[pkg] merge centos 3.17.1-2 in stable (prepare 3.17.4)

Added tag cubicweb-centos-version-3.17.1-2 for changeset 965f894b63cb

[rqlrewrite] fix rqlrewrite unpredictability vs relation sharing. Closes #3036144 The relations index used to determine if relation may be shared only considered a single relation node per relation type. So when the same relation type occurs several time, dict order give unpredictable result. We shall properly consider all relations instead. Tentative test case included (but bug reproduction is by definition unpredictable...)

[rqlrewrite] rewrite doesn't need a solutions argument, always use the Select'ones Also, do a copy systematically else it may lead to empty solutions depending on the rewrite path. Test case tentative included in the following changeset.

[rql rewrite] equivalent but much simpler flow

[pkg] Prepare a new RPM release

[pkg] Add log directory to rpm

[repo] normalize ValidationError on edited entity (closes #2509729) In CubicWeb, ValidationError.entity MUST be the eid of the involved entity, not the entity iteself (as done by yams).

[doc] one must now manipulate the req.session.data dict (closes #2842345) instead of deprecated get/set/del_session_data methods.

[datafeed] add a timeout config option (closes #2745677) So a HTTP GET do not hang forever in the datafeed looping task.

[dataimport] ucsvreader should skip empty lines unless specified otherwise. Closes #3035944

[querier] Add timings to debug prints (DBG_RQL) this may ease to spot some problematic queries

[editcontrollers] Ensure entities are created in an order satisfying schema constraints. Closes #3031719 changes below are also necessary to make the whole thing works: * stop considering InlinedFormFile as eidparam field since they don't hold any value * rework 'pendingfields' handling to have separate processing of inlined fields whose subject entity is created during the edition

[editcontroller] extract RQLQuery.set_attribute/set_inlined methods from handle_field/handle_inlined_relation

[editcontroller] req=self._cw makes things easier to read

[facets,js] fix bogus checkbox icon appearing after the first interaction (closes #2790332) This is because of the hardcoding of the data/ url.

fix typos in docstring, doc and comments

[debian] remove workaround for broken sphinx/jquery combination sphinx in squeeze was updated: http://packages.qa.debian.org/s/sphinx/news/20130112T154737Z.html

[schema] mark CWDataImport as an internal type. Closes #3025536

[pkg] remove deprecated dependency on pysqlite2

Fix two crashes in db-check (closes #3024964) Also improves various debugging messages.

[rql rewrite] may_be_shared_with should consider relation's scope (closes #3024730) When a relation's scope is not the current statement, it must not be shared (instead of comparing the statement).

[rql rewrite] fix may_be_shared_with method so that it actually considers all variable infos and not only the first one

[schema/security] add __hash__ to rql expression. Closes #3013535 This is required so that when some rql expression participate to a dictionary key, only the expression is considered (consistent with comparison). This behaviour is expected in security at least, see the related bug for instance. This scramble msplanner unit tests a bit.

[rql rewrite] move some code from querier to rqlrewrite where it makes more sense. Also, make some minor cleanup/refactoring on the way and try to enhance docstrings. Relates to #3013535

Do not compute actions list in TableLayout view when display_actions attribute is unset Closes #3007281

Use the list of cubes from the filesystem when reading the schema from the filesystem Closes #3007259

[migration] when adding a cube, skip infered relations (closes #3005576) Skipped infered relations are recalculated later.

[3.17 migration] when some cube is missing, add_cube raise ConfigurationError, not ImportError. Closes #2981477

[migraction] rename_entity_type simply warn if old entity type isn't in the schema. Closes #3004069

[deprecation] fix uihelper deprecation warning, use the rtag name, not the class, so it may be copy/pasted

[test] makes fti test order-agnostic (closes #3005633)