backport stable

DEFAULT_ATTRPERMS is now public

[inlined formos] don't pop attributes from kwargs, so they end-up in cw_extra_kwargs which is then passed to the edition form

[notification] we should use unsafe_execute here

cleanup

[calendar view] fix related to the date_range mess (max value exclusion). Calendar views authors should check if more similar fixes needed

[forms] fix RelationField bug with pending eids (eg inlined forms): we have to override process_form_value to avoid caching when recomputing is needed

[shell] rql() should use unsafe_execute when possible

[autoform] we should consider role when checking delete permission

don't update modification_date in repairing_mode (eg upgrade or shell)

[security] we should save back edited_attributes in case of multiple modification of an entity during the same transaction

tune startup logging

[schema] fix RelationSchema.has_perm to properly works with attribute relations

[schema] refactor/cleanup check_permissions

use a close_form method for consistency w/ open_form and allow overriding

add version info to deprecation warning

remove debug print

follow yams api change: attributes permissions are now defined for an 'update' action, no more 'add' / 'delete' which makes no sense in such case. fix afs.relations_by_section permissions checking of object relation on the way.

fix dumb name error causing error when non managers try to change his properties

fix typo in deprecation warning

remove crappy basket specific code

[views] toolbox must be rendered before the title to appear at top right

[uicfg] make deprecation message easier to understand

[uicfg] allow autoform_section's formtype parameter to be a tuple Typical usage is: afs.tag_attribute(('X', 'y'), formtype=('main', 'muledit'), section='attributes')

merge

branch oldstable

create oldstable branch, remove 3.5 tag

merge 3.6 default/stable

merge 3.5/stable heads

Added tag 3.5 for changeset a0571ff0cb5d

[http cache/json controller] ensure json_view does proper cache validation #390986

merge

typo fix in help string

use subprocess instead of os.popen to run diff this avoids argument quoting issues (spaces in path, etc)

[migration] abort becomes possible when asked for confirmation before migration script

sort on sortvalue, not dc_title to allow customization

merge

fix bug when value is None

Added tag cubicweb-debian-version-3.6.0-1 for changeset 17e88f2485d1

Added tag cubicweb-version-3.6.0 for changeset 5f957e351b0a

added jquery.ui image files

provides a fallback_on_none_attribute field attribute, allowing to specify default value for attributes of *existing* entities

fix match_transition selector: for instance w/ forge version change state form overiding, we want the form with the publication_date selected by the editcontroller, whatever the transition

split introduced in the wrong place

packaging fix: no more test in common (2).

packaging fix: no more test in common. Should remove test from the hooks subpackage in rules for hardy

drop cubicweb-client package (there was only hercule.py inside)

remove ref to hercule.py

cw 3.6, itself

i18nupdate

i18n msg fix

cleanup 2.5 __future__ imports

update translatable docstring

fix test

rephrase

fix auto-rollback on unauthorized error: * don't rollback if a commit is processing * don't free pool on rollback

fix ldap test, our ldap server name has changed

fix name error introduced by bad merging of stable

fix 'click here to see the created entity' link, which may not appear according to language settings

quick fix for typed formfield validation: accept empty string, split

[R] devtools: improve default data import mechanism Validation chain is now possible with checkers Before that the expected values needed to be coherent. Now, we can use ObjectStore to validate the input data * add new input transformers: - uppercase - lowercase * add new input checkers (raise AssertionError on error): - decimal: take care of possible comma character as number separator - integer: cast to int() - yesno: to validate boolean value - isalpha - required: input value *must* not be empty * new control checker: - optional: block possible exception we delete field in the returned dict instead of raising AssertionError (exclusive with required) Helper methods to manipulate indexes: * build_rqlindex() is used to build index based on already created entities * fetch() replace get_one()/get_many() methods by factorizing code Minor changes in reporting: * use tell() for all printing * let new value for askerrors to display automatically the report (used in crontab)

merge

backport stable branch

[cw-shell] do not set arbitrary limit on number of arguments

use 'edit' as defaut path for expect_redirect_publish

propagation hooks cleanup

missing import for 2.4 compat

don't play with the custom_workflow relation in autofill

fix bug w/ object relation used as inlined form

cleanup

backport stable

turn down some info logs into debug

turn down some info logs into debug

fix constraint hook: a session may be involved in multiple transaction at time

fix and begin to document autofill algorithm

unset_attribute_storage, for testing purpose at least

oops

backport stable

when no context specified, should return non-zero score as documented

typo

no need for a specific vid argument here

backport stable

merge

merge

add a note about modifying read only attribute cf. syt's commit message in 81021d62138f

merge

use calendar.gmtime instead of time.mktime to avoid UTC/localtime issues with Flot This is suggested by the Flot documentation on http://people.iola.dk/olau/flot/API.txt

must use in-place modification, pending_operations is a read-only property

avoid to see the same warning twice, one for 3.6 another for 3.3

backport stable

fix session's commit handling: when an operation failed in [pre]commit, call rollback_event as well for already processed operation, not only the revert<event> callback

race condition may makes this assertion fail, we don't want that

unvisible field may have been modified, for instance relation added by an hidden input

AdaptedList -> SameETypeList *NO BW COMPAT*, benefit from cw 3.6 releasing of folder,file and blog which use it to get update at the same time. CMHN and PEGASE will need update (but won't go to 3.6 without update, so seem fine).

now raise UnknownProperty, not KeyError

fix test broken by introduction of ViewSameEType action

kill register_appobject_class method, simplifying autoregistration. Update test accordingly (test which manually registers object should also properly call there __register__ method). Drop the disable-appobjects config file entry: no one used it since its introduction years ago.

backport stable branch and some vreg cleanups: * move initialization_completed from cwvreg to base VRegistry class allowing simplification of CWVregistry * cleanup initialization process: __registered__ is now called after initialization completed, by the relevant registry. * fix/remove deprecated tests

fix security issue #36564 (integrity error w/ RQLUniqueConstraint) by using a global lock serializing commit when some RQLUniqueConstraint is encountered.

errors on postcommit should never occurs, log it as critical

missing test data

reminder, cleanup

initialize property once initialization has been fully completed, close #666573

fix duplicated vregistry initialization during tests

cleanup, don't fail if no context set on the sourcecode directive

we should not use inheritance there but proper view redirection, so if one adapt listitem view, he gets the same thing by default with the adapted list

hidden usage of datetime function which has been moved to lgc

i18nupdate

replace DevCubeConfiguration/DevDepsConfiguration by a single DevConfiguration, fixing recently introduced by w/ i18ncube command on the way

moved to lgc.date as well

oops, forgot that part of 3.2 bw compat

hercule (aka c-c client command) have not (ever?) been used for years, stop supporting it

kill most pre 3.2 bw compat code

kill mx compat code (dropped since 3.2), more efficient merge_args implementation

this test is now useless here

3.6 test api update

db fill api has changed, update test accordingly

no needs for a views package, 3.6 api update

need post_build_callbacks attribute with earlier yams version

update yams/rql dependencies

fix long-waiting symetric typo: should be spelled symmetric. Add auto database migration on schema deserialization

moved generic datetime manipulation function to lgc

fix wrong autoclass inclusion

fix rest syntax error

more selectors doc in the book, mostly extracted from code

reorganize, cleanup and properly document base selectors. Kill the may_add_relation selector.

3.6 api update

clear rest_path __unique cache in clear_all_caches, fix related test

backport stable

copy back __doc__ for proper sphinx doc generation

mimick actual request set_header method

fix typo

[views/treeview] fix non-ITree nodes last element computation

[request] add a note about the encoding mgmt (or lack thereof)

[views/plot] replace bogus jsoncall computation by something (statistically) more reliable (json_request being set to True in the JSoncontroller

[doc] small fixes on security part

misc typos found on my way: NameError / SyntaxError / double __select__, <sigh ...>

fix bad rest

fix bad rest

more details on selectors debugging

add_state/add_transition are now deprecated, only add_workflow remaining

damn me, more stupid sed fix...

add note telling you don't et an UStreamIO for binary views

document the ~ operator

start documenting forms

fix stupid sed

moved content to the dbapi section

nicer docstring

live-server doesn't work, don't make think it does

more stuff in doc's README

update exemples using the 3.6 api and add/fix some sections (schema, vreg, talk about CW_MODE in concepts...). So much to do :'(

backport stable

discriminate against None, not False-ish values

Improve the schema command with filtering option. meta data and workflow data are hidden by default. show-meta and show-workflow option are added to display them. Another "hide-type" option allow to hide additional entities type.

Added tag cubicweb-debian-version-3.5.12-1 for changeset 4281e1e2d76b

Added tag cubicweb-version-3.5.12 for changeset 98af3d02b83e

Updated CW tutorial. * Summarized the list of steps to create a new cube. * Added a subsection on writing entities. * Added a subsection on modifying the schema and updating the corresponding instance.

le patch added_tips a été importé

le patch admin_tips_for_backup a été importé

Add a ``schema`` command to cmd ctrl to generate schema image. This changeset add the new commande and do some refactoring in cwconfig and schema.py to allow the use of CubicWebNoAppConfiguration with CubicWebSchemaLoader.

fix rset.related_entity with variables coming from subquery while some others not

3.5.12

cleanup

fix spurious http cache bug: sometimes last-modified headers is generated using non-english local, which ends up in a date that twisted can't parse and make it feels the page may be cached while it may not

simpler last_modified implementation

merge

fetch all forms, avoid potential initialization pb

missing jsonize, avoid spurious error w/ ff

refactor to use the 3.6 API

merge

merge

partial backport of support for adbh's new interface to backup_command and restore_command these methods can return lists or strings which are passed to subprocess.call without or with the shell=True parameter respectively.

more escapes for the backslashes

escape the backslashes

bugfix

add action to list entities of the CWEType currently displayed

should take arbitrary arguments

i18nupdate

calling unicode instead of gettext...

[entity] small optimization: once an entity has been completed, don't redo it (for nothing)

[treeview] ensure items are ordered according to their dc_title, we can't rely on an ordered rset * is_last has to be computed on the caller now, not possible in the callee * rely on b3f61c38526b to use entity.view() with vid in *view*'s arguments

[entity] rename vid argument of .view(...) into __vid to avoid potential conflicts with kwargs

eschema view is gone away

remove misnamed argument

add version number to deprecation warning

[entities] fix CWUser.is_in_group implementation use self.groups property instead of self._groups which might not be initialized.

missing import

must check for relation inlined in main form type

skip final relations at this point

add missing errno import in etwist/server.py

should not raise KeyError if key's missing

choices function should now take the field as argument

cleanup

missing import

systematically call bool on BooleanField values

by default this is not the widget responsability to turn empty string into None, move this behaviour to the field.

ensure we get a correct key

test fixes

useless, same as test_nonregr_eetype_etype_editing

[forms] propagate aguments given to .vocabulary to .choices if it's a callable. So we can properly uses vocabulary(limit=xxx) in places where we used to call .choices(limit), since field.choices is not meant to be accessed directly. Also, document expected behaviour for custom choices method for relation fields.

reorganize code: move everything related to inlined forms and generic relation from editviews/editforms where there are used, eg in autoforms

[forms] fix backward compat for field.choices arguments

use field.vocabulary instead of field.choices, skip already related eids. Also consider vocabulary when no etype specified

[primary view] activate toolbox

it seems that prefixiter is expected to return child *folder* but was relying on a specific .children implementation, fix this

3.6 api update, cleanup

[forms] let the autoform specify the generic relation field label, no need for a custom fieldset, may be handled as for other fields by the renderer.

simplify messages used for inlined forms

should have been in previous commit (moved below in the module)

[forms] the last touch: handle inlined relation forms as fields by introducing a simple InlinedRelationField. This makese things more flexible while removing a lost of overriding necessary.

fix label handling: when label is explicitly set to None, don't try to guess it

surrounding div necessary for proper error localization

cleanup base class

by default use new jquery widgets for Time/Date/DateTime fields

new jquery based widgets for Time/Date/DateTime, backported from crm

introduce a default render implementation on the base widget, which add_media and then call the newly introduced method _render(form, field, formrenderer), abstract of the base widget class. This allow a litle code factorisation and avoid to forget the call to add_media in subclasses.

[forms] new optional suffix attribute on widget objects, used to generage input name / dom id. Ease reuse of widgets from another widget. Also move base values computing to a separated values(form, field) method.

introduce new _ensure_correctly_typed method on fields, responsible to ensure the value return by the widget is suitable for that kind of field. Also factorize date and time fields.

fix typed_value implementation w/ eidparam field which are actual entity's relation and have .value set for their initial value

cleanup module namespace but only importing the formwidgets module

fix deprecation warnings

fif inlined relation forms pb w/ new ajax forms. We've to ensure the varmaker is properly initialized when generating a form where we may add new subentities, which will use ajax call to generate the subform while this subform expect a varmaker to be set.

now that we are property initializing autoform_section for the 'inlined' form type, use it...

[uicfg] fix autoform_section rtags initialization 1. when a relation is marked as inlined in the 'main' form type, we want its opposite (eg when one the other side of the relation) to be marked as hidden in the 'inlined' form type 2. when no section is specified for the 'inlined' form type, use the same as in the 'main' form type to do this properly, we need two initialization stages. The first one to handle 1., the second to handle what was done before and 2. We can't do this in a single stage because we've to know the bare value of the "opposite" tag.

cache result of the COUNT query used to see if a non unique attribute may be used as rest path on the entity instance to avoid recomputing it later

incontext view of EmailAddress is a mailto: link

add CWProperty for_user CWUser when on the user primary view

move CWProperty rtags from autoform to cwproperties module.

fix Bytes submission pb on POST, due to multiple call to field.process_form_value for a same field. In the case of file, stream.read() doesn't return file's data the second time (since the cursor is at the end of the file). Fix this by having a generic process_form_value method that cache field's value in form.formvalues[field] and so _process_form_value is only called once. You should then override that later method on custom fields.

must skip internal field value from the generic relations combobox

case where 'w' argument is None should be handled in do_paginate

[uicfg] clean dead code, fix indexview_etype_section for ExternalUri and Bookmark

[uicfg] fix reloading pb with the indexview_etype_section

must catch UnmodifiedField exception in has_been_modified

command may now officially be either a string or a list, don't make think it's for backward compat

nicer usage for cubicweb-ctl

get back iterparents implementation from folder cubes

eschema view has been removed

[bfss] make it works when adding/updating entities with an attribute using bfss

d-t-w

fix copyright

XXX quick fix html validation by removing <canvas>

skil internal field value, else we get an error

add source in backup/restore failure message

introduced dependency to lgc > 0.47 for new stacklevel argument to deprecated + new db backup/restore api

handle lgc.adbh api changes within regards of backup/restore: * new backup_commands method instead of backup_command * some backend still return command as string as expected by os.system, not subprocess.call

match_rtype hook selector now accepts frometypes/toetypes optional arguments to match only given subject/object entity types

a selector MUST NO return None

some fixes so that deprecation warning are properly localized

to ease overriding of field.choices using *function* (set using autoform_field_kwargs), give the field as named argument

refactor attributes'value generator to deal with yams special Attribute/NOW/TODAY values that may be used on [Interval]BoundConstraint (BoundConstraint wasn't handled btw, fix this). Besides passing 'entity' to generate_* function (needed to get value for another attribute when computing one for instance): * set nicer method names and makes clearer what's private/protected/public * properly handle choices, constraints on float/date/datetime fields

backport stable

backport stable

3.6 api update

[vreg] object_by_id should not have been deprecated. Fix facets. it was *designed* to be able to force an object retreival when selection is potentially not possible. For instance using select instead of object_by_id in get_facet totally breaks the implementation since in get_facet we've not context result set, and we ends up with a NoSelectableObject exception.

fix stupid typo ...

fix condition to paginate a result set

make rollback available in shell (as commit is)

add a reminder

fix indentation / tab pbs

fix name error

add approximated version number to deprecation warning

oops, missing return

we need extrapath when trying to import site_cubicweb files

move extra path computing as a configuration class property so we can reused it

first draft for a simple hooks based custom attribute storage, with a BytesFileSystemStorage POC implementation. Basically: * a dictionary contains maps from which attribute of which entity types are mapped to which custom storage * hooks check for one of these entity type being added/modified/deleted * read is based on the sql generator callback mecanism (used in vcsfile for instance) * all storages have the same basic interface (read, add, update, delete), and should be pluggable in a transparent way (except at migration time when one want to change from a storage to another) * the sample BytesFileSystemStorage: * may store Bytes attributes content of any entity type as file on the file system * is based on one FSPATH rql/sql function and another _fsopen only available in sql * has a dumb file name allocation algorithm

d-t-w

closes #345410: two_lines_rset, two_cols_rset, two_etypes_rset are badly named replaced by multi_lines_rset, multi_columns_rset, multi_etypes_rset, all taking optional 'nb' argument allowing to specify and exact count. Bw compat is kept, usage of deprecated function updated in cw itself. Also drop custom at_least_two_columns selector (in views.plots) since it was the same as multi_columns_rset. Hopefuly the renaming will help in avoid such pb... We should now decide if we want to deprecate one_* selector in favor of multi_*(nb=1). See note on this at then end of selectors.py

fix #636936: read_permission relation should not be mandatory

fix #615848: undefined 'new' variable was refering to the newly introduced subquery / remove_subquery doesn't take an undefine argument (probable copy/paste from the remove_node call)

some docstring

[doc] reformulate dependencies

[views] FinalView should display an empty string if value is None

merge

move base form renderer label handling into a function to reuse it in CompositeFormRenderer and avoid seeing a tuple in column headers...

minor schema deserialization optimizations

oopb, widget has been renamed

specify formvid=edition to get multiple edition working w/ 3.6

reminder

this clutter startup screen for little benefit, remove it

update copyright

move bookmark related uicfg stuff into bookmark specific module, use the EditableURLWidget widget for Bookmark.path

new field's responsibility POC: EditableURLWidget allow to edit Bookmark.path in two separated fields, displaying unquoted values which are requoted on form post processing

refactor field's value retreiving from the widget (eg 'display value' concept): * kill field's display_value method, since getting value specified in a previous post or in req.form should be the widget responsibility (the widget nows what inputs to display, not the field) * new typed_value / format_value on widgets, mainly to ease overriding

dom_id now accepts a suffix argument as input_name

cleaner input_name implementation

when the vregistry is initialized, no need to call vreg.set_schema (else we trigger appobjects registration twice. This allows significant startup speedup.

new url_parse_qs method on base request to parse unicode url formatted query strings

merge

enable cubicweb-ctl db-dump and db-restore on Windows with SQL Server os.system has some issues on Windows, especially regarding whitespace handling, which subprocess handles correctly.

3.6 api update

merge

allow control of generic relations field display through display_fields

no self._cw on widget objects while '_' is defined locally

delete remaining stuff w/ handling of the generic relations

test update

clear vreg's parent_classes cache to fix etype cache / auto-reloading bug EtypeRegistry.parent_classes()'s is cached. When the registry is reloaded, if the cache is not cleared, we might get the old list of parent classes and that leads to strange bugs. Typical scenario : 1/ start application in debug mode 2/ go to add/Bookmark -> everything's fine 3/ change the source 4/ reload add/Bookmark -> 'creation' not selectable because : a/ cw detects a change in the source and reloads vreg b/ cw tries to select 'creation' view for 'Bookmark' - 'creation' view uses specified_etype_implements('Any') - cw compares Bookmark.__bases__[0] (i.e. new version of AnyEntity) with the old version of AnyEntity cached in the vreg. They don't match => view is not applicable.

merge

fix bad indent

backport stable

cleanup

fix bug: the 'table' argument is overwritten by the loop variable, causing latter crash due to this missing information

don't record deambiguifier when it has no type restriction using 'is', so we don't miss later some available constraints

fix warning stack level

cw 3.6 api update

merge

merge

more input/button fixes

backport stable

cleanup

turn all the stuff used to handle 'generic relations' in forms into proper field / widget. -> regroup code from web.request, web.controller, web.views.autoform, web.views.formrenderers, web.views.editcontroller (!) into GenericRelationsField, GenericRelationsWidget in the editviews module (together with the UnrelatedDiv view). So: * almost everything in one place * no more specific behaviour in the form renderer * almost no custom behaviour in autoform (simply add the field when it think it should) Also, the form renderer now display field's value with colspan=2 when field.label is None.

fix tests

[rql2sql] added test that shows crash during sql generation Original query in cubicweb-drh: Any A,E,CV,group_concat(TN) GROUPBY A,E,CV WHERE A is Application, A for_person P, P has_studied_in E?, T? tags A, T name TN, A curriculum CV? Corresponding bug in tracker: #628798

apply patch proposed by 2web to properly remove hidden input set to represent the button used to submit the form, so forms may be properly reused after a post

fix dumb name error

missing import

remove input from jquery filter, so it works as well with buttons

unicode(arg) else we may get some encoding exception

remove deprecation warnings + name error

remove common/test/data directory from __pkginfo__, MANIFEST and debian rules to allow cubicweb installation

fix syntax error

test fixes

3.6 api update

update selectors.py: change entity.schema with entity._cw.vreg.schema

merge

we should access to __regid__ instead of id since in that case bw compat fail: access to id throuhg a class returns the property object instead of calling the property's method

fix autoform_is_inlined bw compat

[forms] fix implementation of backward compat method form_render previous implemtation used form_default_renderer() which is now replaced by default_renderer()

inherit from FieldWidget to inherit from the process_field_data required by the edit controller

with cw 3.6 we've to specify entity's relation role

fix bad merge

import uicfg.auto_form doc

update i18n doc

Fix element tree import for python2.4

forgotten buffer

backport stable branch

fix pb with bytes field processing: currently when an existing file is edited and no new file is specified, it finds no value in req.form and so try to detach the current file, which is wrong. In that case, nothing should be done. So introduce a new UnmodifiedField exception that may be raised in field's process_form_value method (catched in field.process_posted).

fix NameError

don't call handle_inlined_relation where we're not on the relation'subject

don't add another _cw_edited_field hidden when build_context is called twice (may occurs with inlined relation forms)

cleanup

don't use matching_groups() for is_in_group implementation

a few french translation typo corrections

rename need_navigation into paginable paginable has been recently introduced by changeset d4eeaf0393e2. We already had the need_navigation attribute to control the same thing (pagination). As paginable sounds like a better name, deprecates need_navigation (bw compat ok). Also: * fix main pagination functions responsability * deprecate limit_rset_using_nav in favor of paginate/do_paginate * navigation module should be renamed to pagination at some point

backport tracker permission utility functions into the cw.schemas package

Backed out changeset b8ea99f5e8ea, tags.a escape its content.

fix dumb name errors

fix global errors display when there is more than one error

fix error key / field input detection (field's name include -role when specified, the error key *may* include it

missing xml escapes (but textoutofcontext probably ought to be just outofcontext, with no escaping then)

3.6 api update

move autoform's field_by_name implementation on base EntityFieldsForm

specify name *and* role when raising FieldNotFound helps debugging

if specified, benefit from target_etype information in the [partial_]add_relation selector

ouch, has_perm was always returning False...

more 3.6 api update

iirk reminder

3.6 api update

has_been_modified shouldn't raise ProcessFormError

related entities are returned as tuple since a few months ago

should skip INTERNAL_FIELD_VALUE

should use a set for pending fields

fix process_form_values: we should handle value errors and properly raise ProcessFormError

refactor form error handling: * almost everything is in the base form class, with other form errors related code * deprecate form_field_error in favor of field_error and format_error on the renderer * fix pb with js localization of errors by properly using field.role_name() * rename 'errors' into 'error' in forminfo to avoid confusion

new DBG_ALL debug level

return localized message on parse_datetime errors

raising ConfigurationError is cleaner; remove some unused imports

[javascript] upgrade excanvas.js (rev. M3, mars 2009): this fixes plot widgets rendering in IE

do not attempt to daemonize under windows

[cwctl] reset-admin-pwd was failing when db-host was not set

[javascript] IE doesn't support dynamic enctype modification, we have to set encoding too

[forms] make it easier to override CWProperty form buttons

[forms] make it easier to override StringField's mininum size

[skeleton] remove unused os.path.dirname import, use dname instead of dirname as an additional safetly belt

[web/views] if view has attribute paginable=False, do not try to paginate it

update copyright: welcome 2010

[views] backport progressbar template from shingouz

[forms] fix multiple edit bug introduced by changeset cebdf8ee5ad7 Changeset cebdf8ee5ad7 assumed that edited entity was either explicitly passed to the form or found in first row / first col of the resultset because of the one_line_rset() selector. But one_line_rset() is actually trickier : the selector returns 1 if the resultset has more than one row **if** an explicit row is specified during selection process. This is exactly what is done in the 'muledit' form : for row in xrange(len(self.rset)): form = self.vreg['forms'].select('edition', self.req, rset=self.rset, row=row, #... This changeset thus takes row in consideration to find edited entity if 'row' was explcitly specified.

do not recompute the plot on subsequent onload events #615338

fix stripEmptyTextNodes for IE (part of #615337) this does not seem to make the function really work but it does not fail any more

[javascript] upgrade jquery.flot.js to v0.6 to fix IE6-related bugs

[forms] edited_entity must be set before calling session_key() EntityFieldsForm.__init__ calls FieldsForm.__init__ which in turn calls self.session_key(). self.session_key() uses self.edited_key to compute the correct session key. We thus need edited_entity to be set **before** calling super's __init__. The bug only occured when EntityFieldsForm was selected because of its one_line_rset() selector instead of match_kwargs('entity').

must escape user_rql_callback

backport stable branch