programs are automatically searched by looking up $PATH and other prefixed directories.
       Since "postmaster" is deprecated in newer PostgreSQL versions "postgres" is used in
       preference to "postmaster".

       Defaults to "-h -F".

       Extra args to be appended to "postmaster_args". Default is empty.

       Path to "psql" client which is part of the PostgreSQL distribution.

       "psql" can be used to run SQL scripts against the temporary database created by "new":

           my $pgsql = Test::PostgreSQL->new();
           my $psql = $pgsql->psql;

           my $out = `$psql -f /path/to/script.sql 2>&1`;

           die "Error executing script.sql: $out" unless $? == 0;

       Command line arguments necessary for "psql" to connect to the correct PostgreSQL instance.

       Defaults to "-U postgres -d test -h -p $self->port".

       See also "db_owner", "dbname", "host", "base_port".

       Extra args to be appended to "psql_args".

       Arguments specific for "run_psql" invocation, used mostly to set up and seed database
       schema after PostgreSQL instance is launched and configured.

       Default is "-1Xqb -v ON_ERROR_STOP=1". This means:

       ·   1: Run all SQL statements in passed scripts as single transaction

       ·   X: Skip ".psqlrc" files

       ·   q: Run quietly, print only notices and errors on stderr (if any)

       ·   b: Echo SQL statements that cause PostgreSQL exceptions (version 9.5+)

       ·   -v ON_ERROR_STOP=1: Stop processing SQL statements after the first error

       Arrayref with the list of SQL scripts to run after the database was instanced and set up.
       Default is "[]".

       NOTE that "psql" older than 9.6 does not support multiple "-c" and "-f" switches in
       arguments so "seed_scripts" will be executed one by one. This implies multiple
       transactions instead of just one; if you need all seed statements to apply within a single
       transaction, combine them into one seed script.

       Integer value that controls whether PostgreSQL server is started and setup after creating
       "Test::PostgreSQL" instance. Possible values:

       0   Do not start PostgreSQL.

       1   Start PostgreSQL but do not run "setup".

       2   Start PostgreSQL and run "setup".

           Default is 2.


       Create and run a PostgreSQL instance. The instance is terminated when the returned object
       is being DESTROYed.  If required programs (initdb and postmaster) were not found, the
       function returns undef and sets appropriate message to $Test::PostgreSQL::errstr.

       Builds and returns dsn by using given parameters (if any).  Default username is
       "postgres", and dbname is "test" (an empty database).

       Builds and returns a connection URI using the given parameters (if any). See URI::db for
       details about the format.

       Default username is "postgres", and dbname is "test" (an empty database).

       Returns process id of PostgreSQL (or undef if not running).

       Returns TCP port number on which postmaster is accepting connections (or undef if not

       Starts postmaster.

       Stops postmaster.

       Setups the PostgreSQL instance. Note that this method should be invoked before "start".

       Execute "psql" program with the given list of arguments. Usually this would be something

           $pgsql->run_psql('-c', q|'INSERT INTO foo (bar) VALUES (42)'|);


           $pgsql->run_psql('-f', '/path/to/script.sql');

       Note that when using literal SQL statements with "-c" parameter you will need to escape
       them manually like shown above. "run_psql" will not quote them for you.

       The actual command line to execute "psql" will be concatenated from "psql_args",
       "extra_psql_args", and "run_psql_args".

       NOTE that "psql" older than 9.6 does not support multiple "-c" and/or "-f" switches in

       Given a list of script file paths, invoke "run_psql" once with "-f 'script'" for every
       path in PostgreSQL 9.6+, or once per "-f 'script'" for older PostgreSQL versions.


       If your postgres installation is not located in a well known path, or you have many
       versions installed and want to run your tests against particular one, set this environment
       variable to the desired path. For example:

           export POSTGRES_HOME='/usr/local/pgsql94beta'

       This is the same idea and variable name which is used by the installer of DBD::Pg.


       Toby Corkindale, Kazuho Oku, Peter Mottram, Alex Tokarev, plus various contributors.


       Current version copyright X 2012-2015 Toby Corkindale.

       Previous versions copyright (C) 2009 Cybozu Labs, Inc.


       This module is free software, released under the Perl Artistic License 2.0.  See
       <> for more information.