fix(db): Make using pg_service work again. When I did the merge of schema support I broke pg_service.conf support by replacing get_database_name with db_schema_split. This commit fixes it. Also this commit returns the schema if one is specified in pg_service.conf. back_postgresql.py: Replace calls to db_schema_split() with get_database_schema_names() (new name for get_database_name()). Rename db_schema_split to _db_schema_split. It now returns a tuple (dbname, schema) rather than a list. It is used only by get_database_schema_names() which also returns tuples. get_database_schema_names() can also get schema info for the service (if present) as specified by pg_service.conf. Add get_database_user() to get the user from either RDBMS_USER or pg_service.conf. (User needed for creating schema, so not needed before schema patch. import re at the top of file and remove lower import. Remove some schema code from db_command as it's not needed. The database conection is done to either postgresql or template1 existing databases. This command never connects to the roundp specified db. test/test_postgresql.py: Reorganize top level imports, add import os. Replace import of db_schema_split with get_database_schema_names. Also replace calls to db_schema_split. Create new Opener for the service file. Set PGSERVICEFILE to point to test/pg_service.conf. Add three new classes to test Service: 1) using regular db 2) using schema within db 3) Unable to parse schema name from pg_service.conf. The last doesn't need a db. Number 1 and 2 reuse the tests in ROTest to verify db connectivity. test/pg_service.conf: three service connections for: db only, db and schema, and incorrectly specified schema test cases. doc/upgrading.txt: updated to current status. Included example schema definition in service file.

#-*- encoding: utf-8 -*-
""" Testing the jinja2 templating engine of roundup-tracker.

Copyright: 2016 Intevation GmbH.
Author: Bernhard E. Reiter <bernhard@intevation.de>

This module is Free Software under the Roundup licensing of 1.5,
see the COPYING.txt file coming with Roundup.

Just a test file template for now.
"""
import shutil # only, needed for tearDown. TODO: Remove when refactored.
import unittest

from . import db_test_base

TESTSUITE_IDENTIFIER='jinja2'

class TestCase_Zero(unittest.TestCase):
    def test_zero(self):
        self.assertEqual(True, True)


class Jinja2Test(object):
    """Sets up and tears down an instance with database contents.

    Setup and teardown modelled after the use of db_test_base
    by several modules like test_xmlrpc and test_userauditor.

    TODO: Should probably be moved to a base case in db_test_base.py.
    """

    backend = None  # can be used to create tests per backend, see test_xmlrpc

    def setUp(self):
        self.dirname = '_test_' + TESTSUITE_IDENTIFIER
        self.instance = db_test_base.setupTracker(self.dirname, self.backend)
        self.db = self.instance.open('admin')

    def tearDown(self):
        self.db.close()
        try:
            shutil.rmtree(self.dirname)
        except OSError as error:
            if error.errno not in (errno.ENOENT, errno.ESRCH): raise

    def test_zero(self):
        """Do nothing just make sure that setup and teardown works."""
        pass


# only using one database backend for now, not sure if doing all
# backends will keep the test focussed enough to be useful for the used
# computing time. Would be okay to change in the future.
class anydbmJinja2Test(Jinja2Test, unittest.TestCase):
    backend = 'anydbm'

# vim: ts=4 et sts=4 sw=4 ai :