- Enables connections saving & auto cleaning
- Preparing possibility to declare a datasource as read only

- Now delete & update methods use filters & rel_filters

- deleting most of **datas to replace it with a standart dict argument
- using filters in childs of LeFilteredQuery

- Added missing allowed_classes options to some references in em_test

- Modified again the em_test to fit with testing purpose
- First test on multi datasource query written
- debugging of query.py

- adding an EmClass to make a double link between indexes and texts
- adding subqueries infos to dump_infos method of LeFilteredQuery

- defining a new datasource named dummy2 in both tests conf and install/conf.d
- adding two class using the new datasource in em_test.py
 - index : abtract class representing indexes (two fields, name & content)
 - indextheme : thematic index, inheriting from index but with one more field "theme"

- The set_query_filter() method handles subquery creation
- The run isn't written yet...

The _datasource_name attribute replace the old _datasource attribute. The goal of the modification is to maintain easyly accessible datasource UID

Adding tests for the Plugin class

They have to implement a _activate() method in their __init__.py file

- renamed Plugins class to Plugin
- an instance represent a loaded plugin
- classmethod allows to preload & load plugins

Using the conf to load emtests.pickle