Improve error handling on Workflow_statusModify. If divergence is solved by other user, then show a nice message.

Uid must be always consistent. If uid is changed, then workflow action which user selected is applied to different workflow and it causes different result.

Improve error handling on Workflow_statusModify. If requested transition is not available, then show a nice message.

PasswordTool_viewResetPassword/text_content from erp5_core wrongly requires
Field Library in erp5_web and EmbeddedFile_view/my_description from erp5_base
wrongly requires Field Library in erp5_dms.

If an user has set SVN working copies in its preferences, it cannot be
modified anymore after upgrade.

Now both ascending arrow and descending arrow are displayed in title column
and user can sort in specific order explicitly. User also can sort multiple
columns by click without using sort-on-dialog.

Fix error message of StringAttributeMatch Constraint incorrectly renamed after migration to ZODB (78e2ae32).

also raise an exception in case of error, like cloudooo is not running.

Also remove copy pasted comments

Update stock report and field library to use parameters named node_category & section_category instead of site & group

* use translationContext instead of low level hacks
* introduce a script ERP5Site_notifyReportComplete to do the actual notification
* do not activate a useless getTitle
* store strings encoded as zlib to prevent database growth

Signed-off-by: Aurélien Calonne <aurel@nexedi.com>

usually generated SQL query can use a better index this way

We were not able to export and import back roles of a document when more than
one local role group id was used.

It can only work with default id generator and anyway catalog does not
understand this syntax, so this has no effect

you can apply this change on the existing table by the following query :

ALTER TABLE stock DROP INDEX `section_uid_portal_type`, ADD INDEX `section_uid_portal_type_mirror_section_uid` (`section_uid`,`portal_type`, `mirror_section_uid`);

Do not try to translate user input by using Localizier as it will fill up its internal structure (very inefficient) and lead to increase of ZODB (every modification is a new Localizer object holding all entries).