Added tag cubicweb-version-3.13.3 for changeset 0e82e7e5a34f

[rql2sql] take care to boolean_value returning an int (eg sqlite/mysql)

[repo test] remove debug

[repo ms] commit/rollback session if it has been created in the method

3.13.3

i18nupdate

[sql] more test and fix boolean handling

backport oldstable

[repo test] Fix test broken by BFSSTestable changeset bd56a29acaa8 broke several fragile test.

[sql] fix querying of boolean value w/ sqlite

backport oldstable

[datafeed] for datafeed source, we don't want commit in extid2eid but explicitly handled by the source. Also, we should use 'safe' internal session

[sql gen] handle optional on comparison node (eg HAVING expression) and on rhs of final relation. Closes #1859609

[session] safe internal sessions don't deactivate integrity hooks

[sql test] lgdb > 1.6 use int for bool with sqlite backend

[rql2sql] don't introduce substitution for Int and Boolean constants. This is needed for COUNT(DISTINCT) support

[css] syntax fix in cubicweb.css

[entities] skip_copy_for should be considered for object relations too (closes #1857474)

[datafeed] case insensitive etype. Closes #1877014

[datafeed] closes #1875322: empty tag means None value

[datafeed] properly take care to cwsource=system in imported xml. Closes #1877017

[datafeed / fti] rather control a 'complete' parameter than setting empty attribute values

[repo api] deprecates source.extid2eid, only the system source should implement it, and one should directly use the repo api instead of this confusing and error-prone short-cut. Deprecates eid2ext as well for consistency.

backport stable

backport oldstable

[bfss] Fix update of BFSS attribute to None (close #1875289) The current code assume attr value is alway a valide Binary value. But when the attr is optional, the value might be None. The change handle this case.

[skeleton/pytestconf] provide an acceptable solution for windows, closes #1875488

[facets] make sure has_text facet data is posted correctly (closes #1857405) Posted data was associated to facet id (e.g. has-text-123098dasdf898adssdd) rather than facet name (i.e. has-text).

[rql] fix min/max args definition of [TEXT_]LIMIT_SIZE registered procedures (long outstanding bug discovered after recent fix in lgdb)

[schema] rephrase relation definition since constraint doesn't apply to CWRelation

[datafeed] closes #1873616 (user's url corruption when etype is not recognized

[schema] restrict source mapping 'cw_schema' relation to non final relation, we currently have no use for attributes

[c-c db-check] add some sql queries to fix cw_source/is/is_instance relations according to data in the entities table

[c-c db-check] missing comma ending print statement to get expected output

[c-c db-check] disable write security as well so we can delete stuff when needed (e.g. orphan TrInfo)

[c-c db-check] having no owner is allowed, don't try to fix that

[metadata, ldap] make backport of ldap user actually working

[c-c db-check] drop some deprecation warning

[index page] use 'New xxx' generated labels, 'add a xxx' labels are only generated according to inlined form configuration and with an associated context

[doc packaging] prune _sources containing generated text file, include js else search doesn't work

[boxes] 3.13.2 regression: should specify dom id else we may generate bad xhtml

[facets] fix missing 'warnings' import and missing self reference on 'filtered_variable'

stable is now 3.13

oldstable is now 3.12

Added tag cubicweb-debian-version-3.13.2-1 for changeset 9ad5411199e0

Added tag cubicweb-version-3.13.2 for changeset f9227b9d6183

3.13.2

[rql2sql] closes #1832859: fake HAVING terms w/ EXISTS terms there are cases where we want to use the HAVING trick on terms that should belong to an EXISTS subquery. Currently , this causes bad sql. We should fix that until we've a proper rql grammar avoiding the need for this trick. Notice at some point, we may want **actuall** HAVING clauses and GROUPBY used in EXISTS subquery, but that's another story.