Drag_and_drop_between_two_lists.md

Rationale

Tryton current support of drag and drop allows to re-order or re-parent items inside the same list.

The idea of this blueprint is to handle a generic drag and drop between arbitrary lists. Those lists can be two x2many on the same form view, two tree views inside a dashboard or even between two different tabs.

There is two distinct situation. The first one is when the user drop records on a line (and so on an existing record). The other is when the user drop the records on a list view but not on a particular line (for example on an empty list or under the existing lines).

DnD on a record

A new model ir.ui.view.droppable allows to define which records can be dropped on a record of a particular view. It contains the following fields:

• model: A char that contains a model name, E.G: party.party (required).
• condition: A char that contains a Pyson expression.
• view: A many2one that reference a tree view (required).

There is no unique constrain on (model, view) to allow any module to extend the behavior.

A new one2many field droppables on ir.ui.view express the relation supported by the view field.

When the client makes the fields_view_gets call, the server will collect the droppables lines linked to the view and add a droppable property on the view definition. The droppable property contains a dictionary. The keys of this dictionary are the model and the values are list of pysons expression.

So when the user starts to drag records, the client is able to know on which list the user can drop them based on the model and the value of the records. All the records must fit the Pyson expression to be accepted.

When the records are dropped, the client will call the on_drop method of the model that receive the records and will pass the model and ids has arguments. The rpc class on the server will convert those arguments into a list of records. So the on_drop method will look like this:

class MyModel(ModelView):

    def on_drop(self, records):
        # Ideally the method should first test the model
        # of the records 
        for record in records:
            record.some_field = self
            record.save()

As this method must work on the correct values, the manipulated records must be saved prior to be dragged or dropped.

After the on_drop called is finished, the client will reload the records that where dragged and the record on which they are dropped.

DnD on a list

A new field on_drop on the view model defines which method must be called when the records are dropped.

Example of view definition:

        <record model="ir.ui.view" id="production_view_list_done">
            <field name="model">production</field>
            <field name="type">tree</field>
            <field name="name">production_list</field>
            <field name="on_drop">done</field>
        </record>

For reference, the done method is:

    @classmethod
    @ModelView.button
    @Workflow.transition('done')
    def done(cls, productions):
        pool = Pool()
        Move = pool.get('stock.move')
        Date = pool.get('ir.date')
        Move.do([m for p in productions for m in p.outputs])
        cls.write(productions, {
                'effective_date': Date.today(),
                })