[selectors] match_kwargs and match_form_params selectors both accept a new `once_is_enough` argument (closes #1990438) also fix docstring of match_exception and references to selectors in the book.

[cache] replace lgc.cache with something more appropriate (closes #1921713)

merge and fix cubicwebSortValueExtraction pb which disappeared when tablesorter.js has been updated

[table] use explicit add_onload() to setup table sorting (closes #2013493)

[schema display] display_name shouldn't call .lower(). Note the only proper way is to have different msgid for upper/lower cases. Closes #1105433

[ms repository] when an exception is raised during extid2eid and no rollback is done, some manual cleanups have to be done (closes #1993420)

[facets] on the way to vidargs removal our roving on the table view, which leads to introduction of this feature, proved it was not the way to go.

[migraction] fix cmd_change_attribute_type command (closes #2015073) use dbhelper to avoid issuing unsupported SQL

[action] has_editable_relation first argument is once_is_enough flag, not an action name

[table] Only apply sorting event on the first row of a table header (closes #1988780) We use the new selectorHeaders from table-sorter 2.0.5

[table js] Update jquery-tablesorter to 2.0.5 (closes #1989018) The cubicweb specific code is moved to the table view. This version add a new useful option for #1988780

init_facets: Add documentation init_facets is not included in the default sphinx documentation but this docstring should help people digging into cubicweb core. This could go in an Advanced developer section of the book.

[facets] try to get rid of arbitrary constants, be prettier and eat less space (closes #1988706)

[edit controller] add form to pre_web_edit signature to ease linkto usage (used by vcsfile cube)

pylint extension

[printers] fix bug introduced by final view refactoring:   shouldn't be escaped. Also remove some msgids which should have been removed during this refactoring

Add a hidden input:submit field in facets form (closes #1988784) <Enter> is supposed to submit the form only if there is a single input:text field. However most browsers will submit the form on <Enter> anyway if there is an input:submit field. see: http://www.w3.org/MarkUp/html-spec/html-spec_8.html#SEC8.2 Firefox 7.0.1 does not submit form on <Enter> if there is more than a input:text field and not input:submit but does it if there is an input:submit. IE 6 or Firefox 2 behave the same way.

AnyRsetView: Refactor the columns_name method and update doc. We drop the 'label_column_%s' % colidx pattern and add a clean column_label method that might be overwritten by children class to alter the way default columns headers are computed.

[facets] FacetFilterMixIn: add documention for class and generate_form There is also a small refactoring to make the code clearer. An intermediate variable is introduced and the call is reindented to make it clearer. No logic change.

[facet] Add infered type restriction to Filtered result set (closes #1988776) This does not alter the result while allowing more selection option on other variable than the filtered one.

[source] remove uicfg rule for dropped synchronizing attribute

backport stable

[datafeed] remove remaining uses of 'synchronizing' attribute replaced by 'in_synchronization' in 3.13.8. Closes #1989131

[datafeed parser] only update an entity attributes when needed. Closes #1989142 This may avoid error with some entities which have immutable attributes and don't check something actually changed. Also, avoid modifying the cached item representation, this may cause pb when an item is seen several times during an import.

backport stable

[schema] drop erroneous default on CWSource.in_synchronization. Closes #1990585

[forms] fix bug in linkto refactoring (misordered arguments)

[migration] update old yams api in change_attribute_type

[schema] add missing import (trigger warning)

Added tag cubicweb-debian-version-3.13.8-1 for changeset 43f83f5d0a4d