When a process change the list of activated modules, the others must refresh
their pool. We do that by just stopping the pool for that database, the next
request will load the new pool.
issue8988
review302861002
The browse list keep a reference to the deleted modules, so a read error
may be raised when accessing an attribute of any record of the list.
issue9567
review318151002
We add the flake8 configuration used so we ensure everyone uses the same.
We remove the usage of __all__ for non public API.
When possible, we rationalize the class name according to its __name__ and module.
issue9082
review297061002
There is no need to create the graph step by step with only dependencies
already registered as far as the depth computation is propagated to all the
children.
Instead of searching level by level the nodes to iterate, it is faster to use
the sort of Python.
create_graph does not need to return its internal states but only the graph.
The ir.module activate and update method does not need to check for missing
dependencies because it is already checked at startup.
issue7466
review45211002
changeset 83757dc87808 moved the buttons from res to ir module but it missed
the migration of the model data to allow updates from previous databases
issue7186
review38031003
Changeset9bb9fe4cd312 modified the behaviour to mark them as not activated but
they should be marked as to remove so their data is correctly removed with the
install/upgrade wizard.
issue6019
review31691002
Cédric Krier