[autoreload] recompute urlresolver / urlrewriter after autoreload

[autoreload] remove all modules under vregistry_path() from sys.modules before reloading

introduce a new CubicwebEventManager Typical emit usage: >>> from cubicweb import CW_EVENT_MANAGER >>> CW_EVENT_MANAGER.emit('after-source-reload') Typical bind usage: >>> from cubicweb import CW_EVENT_MANAGER >>> CW_EVENT_MANAGER.bind('after-source-reload', mycallback)

[F dbdump] dark and dirty fix for db-copy command

cleanup

[R] use req.entity_from_eid

[reledit] refactor, split into reledit/doreledit where the former try to tell if edition *should* be done, and the later tell if edition *can* be done

use yams api

[table view] vreg api update

[navigation] split tests, add test for __start out of bound

merge

merge (after jquery.tools backout)

Backed out changeset 3e1d2ab5f8c0 (restore ui.tabs.js instead of jquery tools)

[reledit] don't try to use rel-edit on attributes without assicated fields (e.g _format field)

[uicfg] fix metadata detection in primaryview_section

[vreg] api update

to avoid confusions, implements iteritems/itervalues as well

[testlib] take care of re-monkeypatching which'll cause infinite recursion error

[repo] explicitly specifies edited attributes to glob_update_entity

[webapp] missing renaming

[pyro] use lgc.pyro_ext, simplify pyro related options

[we, test] more api update

[controller] should catch and set appli argument here now that the old AppObject class is gone

[uicfg reloading] don't reload uicfg when appobjects modules won't be reloaded as well

[vreg api update] remove some deprecation warnings

[test] update test: class initialization done in etype_class now

[vreg] fix deprecation messages

[appobject selection process] drop the need for the .selected method

[appobject] kill VObject class, move base selector classes to appobject

R [base classes] only AppObject remaning, no more AppRsetObject

merge

[twisted] don't use twistd anymore, all-in-one.py file is needed anymore

[uicfg] reload uicfg when vregistry is being reset

[refactoring] moved rtags specification in web/views

[registry] when a source file is changed, reset and reload the whole registry

R [vreg] important refactoring of the vregistry, moving behaviour to end dictionnary (and so leaving room for more flexibility ; keep bw compat ; update api usage in cw

B [supervising] ignore updated event for entities created in the same transaction

F [rset repo cache] set entity.rset when no set on entities retreived from the cache

enable server side entity caching, 25% speedup on codenaf insertion. ALL CW TESTS OK

#F [querier] debugging tweak

F [adapted list view] fix selector

F [view] missing import

web test update

cleanup

R [schema hooks] big refactoring / reorganization for clearer code, a few fixes on the way

uicfg bugs detected by tests

R import to avoid warning

oops

R [test] update and fix deprecation warnings

R [web test] rewrite schema with a single python module

F [cw.test] OK

explicit imports

get_csv is being renamed to splitstrip

B [schema update] may remove has_text unexpectedly, give another argument to fix it

cleanup

R [session] property for quick access to repo.schema

R [querier] debugging tweaks

document and replace debugged by a contextmanager/decorator class

manual_insert is a public method

F [server test] avoid warning by adding __init__.py file to directory

R [server debug] more server side debugging tweaks

cleanup

oops, should have been in a earlier commit (give rollback_on_failure to doexec)

oops, should have been in a earlier commit (META_RELATION_TYPES renaming)

R [querier] debugging tweaks

R [sql source] cnx._cnx checked in cnx.close, don't do it here

debugging tweaks

[F repo sqlite schema changes] don't rollback on potentially expected schema changes failure

[R schema hooks] rename META_RELATIONS_TYPES into META_RTYPES, use it and other schema consts to avoid errors

merge

F [dialog] offer to create backup. refactor to use l.c.shellutils.ASK

F [doc] mark code block as python source in cookbook

R [all] logilab.common 0.44 provides only deprecated

[doc] update generate_modules to match new sphinxutils API

[R source cnx] use connection wrapper's close method

[F native source] fix attempts to clear cache key

[F workflow repo] don't try to get previous state of newly created entities

[F repo tests] tests have to be updated: * fix schema (stuff missed while rewriting old .sql / .rel files to python) * '* in_basket Basket' now consider all entity types * new ExternalUri / cwuri ertypes * group eids have changed

[F hooks] skip new eids as well as pending ones in DelayedDelete operation

[R hooks] use DONT_CHECK_RTYPES_ON_[ADD|DEL] don't compute rset if not necessary

[R hooks] use DONT_CHECK_RTYPES_ON_[ADD|DEL] constant, don't check wf related internal relations

[R repo session] add_relation method use in hooks

[R hooks] optmize insertion by using direct repository methods when possible

[F schema] set inlined=False to ensure it'is not inlined by accident

[R querier] new READ_ONLY_RTYPES constant

[F repo debugging] log repo event on DBG_REPO debug level

[R schema hooks] use new accessors on CWRelation and CWAttribute entities

[R schema] dont give an iterator, it's a trap

[C schema seria] cleanup

[R repo, schema] use VIRTUAL_RTYPES const

[F server test] sort for test predictability

[R schemaobjs] property accessors for CWRelation/CWAttribute

F: start to handle binary debug log level on the server side

rewrite migrscheam

C remove no more used copy_schema

[D server test] remove deprecated test on application_hooks file, and associated data

backport stable branch

[F schema serial] #344876: missing some 'is'/'is_instance_of' relation for newly created instances

[vreg] consider __abstract__ in register_all

[xhtml form renderer] fix html structure, move header where we are sure to display it when necesary (_render_fields not displayed if main form has no visible fields

missing new renderer argument

merge

Add missing parameter to widget render call.

merge

merge

always explicitly close <base> tags because of an old IE6 bug

[hooksmanager] nicer unregister_hook implementation, allowing to give a Hook subclass as argument

[migration] reset do-not-insert-cwuri once it's properly in the schema

cleanup

[F delconf form] should give rset to select()

[form renderer] fix composite form renderer in case the main form has some visible fields

[form renderer] consider form.fieldsets_in_order if specified. Warn about unexistant or unused fieldsets.

add default values for fielsets_in_order, document base form attributes

[notification view] consider row/col

oops, this is a method

[repo session] ability to ask session to keep it's pool set even when only read queries are done, necessary at least during migration

no rset hooks should have empty rset now

d-t-w

avoid execution of queries which are known to be wrong by letting error propagates

missing db-port option to source's option definitions

make notification views working with no rset

[F plot] make jquery plots accept a fourth column that specifies a date format for tooltips

merge

[doc] #120075 quickstart (was proposed structure)

[F xhml] use a class to avoid duplicated ids when editing several entities at once

merge

[R xhtml] remove xhtml_wrap* function, use instead a single req.document_surrounding_div method

move test on force-html-content-type to xhtml_browser method