[IMP] improved meta data on __openerp__.py
bzr revid: fp@tinyerp.com-20101011223554-ad0mvm8s45inhmg3
This commit is contained in:
@@ -74,6 +74,15 @@ Technically, Pull flows allow to process procurement orders differently, not onl
|
||||
the product being considered, but also depending on which location holds the "need" for that
|
||||
product (i.e. the destination location of that procurement order).
|
||||
|
||||
Use-Case
|
||||
--------
|
||||
|
||||
You can use the demo data as follow:
|
||||
CPU1: Sell some CPU1 from Shop 1 and run the scheduler
|
||||
- Warehouse: delivery order, Shop 1: reception
|
||||
CPU3:
|
||||
- When receiving the product, it goes to Quality Control location then stored to shelf 2.
|
||||
- When delivering the customer: Pick List -> Packing -> Delivery Order from Gate A
|
||||
""",
|
||||
'author': 'OpenERP SA',
|
||||
'depends': ['procurement','stock','sale'],
|
||||
@@ -91,4 +100,3 @@ product (i.e. the destination location of that procurement order).
|
||||
'active': False,
|
||||
'certificate': '0046505115101',
|
||||
}
|
||||
# vim:expandtab:smartindent:tabstop=4:softtabstop=4:shiftwidth=4:
|
||||
|
||||
Reference in New Issue
Block a user