when instance is started normally (eg not in debug mode), check versions before daemonization

doc

close #511810: bad rql generated when looking for vocabulary for a relation on an entity which doesn't exist (yet) technical details: * add a graph of linked variables in the RRQLExpression instance * modify rql rewriter to remove relations where an unexistant variable and or a variable linked to an unexistant variable is used. * activate this feature in Entity.unrelated_rql if the entity doesn't exists yet

typo

consider U when computing main variables

fix cut and paste

do not fail if anon user == manager, simply emit warning that it should not be done in production. close #469395

oops, junk

c-c create stopped asking if anonymous access should be activated introduced by 2654:6512522860aa, close #469418 Generic behaviour of WebCreateHandler should be kept for twisted and all-in-one configurations.

no more dependency on mx date time. Better description of optional dependencies

[web ui] move help from header to footer

merge

[web] set pageid at request instanciation rather than in htmlheader template In some occasions, we need to force a pageid: typically an ajax-rendererd creation form would need to use the pageid of the main page to correclty set it's page_data (e.g. rql varmker). If pageid is found in req.form, it's used, otherwise we build a new one with. make_uid(). The behaviour should be consistent with the former one in JSonController.

test and fix http://www.logilab.org/ticket/499838 refactor nicely on the way

merge

Backed out changeset 8d0b5de8002b

[B] migration: normalize migration script path

small fixes

#509981: cw-ctl shell <instance> file.txt runs doctest.testfile

[web ui] PrimaryView.render_attributes now uses a table

[web ui] primaryview_section tags for CWGroup

[web ui] change "table actions" icon to arrow-up

[web ui] remove "trame_grise" from forms

[web ui] rename form buttons to ok/cancel where workflow change state

[i18n] fix typos

[web ui] decorate form buttons with icons (at last)

[web ui] very simple default configuration for ckeditor

[clean] remove print polluting test outputs

[web ui] fix the editable fields: better look and behaviour.

[icons] make nomail icon more recognizable

[doc] added note about the required js and css in TableWidget

style: s/pourcent/percent/

added SimpleModel for TableWidget

[doc] fix inaccurate docstring

[B] views: fix wrong xhtml when empty rset

do not display anything when bein asked something (fix #504759)

fix wrong answer when the rset is empty but did not raise

put back url info started at the end of initialisations

[B] web: weird display with internationalizable (Closes: #503071)

update docstring, xml escape exception description

[migration] warn if 3.5 workflow migration is going to fail

C - fix typo. make `lgp check` happy.

force debug mode on cubicweb-ctl start on win32 this prevents a crash (daemon mode not yet supported)

read form arg specifying json call context (to make it work under json controller)

[javascript] handle cases where vid parameter is hijacked by parameter-injection

[views] consider fallbackvid request parameter in ViewController

[cleanup] reorder cubicweb namespace attributes declaration

[javascript] asURL now escapes request arguments I can't see no good reason for not escaping parameters. The two main locations where asURL is used are : - edition view: here, the change should have no effect - facets: escaping is clearly needed in that case

merge

remove obsolete caveat

so bad tests hasn't find that one...

backport stable branch

Added tag cubicweb-debian-version-3.5.5-1 for changeset aad818d9d9b6

Added tag cubicweb-version-3.5.5 for changeset 1eca47d59fd9

on precommit / rollback of schema update op, clear order_relations cache of entity schemas

don't allow that even with cow powers

fix reledit related test breakage

test fixes

merge

test fix

nicer relation|entity schema view

reledit: stuff the value into its own div and properly hide it when necessary (but dont lump it with the landingzone div for it switches the form on when one clicks on a value to traverse it)

work around a pb. with subquery_selection_index

latst_trinfo : do not crash on absent wf history

drop _prepare_copy method from rset in favor of a more generic copy method

i18nupdate

prefix sql aliases for entity table by '_' to avoid pb with variable such as 'AS' (eg a keyword in SQL)

small cleanup & revert to less intrusive icon

actually hide the value when editing, cleanup spurious arg

merge

yet another brain fart

2.4 compat

remove spurious wview

deprecated api update

document

merge

when a query is grouped, ensure sort terms are grouped as well

take care to outer joins

display title in the component to avoid relying on view supporting title argument

turn make_relations_query into a true generator (this allows interleaving of actual entities linking and query generation) restore check botched in previous inadvertent commit

node ids shall not begin with a number

remove debug print

accepts single values as create_entity

move these definitions below cell_call

reledit: load actual edition form on demand #471799

merge

kill separate attribute client-side handling #473636

use cuibicweb json encoder for json_dumps, still catch the TypeError in the request-less jsonize

[doc] remove toc.rst which is not used and conflicts with index.rst and fix table of contents numbering

oops

disable syncschema hooks during db initialization

finally

take care to already existant solutions local variable

bad merge

ignored_relations should be a set

remove some warnings

really need some holidays

mark class as abstract

editable-attributes return a list of 2-uple

remove some warnings

dumb name error

typo

forgotten some part of a merge

bad merge

.id has been renamed

be quiet w/ apycot

2.4 compat

2.4 compat

backport stable

typo

restrict follow action to one line rset

cleanup

merge

[shell] make process_script available throuhg c-c shell / migration script context

ugly workaround to javascript serialisation problem #481437 (not considered fixed by this changeset)

merge

revert change, but there's something fishy here

work around a problem in constraint application

should remove entities from kwargs (won't go through pyro)

properly close remote connections

update prototype

[c-c] fixes for shell w/ pyro instance

oops

get dispctrl through extra_kwargs to avoid error when a custom view is specified (potentially not taking any argument)

a way to distinguish __init__ vs render arguments

merge

skip final type

check we've some message to display

#481017: cubicweb-ctl shell on remote instance

initialize entity classes in repository initialization

revert to adim fix

merge

missing args

[workflow] fix NameError in fire_transition()

cleanup

[optimisation] insert is / is_instance_of using sql

clear eschema caches on schema changes

reimplements some methods to traverse operations in reverse order, much efficient w/ transactions where there is a lot of pending operation

skip tidy hook for super session

follow yams 0.25 api changes to improve performance

merge

update for new create_entity implementation

[doc] document rql operator ILIKE

no integrity check for super session

support _cw_unsafe as set_attributes/set_relations

fix related_rql when target_type specified

add kwargs to some base views

fix max-related of related entities handling in primary view

oops

only copy when necessary

add note

fix i18ncube

optional nbsessions arg

fix merge

backport stable branch

[testlib] hummm

[entity] make related_rql honors its targettypes argument

new set_relations method on entity

allow to link to multiple entities

nicer create_entity implementation (and test)

merge

use unicode in the example of create_entity

fix migration script

fix clear_all_cache: ensure we don't remove eid from entity's dict

use typed_eid

simpler implementation

defer trinfo creation when cancelling subworkflow

add a note

fix/prepare propagation hooks usage

3.6 update

pluggable authentication information retreiver

factor out code reusable for authentication tests

missing authorized value

3.6 updates

composite relation, should'nt be needed

[req] new expires argument to set_cookie

[testlib] nicer http headers and cookie faking in fake request

fixed cut and pasted docstring

ignore some internal warnings prevent the display of warnings caused by the definition of deprecated functions so that the output of cubicweb-ctl is not polluted.

3.6 updates

on entity deletion query, return eids of deleted entities

refactor repo authentication to allow pluggable authentifier to login with something else than a password

fix bad merges

3.6 updates

stronger wording

oops

...

fix debug attribute conflict on configuration

oops

more doc

system/user modes + CWDEV instead of installed/dev mixed modes. Fix behaviour when setting CW_MODE explicitly

unused

[reledit] display value anyway when should not edit attribute/relation

do not override __iter__ on entities

don't duplicates restrictions

[multi-source] add test case for an encountered bug, fixed by adding state_of to dont_cross_relations. Added a note to remind me that something should be done to be smarter in handling this

[rql2sql] test we don't crash on dumb queries

give correctly typed eid

oops, feature killed by merge...

backport stable branch

[wf engine] support for subwf exit point with no destination state: go back to state from which we entered into the subworkflow

fix title of CopyFormView

[doc] updated docstrings

fix bug w/ get_entity and subqueries

fix buggy assertion

avoid progress bar when test launched from apycot

merge

[wf] fix detection of transition where we've entered a subworkflow

missing translation

enhance docstring

oops

add a commit point, a new transition exhibits the problem again

a problem with consecutive transition firing when subworkflows are involved

merge

Added tag cubicweb-debian-version-3.5.4-1 for changeset 37d025b2aa77

Added tag cubicweb-version-3.5.4 for changeset 312349b3712e

[migration] make it more explicit that BaseTransition upgrade should only be done if migrating from version prior to 3.5

[migration] fix 3.5.3_Any.py for migrations starting from before 3.5

3.5.4

ensure relation is properly created

[editcontroller] backout (sort of) removal of entity.complete() in validate_form entity is passed to js callbacks when there's one. In that case, we want as much information as we can. The removal was there to avoid complete to fail when all constraints (required relations / attributes) aren't satisfied. This fix only does the complete() after the commit is done, any ValidationError should have been raised at this point.

don't crash if specified page size is greater than rset size

fix startup as a daemon

fix debug msg