Modifying product attributes involves discarding product variants via
the method _unlink_or_archive(), which tries to delete variants, and
archive the ones for which the deletion fails. This process is very
slow when most product variants must be archived. Instead, we filter
out the variants that must be archived anyway (like the ones used in
sales or stock), in order to avoid the variants that cannot be deleted.
On a product with ~200 variants, the time goes from more than 15 minutes
(the request times out) to around 5 seconds!
opw-2440417
closesodoo/odoo#65027
X-original-commit: 63b2f85079404eafac918a74c95e27051539c133
Signed-off-by: Raphael Collet (rco) <rco@openerp.com>
With this commit, all instances of errors being raised inside
`BaseModel.unlink` overrides are moved into methods decorated with
`api.ondelete` which is safer.
Behavior prior to this fix:
When archiving the last variant on a template, the product template
stays active (contrary to the comment indicating that `toggle_active`
will archive the related product.template if there is only one active
`product.product`).
Behavior after the fix:
When archiving the last variant on a template, the product template is
archived. When un-archiving that variant, the product template is
unarchived as well, without un-archiving the other variants.
Note: this is a remake of a fix implemented in 77e5472c041e, as that fix
did not correctly count the inactive variants.
opw-2349862
closesodoo/odoo#60641
X-original-commit: 01160ee5e7f726f3035c566a13e26825236dedfd
Signed-off-by: Nicolas Galler <nicocrm@users.noreply.github.com>
This provides a better API to search for records by name without the
formatting part (`display_name`).
This also simplifies all the overridings of `_name_search` that no
longer need to call `name_get()`. The call to `name_get()` is done in
method `name_search` in a generic way.
Issue
- Install Sales with Product Configurator
- Create a product with 2 dynamic variants:
- Color: blue, red
- Size: S, M
- Create a quotation with that product (any variant)
- Sales > Products > Products Variants
Your product variant is there
- Modify your product template and add "green"
to colors
- Sales > Products > Products Variants
Your product variant is archived
Cause
The logic which determines which variant
should not be archived is only applied
on products that don't have dynamic variants
See https://github.com/odoo/odoo/blob/820fe9a025e34e81368a189ee295b4abcf9fcbda/addons/product/models/product_template.py#L576
Solution
Do a separate logic for dynamic variants that doesn't compute
all possible combinations.
But being sure that it:
- doesn't archive when adding/removing a value
to an existing attribute line
- archive when adding/removing a new attribute line
OPW-2293346
closesodoo/odoo#55016
X-original-commit: 8517fdfa2605e8d15f5e38f2074ece7ba543d56b
Signed-off-by: Sébastien Theys (seb) <seb@odoo.com>
Avoid concatenation of translated strings & raw values and/or other
translated values, and use the new _ API (by providing the formatting
values to ensure the fallback on the english terms in case of wrong
translation).
By including the expected values inside the string to translate, we
ensure that the formatted values are included at the right place in the
resulting translated text.
This also provides more information on the string to translate to
the translators, s.t. the chosen terms for the translations can be more
adapted to the translation context.
closesodoo/odoo#54635
Related: odoo/enterprise#11925
Signed-off-by: Victor Feyens (vfe) <vfe@odoo.com>
Before this fix, the search on the supplierinfo had
an undeterministic result.
When testing odoo/addons/purchase_stock/tests/test_reordering_rule.py:66
`def test_procure_not_default_partner`, it may fail at line 120.
It sets 2 suppliers:
- Firstly, the `default_vendor` (let's assume id=120)
- Then, the `secondary_vendor` (let's assume id=121)
It then runs a procurement without setting the supplier and assert
that the partner of the resulting PO line should be `default_vendor`.
The return of
odoo/addons/product/models/product.py:574
`def _prepare_sellers` depends on Postgres when the supplierinfo
sequence, min_qty and price are the same for multiple records.
By adding the `id` to the `.sorted`, the result will order correctly in that
case.
opws-2272627
-2282081
-2287079
And PS-BE via CSC reported more cases without opw.
closesodoo/odoo#53951
X-original-commit: a733feb8e022084148685599840fe542ee9dafd1
Signed-off-by: Nicolas Martinelli (nim) <nim@odoo.com>
Missing super call for product category deletion.
closesodoo/odoo#53699
X-original-commit: ebc75df2b3d3f80aebc2fc1d8500fa57c1755bbd
Signed-off-by: Victor Feyens (vfe) <vfe@odoo.com>
Updating the uom_id on a product_product will eventually change also the
uom_po_id to have them belonging to the same category. Saving the update
call write() on product with a dict of two values (uom_id and uom_po_id)
As product.product is inherited from product.template, the write() on
the corresponding product_template is called but in this case 2 times
for each values. The first time, only uom_id is updated. The _check_uom
constraint is triggered and obviously failed. The uom_po_id on the
template is not yet updated.
This commit makes the sync in the beginning of write() of product.template.
Which means before calling the constraint.
This commit also makes the onchanges on uom_id acts the same between
product.product and product.template.
Task : 2197818
closesodoo/odoo#48714
Signed-off-by: Arnold Moyaux <amoyaux@users.noreply.github.com>
The business case is the following:
1) The Child company can order the product to the Parent company;
2) The Parent company order the product to a 3rd party supplier.
Settings:
- Enable ‘Inter-Company Transactions’ transactions with the option
‘Synchronize sales/purchase orders’ checked;
- Enable 'dropshipping'.
Set 2 companies:
- Parent company;
- Child company: which belongs to the Parent company.
Set a product 'my purchased product':
- Can be Purchased;
- Product Type: Storable Product.
- ‘Purchase’ tab:
- Set a Vendor price list line 1: which is your Parent company + set
your Child company who can buy the product;
- Set a Vendor price list line 2: which is your 3rd party Supplier +
set your Parent company who can buy the product.
- Log in as the Child company (disable the login in the Parent company,
select only the Child one);
- Place a Purchase Order for ‘my purchased product’ to vendor 'Parent
company';
- Confirm the PO.
Before this commit, an error message appears. The issue arises because
seller_ids is retrieved from the cache and not re-read as superuser.
Therefore, records are missing from the list and the correct seller is
not found.
Now, the correct vendor for the purchase order of the parent company is
selected.
opw-2261971
closesodoo/odoo#51887
X-original-commit: 0a9f0c5d2c85e80dd5dceaeacf3f04bfc8aa0be1
Signed-off-by: Jorge Pinna Puissant (jpp) <jpp@odoo.com>
When the value is the same for all records, the write can be done in
batch, this can lead to a "slight" performance improvement, depending on
write overrides, and how the value was computed for each line.
Steps to reproduce:
* Create a template with 2 variants
* Archive one variant
* Unarchive this same variant
--> the template is archived
This is due to commit 6a13b565cbbad89ff1c4177b171b01f30c83edf9. If the
variant is the only one, toggle_active on the variant should be
reflected on the template. The issue is that toggle active do not count
archived variant. One active variant and one archived variant are
counted as 1.
This commit make toggle_active count all variants
closesodoo/odoo#34769
Signed-off-by: Simon Lejeune (sle) <sle@openerp.com>
- Activate variants and pricelists
- Go to Product > Products Variants
- Search for anything on the 'Pricelist' filter
A traceback is raised: 'Can only create cache keys from hashable
values...'.
This comes from the following change:
https://github.com/odoo/odoo/blob/4b06fe19fa68255b7982d15e5847da2f6d6209fd/addons/web/static/src/js/views/control_panel/control_panel_model.js#L962
It returns a list instead of a string. Since a list is not hashable, it
causes the issue.
There are not much solutions since the context key `pricelist` can be a
`list` or an `int` (an ID). We force the cache key conversion to a tuple
to avoid the `TypeError` and handle the `list` use case (which was
broken on top of crashing).
opw-2187757
closesodoo/odoo#45064
X-original-commit: ec0a8767143b4c4818418281803c4f52fefa8d2a
Signed-off-by: Nicolas Martinelli (nim) <nim@odoo.com>
Use the 'Volume' decimal precision on all `volume` fields.
Complement of commit c1a5221ba2
opw-2185374
closesodoo/odoo#44473
X-original-commit: dc0111d6cdf4032fa840f9ff95f4fa687d33f77c
Signed-off-by: Nicolas Martinelli (nim) <nim@odoo.com>
Since [1] `computer_hdd_attribute_lines` is archived instead of unlinked in
`TestSaleProductAttributeValueConfig` which leads to indeterminism in tests due
to the missing `flush`.
Indeed `flush` is necessary before `invalidate_cache` since [2] if the data
that are pending being written must be saved to the database before the cache
invalidation.
[1] which highlighted the issue was merged after v13.0, but it still makes sense
to merge the current fix in the earliest version possible.
[1] 5361706e5e
[2] 9920f20e4cclosesodoo/odoo#43483
X-original-commit: f721bfe2860293d1456770d57be7805920929ca0
Signed-off-by: Sébastien Theys (seb) <seb@odoo.com>
Variants are generated depending on the configuration of attributes
and values on the template, so copying them does not make sense.
For convenience the template is copied instead and its first variant is
returned.
closes#38151
X-original-commit: 259f3b7fcdb88d1f458ee7ea3667964412b88ea1
Steps to reproduce the bug:
- Let's consider the storable product P and the vendor V
- Set V as a supplier of P
- Archive V
- Go to product P and click on button "Replenish"
Bug:
A RFQ was created with the vendor V
Now when archiving a vendor V, all the product.supplierinfo records linked to V
will be archived.
opw:2066616
closesodoo/odoo#36672
X-original-commit: 5040c17e873b8476c754f1f3e1686c2a02b36e48
Signed-off-by: Nicolas Martinelli (nim) <nim@odoo.com>
To avoid slowing down search like "{relational_field}" contains "{value}"
we always need to return the lazy name_get() for each [_]name_search method
closesodoo/odoo#36735
Signed-off-by: Raphael Collet (rco) <rco@openerp.com>
There are too many image sizes. Since they are stored resized this takes time to
generate when saving a new image, it's more rows on the attachment table, more
files on the disk, ...
64px is close enough to 128px that it can be removed without a big impact on
download size.
It will even reduce download and number of requests when both images are
displayed because now only one has to be downloaded and then benefit from cache.
The difference between the two is typically around 1.5kB which is negligible
these days, especially when the request overhead is around 0.5kB already, not
even taking into account other factors such as latency.
If a 64px image must absolutely be returned, it is still possible to pass the
size parameters to the image route. But the current guideline is to handle
resizing in the views when necessary.
Views
=====
- remove width and height attributes when existing CSS rules are overriding them
(eg. `.oe_kanban_avatar` in the right context)
- add CSS rules instead of width and height attributes when possible
- use `object-fit: cover;` where width and height are forced to avoid distortion
of non-square images
- for products, use `object-fit: contain;` instead, keep ratio but without crop
- add new CSS rules where the expected size was max 64px*64px before due to the
image size itself
- remove `img-fluid` where using size classes to avoid conflicting rules
task-2060865
closesodoo/odoo#36147
Signed-off-by: Sébastien Theys (seb) <seb@odoo.com>
Same reasons as in c3adeb8420
You can not have a field that is:
- computed
- stored
- translatable
One of the three had to be removed
closesodoo/odoo#36717
Signed-off-by: Martin Trigaux (mat) <mat@odoo.com>
With the new ORM, x2m records are not ordered by default.
The _select_sellers implementation requires to loop on the supplier info
records ordered by their _order. We fix that by sorting with a lambda
with the same order set on the model. Simply using .sorted() isn't
enough since it does a sort in sql, so newId records are completely
removed from the resulting recordset.
This is related to task 1947351 and PR 34049.
We move the tests into purchase_stock and actually implement the tests
proposed in the task.
* account, hr_expense, mrp, sale, sale_product_configurator, stock_account,
website_sale, website_sale_comparison
Before this commit, the `product.attribute.value` were stored on the variants.
This required filtering of attribute lines to find the appropriate matching
`product.template.attribute.value` that were used in most of the business code,
such as when computing the `price_extra`.
This also prevented to have multiple attribute lines for the same attribute,
which is needed to handle use cases such as grape varieties for wine products.
This will be done in the following commit.
After this commit, the combination of `product.template.attribute.value` will be
directly stored on the product variant.
Other changes
=============
Add `combination_indices` on product, which allows to quickly find a variant
matching a combination (1 simple indexed equality query as opposed to 1 query
with as many joins as there are attribute lines), and to add an easy
SQL constraint to ensure active combination uniqueness.
Add active field on `product.template.attribute.line` and
`product.template.attribute.value`, with the same behavior as variants:
They become archived if they can't be unlinked. This allows to keep the database
consistent, such as archived variants correctly keeping all their values, sales
order lines keeping their custom and no_variant. This is done with the help of
`ondelete=restrict` on the corresponding m2m fields, and the `unlink` methods
falling back to archiving when `unlink` is restricted.
Part of task-1912579
PR: #32946
Each model of the relation can define `ondelete` for its `column2` foreign key.
This way the behavior is similar to m2o where the target of the `ondelete` is
the `comodel` foreign key.
Some models were already defining `ondelete` 'restrict' on m2m fields and after
review it appears the behavior introduced in this commit is the expected
behavior for them.
The 'restrict' from product.product has to be removed however because the rest
of the code needs adaptation now that it is properly restricted. These changes
will be done on a separate PR.
closesodoo/odoo#35654
Signed-off-by: Xavier Morel (xmo) <xmo@odoo.com>
This branch is the combination of several optimizations in the ORM:
* store field values once in the cache: the cache reflects more
faithfully the database, only fields that explicitly depend on the
context have an extra indirection in the cache;
* delay recomputations by default: use method `recompute` to explicitly
flush out pending recomputations;
* delay updates in method `write`: updates are stored in a data
structure that can be flushed efficiently to the database with method
`flush` (which also flush out recomputations);
* make method `modified` take advantage of inverse fields to inverse
dependencies;
* filter records by evaluating a domain on records in Python;
* a computed field with `readonly=False` behaves like a normal field
with an onchange method;
* computed fields are computed in superuser mode by default.
Work done by Toufik Ben Jaa, Raphael Collet, Denis Ledoux and Fabien
Pinckaers.
closesodoo/odoo#35659
Signed-off-by: Denis Ledoux <beledouxdenis@users.noreply.github.com>
Manage pricelists through a common configuration between sale and point_of_sale
Remove lots of technical fields
Improved Pricelist Flow
-----------------------
if pricelists enabled (product.group_product_pricelist), two modes are proposed:
* basic(default): pricelists enabled, but only with basic rules
* advanced: pricelists enabled, but with advanced rules (product.group_sale_pricelist)
When pricelists are enabled, we can always access them through the menu OR access their product-targeted rules from the products/templates views.
Follow up of 58a2ffa26f
Variant specific field `image_raw_original` was meant to be renamed to
`image_variant_1920` instead of `image_variant_max` to follow the same naming
convention as the other fields (having the pixel number in the name).
closesodoo/odoo#35493
Signed-off-by: Jérémy Kersten (jke) <jke@openerp.com>
This attribute is misleading as it is insufficient to correctly upgrade
the database. It only renames the column in the database, but other
operations are needed, like updating the corresponding `ir.model.fields`
record (and its xmlid). The default values and the translations are also
lost during the upgrade.
Moreover, this feature was misused. It was:
- left on fields during multiple versions.
- used on reports (SQL views). This would be ok if the feature was
complete, but, as is, it was useless.
- kept unchanged after a second renaming of the field (which can happen
versions later the first rename).
- used, even when the meaning of the field changed. i.e. the field
`archived` has been renamed to the classic `active`, but the value
in the database should be switched.
image_original => image_1920 (now resized to 1920)
image_big => image_1024
image_large => image_256
image_medium => image_128
image_small => image_64
image replaced by image_1920 (when writing) or by image_1024 (when displaying
what was previously the big size)
+ add new intermediate format:
image_512
PR: #34925
This table tracked standard price changes.
It was used to compute the valuation at date in AVCO and standard
through `get_history_price`.
The next commits will introduce the stock valuation layers, separating
the valuation from the stock move. As a valuation layer will be created
when a standard price is updated on the product or when a stock move
impacts the valuation, the information from this table will be
duplicated.
We don't plan to adapt `get_history_price` to work with the valuation
layer since the valuation report at date and today will be the same: a
grouped list on the valuation layers with eventually a domain on the
date.
task-1875873
Ivan Yelizariev