New test :
- Create a MO with a finish product tracked by SN with a dozen as
quantity
- Click on produce button in order to open the produce produce wizard
The quantity to do is 1 dozen although the tracking is by serial number
Task #55494.
Refactor resource calendar and resource mixin:
- add a required timezone field on models 'resource.calendar' and 'resource.resource';
- resource calendar attendances are now computed in the timezone of the resource
or the calendar;
- improve API of 'resource.calendar' and 'resource.mixin';
- avoid other modules from using the implementation methods of 'resource.calendar';
- refactor implementation of 'resource.calendar' to make it more efficient.
Moves UoM models, test and data to a new addon in
order to be able to use uom without product.
A simple example is be to be able to use UoM for
timesheets.
This commit only move code, and adapt xml ids
without chaging any feature or functionnal
behavior.
Note: 'product' module now depends on new
'uom' module.
It happens due to workcenter working hours,
the workcenter is not active during the weekend thus
the productivity loss are not counted in the oee.
This commit removes 2 days to the test date if it is
during the weekend.
Usecase to reproduce:
- Set a product with the same lot in different locations
- Create a MO that will uses theses products and that require more than one location
- Produce it.
-> Traceback Expected singleton: stock.move.line
It happens because product produce only expect one move line
by lot. Howeever it could exist more than one move line if the
product for a specific lot are shared among multiple under location.
This commit try to fill the move line with the correct quantity
one by one and add the extra in a new move line.
Use case to reproduce:
- Create a MO with a product tracked and a qty to produce greater than 1.
- The Bom should also use a routing with at least 2 operations.
- On the first workorder only produce 1 unit and record the production
- Go to the next workorder -> The final lot is not set.
It happens because the lot for the next workorder is only set if the
total quanity for the mo was produced.
This commit fixes the bug by removing the code for the workorder lot
assignation outside the wrong condition. This commit also update the
test test_00b_workorder_process that allows us to find the above bug.
You start with an MO with a quantity of 2 with routing and we
produce just 1. At least one of the components is tracked by lots.
When we update the MO to a quantity of 1 again, the temporary
stock.move.lots continued to exist which led to processing a the consume
d move of e.g. 1 piece with 2 pieces of the quants processed when
posting inventory.
A test is added in order to avoid this case
opw-767526 and opw-772786
When producing a finished product with a lot, also the raw materials
without lot tracking should split the move lines to at least know the
quantities consumed for every lot produced. This way, the traceability
report will not confuse the user.
Use case to reproduce:
- Create a product with a BOM needing one component tracked by SN
- Create a MO for this product with 2 to produce
- Produce 1
- Produce the other -> 2 lines and a wrong lot proposed
This happens when the quantity to consume is equal to 0
or if the qty_done on a move_line is equal to the reserved
quantity.
This commit change the condition after the float_compare in
order to take into account the equal comparison.
Since the function in order to generate a MO is usefull
for mrp test. It could be great if we could use it in
every test.
This commit also add 3 arguments in order to modify the
quantity needed for each product.
Before, in case of non-tracked components, when everything foreseen had
already been consumed, it would not add the theoretical quantity on
producing through the produce wizard. This makes little sense as the
behaviour of the wizard is not the same anymore for all cases and this
can only confuse the user.
This revision changes the BOM and routing of what was previously a
"computer desk". The aim is to provide a complete example which can be
used to showcase what can be accomplished with the new tablet view.
This implements a warning when the user tries to do any of the
following operations on a product that is not in stock:
- Scrapping
- Unbuilding
- Repairing
The warning displays the current stock of the product in question per
location which can be helpful in case the user selected the wrong
location.
Suppose you have a product tracked by serial number. In your inventory,
you have two quants: -1 for serial1, +1 for serial2.
If you create a move for 1 product and try to run `action_assign`, it
didn't work before this patch because stock.move's
`_update_reserved_quantity` is guarded by a call to
an unstricted `_get_available_quantity` (unstricted meaning considering
all quants regardless of their lot_id/package_id/owner_id). This call
resulted in an available quantity of 0, thus the stock move could not be
reserved
To fix this issue, we patched `_get_available_quantity` so that it
rightly considers products of different lot_id as different products,
i.e. it returns an available quantity of +1 in the initial scenario. Of
course, we consider all quants of the same lot_id, i.e. if you have -1
for serial1 and +1 for serial2, the available quantity amounts to 0.
In another scenario, let's say we have only -1 serial1 in stock, if we
run `_get_available_quantity`, the method should return 0: nothing is
reservable. We continue this reasonment by returning 0 everytime the
available quantity is negative, tracked or not. This is probably what a
method named `_get_available_quantity` should have done since the
beginning.
There's only a single case where we need to know the quantity even if
it's negative: it's just after we move a quant to its destination
location. If this quant was tracked and resulted in a negative quant
(i.e. we moved something we didn't have), we try to compensate with an
untracked quant. In this special case, we do not want to ignore the
negative quants when calling `_get_available_quantity`. That's why we
introduce a new argument to this method. We could have added a new method
`_get_quantity` for clarity sake.
The production wizard raised a traceback when you were trying to consume
product tracked by lot without reservation.
To fix this issue and some potential others, we have removed the
move lines from the wizard and create produce lines that are used on
validation to create the move lines.
This also let us show the 'To Consume' quantity of a lot instead of the
reserved quantity which has no sense when there is no reservation.
This removes the procurement.order model. To fufill their needs SO, PO, MO and
stock moves now call the _run method of the relevant procurement.group.
This mecanism is now only used for stockable product, tasks now uses their own
independent mecanism.
The _run method will check all the applicable rules and create directly the
needed model to fufill the need.
The modules stock, purchase, mrp, extends the _run method to implement their
specific strategy relevant for the rule type they define.
If an exception happens the message will be logged as a mail messsage on the
source model, for example, if a sales order cannot be fufilled the salesperson
will now see directly the reason.
OLD commit messages:
[WIP] procurement: removing procurement.order in stock, sale, purchase, sale_stock. WIP
fixup! [WIP] procurement: removing procurement.order in stock, sale, purchase, sale_stock. WIP
[IMP] Basic tests
[FIX] test not necessary anymore
[FIX] remove unnecessary print statement
[FIX] unnecessary test + why passing warehouse worked before?
[IMP] purchase: one move by purchase order line
[FIX] purchase: correct inventory tests and pass move_dest_ids among procurements
[FIX] because of bad cherry-pick merge
[IMP] make mrp pass by adding move_dest_ids there too
[IMP] tests of sale_mrp, no need for cancelpropagation then
[IMP] better to consistently use recordsets also for one2many
[FIX] purchase_requisition
[FIX] Exceptions should trigger errors, which should be caught in the tests
[FIX] sale_mrp: remove usage of procurement.order and use sale order name instead of sol
[FIX] stock_dropshipping: add sale_line_id on purchase_line_id
[FIX] Remove pdb
[IMP] add stock_dropshipping files
[IMP] stock: search carrier through sale line instead of procurement group
[IMP] add procrule test and preision needed when updating sol
[FIX] sale_order_dates + [IMP] procurement exceptions by scheduler
[FIX] No need to return task
[IMP] move file as name changes and add corrections
[FIX] Continue Run Schedulers wizard fix
[FIX] name issues of takss
[FIX] updating sale order line, but there is still a problem with the recompute
When producing multiple products tracked by serial number only the
serial number was only set for the first one in every work order. This
makes it so that when you create SN1, SN2 and SN3 in the first work
order they show up as default in the following work orders.
Since refactoring in saas-17 quant_ids are removed and
the logic is instead locate on stock_move_line. This commit
adapt the method by trying to undo the created move_line
during the MO.
This commit also add a default_uom_id in the view's context in order
to remove a traceback. Since 0469d2e
uom is mandatory on stock_move_line.
It also add a bunch of tests in order to check the behavior of unbuild
with 4 test: without tracking, consumed product tracking, produced product
tracking and everything tracking.
Know limitation of this commit:
-> No warning message when unbuild more than in stock.
-> Wrong result if trying to unbuild twice the same MO
with consumed product tracked with different lot/SN
Let's consider 4 products P0, P1, P2 and P3 where:
- P0 depends on P1, P2 and P3
- P1 depends on P2 and P3
- P2 depends on P3
NB: "x depends on y" means that there is a phantom BOM for x where
y is in a line.
With this example, before the fix, when creating a MO for P0,
the function "explode" raised a user error saying that there
was a recursion error.
Even if in this case there was no recursion.
Now to check if there is a recursion, a graph of the dependencies of
the bom is created while exploding the boms and a function "isCyclic"
is applied on the graph to verify that there is no cycle which means
no recursion in the dependencies.
The algorithm was taken from this site: http://www.geeksforgeeks.org/detect-cycle-in-a-graph/
opw:745453
Currently partners have a boolean field to choose whether to receive
notifications only in their Odoo inbox or to receive them in their inbox
and by email. This leads to several issues :
* if a customer is configured to not receive emails he will not receive
any notification on sales orders, leads, ... This is not clearly
indicated to the salesman and it is not easy to know how to change
that behavior
* if an user chooses to receive emails and does not use its inbox a lot
of notifications stay in Odoo. The user has to manually set them as
done to make them disappear which is redundant.
This commit changes that behavior. From now on customers will always
receive all notifications by email. Indeed Odoo is not a customer oriented
mailbox. Moreover sales orders or discussions on leads send to customers
should always be sent by email as it is the standard communication
mechanism. Users will be able to choose to receive notifications in Odoo
or by email. The choice is no longer inbox or inbox + email, but inbox
or email. Choosing one option or the other one depends on the way the
user wants to work.
Technically the field is moved on the users model and selection keys
are renamed. Notification process is modified
* notified_partner_ids contains as before specified recipients as well
as followers matching the subtype
* customers and users working with emails are notified. During that
process customers notifications are marked as done to be able to
track the email state without having needaction. Users notifications
are currently deleted as we do not track their email state.
The removal of partner field implies changes in various addons that
define partner data with this field set in the values.
Before, when you would put a quantity in a BoM line that was not rounded
to its UoM (e.g. use 0.3 piece with rounding 1.0) it would not round this
quantity in the MO also (when exploding the BoM), and the user could also
enter a not rounded quantity himself for the quantity produced or
consumed.
The moves were split with the exact quantities, but the moves were validated
taking into account rounded quantities. This resulted in either moving 2
pieces instead of 1 (entering 0.7 or 1 piece) or raising an error telling
you can not process moves with 0 quantity (0.3 piece).
As product_efficiency and product_rounding was removed from the BoM/BoM
line in v10, if the rounding is correctly handled, the feature could work
like in v9, by e.g. putting 1.03 in the bom line (97% efficiency) and
rounding up when exploding the BoM (e.g. producing 10 pieces would put
it to 11).
When "producing", we round the produced qty to the product uom (it does
not make sense to produce 1.32 if your uom rounding is 1.0). The user
might now change the produced quantity, that's why we round-up when
validating the move: indeed, if the user slightly increase the produced
value (below the uom unit), it actually means he produced more.
This way, instead of not rounding anything and doing rounding behind the
scenes, leading to errors, we round everything and the user sees the
result directly.
When posting inventory during a manufacturing order,
all raw material and finished product moves are used,
also those done from a previous post inventory.
This is no problem for the moving of quants because
a filtering is done in the move_validate method,
but it is a problem for the price calculation in
mrp_account and for linking the consumed/produced_quant_ids
of the quants of the raw and finished moves.
Before this fix, it would calculate already consumed materials
a second time in the price of the finished products and would
link quants produced with previously consumed quants (and vice versa)
when doing a second post inventory.
We add a test to show the issue.
According to the routes set on the products of the raw materials, a
procurement may have to be created for the consumed lines of the MO.
A route may also be set on a product category. Before this patch, we
only look for procurement rules for the routes defined on the products
and on its category, ignoring the inherited route of the parent category.
Fortunetaly, we have a computed field on the product model returning
all the routes from the current category, so we use this one in the
search for procurement rules.
A test has been added to show the failure.
Before, when the routing of a production order was empty (no operations),
it got into an inconsistent state (you click on "plan" and nothing happens).
Empty routings were used before to do subcontracting or 2-steps inventeory
(we did not set any operations on the routing but we set a location).
Here, we change the related routing field on the production order into a
calculated field which becomes empty when the routing has no operations.
This way, the "produce" button is displayed (as for production orders
without routing) and working/
This is a temporary patch to workaround the issuein a stable release.
In master, we might remove the location on routing and place it on the
workcenters and put a constraint that does not allow an empty routing.
Before this patch, the inventory was automatically posted when you click
on "done" on the last planned work order. This behavior was preventing
the user from correcting, for example, the consumed quantity on the raw
materials move on the manufacturing order if it was different from the
theoretical quantity. Note that finishing anyone but the last work order
still let you edit the consumed quantity. Same with changing the lot.
Now you have to click on "mark as done" on the manufacturing order to
post the inventory. Note that this is the same behavior when working
without work order (produce > mark as done will post).
We adapt the tests to reflect this change.
When (re)generating the stock.move.lots in the workorder,
it should use the unit_factor on the move like everywhere else.
This makes sure we do well with the explosion of phantom boms too.
It is also important that if there is no operation_id given
explicitly in the "consumed in operation" on the bom line,
that it still generates the necessary stock.move.lots in the last work order.
Also, in the work order it is better to set the quantities done already.
That way we don't forget to consume them.
This commit contains the core of the new MRP. It contains a whole refactoring
of the MRP application, with improved and new features, written in new API.
Among other here are the main manufacturing workflow improvements :
- Picking type not only for pickings but also for manufacturing orders
- Properties replaced by picking type
- BoM can only be produced with its routing (no other)
- Either produce without routing with only production orders, or produce with
routing
- By default, there is an order in the work orders (serially), but you can
override it to be able to work in parallel
- Time clocking on work orders and block time on work centers with reporting
on OEE, performance, losses, ...
- Real-time adaptation of timings on operations
- Lots/serial numbers can be inputted like in the pickings on manufacturing
orders. It is also possible to input them in the work orders.
- Material availability independent of production order state (possibility to
start production when only part of it is there)
- Work sheets on work orders
- Put messages on work orders to make your workers pay attention to something
- Full traceability link to see for each produced piece of stock, the
consumed pieces, ...
- Separate scrap object (a scrap is not done based on an original move
anymore)
- Separate unbuild system (if you want to unbuild into its original
components)
Thanks to all people that helped during this development, notably but not
limited to Chirag A Dodiya (cod@odoo.com), Gaurav Panchal (gan@odoo.com),
Jignesh Rathod (jir@odoo.com), Mansi Trivedi (mtr@odoo.com), Pariket Trivedi
(ptr@odoo.com).
- Preserved explicit 3rd-party copyright notices
- Explicit boilerplate should not be necessary - copyright law applies
automatically in all countries thanks to Berne Convention + WTO rules,
and a reference to the applicable license is clear enough.
Pratima Gupta