[multi-sources-removal] Drop pyrorql and zmqrql sources
After a few years experementing "true" multi-sources, we're now
moving to "copy-based" source à la datafeed.
As pyro and zmq sources have no more known customers and the related
code is in the way of future refactoring of cubicweb's core, we decided
to drop support for those sources without backward compatibility.
If you're still using a zmqrql or pyrorql source and you want to upgrade,
ask support to move it to datafeed using a pre-3.19 version first.
Related to #2919300 (first step)
# copyright 2003-2013 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/>.
"""this module contains base classes and utilities for cubicweb tests"""
__docformat__ = "restructuredtext en"
import sys
import re
import urlparse
from os.path import dirname, join, abspath
from urllib import unquote
from math import log
from contextlib import contextmanager
from warnings import warn
from types import NoneType
from itertools import chain
import yams.schema
from logilab.common.testlib import TestCase, InnerTest, Tags
from logilab.common.pytest import nocoverage, pause_tracing, resume_tracing
from logilab.common.debugger import Debugger
from logilab.common.umessage import message_from_string
from logilab.common.decorators import cached, classproperty, clear_cache, iclassmethod
from logilab.common.deprecation import deprecated, class_deprecated
from logilab.common.shellutils import getlogin
from cubicweb import ValidationError, NoSelectableObject, AuthenticationError
from cubicweb import cwconfig, dbapi, devtools, web, server, repoapi
from cubicweb.utils import json
from cubicweb.sobjects import notification
from cubicweb.web import Redirect, application
from cubicweb.server.hook import SendMailOp
from cubicweb.server.session import Session
from cubicweb.devtools import SYSTEM_ENTITIES, SYSTEM_RELATIONS, VIEW_VALIDATORS
from cubicweb.devtools import fake, htmlparser, DEFAULT_EMPTY_DB_ID
from cubicweb.utils import json
# low-level utilities ##########################################################
class CubicWebDebugger(Debugger):
"""special debugger class providing a 'view' function which saves some
html into a temporary file and open a web browser to examinate it.
"""
def do_view(self, arg):
import webbrowser
data = self._getval(arg)
with file('/tmp/toto.html', 'w') as toto:
toto.write(data)
webbrowser.open('file:///tmp/toto.html')
def line_context_filter(line_no, center, before=3, after=None):
"""return true if line are in context
if after is None: after = before
"""
if after is None:
after = before
return center - before <= line_no <= center + after
def unprotected_entities(schema, strict=False):
"""returned a set of each non final entity type, excluding "system" entities
(eg CWGroup, CWUser...)
"""
if strict:
protected_entities = yams.schema.BASE_TYPES
else:
protected_entities = yams.schema.BASE_TYPES.union(SYSTEM_ENTITIES)
return set(schema.entities()) - protected_entities
class JsonValidator(object):
def parse_string(self, data):
return json.loads(data)
# email handling, to test emails sent by an application ########################
MAILBOX = []
class Email(object):
"""you'll get instances of Email into MAILBOX during tests that trigger
some notification.
* `msg` is the original message object
* `recipients` is a list of email address which are the recipients of this
message
"""
def __init__(self, fromaddr, recipients, msg):
self.fromaddr = fromaddr
self.recipients = recipients
self.msg = msg
@property
def message(self):
return message_from_string(self.msg)
@property
def subject(self):
return self.message.get('Subject')
@property
def content(self):
return self.message.get_payload(decode=True)
def __repr__(self):
return '<Email to %s with subject %s>' % (','.join(self.recipients),
self.message.get('Subject'))
# the trick to get email into MAILBOX instead of actually sent: monkey patch
# cwconfig.SMTP object
class MockSMTP:
def __init__(self, server, port):
pass
def close(self):
pass
def sendmail(self, fromaddr, recipients, msg):
MAILBOX.append(Email(fromaddr, recipients, msg))
cwconfig.SMTP = MockSMTP
class TestCaseConnectionProxy(object):
"""thin wrapper around `cubicweb.dbapi.Connection` context-manager
used in CubicWebTC (cf. `cubicweb.devtools.testlib.CubicWebTC.login` method)
It just proxies to the default connection context manager but
restores the original connection on exit.
"""
def __init__(self, testcase, cnx):
self.testcase = testcase
self.cnx = cnx
def __getattr__(self, attrname):
return getattr(self.cnx, attrname)
def __enter__(self):
# already open
return self.cnx
def __exit__(self, exctype, exc, tb):
try:
return self.cnx.__exit__(exctype, exc, tb)
finally:
self.testcase.restore_connection()
# Repoaccess utility ###############################################3###########
class RepoAccess(object):
"""An helper to easily create object to access the repo as a specific user
Each RepoAccess have it own session.
A repo access can create three type of object:
.. automethod:: cubicweb.testlib.RepoAccess.repo_cnx
.. automethod:: cubicweb.testlib.RepoAccess.client_cnx
.. automethod:: cubicweb.testlib.RepoAccess.web_request
The RepoAccess need to be closed to destroy the associated Session.
TestCase usually take care of this aspect for the user.
.. automethod:: cubicweb.testlib.RepoAccess.close
"""
def __init__(self, repo, login, requestcls):
self._repo = repo
self._login = login
self.requestcls = requestcls
# opening session
#
# XXX this very hackish code should be cleaned and move on repo.
with repo.internal_cnx() as cnx:
rset = cnx.execute('CWUser U WHERE U login %(u)s', {'u': login})
user = rset.get_entity(0, 0)
user.groups
user.properties
self._session = Session(user, repo)
repo._sessions[self._session.id] = self._session
self._session.user._cw = self._session
@ contextmanager
def repo_cnx(self):
"""Context manager returning a server side connection for the user"""
with self._session.new_cnx() as cnx:
yield cnx
@ contextmanager
def client_cnx(self):
"""Context manager returning a client side connection for the user"""
with repoapi.ClientConnection(self._session) as cnx:
yield cnx
@ contextmanager
def web_request(self, url=None, headers={}, **kwargs):
"""Context manager returning a web request pre-linked to a client cnx
To commit and rollback use::
req.cnx.commit()
req.cnx.rolback()
"""
req = self.requestcls(self._repo.vreg, url=url, headers=headers, form=kwargs)
clt_cnx = repoapi.ClientConnection(self._session)
req.set_cnx(clt_cnx)
with clt_cnx:
yield req
def close(self):
"""Close the session associated to the RepoAccess"""
if self._session is not None:
self._repo.close(self._session.id)
self._session = None
# base class for cubicweb tests requiring a full cw environments ###############
class CubicWebTC(TestCase):
"""abstract class for test using an apptest environment
attributes:
* `vreg`, the vregistry
* `schema`, self.vreg.schema
* `config`, cubicweb configuration
* `cnx`, dbapi connection to the repository using an admin user
* `session`, server side session associated to `cnx`
* `app`, the cubicweb publisher (for web testing)
* `repo`, the repository object
* `admlogin`, login of the admin user
* `admpassword`, password of the admin user
* `shell`, create and use shell environment
"""
appid = 'data'
configcls = devtools.ApptestConfiguration
tags = TestCase.tags | Tags('cubicweb', 'cw_repo')
test_db_id = DEFAULT_EMPTY_DB_ID
_cnxs = set() # establised connection
# stay on connection for leak detection purpose
def __init__(self, *args, **kwargs):
self._admin_session = None
self._admin_clt_cnx = None
self._current_session = None
self._current_clt_cnx = None
self.repo = None
self._open_access = set()
super(CubicWebTC, self).__init__(*args, **kwargs)
# repository connection handling ###########################################
def new_access(self, login):
"""provide a new RepoAccess object for a given user
The access is automatically closed at the end of the test."""
access = RepoAccess(self.repo, login, self.requestcls)
self._open_access.add(access)
return access
def _close_access(self):
while self._open_access:
self._open_access.pop().close()
@deprecated('[3.19] explicitly use RepoAccess object in test instead')
def set_cnx(self, cnx):
""""""
# XXX we want to deprecate this
assert getattr(cnx, '_session', None) is not None
if cnx is self._admin_clt_cnx:
self._pop_custom_cnx()
else:
self._cnxs.add(cnx) # register the cns to make sure it is removed
self._current_session = cnx._session
self._current_clt_cnx = cnx
@property
@deprecated('[3.19] explicitly use RepoAccess object in test instead')
def cnx(self):
# XXX we want to deprecate this
clt_cnx = self._current_clt_cnx
if clt_cnx is None:
clt_cnx = self._admin_clt_cnx
return clt_cnx
def _close_cnx(self):
"""ensure that all cnx used by a test have been closed"""
for cnx in list(self._cnxs):
if cnx._open and not cnx._session.closed:
cnx.rollback()
cnx.close()
self._cnxs.remove(cnx)
@property
@deprecated('[3.19] explicitly use RepoAccess object in test instead')
def session(self):
"""return current server side session"""
# XXX We want to use a srv_connection instead and deprecate this
# property
session = self._current_session
if session is None:
session = self._admin_session
# bypassing all sanity to use the same repo cnx in the session we
# can't call set_cnx as the Connection is not managed by the
# session.
session._Session__threaddata.cnx = self._admin_clt_cnx._cnx
session.set_cnxset()
return session
@property
@deprecated('[3.19] explicitly use RepoAccess object in test instead')
def websession(self):
return self.session
@property
@deprecated('[3.19] explicitly use RepoAccess object in test instead')
def adminsession(self):
"""return current server side session (using default manager account)"""
return self._admin_session
@deprecated('[3.19] explicitly use RepoAccess object in test instead')
def login(self, login, **kwargs):
"""return a connection for the given login/password"""
__ = kwargs.pop('autoclose', True) # not used anymore
if login == self.admlogin:
# definitly don't want autoclose when used as a context manager
clt_cnx = repoapi.ClientConnection(self._admin_session)
else:
if not kwargs:
kwargs['password'] = str(login)
clt_cnx = repoapi.connect(self.repo, login, **kwargs)
self.set_cnx(clt_cnx)
clt_cnx.__enter__()
return TestCaseConnectionProxy(self, clt_cnx)
@deprecated('[3.19] explicitly use RepoAccess object in test instead')
def restore_connection(self):
self._pop_custom_cnx()
def _pop_custom_cnx(self):
if self._current_clt_cnx is not None:
if self._current_clt_cnx._open:
self._current_clt_cnx.close()
if not self._current_session.closed:
self.repo.close(self._current_session.id)
self._current_clt_cnx = None
self._current_session = None
#XXX this doesn't need to a be classmethod anymore
def _init_repo(self):
"""init the repository and connection to it.
"""
# setup configuration for test
self.init_config(self.config)
# get or restore and working db.
db_handler = devtools.get_test_db_handler(self.config)
db_handler.build_db_cache(self.test_db_id, self.pre_setup_database)
db_handler.restore_database(self.test_db_id)
self.repo = db_handler.get_repo(startup=True)
# get an admin session (without actual login)
sources = db_handler.config.sources()
login = unicode(sources['admin']['login'])
self.admin_access = self.new_access(login)
self._admin_session = self.admin_access._session
self._admin_clt_cnx = repoapi.ClientConnection(self._admin_session)
self._cnxs.add(self._admin_clt_cnx)
self._admin_clt_cnx.__enter__()
self.config.repository = lambda x=None: self.repo
# db api ##################################################################
@nocoverage
@deprecated('[3.19] explicitly use RepoAccess object in test instead')
def cursor(self, req=None):
if req is not None:
return req.cnx
else:
return self.cnx
@nocoverage
@deprecated('[3.19] explicitly use RepoAccess object in test instead')
def execute(self, rql, args=None, req=None):
"""executes <rql>, builds a resultset, and returns a couple (rset, req)
where req is a FakeRequest
"""
req = req or self.request(rql=rql)
return req.execute(unicode(rql), args)
@nocoverage
@deprecated('[3.19] explicitly use RepoAccess object in test instead')
def commit(self):
try:
return self.cnx.commit()
finally:
self.session.set_cnxset() # ensure cnxset still set after commit
@nocoverage
@deprecated('[3.19] explicitly use RepoAccess object in test instead')
def rollback(self):
try:
self.cnx.rollback()
except dbapi.ProgrammingError:
pass # connection closed
finally:
self.session.set_cnxset() # ensure cnxset still set after commit
requestcls = fake.FakeRequest
@deprecated('[3.19] explicitly use RepoAccess object in test instead')
def request(self, rollbackfirst=False, url=None, headers={}, **kwargs):
"""return a web ui request"""
if rollbackfirst:
self.cnx.rollback()
req = self.requestcls(self.vreg, url=url, headers=headers, form=kwargs)
req.set_cnx(self.cnx)
return req
# server side db api #######################################################
@deprecated('[3.19] explicitly use RepoAccess object in test instead')
def sexecute(self, rql, args=None):
self.session.set_cnxset()
return self.session.execute(rql, args)
# config management ########################################################
@classproperty
def config(cls):
"""return the configuration object
Configuration is cached on the test class.
"""
try:
assert not cls is CubicWebTC, "Don't use CubicWebTC directly to prevent database caching issue"
return cls.__dict__['_config']
except KeyError:
home = abspath(join(dirname(sys.modules[cls.__module__].__file__), cls.appid))
config = cls._config = cls.configcls(cls.appid, apphome=home)
config.mode = 'test'
return config
@classmethod
def init_config(cls, config):
"""configuration initialization hooks.
You may only want to override here the configuraton logic.
Otherwise, consider to use a different :class:`ApptestConfiguration`
defined in the `configcls` class attribute"""
source = config.sources()['system']
cls.admlogin = unicode(source['db-user'])
cls.admpassword = source['db-password']
# uncomment the line below if you want rql queries to be logged
#config.global_set_option('query-log-file',
# '/tmp/test_rql_log.' + `os.getpid()`)
config.global_set_option('log-file', None)
# set default-dest-addrs to a dumb email address to avoid mailbox or
# mail queue pollution
config.global_set_option('default-dest-addrs', ['whatever'])
send_to = '%s@logilab.fr' % getlogin()
config.global_set_option('sender-addr', send_to)
config.global_set_option('default-dest-addrs', send_to)
config.global_set_option('sender-name', 'cubicweb-test')
config.global_set_option('sender-addr', 'cubicweb-test@logilab.fr')
# default_base_url on config class isn't enough for TestServerConfiguration
config.global_set_option('base-url', config.default_base_url())
# web resources
try:
config.global_set_option('embed-allowed', re.compile('.*'))
except Exception: # not in server only configuration
pass
@property
def vreg(self):
return self.repo.vreg
# global resources accessors ###############################################
@property
def schema(self):
"""return the application schema"""
return self.vreg.schema
def shell(self):
"""return a shell session object"""
from cubicweb.server.migractions import ServerMigrationHelper
return ServerMigrationHelper(None, repo=self.repo, cnx=self.cnx,
interactive=False,
# hack so it don't try to load fs schema
schema=1)
def set_option(self, optname, value):
self.config.global_set_option(optname, value)
def set_debug(self, debugmode):
server.set_debug(debugmode)
def debugged(self, debugmode):
return server.debugged(debugmode)
# default test setup and teardown #########################################
def setUp(self):
# monkey patch send mail operation so emails are sent synchronously
self._patch_SendMailOp()
pause_tracing()
previous_failure = self.__class__.__dict__.get('_repo_init_failed')
if previous_failure is not None:
self.skipTest('repository is not initialised: %r' % previous_failure)
try:
self._init_repo()
self.addCleanup(self._close_cnx)
except Exception as ex:
self.__class__._repo_init_failed = ex
raise
resume_tracing()
self.setup_database()
self.commit()
MAILBOX[:] = [] # reset mailbox
def tearDown(self):
# XXX hack until logilab.common.testlib is fixed
if self._admin_clt_cnx is not None:
if self._admin_clt_cnx._open:
self._admin_clt_cnx.close()
self._admin_clt_cnx = None
if self._admin_session is not None:
if not self._admin_session.closed:
self.repo.close(self._admin_session.id)
self._admin_session = None
while self._cleanups:
cleanup, args, kwargs = self._cleanups.pop(-1)
cleanup(*args, **kwargs)
def _patch_SendMailOp(self):
# monkey patch send mail operation so emails are sent synchronously
_old_mail_postcommit_event = SendMailOp.postcommit_event
SendMailOp.postcommit_event = SendMailOp.sendmails
def reverse_SendMailOp_monkey_patch():
SendMailOp.postcommit_event = _old_mail_postcommit_event
self.addCleanup(reverse_SendMailOp_monkey_patch)
def setup_database(self):
"""add your database setup code by overriding this method"""
@classmethod
def pre_setup_database(cls, session, config):
"""add your pre database setup code by overriding this method
Do not forget to set the cls.test_db_id value to enable caching of the
result.
"""
# user / session management ###############################################
def user(self, req=None):
"""return the application schema"""
if req is None:
return self.request().user
else:
return req.user
@iclassmethod # XXX turn into a class method
def create_user(self, req, login=None, groups=('users',), password=None,
email=None, commit=True, **kwargs):
"""create and return a new user entity"""
if isinstance(req, basestring):
warn('[3.12] create_user arguments are now (req, login[, groups, password, commit, **kwargs])',
DeprecationWarning, stacklevel=2)
if not isinstance(groups, (tuple, list)):
password = groups
groups = login
elif isinstance(login, tuple):
groups = login
login = req
assert not isinstance(self, type)
req = self._orig_cnx[0].request()
if password is None:
password = login.encode('utf8')
user = req.create_entity('CWUser', login=unicode(login),
upassword=password, **kwargs)
req.execute('SET X in_group G WHERE X eid %%(x)s, G name IN(%s)'
% ','.join(repr(str(g)) for g in groups),
{'x': user.eid})
if email is not None:
req.create_entity('EmailAddress', address=unicode(email),
reverse_primary_email=user)
user.cw_clear_relation_cache('in_group', 'subject')
if commit:
try:
req.commit() # req is a session
except AttributeError:
req.cnx.commit()
return user
# other utilities #########################################################
@contextmanager
def temporary_appobjects(self, *appobjects):
self.vreg._loadedmods.setdefault(self.__module__, {})
for obj in appobjects:
self.vreg.register(obj)
registered = getattr(obj, '__registered__', None)
if registered:
for registry in obj.__registries__:
registered(self.vreg[registry])
try:
yield
finally:
for obj in appobjects:
self.vreg.unregister(obj)
@contextmanager
def temporary_permissions(self, *perm_overrides, **perm_kwoverrides):
"""Set custom schema permissions within context.
There are two ways to call this method, which may be used together :
* using positional argument(s):
.. sourcecode:: python
rdef = self.schema['CWUser'].rdef('login')
with self.temporary_permissions((rdef, {'read': ()})):
...
* using named argument(s):
.. sourcecode:: python
rdef = self.schema['CWUser'].rdef('login')
with self.temporary_permissions(CWUser={'read': ()}):
...
Usually the former will be prefered to override permissions on a
relation definition, while the latter is well suited for entity types.
The allowed keys in the permission dictionary depends on the schema type
(entity type / relation definition). Resulting permissions will be
similar to `orig_permissions.update(partial_perms)`.
"""
torestore = []
for erschema, etypeperms in chain(perm_overrides, perm_kwoverrides.iteritems()):
if isinstance(erschema, basestring):
erschema = self.schema[erschema]
for action, actionperms in etypeperms.iteritems():
origperms = erschema.permissions[action]
erschema.set_action_permissions(action, actionperms)
torestore.append([erschema, action, origperms])
yield
for erschema, action, permissions in torestore:
if action is None:
erschema.permissions = permissions
else:
erschema.set_action_permissions(action, permissions)
def assertModificationDateGreater(self, entity, olddate):
entity.cw_attr_cache.pop('modification_date', None)
self.assertTrue(entity.modification_date > olddate)
def assertItemsEqual(self, it1, it2, *args, **kwargs):
it1 = set(getattr(x, 'eid', x) for x in it1)
it2 = set(getattr(x, 'eid', x) for x in it2)
super(CubicWebTC, self).assertItemsEqual(it1, it2, *args, **kwargs)
def assertMessageEqual(self, req, params, msg):
msg = req.session.data[params['_cwmsgid']]
self.assertEqual(msg, msg)
# workflow utilities #######################################################
def assertPossibleTransitions(self, entity, expected):
transitions = entity.cw_adapt_to('IWorkflowable').possible_transitions()
self.assertListEqual(sorted(tr.name for tr in transitions),
sorted(expected))
# views and actions registries inspection ##################################
def pviews(self, req, rset):
return sorted((a.__regid__, a.__class__)
for a in self.vreg['views'].possible_views(req, rset=rset))
def pactions(self, req, rset,
skipcategories=('addrelated', 'siteactions', 'useractions',
'footer', 'manage')):
return [(a.__regid__, a.__class__)
for a in self.vreg['actions'].poss_visible_objects(req, rset=rset)
if a.category not in skipcategories]
def pactions_by_cats(self, req, rset, categories=('addrelated',)):
return [(a.__regid__, a.__class__)
for a in self.vreg['actions'].poss_visible_objects(req, rset=rset)
if a.category in categories]
def pactionsdict(self, req, rset,
skipcategories=('addrelated', 'siteactions', 'useractions',
'footer', 'manage')):
res = {}
for a in self.vreg['actions'].poss_visible_objects(req, rset=rset):
if a.category not in skipcategories:
res.setdefault(a.category, []).append(a.__class__)
return res
def action_submenu(self, req, rset, id):
return self._test_action(self.vreg['actions'].select(id, req, rset=rset))
def _test_action(self, action):
class fake_menu(list):
@property
def items(self):
return self
class fake_box(object):
def action_link(self, action, **kwargs):
return (action.title, action.url())
submenu = fake_menu()
action.fill_menu(fake_box(), submenu)
return submenu
def list_views_for(self, rset):
"""returns the list of views that can be applied on `rset`"""
req = rset.req
only_once_vids = ('primary', 'secondary', 'text')
req.data['ex'] = ValueError("whatever")
viewsvreg = self.vreg['views']
for vid, views in viewsvreg.items():
if vid[0] == '_':
continue
if rset.rowcount > 1 and vid in only_once_vids:
continue
views = [view for view in views
if view.category != 'startupview'
and not issubclass(view, notification.NotificationView)
and not isinstance(view, class_deprecated)]
if views:
try:
view = viewsvreg._select_best(views, req, rset=rset)
if view is None:
raise NoSelectableObject((req,), {'rset':rset}, views)
if view.linkable():
yield view
else:
not_selected(self.vreg, view)
# else the view is expected to be used as subview and should
# not be tested directly
except NoSelectableObject:
continue
def list_actions_for(self, rset):
"""returns the list of actions that can be applied on `rset`"""
req = rset.req
for action in self.vreg['actions'].possible_objects(req, rset=rset):
yield action
def list_boxes_for(self, rset):
"""returns the list of boxes that can be applied on `rset`"""
req = rset.req
for box in self.vreg['ctxcomponents'].possible_objects(req, rset=rset):
yield box
def list_startup_views(self):
"""returns the list of startup views"""
req = self.request()
for view in self.vreg['views'].possible_views(req, None):
if view.category == 'startupview':
yield view.__regid__
else:
not_selected(self.vreg, view)
# web ui testing utilities #################################################
@property
@cached
def app(self):
"""return a cubicweb publisher"""
publisher = application.CubicWebPublisher(self.repo, self.config)
def raise_error_handler(*args, **kwargs):
raise
publisher.error_handler = raise_error_handler
return publisher
def remote_call(self, fname, *args):
"""remote json call simulation"""
dump = json.dumps
args = [dump(arg) for arg in args]
req = self.request(fname=fname, pageid='123', arg=args)
ctrl = self.vreg['controllers'].select('ajax', req)
return ctrl.publish(), req
def app_handle_request(self, req, path='view'):
return self.app.core_handle(req, path)
@deprecated("[3.15] app_handle_request is the new and better way"
" (beware of small semantic changes)")
def app_publish(self, *args, **kwargs):
return self.app_handle_request(*args, **kwargs)
def ctrl_publish(self, req, ctrl='edit', rset=None):
"""call the publish method of the edit controller"""
ctrl = self.vreg['controllers'].select(ctrl, req, appli=self.app)
try:
result = ctrl.publish(rset)
req.cnx.commit()
except web.Redirect:
req.cnx.commit()
raise
return result
def req_from_url(self, url):
"""parses `url` and builds the corresponding CW-web request
req.form will be setup using the url's query string
"""
req = self.request(url=url)
if isinstance(url, unicode):
url = url.encode(req.encoding) # req.setup_params() expects encoded strings
querystring = urlparse.urlparse(url)[-2]
params = urlparse.parse_qs(querystring)
req.setup_params(params)
return req
def url_publish(self, url, data=None):
"""takes `url`, uses application's app_resolver to find the appropriate
controller and result set, then publishes the result.
To simulate post of www-form-encoded data, give a `data` dictionary
containing desired key/value associations.
This should pretty much correspond to what occurs in a real CW server
except the apache-rewriter component is not called.
"""
req = self.req_from_url(url)
if data is not None:
req.form.update(data)
ctrlid, rset = self.app.url_resolver.process(req, req.relative_path(False))
return self.ctrl_publish(req, ctrlid, rset)
def http_publish(self, url, data=None):
"""like `url_publish`, except this returns a http response, even in case of errors"""
req = self.req_from_url(url)
if data is not None:
req.form.update(data)
# remove the monkey patched error handler
fake_error_handler = self.app.error_handler
del self.app.error_handler
try:
result = self.app_handle_request(req, req.relative_path(False))
finally:
self.app.error_handler = fake_error_handler
return result, req
@staticmethod
def _parse_location(req, location):
try:
path, params = location.split('?', 1)
except ValueError:
path = location
params = {}
else:
cleanup = lambda p: (p[0], unquote(p[1]))
params = dict(cleanup(p.split('=', 1)) for p in params.split('&') if p)
if path.startswith(req.base_url()): # may be relative
path = path[len(req.base_url()):]
return path, params
def expect_redirect(self, callback, req):
"""call the given callback with req as argument, expecting to get a
Redirect exception
"""
try:
callback(req)
except Redirect as ex:
return self._parse_location(req, ex.location)
else:
self.fail('expected a Redirect exception')
def expect_redirect_handle_request(self, req, path='edit'):
"""call the publish method of the application publisher, expecting to
get a Redirect exception
"""
result = self.app_handle_request(req, path)
self.assertTrue(300 <= req.status_out <400, req.status_out)
location = req.get_response_header('location')
return self._parse_location(req, location)
@deprecated("[3.15] expect_redirect_handle_request is the new and better way"
" (beware of small semantic changes)")
def expect_redirect_publish(self, *args, **kwargs):
return self.expect_redirect_handle_request(*args, **kwargs)
def set_auth_mode(self, authmode, anonuser=None):
self.set_option('auth-mode', authmode)
self.set_option('anonymous-user', anonuser)
if anonuser is None:
self.config.anonymous_credential = None
else:
self.config.anonymous_credential = (anonuser, anonuser)
def init_authentication(self, authmode, anonuser=None):
self.set_auth_mode(authmode, anonuser)
req = self.requestcls(self.vreg, url='login')
sh = self.app.session_handler
authm = sh.session_manager.authmanager
authm.anoninfo = self.vreg.config.anonymous_user()
authm.anoninfo = authm.anoninfo[0], {'password': authm.anoninfo[1]}
# not properly cleaned between tests
self.open_sessions = sh.session_manager._sessions = {}
return req, self.websession
def assertAuthSuccess(self, req, origsession, nbsessions=1):
sh = self.app.session_handler
session = self.app.get_session(req)
clt_cnx = repoapi.ClientConnection(session)
req.set_cnx(clt_cnx)
self.assertEqual(len(self.open_sessions), nbsessions, self.open_sessions)
self.assertEqual(session.login, origsession.login)
self.assertEqual(session.anonymous_session, False)
def assertAuthFailure(self, req, nbsessions=0):
with self.assertRaises(AuthenticationError):
self.app.get_session(req)
# +0 since we do not track the opened session
self.assertEqual(len(self.open_sessions), nbsessions)
clear_cache(req, 'get_authorization')
# content validation #######################################################
# validators are used to validate (XML, DTD, whatever) view's content
# validators availables are :
# DTDValidator : validates XML + declared DTD
# SaxOnlyValidator : guarantees XML is well formed
# None : do not try to validate anything
# validators used must be imported from from.devtools.htmlparser
content_type_validators = {
# maps MIME type : validator name
#
# do not set html validators here, we need HTMLValidator for html
# snippets
#'text/html': DTDValidator,
#'application/xhtml+xml': DTDValidator,
'application/xml': htmlparser.XMLValidator,
'text/xml': htmlparser.XMLValidator,
'application/json': JsonValidator,
'text/plain': None,
'text/comma-separated-values': None,
'text/x-vcard': None,
'text/calendar': None,
'image/png': None,
}
# maps vid : validator name (override content_type_validators)
vid_validators = dict((vid, htmlparser.VALMAP[valkey])
for vid, valkey in VIEW_VALIDATORS.iteritems())
def view(self, vid, rset=None, req=None, template='main-template',
**kwargs):
"""This method tests the view `vid` on `rset` using `template`
If no error occurred while rendering the view, the HTML is analyzed
and parsed.
:returns: an instance of `cubicweb.devtools.htmlparser.PageInfo`
encapsulation the generated HTML
"""
if req is None:
if rset is None:
req = self.request()
else:
req = rset.req
req.form['vid'] = vid
viewsreg = self.vreg['views']
view = viewsreg.select(vid, req, rset=rset, **kwargs)
# set explicit test description
if rset is not None:
self.set_description("testing vid=%s defined in %s with (%s)" % (
vid, view.__module__, rset.printable_rql()))
else:
self.set_description("testing vid=%s defined in %s without rset" % (
vid, view.__module__))
if template is None: # raw view testing, no template
viewfunc = view.render
else:
kwargs['view'] = view
viewfunc = lambda **k: viewsreg.main_template(req, template,
rset=rset, **kwargs)
return self._test_view(viewfunc, view, template, kwargs)
def _test_view(self, viewfunc, view, template='main-template', kwargs={}):
"""this method does the actual call to the view
If no error occurred while rendering the view, the HTML is analyzed
and parsed.
:returns: an instance of `cubicweb.devtools.htmlparser.PageInfo`
encapsulation the generated HTML
"""
try:
output = viewfunc(**kwargs)
except Exception:
# hijack exception: generative tests stop when the exception
# is not an AssertionError
klass, exc, tcbk = sys.exc_info()
try:
msg = '[%s in %s] %s' % (klass, view.__regid__, exc)
except Exception:
msg = '[%s in %s] undisplayable exception' % (klass, view.__regid__)
raise AssertionError, msg, tcbk
return self._check_html(output, view, template)
def get_validator(self, view=None, content_type=None, output=None):
if view is not None:
try:
return self.vid_validators[view.__regid__]()
except KeyError:
if content_type is None:
content_type = view.content_type
if content_type is None:
content_type = 'text/html'
if content_type in ('text/html', 'application/xhtml+xml') and output:
if output.startswith('<!DOCTYPE html>'):
# only check XML well-formness since HTMLValidator isn't html5
# compatible and won't like various other extensions
default_validator = htmlparser.XMLSyntaxValidator
elif output.startswith('<?xml'):
default_validator = htmlparser.DTDValidator
else:
default_validator = htmlparser.HTMLValidator
else:
default_validator = None
validatorclass = self.content_type_validators.get(content_type,
default_validator)
if validatorclass is None:
return
return validatorclass()
@nocoverage
def _check_html(self, output, view, template='main-template'):
"""raises an exception if the HTML is invalid"""
output = output.strip()
validator = self.get_validator(view, output=output)
if validator is None:
return output # return raw output if no validator is defined
if isinstance(validator, htmlparser.DTDValidator):
# XXX remove <canvas> used in progress widget, unknown in html dtd
output = re.sub('<canvas.*?></canvas>', '', output)
return self.assertWellFormed(validator, output.strip(), context= view.__regid__)
def assertWellFormed(self, validator, content, context=None):
try:
return validator.parse_string(content)
except Exception:
# hijack exception: generative tests stop when the exception
# is not an AssertionError
klass, exc, tcbk = sys.exc_info()
if context is None:
msg = u'[%s]' % (klass,)
else:
msg = u'[%s in %s]' % (klass, context)
msg = msg.encode(sys.getdefaultencoding(), 'replace')
try:
str_exc = str(exc)
except Exception:
str_exc = 'undisplayable exception'
msg += str_exc
if content is not None:
position = getattr(exc, "position", (0,))[0]
if position:
# define filter
if isinstance(content, str):
content = unicode(content, sys.getdefaultencoding(), 'replace')
content = validator.preprocess_data(content)
content = content.splitlines()
width = int(log(len(content), 10)) + 1
line_template = " %" + ("%i" % width) + "i: %s"
# XXX no need to iterate the whole file except to get
# the line number
content = u'\n'.join(line_template % (idx + 1, line)
for idx, line in enumerate(content)
if line_context_filter(idx+1, position))
msg += u'\nfor content:\n%s' % content
raise AssertionError, msg, tcbk
def assertDocTestFile(self, testfile):
# doctest returns tuple (failure_count, test_count)
result = self.shell().process_script(testfile)
if result[0] and result[1]:
raise self.failureException("doctest file '%s' failed"
% testfile)
# notifications ############################################################
def assertSentEmail(self, subject, recipients=None, nb_msgs=None):
"""test recipients in system mailbox for given email subject
:param subject: email subject to find in mailbox
:param recipients: list of email recipients
:param nb_msgs: expected number of entries
:returns: list of matched emails
"""
messages = [email for email in MAILBOX
if email.message.get('Subject') == subject]
if recipients is not None:
sent_to = set()
for msg in messages:
sent_to.update(msg.recipients)
self.assertSetEqual(set(recipients), sent_to)
if nb_msgs is not None:
self.assertEqual(len(MAILBOX), nb_msgs)
return messages
# auto-populating test classes and utilities ###################################
from cubicweb.devtools.fill import insert_entity_queries, make_relations_queries
# XXX cleanup unprotected_entities & all mess
def how_many_dict(schema, cursor, how_many, skip):
"""given a schema, compute how many entities by type we need to be able to
satisfy relations cardinality.
The `how_many` argument tells how many entities of which type we want at
least.
Return a dictionary with entity types as key, and the number of entities for
this type as value.
"""
relmap = {}
for rschema in schema.relations():
if rschema.final:
continue
for subj, obj in rschema.rdefs:
card = rschema.rdef(subj, obj).cardinality
# if the relation is mandatory, we'll need at least as many subj and
# obj to satisfy it
if card[0] in '1+' and card[1] in '1?':
# subj has to be linked to at least one obj,
# but obj can be linked to only one subj
# -> we need at least as many subj as obj to satisfy
# cardinalities for this relation
relmap.setdefault((rschema, subj), []).append(str(obj))
if card[1] in '1+' and card[0] in '1?':
# reverse subj and obj in the above explanation
relmap.setdefault((rschema, obj), []).append(str(subj))
unprotected = unprotected_entities(schema)
for etype in skip: # XXX (syt) duh? explain or kill
unprotected.add(etype)
howmanydict = {}
# step 1, compute a base number of each entity types: number of already
# existing entities of this type + `how_many`
for etype in unprotected_entities(schema, strict=True):
howmanydict[str(etype)] = cursor.execute('Any COUNT(X) WHERE X is %s' % etype)[0][0]
if etype in unprotected:
howmanydict[str(etype)] += how_many
# step 2, augment nb entity per types to satisfy cardinality constraints,
# by recomputing for each relation that constrained an entity type:
#
# new num for etype = max(current num, sum(num for possible target etypes))
#
# XXX we should first check there is no cycle then propagate changes
for (rschema, etype), targets in relmap.iteritems():
relfactor = sum(howmanydict[e] for e in targets)
howmanydict[str(etype)] = max(relfactor, howmanydict[etype])
return howmanydict
class AutoPopulateTest(CubicWebTC):
"""base class for test with auto-populating of the database"""
__abstract__ = True
test_db_id = 'autopopulate'
tags = CubicWebTC.tags | Tags('autopopulated')
pdbclass = CubicWebDebugger
# this is a hook to be able to define a list of rql queries
# that are application dependent and cannot be guessed automatically
application_rql = []
no_auto_populate = ()
ignored_relations = set()
def to_test_etypes(self):
return unprotected_entities(self.schema, strict=True)
def custom_populate(self, how_many, cursor):
pass
def post_populate(self, cursor):
pass
@nocoverage
def auto_populate(self, how_many):
"""this method populates the database with `how_many` entities
of each possible type. It also inserts random relations between them
"""
with self.session.security_enabled(read=False, write=False):
self._auto_populate(how_many)
def _auto_populate(self, how_many):
cu = self.cursor()
self.custom_populate(how_many, cu)
vreg = self.vreg
howmanydict = how_many_dict(self.schema, cu, how_many, self.no_auto_populate)
for etype in unprotected_entities(self.schema):
if etype in self.no_auto_populate:
continue
nb = howmanydict.get(etype, how_many)
for rql, args in insert_entity_queries(etype, self.schema, vreg, nb):
cu.execute(rql, args)
edict = {}
for etype in unprotected_entities(self.schema, strict=True):
rset = cu.execute('%s X' % etype)
edict[str(etype)] = set(row[0] for row in rset.rows)
existingrels = {}
ignored_relations = SYSTEM_RELATIONS | self.ignored_relations
for rschema in self.schema.relations():
if rschema.final or rschema in ignored_relations:
continue
rset = cu.execute('DISTINCT Any X,Y WHERE X %s Y' % rschema)
existingrels.setdefault(rschema.type, set()).update((x, y) for x, y in rset)
q = make_relations_queries(self.schema, edict, cu, ignored_relations,
existingrels=existingrels)
for rql, args in q:
try:
cu.execute(rql, args)
except ValidationError as ex:
# failed to satisfy some constraint
print 'error in automatic db population', ex
self.session.commit_state = None # reset uncommitable flag
self.post_populate(cu)
self.commit()
def iter_individual_rsets(self, etypes=None, limit=None):
etypes = etypes or self.to_test_etypes()
for etype in etypes:
if limit:
rql = 'Any X LIMIT %s WHERE X is %s' % (limit, etype)
else:
rql = 'Any X WHERE X is %s' % etype
rset = self.execute(rql)
for row in xrange(len(rset)):
if limit and row > limit:
break
# XXX iirk
rset2 = rset.limit(limit=1, offset=row)
yield rset2
def iter_automatic_rsets(self, limit=10):
"""generates basic resultsets for each entity type"""
etypes = self.to_test_etypes()
if not etypes:
return
for etype in etypes:
yield self.execute('Any X LIMIT %s WHERE X is %s' % (limit, etype))
etype1 = etypes.pop()
try:
etype2 = etypes.pop()
except KeyError:
etype2 = etype1
# test a mixed query (DISTINCT/GROUP to avoid getting duplicate
# X which make muledit view failing for instance (html validation fails
# because of some duplicate "id" attributes)
yield self.execute('DISTINCT Any X, MAX(Y) GROUPBY X WHERE X is %s, Y is %s' % (etype1, etype2))
# test some application-specific queries if defined
for rql in self.application_rql:
yield self.execute(rql)
def _test_everything_for(self, rset):
"""this method tries to find everything that can be tested
for `rset` and yields a callable test (as needed in generative tests)
"""
propdefs = self.vreg['propertydefs']
# make all components visible
for k, v in propdefs.items():
if k.endswith('visible') and not v['default']:
propdefs[k]['default'] = True
for view in self.list_views_for(rset):
backup_rset = rset.copy(rset.rows, rset.description)
yield InnerTest(self._testname(rset, view.__regid__, 'view'),
self.view, view.__regid__, rset,
rset.req.reset_headers(), 'main-template')
# We have to do this because some views modify the
# resultset's syntax tree
rset = backup_rset
for action in self.list_actions_for(rset):
yield InnerTest(self._testname(rset, action.__regid__, 'action'), self._test_action, action)
for box in self.list_boxes_for(rset):
w = [].append
yield InnerTest(self._testname(rset, box.__regid__, 'box'), box.render, w)
@staticmethod
def _testname(rset, objid, objtype):
return '%s_%s_%s' % ('_'.join(rset.column_types(0)), objid, objtype)
# concrete class for automated application testing ############################
class AutomaticWebTest(AutoPopulateTest):
"""import this if you wan automatic tests to be ran"""
tags = AutoPopulateTest.tags | Tags('web', 'generated')
def setUp(self):
assert not self.__class__ is AutomaticWebTest, 'Please subclass AutomaticWebTest to prevent database caching issue'
super(AutomaticWebTest, self).setUp()
# access to self.app for proper initialization of the authentication
# machinery (else some views may fail)
self.app
## one each
def test_one_each_config(self):
self.auto_populate(1)
for rset in self.iter_automatic_rsets(limit=1):
for testargs in self._test_everything_for(rset):
yield testargs
## ten each
def test_ten_each_config(self):
self.auto_populate(10)
for rset in self.iter_automatic_rsets(limit=10):
for testargs in self._test_everything_for(rset):
yield testargs
## startup views
def test_startup_views(self):
for vid in self.list_startup_views():
req = self.request()
yield self.view, vid, None, req
# registry instrumentization ###################################################
def not_selected(vreg, appobject):
try:
vreg._selected[appobject.__class__] -= 1
except (KeyError, AttributeError):
pass
# def vreg_instrumentize(testclass):
# # XXX broken
# from cubicweb.devtools.apptest import TestEnvironment
# env = testclass._env = TestEnvironment('data', configcls=testclass.configcls)
# for reg in env.vreg.itervalues():
# reg._selected = {}
# try:
# orig_select_best = reg.__class__.__orig_select_best
# except Exception:
# orig_select_best = reg.__class__._select_best
# def instr_select_best(self, *args, **kwargs):
# selected = orig_select_best(self, *args, **kwargs)
# try:
# self._selected[selected.__class__] += 1
# except KeyError:
# self._selected[selected.__class__] = 1
# except AttributeError:
# pass # occurs on reg used to restore database
# return selected
# reg.__class__._select_best = instr_select_best
# reg.__class__.__orig_select_best = orig_select_best
# def print_untested_objects(testclass, skipregs=('hooks', 'etypes')):
# for regname, reg in testclass._env.vreg.iteritems():
# if regname in skipregs:
# continue
# for appobjects in reg.itervalues():
# for appobject in appobjects:
# if not reg._selected.get(appobject):
# print 'not tested', regname, appobject