[win32] fix deadlock occuring on the sequence tables with SQLServer
actually, this deadlock would occur with any db backend other that PostgreSQL
as the previous code was heavily relying on PG's SEQUENCE facility, not available
elsewhere.
Deadlock description:
Thread1 starts creating entities (and therefore calls create_eid):
-> this creates a DB-level lock on the entities_id_seq table, which
will last until end of transaction
Thread2 calls create_eid, which acquires the Python lock object, but
updating the entities_id_seq is held by the DB lock
Thread1 wants to create a new entity, calls create_eid, and is stuck by
the Python lock object held by Thread2.
Solution:
use a separate connection to read and write the entities_id_seq table.
# copyright 2003-2010 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/>.
"""some utilities to ease repository testing
This module contains functions to initialize a new repository.
"""
__docformat__ = "restructuredtext en"
from pprint import pprint
from logilab.common.decorators import clear_cache
def tuplify(list):
for i in range(len(list)):
if type(list[i]) is not type(()):
list[i] = tuple(list[i])
return list
def snippet_cmp(a, b):
a = (a[0], [e.expression for e in a[1]])
b = (b[0], [e.expression for e in b[1]])
return cmp(a, b)
def test_plan(self, rql, expected, kwargs=None):
plan = self._prepare_plan(rql, kwargs)
self.planner.build_plan(plan)
try:
self.assertEquals(len(plan.steps), len(expected),
'expected %s steps, got %s' % (len(expected), len(plan.steps)))
# step order is important
for i, step in enumerate(plan.steps):
compare_steps(self, step.test_repr(), expected[i])
except AssertionError:
pprint([step.test_repr() for step in plan.steps])
raise
def compare_steps(self, step, expected):
try:
self.assertEquals(step[0], expected[0], 'expected step type %s, got %s' % (expected[0], step[0]))
if len(step) > 2 and isinstance(step[1], list) and isinstance(expected[1], list):
queries, equeries = step[1], expected[1]
self.assertEquals(len(queries), len(equeries),
'expected %s queries, got %s' % (len(equeries), len(queries)))
for i, (rql, sol) in enumerate(queries):
self.assertEquals(rql, equeries[i][0])
self.assertEquals(sorted(sol), sorted(equeries[i][1]))
idx = 2
else:
idx = 1
self.assertEquals(step[idx:-1], expected[idx:-1],
'expected step characteristic \n%s\n, got\n%s' % (expected[1:-1], step[1:-1]))
self.assertEquals(len(step[-1]), len(expected[-1]),
'got %s child steps, expected %s' % (len(step[-1]), len(expected[-1])))
except AssertionError:
print 'error on step ',
pprint(step[:-1])
raise
children = step[-1]
if step[0] in ('UnionFetchStep', 'UnionStep'):
# sort children
children = sorted(children)
expectedchildren = sorted(expected[-1])
else:
expectedchildren = expected[-1]
for i, substep in enumerate(children):
compare_steps(self, substep, expectedchildren[i])
class DumbOrderedDict(list):
def __iter__(self):
return self.iterkeys()
def __contains__(self, key):
return key in self.iterkeys()
def __getitem__(self, key):
for key_, value in list.__iter__(self):
if key == key_:
return value
raise KeyError(key)
def iterkeys(self):
return (x for x, y in list.__iter__(self))
def iteritems(self):
return (x for x in list.__iter__(self))
def items(self):
return [x for x in list.__iter__(self)]
class DumbOrderedDict2(object):
def __init__(self, origdict, sortkey):
self.origdict = origdict
self.sortkey = sortkey
def __getattr__(self, attr):
return getattr(self.origdict, attr)
def __iter__(self):
return iter(sorted(self.origdict, key=self.sortkey))
def schema_eids_idx(schema):
"""return a dictionary mapping schema types to their eids so we can reread
it from the fs instead of the db (too costly) between tests
"""
schema_eids = {}
for x in schema.entities():
schema_eids[x] = x.eid
for x in schema.relations():
schema_eids[x] = x.eid
for rdef in x.rdefs.itervalues():
schema_eids[(rdef.subject, rdef.rtype, rdef.object)] = rdef.eid
return schema_eids
def restore_schema_eids_idx(schema, schema_eids):
"""rebuild schema eid index"""
for x in schema.entities():
x.eid = schema_eids[x]
schema._eid_index[x.eid] = x
for x in schema.relations():
x.eid = schema_eids[x]
schema._eid_index[x.eid] = x
for rdef in x.rdefs.itervalues():
rdef.eid = schema_eids[(rdef.subject, rdef.rtype, rdef.object)]
schema._eid_index[rdef.eid] = rdef
from logilab.common.testlib import TestCase
from rql import RQLHelper
from cubicweb.devtools.fake import FakeRepo, FakeSession
from cubicweb.server import set_debug
from cubicweb.server.querier import QuerierHelper
from cubicweb.server.session import Session
from cubicweb.server.sources.rql2sql import remove_unused_solutions
class RQLGeneratorTC(TestCase):
schema = None # set this in concret test
def setUp(self):
self.repo = FakeRepo(self.schema)
self.rqlhelper = RQLHelper(self.schema, special_relations={'eid': 'uid',
'has_text': 'fti'})
self.qhelper = QuerierHelper(self.repo, self.schema)
ExecutionPlan._check_permissions = _dummy_check_permissions
rqlannotation._select_principal = _select_principal
def tearDown(self):
ExecutionPlan._check_permissions = _orig_check_permissions
rqlannotation._select_principal = _orig_select_principal
def set_debug(self, debug):
set_debug(debug)
def _prepare(self, rql):
#print '******************** prepare', rql
union = self.rqlhelper.parse(rql)
#print '********* parsed', union.as_string()
self.rqlhelper.compute_solutions(union)
#print '********* solutions', solutions
self.rqlhelper.simplify(union)
#print '********* simplified', union.as_string()
plan = self.qhelper.plan_factory(union, {}, FakeSession(self.repo))
plan.preprocess(union)
for select in union.children:
select.solutions.sort()
#print '********* ppsolutions', solutions
return union
class BaseQuerierTC(TestCase):
repo = None # set this in concret test
def setUp(self):
self.o = self.repo.querier
self.session = self.repo._sessions.values()[0]
self.ueid = self.session.user.eid
assert self.ueid != -1
self.repo._type_source_cache = {} # clear cache
self.pool = self.session.set_pool()
self.maxeid = self.get_max_eid()
do_monkey_patch()
self._dumb_sessions = []
def get_max_eid(self):
return self.session.execute('Any MAX(X)')[0][0]
def cleanup(self):
self.session.set_pool()
self.session.execute('DELETE Any X WHERE X eid > %s' % self.maxeid)
def tearDown(self):
undo_monkey_patch()
self.session.rollback()
self.cleanup()
self.commit()
# properly close dumb sessions
for session in self._dumb_sessions:
session.rollback()
session.close()
self.repo._free_pool(self.pool)
assert self.session.user.eid != -1
def set_debug(self, debug):
set_debug(debug)
def _rqlhelper(self):
rqlhelper = self.repo.vreg.rqlhelper
# reset uid_func so it don't try to get type from eids
rqlhelper._analyser.uid_func = None
rqlhelper._analyser.uid_func_mapping = {}
return rqlhelper
def _prepare_plan(self, rql, kwargs=None, simplify=True):
rqlhelper = self._rqlhelper()
rqlst = rqlhelper.parse(rql)
rqlhelper.compute_solutions(rqlst, kwargs=kwargs)
if simplify:
rqlhelper.simplify(rqlst)
for select in rqlst.children:
select.solutions.sort()
return self.o.plan_factory(rqlst, kwargs, self.session)
def _prepare(self, rql, kwargs=None):
plan = self._prepare_plan(rql, kwargs, simplify=False)
plan.preprocess(plan.rqlst)
rqlst = plan.rqlst.children[0]
rqlst.solutions = remove_unused_solutions(rqlst, rqlst.solutions, {}, self.repo.schema)[0]
return rqlst
def user_groups_session(self, *groups):
"""lightweight session using the current user with hi-jacked groups"""
# use self.session.user.eid to get correct owned_by relation, unless explicit eid
u = self.repo._build_user(self.session, self.session.user.eid)
u._groups = set(groups)
s = Session(u, self.repo)
s._threaddata.pool = self.pool
# register session to ensure it gets closed
self._dumb_sessions.append(s)
return s
def execute(self, rql, args=None, build_descr=True):
return self.o.execute(self.session, rql, args, build_descr)
def commit(self):
self.session.commit()
self.session.set_pool()
class BasePlannerTC(BaseQuerierTC):
newsources = 0
def setup(self):
clear_cache(self.repo, 'rel_type_sources')
clear_cache(self.repo, 'rel_type_sources')
clear_cache(self.repo, 'can_cross_relation')
clear_cache(self.repo, 'is_multi_sources_relation')
# XXX source_defs
self.o = self.repo.querier
self.session = self.repo._sessions.values()[0]
self.pool = self.session.set_pool()
self.schema = self.o.schema
self.sources = self.o._repo.sources
self.system = self.sources[-1]
do_monkey_patch()
self._dumb_sessions = [] # by hi-jacked parent setup
def add_source(self, sourcecls, uri):
self.sources.append(sourcecls(self.repo, self.o.schema,
{'uri': uri}))
self.repo.sources_by_uri[uri] = self.sources[-1]
setattr(self, uri, self.sources[-1])
self.newsources += 1
def tearDown(self):
while self.newsources:
source = self.sources.pop(-1)
del self.repo.sources_by_uri[source.uri]
self.newsources -= 1
undo_monkey_patch()
for session in self._dumb_sessions:
session._threaddata.pool = None
session.close()
def _prepare_plan(self, rql, kwargs=None):
rqlst = self.o.parse(rql, annotate=True)
self.o.solutions(self.session, rqlst, kwargs)
if rqlst.TYPE == 'select':
self.repo.vreg.rqlhelper.annotate(rqlst)
for select in rqlst.children:
select.solutions.sort()
else:
rqlst.solutions.sort()
return self.o.plan_factory(rqlst, kwargs, self.session)
# monkey patch some methods to get predicatable results #######################
from cubicweb.rqlrewrite import RQLRewriter
_orig_insert_snippets = RQLRewriter.insert_snippets
_orig_build_variantes = RQLRewriter.build_variantes
def _insert_snippets(self, snippets, varexistsmap=None):
_orig_insert_snippets(self, sorted(snippets, snippet_cmp), varexistsmap)
def _build_variantes(self, newsolutions):
variantes = _orig_build_variantes(self, newsolutions)
sortedvariantes = []
for variante in variantes:
orderedkeys = sorted((k[1], k[2], v) for k, v in variante.iteritems())
variante = DumbOrderedDict(sorted(variante.iteritems(),
lambda a, b: cmp((a[0][1],a[0][2],a[1]),
(b[0][1],b[0][2],b[1]))))
sortedvariantes.append( (orderedkeys, variante) )
return [v for ok, v in sorted(sortedvariantes)]
from cubicweb.server.querier import ExecutionPlan
_orig_check_permissions = ExecutionPlan._check_permissions
_orig_init_temp_table = ExecutionPlan.init_temp_table
def _check_permissions(*args, **kwargs):
res, restricted = _orig_check_permissions(*args, **kwargs)
res = DumbOrderedDict(sorted(res.iteritems(), lambda a, b: cmp(a[1], b[1])))
return res, restricted
def _dummy_check_permissions(self, rqlst):
return {(): rqlst.solutions}, set()
def _init_temp_table(self, table, selection, solution):
if self.tablesinorder is None:
tablesinorder = self.tablesinorder = {}
else:
tablesinorder = self.tablesinorder
if not table in tablesinorder:
tablesinorder[table] = 'table%s' % len(tablesinorder)
return _orig_init_temp_table(self, table, selection, solution)
from cubicweb.server import rqlannotation
_orig_select_principal = rqlannotation._select_principal
def _select_principal(scope, relations):
return _orig_select_principal(scope, relations,
_sort=lambda rels: sorted(rels, key=lambda x: x.r_type))
try:
from cubicweb.server.msplanner import PartPlanInformation
except ImportError:
class PartPlanInformation(object):
def merge_input_maps(self, *args):
pass
def _choose_term(self, sourceterms):
pass
_orig_merge_input_maps = PartPlanInformation.merge_input_maps
_orig_choose_term = PartPlanInformation._choose_term
def _merge_input_maps(*args):
return sorted(_orig_merge_input_maps(*args))
def _choose_term(self, sourceterms):
# predictable order for test purpose
def get_key(x):
try:
# variable
return x.name
except AttributeError:
try:
# relation
return x.r_type
except AttributeError:
# const
return x.value
return _orig_choose_term(self, DumbOrderedDict2(sourceterms, get_key))
def do_monkey_patch():
RQLRewriter.insert_snippets = _insert_snippets
RQLRewriter.build_variantes = _build_variantes
ExecutionPlan._check_permissions = _check_permissions
ExecutionPlan.tablesinorder = None
ExecutionPlan.init_temp_table = _init_temp_table
PartPlanInformation.merge_input_maps = _merge_input_maps
PartPlanInformation._choose_term = _choose_term
def undo_monkey_patch():
RQLRewriter.insert_snippets = _orig_insert_snippets
RQLRewriter.build_variantes = _orig_build_variantes
ExecutionPlan._check_permissions = _orig_check_permissions
ExecutionPlan.init_temp_table = _orig_init_temp_table
PartPlanInformation.merge_input_maps = _orig_merge_input_maps
PartPlanInformation._choose_term = _orig_choose_term