cubicweb/server/test/unittest_migractions.py
changeset 11057 0b59724cb3f2
parent 11039 6424a2e112f5
child 11129 97095348b3ee
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/cubicweb/server/test/unittest_migractions.py	Sat Jan 16 13:48:51 2016 +0100
@@ -0,0 +1,933 @@
+# copyright 2003-2014 LOGILAB S.A. (Paris, FRANCE), all rights reserved.
+# contact http://www.logilab.fr/ -- mailto:contact@logilab.fr
+#
+# This file is part of CubicWeb.
+#
+# CubicWeb is free software: you can redistribute it and/or modify it under the
+# terms of the GNU Lesser General Public License as published by the Free
+# Software Foundation, either version 2.1 of the License, or (at your option)
+# any later version.
+#
+# CubicWeb is distributed in the hope that it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+# FOR A PARTICULAR PURPOSE.  See the GNU Lesser General Public License for more
+# details.
+#
+# You should have received a copy of the GNU Lesser General Public License along
+# with CubicWeb.  If not, see <http://www.gnu.org/licenses/>.
+"""unit tests for module cubicweb.server.migractions"""
+
+from datetime import date
+import os, os.path as osp
+from contextlib import contextmanager
+
+from logilab.common.testlib import unittest_main, Tags, tag
+from logilab.common import tempattr
+
+from yams.constraints import UniqueConstraint
+
+from cubicweb import ConfigurationError, ValidationError, ExecutionError
+from cubicweb.devtools import startpgcluster, stoppgcluster
+from cubicweb.devtools.testlib import CubicWebTC
+from cubicweb.server.sqlutils import SQL_PREFIX
+from cubicweb.server.migractions import ServerMigrationHelper
+
+import cubicweb.devtools
+
+
+HERE = osp.dirname(osp.abspath(__file__))
+
+
+def setUpModule():
+    startpgcluster(__file__)
+
+
+migrschema = None
+def tearDownModule(*args):
+    global migrschema
+    del migrschema
+    if hasattr(MigrationCommandsTC, 'origschema'):
+        del MigrationCommandsTC.origschema
+    if hasattr(MigrationCommandsComputedTC, 'origschema'):
+        del MigrationCommandsComputedTC.origschema
+    stoppgcluster(__file__)
+
+
+class MigrationConfig(cubicweb.devtools.TestServerConfiguration):
+    default_sources = cubicweb.devtools.DEFAULT_PSQL_SOURCES
+    CUBES_PATH = cubicweb.devtools.TestServerConfiguration.CUBES_PATH + [
+        osp.join(HERE, 'data-migractions', 'cubes')]
+
+
+class MigrationTC(CubicWebTC):
+
+    appid = 'data-migractions'
+
+    configcls = MigrationConfig
+
+    tags = CubicWebTC.tags | Tags(('server', 'migration', 'migractions'))
+
+    def _init_repo(self):
+        super(MigrationTC, self)._init_repo()
+        # we have to read schema from the database to get eid for schema entities
+        self.repo.set_schema(self.repo.deserialize_schema(), resetvreg=False)
+        # hack to read the schema from data/migrschema
+        config = self.config
+        config.appid = osp.join(self.appid, 'migratedapp')
+        config._apphome = osp.join(HERE, config.appid)
+        global migrschema
+        migrschema = config.load_schema()
+        config.appid = self.appid
+        config._apphome = osp.join(HERE, self.appid)
+
+    def setUp(self):
+        self.configcls.cls_adjust_sys_path()
+        super(MigrationTC, self).setUp()
+
+    def tearDown(self):
+        super(MigrationTC, self).tearDown()
+        self.repo.vreg['etypes'].clear_caches()
+
+    @contextmanager
+    def mh(self):
+        with self.admin_access.repo_cnx() as cnx:
+            yield cnx, ServerMigrationHelper(self.repo.config, migrschema,
+                                             repo=self.repo, cnx=cnx,
+                                             interactive=False)
+
+    def table_sql(self, mh, tablename):
+        result = mh.sqlexec("SELECT table_name FROM information_schema.tables WHERE LOWER(table_name)=%(table)s",
+                            {'table': tablename.lower()})
+        if result:
+            return result[0][0]
+        return None # no such table
+
+    def table_schema(self, mh, tablename):
+        result = mh.sqlexec("SELECT column_name, data_type, character_maximum_length FROM information_schema.columns "
+                            "WHERE LOWER(table_name) = %(table)s", {'table': tablename.lower()})
+        assert result, 'no table %s' % tablename
+        return dict((x[0], (x[1], x[2])) for x in result)
+
+
+class MigrationCommandsTC(MigrationTC):
+
+    def _init_repo(self):
+        super(MigrationCommandsTC, self)._init_repo()
+        assert 'Folder' in migrschema
+
+    def test_add_attribute_bool(self):
+        with self.mh() as (cnx, mh):
+            self.assertNotIn('yesno', self.schema)
+            cnx.create_entity('Note')
+            cnx.commit()
+            mh.cmd_add_attribute('Note', 'yesno')
+            self.assertIn('yesno', self.schema)
+            self.assertEqual(self.schema['yesno'].subjects(), ('Note',))
+            self.assertEqual(self.schema['yesno'].objects(), ('Boolean',))
+            self.assertEqual(self.schema['Note'].default('yesno'), False)
+            # test default value set on existing entities
+            note = cnx.execute('Note X').get_entity(0, 0)
+            self.assertEqual(note.yesno, False)
+            # test default value set for next entities
+            self.assertEqual(cnx.create_entity('Note').yesno, False)
+
+    def test_add_attribute_int(self):
+        with self.mh() as (cnx, mh):
+            self.assertNotIn('whatever', self.schema)
+            cnx.create_entity('Note')
+            cnx.commit()
+            orderdict = dict(mh.rqlexec('Any RTN, O WHERE X name "Note", RDEF from_entity X, '
+                                        'RDEF relation_type RT, RDEF ordernum O, RT name RTN'))
+            mh.cmd_add_attribute('Note', 'whatever')
+            self.assertIn('whatever', self.schema)
+            self.assertEqual(self.schema['whatever'].subjects(), ('Note',))
+            self.assertEqual(self.schema['whatever'].objects(), ('Int',))
+            self.assertEqual(self.schema['Note'].default('whatever'), 0)
+            # test default value set on existing entities
+            note = cnx.execute('Note X').get_entity(0, 0)
+            self.assertIsInstance(note.whatever, int)
+            self.assertEqual(note.whatever, 0)
+            # test default value set for next entities
+            self.assertEqual(cnx.create_entity('Note').whatever, 0)
+            # test attribute order
+            orderdict2 = dict(mh.rqlexec('Any RTN, O WHERE X name "Note", RDEF from_entity X, '
+                                         'RDEF relation_type RT, RDEF ordernum O, RT name RTN'))
+            whateverorder = migrschema['whatever'].rdef('Note', 'Int').order
+            for k, v in orderdict.items():
+                if v >= whateverorder:
+                    orderdict[k] = v+1
+            orderdict['whatever'] = whateverorder
+            self.assertDictEqual(orderdict, orderdict2)
+            #self.assertEqual([r.type for r in self.schema['Note'].ordered_relations()],
+            #                  ['modification_date', 'creation_date', 'owned_by',
+            #                   'eid', 'ecrit_par', 'inline1', 'date', 'type',
+            #                   'whatever', 'date', 'in_basket'])
+            # NB: commit instead of rollback make following test fail with py2.5
+            #     this sounds like a pysqlite/2.5 bug (the same eid is affected to
+            #     two different entities)
+
+    def test_add_attribute_varchar(self):
+        with self.mh() as (cnx, mh):
+            self.assertNotIn('whatever', self.schema)
+            cnx.create_entity('Note')
+            cnx.commit()
+            self.assertNotIn('shortpara', self.schema)
+            mh.cmd_add_attribute('Note', 'shortpara')
+            self.assertIn('shortpara', self.schema)
+            self.assertEqual(self.schema['shortpara'].subjects(), ('Note', ))
+            self.assertEqual(self.schema['shortpara'].objects(), ('String', ))
+            # test created column is actually a varchar(64)
+            fields = self.table_schema(mh, '%sNote' % SQL_PREFIX)
+            self.assertEqual(fields['%sshortpara' % SQL_PREFIX], ('character varying', 64))
+            # test default value set on existing entities
+            self.assertEqual(cnx.execute('Note X').get_entity(0, 0).shortpara, 'hop')
+            # test default value set for next entities
+            self.assertEqual(cnx.create_entity('Note').shortpara, 'hop')
+
+    def test_add_datetime_with_default_value_attribute(self):
+        with self.mh() as (cnx, mh):
+            self.assertNotIn('mydate', self.schema)
+            self.assertNotIn('oldstyledefaultdate', self.schema)
+            self.assertNotIn('newstyledefaultdate', self.schema)
+            mh.cmd_add_attribute('Note', 'mydate')
+            mh.cmd_add_attribute('Note', 'oldstyledefaultdate')
+            mh.cmd_add_attribute('Note', 'newstyledefaultdate')
+            self.assertIn('mydate', self.schema)
+            self.assertIn('oldstyledefaultdate', self.schema)
+            self.assertIn('newstyledefaultdate', self.schema)
+            self.assertEqual(self.schema['mydate'].subjects(), ('Note', ))
+            self.assertEqual(self.schema['mydate'].objects(), ('Date', ))
+            testdate = date(2005, 12, 13)
+            eid1 = mh.rqlexec('INSERT Note N')[0][0]
+            eid2 = mh.rqlexec('INSERT Note N: N mydate %(mydate)s', {'mydate' : testdate})[0][0]
+            d1 = mh.rqlexec('Any D WHERE X eid %(x)s, X mydate D', {'x': eid1})[0][0]
+            d2 = mh.rqlexec('Any D WHERE X eid %(x)s, X mydate D', {'x': eid2})[0][0]
+            d3 = mh.rqlexec('Any D WHERE X eid %(x)s, X oldstyledefaultdate D', {'x': eid1})[0][0]
+            d4 = mh.rqlexec('Any D WHERE X eid %(x)s, X newstyledefaultdate D', {'x': eid1})[0][0]
+            self.assertEqual(d1, date.today())
+            self.assertEqual(d2, testdate)
+            myfavoritedate = date(2013, 1, 1)
+            self.assertEqual(d3, myfavoritedate)
+            self.assertEqual(d4, myfavoritedate)
+
+    def test_drop_chosen_constraints_ctxmanager(self):
+        with self.mh() as (cnx, mh):
+            with mh.cmd_dropped_constraints('Note', 'unique_id', UniqueConstraint):
+                mh.cmd_add_attribute('Note', 'unique_id')
+                # make sure the maxsize constraint is not dropped
+                self.assertRaises(ValidationError,
+                                  mh.rqlexec,
+                                  'INSERT Note N: N unique_id "xyz"')
+                mh.rollback()
+                # make sure the unique constraint is dropped
+                mh.rqlexec('INSERT Note N: N unique_id "x"')
+                mh.rqlexec('INSERT Note N: N unique_id "x"')
+                mh.rqlexec('DELETE Note N')
+
+    def test_drop_required_ctxmanager(self):
+        with self.mh() as (cnx, mh):
+            with mh.cmd_dropped_constraints('Note', 'unique_id', cstrtype=None,
+                                            droprequired=True):
+                mh.cmd_add_attribute('Note', 'unique_id')
+                mh.rqlexec('INSERT Note N')
+                mh.rqlexec('SET N unique_id "x"')
+            # make sure the required=True was restored
+            self.assertRaises(ValidationError, mh.rqlexec, 'INSERT Note N')
+            mh.rollback()
+
+    def test_rename_attribute(self):
+        with self.mh() as (cnx, mh):
+            self.assertNotIn('civility', self.schema)
+            eid1 = mh.rqlexec('INSERT Personne X: X nom "lui", X sexe "M"')[0][0]
+            eid2 = mh.rqlexec('INSERT Personne X: X nom "l\'autre", X sexe NULL')[0][0]
+            mh.cmd_rename_attribute('Personne', 'sexe', 'civility')
+            self.assertNotIn('sexe', self.schema)
+            self.assertIn('civility', self.schema)
+            # test data has been backported
+            c1 = mh.rqlexec('Any C WHERE X eid %s, X civility C' % eid1)[0][0]
+            self.assertEqual(c1, 'M')
+            c2 = mh.rqlexec('Any C WHERE X eid %s, X civility C' % eid2)[0][0]
+            self.assertEqual(c2, None)
+
+    def test_workflow_actions(self):
+        with self.mh() as (cnx, mh):
+            wf = mh.cmd_add_workflow(u'foo', ('Personne', 'Email'),
+                                     ensure_workflowable=False)
+            for etype in ('Personne', 'Email'):
+                s1 = mh.rqlexec('Any N WHERE WF workflow_of ET, ET name "%s", WF name N' %
+                                etype)[0][0]
+                self.assertEqual(s1, "foo")
+                s1 = mh.rqlexec('Any N WHERE ET default_workflow WF, ET name "%s", WF name N' %
+                                etype)[0][0]
+                self.assertEqual(s1, "foo")
+
+    def test_add_entity_type(self):
+        with self.mh() as (cnx, mh):
+            self.assertNotIn('Folder2', self.schema)
+            self.assertNotIn('filed_under2', self.schema)
+            mh.cmd_add_entity_type('Folder2')
+            self.assertIn('Folder2', self.schema)
+            self.assertIn('Old', self.schema)
+            self.assertTrue(cnx.execute('CWEType X WHERE X name "Folder2"'))
+            self.assertIn('filed_under2', self.schema)
+            self.assertTrue(cnx.execute('CWRType X WHERE X name "filed_under2"'))
+            self.assertEqual(sorted(str(rs) for rs in self.schema['Folder2'].subject_relations()),
+                              ['created_by', 'creation_date', 'cw_source', 'cwuri',
+                               'description', 'description_format',
+                               'eid',
+                               'filed_under2', 'has_text',
+                               'identity', 'in_basket', 'is', 'is_instance_of',
+                               'modification_date', 'name', 'owned_by'])
+            self.assertCountEqual([str(rs) for rs in self.schema['Folder2'].object_relations()],
+                                  ['filed_under2', 'identity'])
+            # Old will be missing as it has been renamed into 'New' in the migrated
+            # schema while New hasn't been added here.
+            self.assertEqual(sorted(str(e) for e in self.schema['filed_under2'].subjects()),
+                             sorted(str(e) for e in self.schema.entities() if not e.final and e != 'Old'))
+            self.assertEqual(self.schema['filed_under2'].objects(), ('Folder2',))
+            eschema = self.schema.eschema('Folder2')
+            for cstr in eschema.rdef('name').constraints:
+                self.assertTrue(hasattr(cstr, 'eid'))
+
+    def test_add_cube_with_custom_final_type(self):
+        with self.mh() as (cnx, mh):
+            try:
+                mh.cmd_add_cube('fakecustomtype')
+                self.assertIn('Numeric', self.schema)
+                self.assertTrue(self.schema['Numeric'].final)
+                rdef = self.schema['num'].rdefs[('Location', 'Numeric')]
+                self.assertEqual(rdef.scale, 10)
+                self.assertEqual(rdef.precision, 18)
+                fields = self.table_schema(mh, '%sLocation' % SQL_PREFIX)
+                self.assertEqual(fields['%snum' % SQL_PREFIX], ('numeric', None)) # XXX
+            finally:
+                mh.cmd_drop_cube('fakecustomtype')
+
+    def test_add_drop_entity_type(self):
+        with self.mh() as (cnx, mh):
+            mh.cmd_add_entity_type('Folder2')
+            wf = mh.cmd_add_workflow(u'folder2 wf', 'Folder2',
+                                     ensure_workflowable=False)
+            todo = wf.add_state(u'todo', initial=True)
+            done = wf.add_state(u'done')
+            wf.add_transition(u'redoit', done, todo)
+            wf.add_transition(u'markasdone', todo, done)
+            cnx.commit()
+            eschema = self.schema.eschema('Folder2')
+            mh.cmd_drop_entity_type('Folder2')
+            self.assertNotIn('Folder2', self.schema)
+            self.assertFalse(cnx.execute('CWEType X WHERE X name "Folder2"'))
+            # test automatic workflow deletion
+            self.assertFalse(cnx.execute('Workflow X WHERE NOT X workflow_of ET'))
+            self.assertFalse(cnx.execute('State X WHERE NOT X state_of WF'))
+            self.assertFalse(cnx.execute('Transition X WHERE NOT X transition_of WF'))
+
+    def test_rename_entity_type(self):
+        with self.mh() as (cnx, mh):
+            entity = mh.create_entity('Old', name=u'old')
+            self.repo.type_and_source_from_eid(entity.eid, entity._cw)
+            mh.cmd_rename_entity_type('Old', 'New')
+            mh.cmd_rename_attribute('New', 'name', 'new_name')
+
+    def test_add_drop_relation_type(self):
+        with self.mh() as (cnx, mh):
+            mh.cmd_add_entity_type('Folder2', auto=False)
+            mh.cmd_add_relation_type('filed_under2')
+            self.assertIn('filed_under2', self.schema)
+            # Old will be missing as it has been renamed into 'New' in the migrated
+            # schema while New hasn't been added here.
+            self.assertEqual(sorted(str(e) for e in self.schema['filed_under2'].subjects()),
+                             sorted(str(e) for e in self.schema.entities()
+                                    if not e.final and e != 'Old'))
+            self.assertEqual(self.schema['filed_under2'].objects(), ('Folder2',))
+            mh.cmd_drop_relation_type('filed_under2')
+            self.assertNotIn('filed_under2', self.schema)
+            # this should not crash
+            mh.cmd_drop_relation_type('filed_under2')
+
+    def test_add_relation_definition_nortype(self):
+        with self.mh() as (cnx, mh):
+            mh.cmd_add_relation_definition('Personne', 'concerne2', 'Affaire')
+            self.assertEqual(self.schema['concerne2'].subjects(),
+                             ('Personne',))
+            self.assertEqual(self.schema['concerne2'].objects(),
+                             ('Affaire', ))
+            self.assertEqual(self.schema['concerne2'].rdef('Personne', 'Affaire').cardinality,
+                              '1*')
+            mh.cmd_add_relation_definition('Personne', 'concerne2', 'Note')
+            self.assertEqual(sorted(self.schema['concerne2'].objects()), ['Affaire', 'Note'])
+            mh.create_entity('Personne', nom=u'tot')
+            mh.create_entity('Affaire')
+            mh.rqlexec('SET X concerne2 Y WHERE X is Personne, Y is Affaire')
+            cnx.commit()
+            mh.cmd_drop_relation_definition('Personne', 'concerne2', 'Affaire')
+            self.assertIn('concerne2', self.schema)
+            mh.cmd_drop_relation_definition('Personne', 'concerne2', 'Note')
+            self.assertNotIn('concerne2', self.schema)
+
+    def test_drop_relation_definition_existant_rtype(self):
+        with self.mh() as (cnx, mh):
+            self.assertEqual(sorted(str(e) for e in self.schema['concerne'].subjects()),
+                              ['Affaire', 'Personne'])
+            self.assertEqual(sorted(str(e) for e in self.schema['concerne'].objects()),
+                              ['Affaire', 'Division', 'Note', 'Societe', 'SubDivision'])
+            mh.cmd_drop_relation_definition('Personne', 'concerne', 'Affaire')
+            self.assertEqual(sorted(str(e) for e in self.schema['concerne'].subjects()),
+                              ['Affaire'])
+            self.assertEqual(sorted(str(e) for e in self.schema['concerne'].objects()),
+                              ['Division', 'Note', 'Societe', 'SubDivision'])
+            mh.cmd_add_relation_definition('Personne', 'concerne', 'Affaire')
+            self.assertEqual(sorted(str(e) for e in self.schema['concerne'].subjects()),
+                              ['Affaire', 'Personne'])
+            self.assertEqual(sorted(str(e) for e in self.schema['concerne'].objects()),
+                              ['Affaire', 'Division', 'Note', 'Societe', 'SubDivision'])
+            # trick: overwrite self.maxeid to avoid deletion of just reintroduced types
+            self.maxeid = cnx.execute('Any MAX(X)')[0][0]
+
+    def test_drop_relation_definition_with_specialization(self):
+        with self.mh() as (cnx, mh):
+            self.assertEqual(sorted(str(e) for e in self.schema['concerne'].subjects()),
+                              ['Affaire', 'Personne'])
+            self.assertEqual(sorted(str(e) for e in self.schema['concerne'].objects()),
+                              ['Affaire', 'Division', 'Note', 'Societe', 'SubDivision'])
+            mh.cmd_drop_relation_definition('Affaire', 'concerne', 'Societe')
+            self.assertEqual(sorted(str(e) for e in self.schema['concerne'].subjects()),
+                              ['Affaire', 'Personne'])
+            self.assertEqual(sorted(str(e) for e in self.schema['concerne'].objects()),
+                              ['Affaire', 'Note'])
+            mh.cmd_add_relation_definition('Affaire', 'concerne', 'Societe')
+            self.assertEqual(sorted(str(e) for e in self.schema['concerne'].subjects()),
+                              ['Affaire', 'Personne'])
+            self.assertEqual(sorted(str(e) for e in self.schema['concerne'].objects()),
+                              ['Affaire', 'Division', 'Note', 'Societe', 'SubDivision'])
+            # trick: overwrite self.maxeid to avoid deletion of just reintroduced types
+            self.maxeid = cnx.execute('Any MAX(X)')[0][0]
+
+    def test_rename_relation(self):
+        self.skipTest('implement me')
+
+    def test_change_relation_props_non_final(self):
+        with self.mh() as (cnx, mh):
+            rschema = self.schema['concerne']
+            card = rschema.rdef('Affaire', 'Societe').cardinality
+            self.assertEqual(card, '**')
+            try:
+                mh.cmd_change_relation_props('Affaire', 'concerne', 'Societe',
+                                             cardinality='?*')
+                card = rschema.rdef('Affaire', 'Societe').cardinality
+                self.assertEqual(card, '?*')
+            finally:
+                mh.cmd_change_relation_props('Affaire', 'concerne', 'Societe',
+                                             cardinality='**')
+
+    def test_change_relation_props_final(self):
+        with self.mh() as (cnx, mh):
+            rschema = self.schema['adel']
+            card = rschema.rdef('Personne', 'String').fulltextindexed
+            self.assertEqual(card, False)
+            try:
+                mh.cmd_change_relation_props('Personne', 'adel', 'String',
+                                             fulltextindexed=True)
+                card = rschema.rdef('Personne', 'String').fulltextindexed
+                self.assertEqual(card, True)
+            finally:
+                mh.cmd_change_relation_props('Personne', 'adel', 'String',
+                                             fulltextindexed=False)
+
+    def test_sync_schema_props_perms_rqlconstraints(self):
+        with self.mh() as (cnx, mh):
+            # Drop one of the RQLConstraint.
+            rdef = self.schema['evaluee'].rdefs[('Personne', 'Note')]
+            oldconstraints = rdef.constraints
+            self.assertIn('S created_by U',
+                          [cstr.expression for cstr in oldconstraints])
+            mh.cmd_sync_schema_props_perms('evaluee', commit=True)
+            newconstraints = rdef.constraints
+            self.assertNotIn('S created_by U',
+                             [cstr.expression for cstr in newconstraints])
+
+            # Drop all RQLConstraint.
+            rdef = self.schema['travaille'].rdefs[('Personne', 'Societe')]
+            oldconstraints = rdef.constraints
+            self.assertEqual(len(oldconstraints), 2)
+            mh.cmd_sync_schema_props_perms('travaille', commit=True)
+            rdef = self.schema['travaille'].rdefs[('Personne', 'Societe')]
+            newconstraints = rdef.constraints
+            self.assertEqual(len(newconstraints), 0)
+
+    @tag('longrun')
+    def test_sync_schema_props_perms(self):
+        with self.mh() as (cnx, mh):
+            nbrqlexpr_start = cnx.execute('Any COUNT(X) WHERE X is RQLExpression')[0][0]
+            migrschema['titre'].rdefs[('Personne', 'String')].order = 7
+            migrschema['adel'].rdefs[('Personne', 'String')].order = 6
+            migrschema['ass'].rdefs[('Personne', 'String')].order = 5
+            migrschema['Personne'].description = 'blabla bla'
+            migrschema['titre'].description = 'usually a title'
+            migrschema['titre'].rdefs[('Personne', 'String')].description = 'title for this person'
+            delete_concerne_rqlexpr = self._rrqlexpr_rset(cnx, 'delete', 'concerne')
+            add_concerne_rqlexpr = self._rrqlexpr_rset(cnx, 'add', 'concerne')
+
+            # make sure properties (e.g. etype descriptions) are synced by the
+            # second call to sync_schema
+            mh.cmd_sync_schema_props_perms(syncprops=False, commit=False)
+            mh.cmd_sync_schema_props_perms(commit=False)
+
+            self.assertEqual(cnx.execute('Any D WHERE X name "Personne", X description D')[0][0],
+                             'blabla bla')
+            self.assertEqual(cnx.execute('Any D WHERE X name "titre", X description D')[0][0],
+                             'usually a title')
+            self.assertEqual(cnx.execute('Any D WHERE X relation_type RT, RT name "titre",'
+                                         'X from_entity FE, FE name "Personne",'
+                                         'X description D')[0][0],
+                             'title for this person')
+            rinorder = [n for n, in cnx.execute(
+                'Any N ORDERBY O,N WHERE X is CWAttribute, X relation_type RT, RT name N,'
+                'X from_entity FE, FE name "Personne",'
+                'X ordernum O')]
+            expected = [u'nom', u'prenom', u'sexe', u'promo', u'ass', u'adel', u'titre',
+                        u'web', u'tel', u'fax', u'datenaiss', u'test', u'tzdatenaiss',
+                        u'description', u'firstname',
+                        u'creation_date', u'cwuri', u'modification_date']
+            self.assertEqual(expected, rinorder)
+
+            # test permissions synchronization ####################################
+            # new rql expr to add note entity
+            eexpr = self._erqlexpr_entity(cnx, 'add', 'Note')
+            self.assertEqual(eexpr.expression,
+                              'X ecrit_part PE, U in_group G, '
+                              'PE require_permission P, P name "add_note", P require_group G')
+            self.assertEqual([et.name for et in eexpr.reverse_add_permission], ['Note'])
+            self.assertEqual(eexpr.reverse_read_permission, ())
+            self.assertEqual(eexpr.reverse_delete_permission, ())
+            self.assertEqual(eexpr.reverse_update_permission, ())
+            self.assertTrue(self._rrqlexpr_rset(cnx, 'add', 'para'))
+            # no rqlexpr to delete para attribute
+            self.assertFalse(self._rrqlexpr_rset(cnx, 'delete', 'para'))
+            # new rql expr to add ecrit_par relation
+            rexpr = self._rrqlexpr_entity(cnx, 'add', 'ecrit_par')
+            self.assertEqual(rexpr.expression,
+                              'O require_permission P, P name "add_note", '
+                              'U in_group G, P require_group G')
+            self.assertEqual([rdef.rtype.name for rdef in rexpr.reverse_add_permission], ['ecrit_par'])
+            self.assertEqual(rexpr.reverse_read_permission, ())
+            self.assertEqual(rexpr.reverse_delete_permission, ())
+            # no more rqlexpr to delete and add travaille relation
+            self.assertFalse(self._rrqlexpr_rset(cnx, 'add', 'travaille'))
+            self.assertFalse(self._rrqlexpr_rset(cnx, 'delete', 'travaille'))
+            # no more rqlexpr to delete and update Societe entity
+            self.assertFalse(self._erqlexpr_rset(cnx, 'update', 'Societe'))
+            self.assertFalse(self._erqlexpr_rset(cnx, 'delete', 'Societe'))
+            # no more rqlexpr to read Affaire entity
+            self.assertFalse(self._erqlexpr_rset(cnx, 'read', 'Affaire'))
+            # rqlexpr to update Affaire entity has been updated
+            eexpr = self._erqlexpr_entity(cnx, 'update', 'Affaire')
+            self.assertEqual(eexpr.expression, 'X concerne S, S owned_by U')
+            # no change for rqlexpr to add and delete Affaire entity
+            self.assertEqual(len(self._erqlexpr_rset(cnx, 'delete', 'Affaire')), 1)
+            self.assertEqual(len(self._erqlexpr_rset(cnx, 'add', 'Affaire')), 1)
+            # no change for rqlexpr to add and delete concerne relation
+            self.assertEqual(len(self._rrqlexpr_rset(cnx, 'delete', 'concerne')),
+                             len(delete_concerne_rqlexpr))
+            self.assertEqual(len(self._rrqlexpr_rset(cnx, 'add', 'concerne')),
+                             len(add_concerne_rqlexpr))
+            # * migrschema involve:
+            #   * 7 erqlexprs deletions (2 in (Affaire + Societe + Note.para) + 1 Note.something
+            #   * 2 rrqlexprs deletions (travaille)
+            #   * 1 update (Affaire update)
+            #   * 2 new (Note add, ecrit_par add)
+            #   * 2 implicit new for attributes (Note.para, Person.test)
+            # remaining orphan rql expr which should be deleted at commit (composite relation)
+            # unattached expressions -> pending deletion on commit
+            self.assertEqual(cnx.execute('Any COUNT(X) WHERE X is RQLExpression, X exprtype "ERQLExpression",'
+                                         'NOT ET1 read_permission X, NOT ET2 add_permission X, '
+                                         'NOT ET3 delete_permission X, NOT ET4 update_permission X')[0][0],
+                              7)
+            self.assertEqual(cnx.execute('Any COUNT(X) WHERE X is RQLExpression, X exprtype "RRQLExpression",'
+                                         'NOT ET1 read_permission X, NOT ET2 add_permission X, '
+                                         'NOT ET3 delete_permission X, NOT ET4 update_permission X')[0][0],
+                              2)
+            # finally
+            self.assertEqual(cnx.execute('Any COUNT(X) WHERE X is RQLExpression')[0][0],
+                             nbrqlexpr_start + 1 + 2 + 2 + 2)
+            cnx.commit()
+            # unique_together test
+            self.assertEqual(len(self.schema.eschema('Personne')._unique_together), 1)
+            self.assertCountEqual(self.schema.eschema('Personne')._unique_together[0],
+                                               ('nom', 'prenom', 'datenaiss'))
+            rset = cnx.execute('Any C WHERE C is CWUniqueTogetherConstraint, C constraint_of ET, ET name "Personne"')
+            self.assertEqual(len(rset), 1)
+            relations = [r.name for r in rset.get_entity(0, 0).relations]
+            self.assertCountEqual(relations, ('nom', 'prenom', 'datenaiss'))
+
+    def _erqlexpr_rset(self, cnx, action, ertype):
+        rql = 'RQLExpression X WHERE ET is CWEType, ET %s_permission X, ET name %%(name)s' % action
+        return cnx.execute(rql, {'name': ertype})
+
+    def _erqlexpr_entity(self, cnx, action, ertype):
+        rset = self._erqlexpr_rset(cnx, action, ertype)
+        self.assertEqual(len(rset), 1)
+        return rset.get_entity(0, 0)
+
+    def _rrqlexpr_rset(self, cnx, action, ertype):
+        rql = 'RQLExpression X WHERE RT is CWRType, RDEF %s_permission X, RT name %%(name)s, RDEF relation_type RT' % action
+        return cnx.execute(rql, {'name': ertype})
+
+    def _rrqlexpr_entity(self, cnx, action, ertype):
+        rset = self._rrqlexpr_rset(cnx, action, ertype)
+        self.assertEqual(len(rset), 1)
+        return rset.get_entity(0, 0)
+
+    def test_set_size_constraint(self):
+        with self.mh() as (cnx, mh):
+            # existing previous value
+            try:
+                mh.cmd_set_size_constraint('CWEType', 'name', 128)
+            finally:
+                mh.cmd_set_size_constraint('CWEType', 'name', 64)
+            # non existing previous value
+            try:
+                mh.cmd_set_size_constraint('CWEType', 'description', 256)
+            finally:
+                mh.cmd_set_size_constraint('CWEType', 'description', None)
+
+    @tag('longrun')
+    def test_add_drop_cube_and_deps(self):
+        with self.mh() as (cnx, mh):
+            schema = self.repo.schema
+            self.assertEqual(sorted((str(s), str(o)) for s, o in schema['see_also'].rdefs),
+                             sorted([('EmailThread', 'EmailThread'), ('Folder', 'Folder'),
+                                     ('Bookmark', 'Bookmark'), ('Bookmark', 'Note'),
+                                     ('Note', 'Note'), ('Note', 'Bookmark')]))
+            try:
+                mh.cmd_drop_cube('fakeemail', removedeps=True)
+                # file was there because it's an email dependancy, should have been removed
+                self.assertNotIn('fakeemail', self.config.cubes())
+                self.assertNotIn(self.config.cube_dir('fakeemail'), self.config.cubes_path())
+                self.assertNotIn('file', self.config.cubes())
+                self.assertNotIn(self.config.cube_dir('file'), self.config.cubes_path())
+                for ertype in ('Email', 'EmailThread', 'EmailPart', 'File',
+                               'sender', 'in_thread', 'reply_to', 'data_format'):
+                    self.assertNotIn(ertype, schema)
+                self.assertEqual(sorted(schema['see_also'].rdefs),
+                                  sorted([('Folder', 'Folder'),
+                                          ('Bookmark', 'Bookmark'),
+                                          ('Bookmark', 'Note'),
+                                          ('Note', 'Note'),
+                                          ('Note', 'Bookmark')]))
+                self.assertEqual(sorted(schema['see_also'].subjects()), ['Bookmark', 'Folder', 'Note'])
+                self.assertEqual(sorted(schema['see_also'].objects()), ['Bookmark', 'Folder', 'Note'])
+                self.assertEqual(cnx.execute('Any X WHERE X pkey "system.version.fakeemail"').rowcount, 0)
+                self.assertEqual(cnx.execute('Any X WHERE X pkey "system.version.file"').rowcount, 0)
+            finally:
+                mh.cmd_add_cube('fakeemail')
+                self.assertIn('fakeemail', self.config.cubes())
+                self.assertIn(self.config.cube_dir('fakeemail'), self.config.cubes_path())
+                self.assertIn('file', self.config.cubes())
+                self.assertIn(self.config.cube_dir('file'), self.config.cubes_path())
+                for ertype in ('Email', 'EmailThread', 'EmailPart', 'File',
+                               'sender', 'in_thread', 'reply_to', 'data_format'):
+                    self.assertIn(ertype, schema)
+                self.assertEqual(sorted(schema['see_also'].rdefs),
+                                 sorted([('EmailThread', 'EmailThread'), ('Folder', 'Folder'),
+                                         ('Bookmark', 'Bookmark'),
+                                         ('Bookmark', 'Note'),
+                                         ('Note', 'Note'),
+                                         ('Note', 'Bookmark')]))
+                self.assertEqual(sorted(schema['see_also'].subjects()), ['Bookmark', 'EmailThread', 'Folder', 'Note'])
+                self.assertEqual(sorted(schema['see_also'].objects()), ['Bookmark', 'EmailThread', 'Folder', 'Note'])
+                from cubes.fakeemail.__pkginfo__ import version as email_version
+                from cubes.file.__pkginfo__ import version as file_version
+                self.assertEqual(cnx.execute('Any V WHERE X value V, X pkey "system.version.fakeemail"')[0][0],
+                                  email_version)
+                self.assertEqual(cnx.execute('Any V WHERE X value V, X pkey "system.version.file"')[0][0],
+                                  file_version)
+                # trick: overwrite self.maxeid to avoid deletion of just reintroduced
+                #        types (and their associated tables!)
+                self.maxeid = cnx.execute('Any MAX(X)')[0][0]
+                # why this commit is necessary is unclear to me (though without it
+                # next test may fail complaining of missing tables
+                cnx.commit()
+
+
+    @tag('longrun')
+    def test_add_drop_cube_no_deps(self):
+        with self.mh() as (cnx, mh):
+            cubes = set(self.config.cubes())
+            schema = self.repo.schema
+            try:
+                mh.cmd_drop_cube('fakeemail')
+                cubes.remove('fakeemail')
+                self.assertNotIn('fakeemail', self.config.cubes())
+                self.assertIn('file', self.config.cubes())
+                for ertype in ('Email', 'EmailThread', 'EmailPart',
+                               'sender', 'in_thread', 'reply_to'):
+                    self.assertNotIn(ertype, schema)
+            finally:
+                mh.cmd_add_cube('fakeemail')
+                self.assertIn('fakeemail', self.config.cubes())
+                # trick: overwrite self.maxeid to avoid deletion of just reintroduced
+                #        types (and their associated tables!)
+                self.maxeid = cnx.execute('Any MAX(X)')[0][0] # XXXXXXX KILL KENNY
+                # why this commit is necessary is unclear to me (though without it
+                # next test may fail complaining of missing tables
+                cnx.commit()
+
+    def test_drop_dep_cube(self):
+        with self.mh() as (cnx, mh):
+            with self.assertRaises(ConfigurationError) as cm:
+                mh.cmd_drop_cube('file')
+            self.assertEqual(str(cm.exception), "can't remove cube file, used as a dependency")
+
+    @tag('longrun')
+    def test_introduce_base_class(self):
+        with self.mh() as (cnx, mh):
+            mh.cmd_add_entity_type('Para')
+            self.assertEqual(sorted(et.type for et in self.schema['Para'].specialized_by()),
+                             ['Note'])
+            self.assertEqual(self.schema['Note'].specializes().type, 'Para')
+            mh.cmd_add_entity_type('Text')
+            self.assertEqual(sorted(et.type for et in self.schema['Para'].specialized_by()),
+                              ['Note', 'Text'])
+            self.assertEqual(self.schema['Text'].specializes().type, 'Para')
+            # test columns have been actually added
+            text = cnx.execute('INSERT Text X: X para "hip", X summary "hop", X newattr "momo"').get_entity(0, 0)
+            note = cnx.execute('INSERT Note X: X para "hip", X shortpara "hop", X newattr "momo", X unique_id "x"').get_entity(0, 0)
+            aff = cnx.execute('INSERT Affaire X').get_entity(0, 0)
+            self.assertTrue(cnx.execute('SET X newnotinlined Y WHERE X eid %(x)s, Y eid %(y)s',
+                                         {'x': text.eid, 'y': aff.eid}))
+            self.assertTrue(cnx.execute('SET X newnotinlined Y WHERE X eid %(x)s, Y eid %(y)s',
+                                         {'x': note.eid, 'y': aff.eid}))
+            self.assertTrue(cnx.execute('SET X newinlined Y WHERE X eid %(x)s, Y eid %(y)s',
+                                         {'x': text.eid, 'y': aff.eid}))
+            self.assertTrue(cnx.execute('SET X newinlined Y WHERE X eid %(x)s, Y eid %(y)s',
+                                         {'x': note.eid, 'y': aff.eid}))
+            # XXX remove specializes by ourselves, else tearDown fails when removing
+            # Para because of Note inheritance. This could be fixed by putting the
+            # MemSchemaCWETypeDel(session, name) operation in the
+            # after_delete_entity(CWEType) hook, since in that case the MemSchemaSpecializesDel
+            # operation would be removed before, but I'm not sure this is a desired behaviour.
+            #
+            # also we need more tests about introducing/removing base classes or
+            # specialization relationship...
+            cnx.execute('DELETE X specializes Y WHERE Y name "Para"')
+            cnx.commit()
+            self.assertEqual(sorted(et.type for et in self.schema['Para'].specialized_by()),
+                             [])
+            self.assertEqual(self.schema['Note'].specializes(), None)
+            self.assertEqual(self.schema['Text'].specializes(), None)
+
+    def test_add_symmetric_relation_type(self):
+        with self.mh() as (cnx, mh):
+            self.assertFalse(self.table_sql(mh, 'same_as_relation'))
+            mh.cmd_add_relation_type('same_as')
+            self.assertTrue(self.table_sql(mh, 'same_as_relation'))
+
+    def test_change_attribute_type(self):
+        with self.mh() as (cnx, mh):
+            mh.cmd_create_entity('Societe', tel=1)
+            mh.commit()
+            mh.change_attribute_type('Societe', 'tel', 'Float')
+            self.assertNotIn(('Societe', 'Int'), self.schema['tel'].rdefs)
+            self.assertIn(('Societe', 'Float'), self.schema['tel'].rdefs)
+            self.assertEqual(self.schema['tel'].rdefs[('Societe', 'Float')].object, 'Float')
+            tel = mh.rqlexec('Any T WHERE X tel T')[0][0]
+            self.assertEqual(tel, 1.0)
+            self.assertIsInstance(tel, float)
+
+    def test_drop_required_inlined_relation(self):
+        with self.mh() as (cnx, mh):
+            bob = mh.cmd_create_entity('Personne', nom=u'bob')
+            note = mh.cmd_create_entity('Note', ecrit_par=bob)
+            mh.commit()
+            rdef = mh.fs_schema.rschema('ecrit_par').rdefs[('Note', 'Personne')]
+            with tempattr(rdef, 'cardinality', '1*'):
+                mh.sync_schema_props_perms('ecrit_par', syncperms=False)
+            mh.cmd_drop_relation_type('ecrit_par')
+            self.assertNotIn('%secrit_par' % SQL_PREFIX,
+                             self.table_schema(mh, '%sPersonne' % SQL_PREFIX))
+
+    def test_drop_inlined_rdef_delete_data(self):
+        with self.mh() as (cnx, mh):
+            note = mh.cmd_create_entity('Note', ecrit_par=cnx.user.eid)
+            mh.commit()
+            mh.drop_relation_definition('Note', 'ecrit_par', 'CWUser')
+            self.assertFalse(mh.sqlexec('SELECT * FROM cw_Note WHERE cw_ecrit_par IS NOT NULL'))
+
+class MigrationCommandsComputedTC(MigrationTC):
+    """ Unit tests for computed relations and attributes
+    """
+    appid = 'datacomputed'
+
+    def setUp(self):
+        MigrationTC.setUp(self)
+        # ensure vregistry is reloaded, needed by generated hooks for computed
+        # attributes
+        self.repo.vreg.set_schema(self.repo.schema)
+
+    def test_computed_relation_add_relation_definition(self):
+        self.assertNotIn('works_for', self.schema)
+        with self.mh() as (cnx, mh):
+            with self.assertRaises(ExecutionError) as exc:
+                mh.cmd_add_relation_definition('Employee', 'works_for', 'Company')
+        self.assertEqual(str(exc.exception),
+                         'Cannot add a relation definition for a computed '
+                         'relation (works_for)')
+
+    def test_computed_relation_drop_relation_definition(self):
+        self.assertIn('notes', self.schema)
+        with self.mh() as (cnx, mh):
+            with self.assertRaises(ExecutionError) as exc:
+                mh.cmd_drop_relation_definition('Company', 'notes', 'Note')
+        self.assertEqual(str(exc.exception),
+                         'Cannot drop a relation definition for a computed '
+                         'relation (notes)')
+
+    def test_computed_relation_add_relation_type(self):
+        self.assertNotIn('works_for', self.schema)
+        with self.mh() as (cnx, mh):
+            mh.cmd_add_relation_type('works_for')
+            self.assertIn('works_for', self.schema)
+            self.assertEqual(self.schema['works_for'].rule,
+                             'O employees S, NOT EXISTS (O associates S)')
+            self.assertEqual(self.schema['works_for'].objects(), ('Company',))
+            self.assertEqual(self.schema['works_for'].subjects(), ('Employee',))
+            self.assertFalse(self.table_sql(mh, 'works_for_relation'))
+            e = cnx.create_entity('Employee')
+            a = cnx.create_entity('Employee')
+            cnx.create_entity('Company', employees=e, associates=a)
+            cnx.commit()
+            company = cnx.execute('Company X').get_entity(0, 0)
+            self.assertEqual([e.eid],
+                             [x.eid for x in company.reverse_works_for])
+            mh.rollback()
+
+    def test_computed_relation_drop_relation_type(self):
+        self.assertIn('notes', self.schema)
+        with self.mh() as (cnx, mh):
+            mh.cmd_drop_relation_type('notes')
+        self.assertNotIn('notes', self.schema)
+
+    def test_computed_relation_sync_schema_props_perms(self):
+        self.assertIn('whatever', self.schema)
+        with self.mh() as (cnx, mh):
+            mh.cmd_sync_schema_props_perms('whatever')
+            self.assertEqual(self.schema['whatever'].rule,
+                             'S employees E, O associates E')
+            self.assertEqual(self.schema['whatever'].objects(), ('Company',))
+            self.assertEqual(self.schema['whatever'].subjects(), ('Company',))
+            self.assertFalse(self.table_sql(mh, 'whatever_relation'))
+
+    def test_computed_relation_sync_schema_props_perms_security(self):
+        with self.mh() as (cnx, mh):
+            rdef = next(iter(self.schema['perm_changes'].rdefs.values()))
+            self.assertEqual(rdef.permissions,
+                             {'add': (), 'delete': (),
+                              'read': ('managers', 'users')})
+            mh.cmd_sync_schema_props_perms('perm_changes')
+            self.assertEqual(self.schema['perm_changes'].permissions,
+                             {'read': ('managers',)})
+            rdef = next(iter(self.schema['perm_changes'].rdefs.values()))
+            self.assertEqual(rdef.permissions,
+                             {'add': (), 'delete': (),
+                              'read': ('managers',)})
+
+    def test_computed_relation_sync_schema_props_perms_on_rdef(self):
+        self.assertIn('whatever', self.schema)
+        with self.mh() as (cnx, mh):
+            with self.assertRaises(ExecutionError) as exc:
+                mh.cmd_sync_schema_props_perms(
+                    ('Company', 'whatever', 'Person'))
+        self.assertEqual(str(exc.exception),
+                         'Cannot synchronize a relation definition for a computed '
+                         'relation (whatever)')
+
+    def test_computed_relation_rename_relation_type(self):
+        with self.mh() as (cnx, mh):
+            mh.cmd_rename_relation_type('to_be_renamed', 'renamed')
+        self.assertIn('renamed', self.schema)
+        self.assertNotIn('to_be_renamed', self.schema)
+
+    # computed attributes migration ############################################
+
+    def setup_add_score(self):
+        with self.admin_access.client_cnx() as cnx:
+            assert not cnx.execute('Company X')
+            c = cnx.create_entity('Company')
+            e1 = cnx.create_entity('Employee', reverse_employees=c)
+            cnx.create_entity('Note', note=2, concerns=e1)
+            e2 = cnx.create_entity('Employee', reverse_employees=c)
+            cnx.create_entity('Note', note=4, concerns=e2)
+            cnx.commit()
+
+    def assert_score_initialized(self, mh):
+        self.assertEqual(self.schema['score'].rdefs['Company', 'Float'].formula,
+                         'Any AVG(NN) WHERE X employees E, N concerns E, N note NN')
+        fields = self.table_schema(mh, '%sCompany' % SQL_PREFIX)
+        self.assertEqual(fields['%sscore' % SQL_PREFIX], ('double precision', None))
+        self.assertEqual([[3.0]],
+                         mh.rqlexec('Any CS WHERE C score CS, C is Company').rows)
+
+    def test_computed_attribute_add_relation_type(self):
+        self.assertNotIn('score', self.schema)
+        self.setup_add_score()
+        with self.mh() as (cnx, mh):
+            mh.cmd_add_relation_type('score')
+            self.assertIn('score', self.schema)
+            self.assertEqual(self.schema['score'].objects(), ('Float',))
+            self.assertEqual(self.schema['score'].subjects(), ('Company',))
+            self.assert_score_initialized(mh)
+
+    def test_computed_attribute_add_attribute(self):
+        self.assertNotIn('score', self.schema)
+        self.setup_add_score()
+        with self.mh() as (cnx, mh):
+            mh.cmd_add_attribute('Company', 'score')
+            self.assertIn('score', self.schema)
+            self.assert_score_initialized(mh)
+
+    def assert_computed_attribute_dropped(self):
+        self.assertNotIn('note20', self.schema)
+        with self.mh() as (cnx, mh):
+            fields = self.table_schema(mh, '%sNote' % SQL_PREFIX)
+        self.assertNotIn('%snote20' % SQL_PREFIX, fields)
+
+    def test_computed_attribute_drop_type(self):
+        self.assertIn('note20', self.schema)
+        with self.mh() as (cnx, mh):
+            mh.cmd_drop_relation_type('note20')
+        self.assert_computed_attribute_dropped()
+
+    def test_computed_attribute_drop_relation_definition(self):
+        self.assertIn('note20', self.schema)
+        with self.mh() as (cnx, mh):
+            mh.cmd_drop_relation_definition('Note', 'note20', 'Int')
+        self.assert_computed_attribute_dropped()
+
+    def test_computed_attribute_drop_attribute(self):
+        self.assertIn('note20', self.schema)
+        with self.mh() as (cnx, mh):
+            mh.cmd_drop_attribute('Note', 'note20')
+        self.assert_computed_attribute_dropped()
+
+    def test_computed_attribute_sync_schema_props_perms_rtype(self):
+        self.assertIn('note100', self.schema)
+        with self.mh() as (cnx, mh):
+            mh.cmd_sync_schema_props_perms('note100')
+        rdef = self.schema['note100'].rdefs['Note', 'Int']
+        self.assertEqual(rdef.formula_select.as_string(),
+                         'Any (N * 100) WHERE X note N, X is Note')
+        self.assertEqual(rdef.formula, 'Any N*100 WHERE X note N')
+
+    def test_computed_attribute_sync_schema_props_perms_rdef(self):
+        self.setup_add_score()
+        with self.mh() as (cnx, mh):
+            mh.cmd_sync_schema_props_perms(('Note', 'note100', 'Int'))
+            self.assertEqual([[200], [400]],
+                             cnx.execute('Any N ORDERBY N WHERE X note100 N').rows)
+            self.assertEqual([[300]],
+                             cnx.execute('Any CS WHERE C score100 CS, C is Company').rows)
+
+
+if __name__ == '__main__':
+    unittest_main()