[dataimport] massive_store: ignore primary key when requesting indexes on a table Primary keys show up in both lists of constraints and indexes. Pick a side and stick to it.

[dataimport] massive_store: reformat sql query

[devtools] Use super() in TestServerConfiguration.__init__() While cleaning up the use of init_config() in WSGIAppTC, I wondered why the method was trying to set the 'https_uiprops' and 'https_datadir_url' attributes on the config, when WebConfiguration's __init__ should be taking care of setting them. WSGIAppTC uses the default configuration - one that derives from ApptestConfiguration (noted as G below). The issue is that the class hierarchy of ApptestConfiguration includes TestServerConfiguration (C in the graph below) which calls ServerConfiguration's __init__ directly (B below). As B does not have an __init__ method, CubicWebConfiguration's __init__ is called directly (A below). This has the unintended consequence of preventing WebConfiguration's __init__ from being called. A / \ B D | | C E \ / F | G | ... A: cwconfig.CubicWebConfiguration B: server.serverconfig.ServerConfiguration C: devtools.TestServerConfiguration D: web.webconfig.WebConfiguration E: etwist.twconfig.WebConfigurationBase F: devtools.BaseApptestConfiguration G: devtools.ApptestConfiguration NB: unittest_wsgi wasn't really broken, as it's the only test in wsgi/test/, but I figured I might as well clean it up. Related to #9297611.

[devtools] Fix database creation issues for CubicWebTestTC tests To reproduce (with a purged repo): * python web/test/test_jscript.py * python web/test/unittest_views_xmlrss.py The '__default_empty_db__' will contain the wrong 'base_url' configuration, thus breaking URLs in the xmlrss tests. Closes #9297611.

[devtools] Fix database creation issues for CubicWebServerTC tests To reproduce (with a purged repo): * python web/test/unittest_web.py MiscOptionsTC * python web/test/unittest_views_xmlrss.py The core issue is that CubicWebServerConfig.default_base_url() is called inside init_config(), which is called during the equivalent of 'db-create'. The '__default_empty_db__' will contain the wrong 'base_url' configuration, thus breaking URLs in the xmlrss tests. Related to #9297611.

[devtools] Move attribute docstring to correct location

[web] Make tests pass if tal isn't available

[py3k] fix cubicweb.mail.parse_message_id base64.b64decode returns bytes. Fixes sobjects/test/unittest_notification.py.

[test/py3k] fix test_recipients_finder AssertionError: <zip object at 0x7f1dfb0230c8> != [('abcd@logilab.fr', 'en'), ('efgh@logilab.fr', 'en')]

[dataimport] fix SQLGenObjectStore's add_info for python3

[dataimport] py3k doesn't like + on dict_items objects

[repository] remove unused check_session method and its test Related to #1381328.

[migration/test] avoid test interdependency by removing fakecustomtype cube There is still a failing test afterward but it doesn't seem related to changes in this series.

[migration/test] use numeric(precision, scale) instead of geometry The latter requires the postgis extension which is overkill for this test.

[server, hooks] allow callable in dbh.TYPE_MAPPING This is necessary for cubes introducing new custom final types whose generated sql may depend on some extra configuration, avoiding ugly-buggy hack as done currently by the postgis cube. Closes #7569998

[migration] consider extra_props in rdef created during attribute creation hook else we may miss important data such as type parametrization (see postgis.Geometry for instance). Related to #7569998

[migration] only define variable if needed

[migration] handle non-existing attribute type when adding an entity type The final type needs to be added to the schema first. Related to #7569998.

[migration] handle the case where new final entity type is not yet in the schema This may occur on new cube addition. Related to #7569998.

[migration] config.add_cubes should ensure site_cubicweb of new cubes is called Related to #7569998

[migration] test addition of a cube providing a custom final type This does not work yet. Related to #7569998

[server] add test demonstrating cache update pb since cw > 3.21 Fixed by the previous changeset.

Bring back the separate web-side entity cache Prior to changeset 635cfac73d28 "[repoapi] fold ClientConnection into Connection", we had two entity caches: one on the client/dbapi/web side, and one on the server/repo side. This is a waste, but it is actually needed as long as we have the magic _cw attribute on entities which must sometimes be a request and sometimes a cnx. Removing the duplication caused weird problems with entity._cw alternating between both types of objects, which is unexpected by both the repo and the web sides. We add an entity cache on ConnectionCubicWebRequestBase, separate from the Connection's, and bring back the _cw_update_attr_cache/dont-cache-attrs mechanism, to let the server/edition code let caches know which attributes have been modified Entity.as_rset can be cached again, as ResultSet no longer modifies the entity when fetching it from a cache. Contrary to the pre-3.21 code, _cw_update_attr_cache now handles web requests and connections in the same way (otherwise the cache ends up with wrong values if a hook modifies attributes), but dont-cache-attrs is never set for (inlined) relations. Closes #6863543

Backout "[entity,storage] remove hackish code from the bfss path The merge of the repo-side and web-side entity caches doesn't actually work, because entities have a _cw attribute which is sometimes a connection and sometimes a web request, and whose nature changes not-really-predictably whenever ResultSet.get_entity() is called and brings it out of the cache. This reverts commit 59a79300f213.

merge with 3.21.5-2

[doc] add documentation about drop_cube

Added tag debian/3.21.5-2 for changeset 228b6d2777e4

[pkg] don't install dataimport/test in both cubicweb-server and cubicweb-dev Closes #9239572

merge with 3.21.5

Added tag 3.21.5, debian/3.21.5-1, centos/3.21.5-1 for changeset e0572a786e6b

[pkg] 3.21.5

[js] Fix a ReferenceError in cubicweb.preferences.js Rename function using proper English along the way. Closes #8864731.

[dataimport] add tests for SQLGenObjectStore (closes #9177565) Stolen from the massive import store, with the irrelevant bits removed. Remove redundant tests from test_pgstore.py.

[dataimport] separate entities table from other metadata in SQLGenObjectStore 'entities' needs to be updated first as every other table has references to its rows. Remove use of thread-local storage in the process. related to #9177565.

[dataimport] remove threading support from SQLGenObjectStore It's untested and does not actually work at the moment with nb_threads > 1. related to #9177565.

[dataimport] prepare_insert_entity should not crash with SQLGenObjectStore grafted from 1079d68130e1e780791ceb2050e8a8b7a910f48d (3.22) related to #9177565.

[dataimport] Fix method signature Since d4bd28d5fca8, _storage_handler expect 4 args. grafted from b089858c49a2806b6ea8a7ee66773436701e947c (3.22) related to #9177565.

[dataimport] add missing import 'eschema_eid' grafted from e3ec46b5c710067a0cecca53cc593a0dd4bd927e (3.22) related to #9177565

merge with 3.21.4

Added tag 3.21.4, debian/3.21.4-1, centos/3.21.4-1 for changeset d3b92d3a7db0

[pkg] 3.21.4

merge with 3.20.12

Added tag 3.20.12, debian/3.20.12-1, centos/3.20.12-1 for changeset 03e8fc9f79a6

[tox] Ensure pytest gets run with virtualenv's python

[tox] Use generating environments Drop the "env" option, which is unknown, and set a generative_ "envlist". Then use conditional dependencies based on factors_ to avoid repeating environment sections. The only singular one is the "cubicweb" environment because of its "commands". At some point, it'd be nice to find a way to have python version prefixes (py27-hooks or py34-web) for environment names but I could not find a way to extract the package name from such names. .. _generative: http://tox.readthedocs.org/en/latest/config.html#generative-envlist .. _factors: http://tox.readthedocs.org/en/latest/config.html#factors-and-factor-conditional-settings

a bit of pep8

[adapters] Use an adapter for serializing entities to a bare python structure which may be used for e.g. JSON export or other and allow to have different serializations for each entity types. Closes #6378035

[pkg] 3.20.12

[dataimport] Fallback from COPY to executemany (closes #9135780) If we can't build a suitable COPY command (e.g. because data has unsupported type), use executemany instead. Regression introduced in 3.20.0.

[web/views] unlink svg temp file after we're done Otherwise we fill up the tempdir. Closes #9133268.

Use cubicweb._

[web/views] Controllers should always return bytes Previous code breaks only on py3k.

[devtools] Fix 'newcube' on py3k

[toolsutils] Small fixes to fill_templated_file() * Use context managers to close _both_ files * Use io.open to enforce text mode on both python 2 and 3

[skeleton] Fix broken spec file template Broken in commit 73c1c9cf6bda.

[devtools] Fix fd leak in tests

[devtools] stop using datetime.now() in cwctl newcube and i18ncube

[hooks] fix timestamp confusion in DataImportsCleanupStartupHook CWDataImport.start_timestamp is inserted as datetime.utcnow(), in server/sources/datafeed.py:DataFeedSourceinit_import_log. Don't compare it to datetime.now().

[hooks] fix timestamp confusion in TransactionsCleanupStartupHook transactions.tx_time is inserted as datetime.utcnow(), in server/sources/native.py:NativeSQLSource.start_undoable_transaction. Don't compare it to datetime.now().

merge with 3.21.3 The change in unittest_serverctl.py is needed because of daef7ce08fea (from 3.20.11) and 3914388b2d0f (from the 3.22 branch). Due to both changes, CubicWebTC.config.repository no longer creates a new repository (and thus, a new connection). Since both DBDumpCommand and CubicWebTC.tearDown try to shutdown the repo, tearDown breaks apart. The solution is to temporarily disable ServerConfiguration's config cache. By forcing DBDumpCommand to get a new configuration object, it then gets its own Repo object, allowing tearDown and DBDumpCommand to work independently.

Added tag 3.21.3, debian/3.21.3-1, centos/3.21.3-1 for changeset 9edfe9429209

[pkg] 3.21.3

merge with 3.20.11

update repo sources on update of source's entities .url attribute

Added tag 3.20.11, debian/3.20.11-1, centos/3.20.11-1 for changeset c44930ac9579

[pkg] 3.20.11

[rest] a bit of pep8

[rest] avoid field on two lines if name is above some size limit. Closes #8961835

action, view and service so managers can start source synchronization from the web ui Closes #5474286

[test] Skip ldap source tests if slapd is not available

[doc] Minor changes to the blog tutorial

[skel] remove (what looks like) a spurious copy/paste

Merge default heads published "concurrently"

[qunit] move from CubicWebServerTC to CubicWebTestTC This test now uses wsgi instead of twisted, and thus passes under python3.

[qunit] no need to enable CORS anymore We serve everything needed for the tests through cubicweb.

[web/test] use GET, not POST, to load html files POST comes with a body. There isn't one.

[dataimport] prepare_insert_entity should not crash with SQLGenObjectStore

[dataimport] Fix method signature Since d4bd28d5fca8, _storage_handler expect 4 args.

[dataimport] add missing import 'eschema_eid'

[test] Remove unused import in unittest_qunit.py

[dataimport] Format strings with % instead of .format() For consistency with CubicWeb code base and to avoid Python 2.6 compatibility traps.

[test] Use datapath() in massive store tests

[devtools] Disable anonymous login right before executing tests CubicWebTC.init_config is called many times, especially when creating the default database. This database is then reused by all tests in the same directory. The issue is that, if a test has 'anonymous_allowed' set to False, the anonymous user won't be created by postcreate.py in the default database. All other tests will fail miserably if they expected anonymous login to work. By actually disabling anonymous login after _init_repo, the default database (and configuration) will be initialized with a working anonymous user. Closes #7485127.

[hooks] build EnsureSymmetricRelations{Add,Delete} dynamically Instead of setting the __select__ attribute at registration time, which confuses the filterable_selectors cache when running different tests (with different schemas) in the same process, just make entirely new classes. Closes #8785708.

[qunit] stop dealing with filesystem paths qunit tests need a few things served by cubicweb: - qunit itself, which is now handled by CWDevtoolsStaticController (serving files in cubicweb/devtools/data) - standard cubicweb or cubes data files, handled by the DataController - the tests themselves and their dependencies. These can live in <apphome>/data or <apphome>/static and be served by one of the STATIC_CONTROLLERS This avoids having to guess in CWSoftwareRootStaticController where to serve things from (some files may be installed, others are in the source tree), and should hopefully make it possible to have these tests pass when using tox, and to write qunit tests for cubes, outside of cubicweb itself. This requires modifying the tests to only declare URL paths instead of filesystem paths, and moving support files below test/data/static.

[devtools] fix QUnitResultController It needs to return something.

[utils] Drop now-unused SizeConstrainedList It's last usage in CubicWeb got removed from changeset '[views] remove breadcrumbs usage to retrieve last visited page' Related to #5456850.

[views] remove breadcrumbs usage to retrieve last visited page Use JavaScript history.back to go to the last visited page. Rename _return_to_lastpage to _redirect as it becomes misnamed. Drop `test_redirect_default` test in `unittest_views_basecontrollers` which essentially tested this (now removed) breadcrumbs "feature". Closes #5456850.

[cwconfig] Explicitly close generated config file

[doc] Add 3.21 release date

[dataimport] fix db-namespace default value handling In a fresh new instance, the "db-namespace" option is set as empty by default in the "sources" file. This seems correct but empty string is not a valid pg_schema and actually means that we should use "public"

[server] unbreak db-restore

[doc] add 3.22 changelog

[rset] remove stray (commented-out) print Left from debugging in af266f27c4d5

Un-hide jshintrc dot-files belong in ~, not anywhere else.

[views] Use 'pyvaltable' instead of hand-crafted <table> tags

[web] Stop using config.repository() Since commit 3914388b2d0f, this method creates a new Repository object and thus reloads all the cubes, taking for ever. Closes #8634417.

[web] Stop using webconfig.vc_config() The main issue is that this creates an entirely new Repository object (since commit 3914388b2d0f) and thus reloads all the cubes, taking for ever. Related to #8634417.

[migration] don't handle data deletion anymore on schema changes In most cases when we want to drop some entity/relation type, we don't care whether hooks are called on their deletion. There is even low chances that some hooks still exists, based on an old version of the schema. Last but not least, this is horribly inefficient. So this should be clearly documented and handled by application's programmer if desired. This patch removes unnecessary deletion (because table or column will be later dropped) and reimplements the case of partial deletion (only one relation definition among several, hence the database structure isn't modified) using sql. Only one test regarding deletion of inlined relation def is added as other cases seem to be covered by existing tests. Closes #7023315

[hooks/syncschema] drop unused import

[hooks/syncschema] indentation fix

[server/test] cosmetic whitespace fix

[hooks/syncschema] don't set NOT NULL on inlined relation columns The NOT NULL constraint was being added on inlined relation cardinality change while it shouldn't. We don't want this, because this forbids to add an entity without the relation set at once (it may be added later in the transaction). This patch fix synchronization hook to stop doing this. Closes #6211101

[hooks/syncschema] Turn DropColumnOp into a data operation While writing a test for a migration issue where "NOT NULL" was added to an inlined relation column, I ran into a case where the same column was deleted twice, which of course results in an error. This is because the "ecrit_par" relation used by the test is ambiguous. I've hence turned the DropColumnOp into a data operation to ensure we'll delete the column only once in such case. Related to #6211101

[server/sources] fix name error in eid_type_source_pre_131 i.e. the eid_type_source implementation used when some pre 3.13.1 database is detected. Regression from 5fbdbbe51867.

[repository] deprecate the extid2eid based multi-sources API This API is cumbersome and lead to obfuscated code because of the callback mecanism implied when some entity has to be created. Since we dropped the "true" multi-source, this mecanism is not needed anymore and one should prefer to use the dataimport API inside its parser instead. Notice the cwxml parser will trigger the deprecation warning, it should not be used anymore in favor of ad-hoc parsers.

[ldapfeed] use source's extid2eid, the repo's one will be deprecated in a near future.

[ldapfeed / dataimport] port ldapfeed parser to dataimport API This makes the code easier to understand and will allow to deprecate the old multi-sources api, based on creation callback through repo.extid2eid. Currently with this patch, modification dates are not checked, hence entities will be systematically updated. We run the import with no hooks, because RQLObjectStore can only add entities to the system source.

[server/test] document expected packages to run ldap tests

[ldapfeed] simplify ldap2attrs * tdict argument is never given * separate value retrieval from setting output (will ease a latter cset)

[ldapfeed] simplify deactivation of unencountered users there is no need for the by_etype intermediary dictionnary since we only consider CWUsers.

[server/test] display rset content on failure