Personal stages of project.stage records are based on two main field:
- personal_stage_type_ids: the list of all personal stages linked to a
task (M2M)
- personal_stage_type_id: a computed field (not stored) indicating the
personal stage of a task for the current user.
When reading a set of project.task records grouped by personal stages,
two options are possible:
- Group the records by personal_stage_type_id (approach used in former
app Notes): in which case the read_group method has to ne rewritten as
it can be used on a non-stored field.
- Group the records by personal_stage_type_ids (approach used in app
Project) in which case, the kanban view has to be overriden to be able
to drag and drop a task between personal stages (which is not possible
by default, when grouping according to a M2M field).
The main evolution proposed by this refactor is to use an hybrid
approach that would:
1. Group the project.task records by personal_stage_type_id
2. Use the read_group with groupby set to 'personal_stage_type_ids' as
this should give the same result.
This would allow to:
- Avoid a complex and costly (performance wise) read_group override
- Avoid an override of the kanban view that is costly to maintain
- Simplify the implementation (and thus readability) of personal stage
management (among which, removal of the model
project.task.stage.personal).
task-3345132
Part-of: odoo/odoo#140050
Hugo Carlier (Huca)
2236e5c97c