* fix quantity computation
* use a payment_state field instead of inovice state one.
* Use proper digits and unit conversion.
* Add a filter "To repair".
* Remove dead code and clean warnings.
In the current implementation of Odoo's _assign_picking() method in stock.move, there's a conditional check that looks at whether all the moves associated with a picking have the same partner_id and origin. If any move doesn't align with these conditions, the origin of the picking is set to False.
if any(picking.partner_id.id != m.partner_id.id or
picking.origin != m.origin for m in moves):
# If a picking is found, we'll append `move` to its move list and thus its
# `partner_id` and `ref` field will refer to multiple records. In this
# case, we chose to wipe them.
picking.write({
'partner_id': False,
'origin': False,
})
In the context of RMA when we have multiple moves associated with a picking, each coming from a different RMA order line, we encounter a problem. Each move has its origin set as the name of the RMA orde line (line.name), so as soon as a second move from a different line is appended to the picking, the origin of the picking is wiped, because it doesn't match the origin of the first move.
In order to prevent the partner_id of the picking from being set to False when there are multiple associated moves, I propose that we change the origin of the procurement from the name of the RMA line to the name of the procurement group (group.name). This way, all moves associated with a picking will share the same origin, preserving the origin of the picking and ensuring it doesn't get inadvertently set to False.
Without this, some policies are not being copied from the
operation selected when creating new rma line from a rma group.
In v16 this patch and the usage of such onchange can be removed
in favor of (pre)computed stored editable fields for all policies
and configuration in the RMA operation.
