[schema migration] import refactoring to fix #1109558 and enhances things on the way the main pb demonstrated by #1109558 was due to the fact that in-memory schema was updated in commit_event of operations. This is undesired in most cases, since we want the modification to be taken into account in the interval between the modification detection and the commit_event. I've fixed this by merging in-memory schema / database alteration operations for most important changes, doing in-memory schema changes as they are detected and implementing a revertcommit_event method to revert them if necessary (with exception for removal of stuff from the schema, where this is simply done in a postcommit_event methods). Also, I've benefited from this to support reverting of database alteration for some operations (more to be done there), and to move so system source alteration code to the native source code for a nicer design. There may be some more stuff in syncschema.py that would benefit from similar changes, but most important things are done (at least to close #1109558, w/ unittest_syncschema and unittest_migration green).

backport stable

[test] running transaction while session is closed are now killed, update test accordingly

[security] fix read rql expression insertion: we should not insert rql expr on variables only referenced in neged relation

[rql2sql] test and fix crash when OR in HAVING where both expresion are 'fake' having terms (eg not using aggregat function)

cleanups

[config] use hashlib to avoid warning w/ py2.6

[req] hashlib has no hash function, but has an hash1 function even in python 2.5

[utils] improve repr() of RepeatList

[doc] fix typo

[selectors] use before-registry-reset event to init is_instance cache: cleaner code and avoid reloading bug (making hooks test fail for instance)

[vreg] new [before|after]-registry-reset event, taking vreg as argument

[doc] update book for adapters

[reledit] fix relation cardinality handling

[selectors] provide a new, optimized, is_instance selector that should at some point replace implements (along with the adaptable selector)

[web/views/massmailing] fix cancel button msg/icon

[testlib] improve test description messages

[web/views/reledit] fix typo (-add we used as id component in delete id)

[web/test] fix test (doreledit must have a formid and this is now checked)

[web ui] add missing image

[ITree] small fixes: parent_entity is a IBreadCrumbs method, we should use .parent on ITree

restore bw compat on Entity.related_rql

reledit refactoring * js handling rewritten to exploit all form capabilities (such as file upload ...) * attribute-like composite relations (one ttype, cardinality in '?1' from composite side) are more cutely handled, with a 'add'/'delete' additional actions/icons * a reledit_ctrl rtag to finely control: reloading, edition and default values * a proper chapter in the documentation (book) * many bugfixes

backport stable

cleanup

[idownloadable] make the download view http cache friendly

[publisher] avoid useless rollback after successful commit, which clutters debug logs and may also not be cost-free

[controller] same method defined on the base class

backport stable

merge stable heads

Added tag cubicweb-debian-version-3.8.6-1 for changeset 607a90073911

Added tag cubicweb-version-3.8.6 for changeset 1a24c62aefc5

[dbapi] refactor so that cw 3.8.6 can still speak with a 3.8.5 server

cleanup

[bfss] fix file update to ensure file's content is available on the fs asap... and not only at commit time. So it's consistent with entity creation behaviour. The new file is created at assignement time and removed if the commit is rollbacked.

[ldap] fix exception handling on authentication error

[ldap] cleanup password checking

cleanup

[css] make body bg color configurable

bump version number to trigger migration

3.8.6

[migration] schema should be accessed through .repo

backport stable

[security] use set_operation for relation permission checking operation

[migration] refactor schema migration fix introduced by 5833:d7256ae7c1d1 so the code looks better and avoid clutering lgdb api. Depends on lgdb 1.0.5.

[schema migration] fix name error introduced in parent changeset

[schema migration] avoid potential duplicated ft reindexation due to multiple UpdateFTIndexOp operations by using set_operation

[sql generation] fix bug with inlined relation and referenced by other neged relation we should check in that the column is not null in such cases

[refactor] refactor rtag keys generations.

[fix] fix a uicfg bug in autoform_section When several rtags matches the same relation (for example "* login *" and "CWUser login *") inconsistent rtag could be merged.

[debian] fix logrotate error introduced by 5806:9ef0e95f4d2b: sh don't want spaces here

[debian] fix logrotate error introduced by 5806:9ef0e95f4d2b: sh don't want spaces here

cleanups

[login form] fix bad login field label by using pgettext

updated docstrings on Operation class

do not allow password-less bind: windows accepts this as Anonymous Logon

[#969169] don't display remove link in edition forms if the cardinality is 1

fix name error

fix schema migration hooks (partial) SQL Server requires some additional steps to alter columns (index drop and restore, etc...) and this in turn requires passing a connection on the DB to find out the indexes. Also partially fix an issue where the null_allowed value was not correctly computed. Still some problems in there (see https://www.logilab.net/cwo/ticket/1109558)

delete trailing whitespace