**Steps to reproduce:**
* Install the **l10n_pe_edi** module.
* Change the company currency to **USD**.
* Configure **Peruvian detraction** settings:
  * Set a **withholding percentage** on a product.
  * Set a **withholding code** on the same product.
* Create a customer invoice in **USD**.
* Set the **Operation Type** to **[1001] Operation Subject to Detraction**.
* Add the configured product to the invoice.
* Confirm the invoice, **send it to SUNAT** by *process now**, and generate the ZIP which has XML.

**Observed behavior:**
* The generated XML contains: `<cbc:Amount currencyID="PEN">116.82</cbc:Amount>` The `currencyID` is `PEN` but the amount is the raw USD value instead of being converted to PEN using the exchange rate.
* The PDF report also shows the withholding amount with the wrong currency symbol.

**Cause:**
* `_l10n_pe_edi_get_spot()` computed `spot['amount']` as `amount_total_signed * percent`, where `amount_total_signed` is stored in the **company currency**. When the company currency is PEN this works correctly, but when it is changed to USD the value is in USD while the XML node hardcodes `currencyID="PEN"`.
* `spot['currency']` was also set to `company_id.currency_id`, so the PDF rendered the amount with the wrong currency symbol.
* In `_add_invoice_payment_terms_nodes`, `spot_amount` was selected via `spot['amount'] if invoice.currency_id == invoice.company_id.currency_id` which relied on company currency being PEN as a proxy for the invoice being in PEN, breaking when company currency is USD.

**Fix:**
* In `_l10n_pe_edi_get_spot()`, compute `spot['amount']` by explicitly converting `amount_total` from the invoice currency to PEN using `currency_id._convert(..., env.ref('base.PEN'), ...)`, ensuring the detraction amount is always in PEN regardless of the company currency.
* Set `spot['currency']` to `env.ref('base.PEN')` so the PDF report always displays the withholding amount with the correct PEN symbol.
* In `_add_invoice_payment_terms_nodes`, always use `spot['spot_amount']` (invoice currency amount) for the `FormaPago`/`Cuota` arithmetic, since `spot['amount']` is now exclusively the PEN amount for the detraction XML node.

opw-6327571

Forward-Port-Of: odoo/enterprise#125063
Forward-Port-Of: odoo/enterprise#123782

**Problem:**
When a template is restricted to a department, appraisals for employees of that department do not use it — a generic (global) template is used instead.

**Steps to reproduce:**
1. Create an appraisal template and restrict it to a department.
2. Make sure at least one global (no department) template also exists.
3. Open a new appraisal and select an employee of that department.
4. Observe the appraisal template stays on the global one.

**Current behavior:**
The global template that was auto-selected before an employee was chosen is kept, ignoring the department-restricted template.

**Expected behavior:**
The template restricted to the employee's department is selected.

**Cause of the issue:**
On a new appraisal the template is computed before an employee (and therefore a department) is set, so a global template is pre-selected. When the employee is then chosen, `_compute_appraisal_template` reruns but the `appraisal.appraisal_template_id or ...` clause short-circuits on the already set global template and never falls through to the department's template.

**Fix:**
Evaluating the department's template before the existing value ensures a department-restricted template takes precedence over a template that was only auto-selected as a default. The existing value is still preserved when the department has no template of its own, so deliberate choices are untouched.

opw-6293936

Forward-Port-Of: odoo/enterprise#121377

***Issue:***
Since https://github.com/odoo/enterprise/commit/8baaad621a1555680cd77ab514f626137b5b584d the div holding the selection box and the Actions menu carries t-key="targetRecords.length" in the documents list and kanban views, so every change of the selection count destroys and recreates everything inside it. Ticking a document checkbox updates both the selection and the focused record, and with these two updates the recreation goes wrong: the new ActionMenus component ends up destroyed while its button is still in the page. The dropdown click listeners are removed with the component, so clicking Actions does nothing until the selection changes again.

That t-key was added because the selection box shows a stale count after a file upload. The real problem is that SelectionBox caches this.root once in setup. The upload flow reloads the model, which replaces the root record list, and the box keeps counting the selection of the old one.

***Fix:***
Remove the t-key from documents_list_controller.xml and documents_kanban_controller.xml and make DocumentsSelectionBox read the current props.root instead of the value cached by the parent setup. The count follows the new record list after an upload without remounting anything, and the Actions menu is not recreated so its dropdown keeps working. The upload scenario stays covered by the existing multi_view.test.js test, a new test checks that the Actions menu opens with two selected documents.

***Steps to reproduce:***
1. Open the Documents app
2. Switch to the list view
3. Tick the checkboxes of two documents
4. Click the Actions button next to the "2 selected" box

=> the Actions dropdown does not open

Ticket [link](https://www.odoo.com/odoo/project.task/6365618)

opw-6365618

Forward-Port-Of: odoo/enterprise#124238
Forward-Port-Of: odoo/enterprise#123575

# How to reproduce
- Have a lot of reusable & locationless packages (e.g. > 10 000)
- Go to any transfer via the barcode application

# The issue
There is a very long loading time, even in local testing. The client of the tickets experiences loadings up to 120 seconds with 50k packages

# Cause
When opening a transfer, we load barcode data by doing an API call to `_get_stock_barcode_data` : https://github.com/odoo/enterprise/blob/fe058ef501767b7ed9758fc9264f664b32c6bae8/stock_barcode/models/stock_picking.py#L85

During this we preload a lot of records, notably packages : https://github.com/odoo/enterprise/blob/fe058ef501767b7ed9758fc9264f664b32c6bae8/stock_barcode/models/stock_picking.py#L128

The issue is that in the fields we read for the packages, two of them (`location_dest_id` & `contained_quant_ids`) have a `_read_group` in their compute (or in the compute of one of the fields they depend on) :
https://github.com/odoo/odoo/blob/625e6bcbd66c45ea2f699df14e2ea12e2e28a893/addons/stock/models/stock_package.py#L65 https://github.com/odoo/odoo/blob/625e6bcbd66c45ea2f699df14e2ea12e2e28a893/addons/stock/models/stock_package.py#L146

Fortunately, this does not mean that we make a query for every records. Instead, in Odoo, we fetch records in batch of 1000.

So, for the case of the client, every time he loads the database, the backend does 50 000 / 1000 x 2 = 100 queries, which hinders performance a lot

A [PERF] commit was done to limit the number of packages that are fetched base on a config parameter. The problem is that this parameter does not have a default value, so clients still end up with the problem.

[PERF]: https://github.com/odoo/enterprise/commit/efe18bc1ea479270e42846986d7ed449b0865617

# Proposed Solution
Add a default value for that config parameter. The exact value is up to discussion

opw-6200730

Forward-Port-Of: odoo/enterprise#124538
Forward-Port-Of: odoo/enterprise#123696

Version - 19.0

Steps to reproduce(runbot):
- Install `l10n_au_hr_payroll`
- Go to Payroll -> Configuration -> Super Accounts
- Create a Super Account by filling all the required fields
- You will encounter the issue
**TraceBack:**
```
('hr.menu_hr_employee_payroll', 485, 'Employees > Employees', 756):
 Traceback (most recent call last):
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 346, in crawl_menu
    self.mock_action(action_vals)
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 377, in mock_action
    return self.mock_act_window(action)
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 537, in mock_act_window
    mock_method(model, view, fields_list, domain, group_by)
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 570, in mock_view_form
    [data] = record.read(fields_list)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 3490, in read
    return self._read_format(fnames=fields, load=load)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 3747, in _read_format
    vals[name] = convert(record[name], record, use_display_name)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 6688, in __getitem__
    return self._fields[key].__get__(self)
   File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 1744, in __get__
    self.compute_value(recs)
   File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 1915, in compute_value
    records._compute_field_value(self)
   File "/home/odoo/src/odoo/19.0/addons/mail/models/mail_thread.py", line 484, in _compute_field_value
    return super()._compute_field_value(field)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 4954, in _compute_field_value
    determine(field.compute, self)
   File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 81, in determine
    return needle(*args)
   File "/home/odoo/src/enterprise/19.0/l10n_au_hr_payroll/models/hr_employee.py", line 118, in _compute_proportion_warnings
    proportions[emp.id] * 100,
 KeyError: 4
```
Issue:
During the refactoring from `read_group()` to `_read_group()`, the `proportions` dictionary became keyed by `hr.employee` records instead of employee IDs. While the condition was updated to use `proportions.get(emp)`, the warning message still accessed the dictionary using `proportions[emp.id]`, resulting in a `KeyError`.

Fix:
Use the employee record consistently when accessing the `proportions` dictionary after switching to `_read_group()`.

Task ID - 6390373
Refrence:
https://github.com/odoo/enterprise/pull/72217/changes#diff-8ae3564d54e47eec919ef273d44f6276d8c6df60e21c5c8c8f7292b1c38cf541R460

Forward-Port-Of: odoo/enterprise#124478

In Multiwebsite settings, when the public user interactions needs email generation (appointment or event flow), the email links are generated with a base url that does not corresponds to the one from which the request started.

Case 1:
- Have website A and website B
- Create an appointment page website A
- Log in via website B
- As public user, make an appointment in Website A
- Check the generated email

Issue: button links in the email will redirect to the wrong website, so users will encounter an issue when managing the appointment. This occurs because when an user log in, the system parameter 'web.base_url' is updated with the current url.
This parameter is then used as fallback when we need to retrieve the base url without an active record

Case 2:
- Have website A and website B
- Create an event and assign it to website B
- As public user, access the event and register to it
- Check the generated email

Issue: button links in the email will redirect to the wrong website, so users will encounter an issue when managing the event.
This occurs because the record `event.registration` has no website_id field and the base url is taken from the company default website (website A)

Backport with improvements of 15bae202d8f1b5bf70bbc63b2d89025e9237e6cf

opw-4146760
opw-4336369

Forward-Port-Of: odoo/enterprise#124714
Forward-Port-Of: odoo/enterprise#122669

**Issue:**

Envia API documentation is very bad. When trying to generate an Amazon Ground shipping with Envia, the API will sometimes reject the payload with the following error:

`Total items weight exceeds package weight. Please refer to API documentation for allowable limits. (D-703)`

Even though the [Envia documentation](https://docs.envia.com/reference/create-shipping-label) doesn't include it, it seems that Amazon requires the weight for each product. If the payload doesn't include the per-item weight, Amazon will set a default weight, which will cause the error to show up if the delivery contains a high quantity. This can be a blocking issue for some customers.

**Steps to reproduce on a fresh DB:**

- Install Inventory, Sales, and Envia Shipping
- Switch company to be based in India (need to provide an address to deliver from, I just used the Odoo India address - InfoCity Gate, Gandhinagar, Gujarat 382007)
- Configure an Amazon delivery carrier with Envia (India seems to be the only country that supports Amazon with Envia)
	- Inventory > Configuration > Delivery Methods > Envia.com method
	- Must be in a Production environment (because Envia's sandbox server does not work for some reason?)
	- Add an Envia Production Access Token (will have to use one with some funds)
	- Change 'Ship From' to India
	- Configure the `Envia.com Service Name`, set Carrier and Service to AMAZON and AMAZON - Amazon Shipping Standard
- Create a demo product, give it an arbitrary weight like 0.1 kg
- Create a sales order to an Indian customer, include the demo product with a high quantity (like 20), confirm (might have to create a warehouse and configure the customer's information)
- On the generated picking, set the Carrier to Envia.com on the Additional Info tab
- When trying to Validate, it should throw the error

**Fix:**
When generating the payload in _get_shipping_lines() in envia_request, include the item weight in the dictionary.

Related ticket: opw-6261682

Forward-Port-Of: odoo/enterprise#125017
Forward-Port-Of: odoo/enterprise#123935

**Steps to reproduce:**
* Install the **l10n_pe_edi** module.
* Change the company currency to **USD**.
* Configure **Peruvian detraction** settings:
  * Set a **withholding percentage** on a product.
  * Set a **withholding code** on the same product.
* Create a customer invoice in **USD**.
* Set the **Operation Type** to **[1001] Operation Subject to Detraction**.
* Add the configured product to the invoice.
* Confirm the invoice, **send it to SUNAT** by *process now**, and generate the ZIP which has XML.

**Observed behavior:**
* The generated XML contains: `<cbc:Amount currencyID="PEN">116.82</cbc:Amount>` The `currencyID` is `PEN` but the amount is the raw USD value instead of being converted to PEN using the exchange rate.
* The PDF report also shows the withholding amount with the wrong currency symbol.

**Cause:**
* `_l10n_pe_edi_get_spot()` computed `spot['amount']` as `amount_total_signed * percent`, where `amount_total_signed` is stored in the **company currency**. When the company currency is PEN this works correctly, but when it is changed to USD the value is in USD while the XML node hardcodes `currencyID="PEN"`.
* `spot['currency']` was also set to `company_id.currency_id`, so the PDF rendered the amount with the wrong currency symbol.
* In `_add_invoice_payment_terms_nodes`, `spot_amount` was selected via `spot['amount'] if invoice.currency_id == invoice.company_id.currency_id` which relied on company currency being PEN as a proxy for the invoice being in PEN, breaking when company currency is USD.

**Fix:**
* In `_l10n_pe_edi_get_spot()`, compute `spot['amount']` by explicitly converting `amount_total` from the invoice currency to PEN using `currency_id._convert(..., env.ref('base.PEN'), ...)`, ensuring the detraction amount is always in PEN regardless of the company currency.
* Set `spot['currency']` to `env.ref('base.PEN')` so the PDF report always displays the withholding amount with the correct PEN symbol.
* In `_add_invoice_payment_terms_nodes`, always use `spot['spot_amount']` (invoice currency amount) for the `FormaPago`/`Cuota` arithmetic, since `spot['amount']` is now exclusively the PEN amount for the detraction XML node.

opw-6327571

Forward-Port-Of: odoo/enterprise#125063
Forward-Port-Of: odoo/enterprise#123782

Issue
-----
When making an international delivery to a partner with different invoice and delivery addresses, we send the delivery address as the `Sold To` address as well.

Problematic case 1
-----
- Create a belgian company
- Setup UPS
- Create a French customer
	- Add a different french delivery address
- Create a product (with some weight)
- Create a SO (with UPS delivery) to the customer & confirm
- Validate the transfer

> Commercial invoice `Sold To` uses the delivery address

Solution for case 1
-----
Use the delivery address' `commercial_partner_id`. This leads to another issue in some edge cases...

Problematic case 2 (caused by case 1 fix)
-----
- Create a belgian company
- Setup UPS
- Create a French customer
	- Add a delivery address in Switzerland
- Create a product (with some weight)
- Create a SO (with UPS delivery) to the customer & confirm
- Validate the transfer

> UPS error `The Sold To party's country code must be the same as the Ship To party's country code with the exception of Canada and satellite countries.`

Solution for case 2
-----
Default back to delivery address for the `Sold To` field when countries don't match, as this is a limitation of the UPS API. Warn the user, either on the SO or the transfer itself (if no SO).

Warning looks like this (on SO):
<img width="1914" height="716" alt="image" src="https://github.com/user-attachments/assets/f7aa73c4-f24c-42da-8f3e-6a58765ef020" />

-----
Ticket:
opw-6200263

Forward-Port-Of: odoo/enterprise#121479
Forward-Port-Of: odoo/enterprise#118031

Version - 19.0

Steps to reproduce(runbot):
- Install `l10n_au_hr_payroll`
- Go to Payroll -> Configuration -> Super Accounts
- Create a Super Account by filling all the required fields
- You will encounter the issue
**TraceBack:**
```
('hr.menu_hr_employee_payroll', 485, 'Employees > Employees', 756):
 Traceback (most recent call last):
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 346, in crawl_menu
    self.mock_action(action_vals)
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 377, in mock_action
    return self.mock_act_window(action)
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 537, in mock_act_window
    mock_method(model, view, fields_list, domain, group_by)
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 570, in mock_view_form
    [data] = record.read(fields_list)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 3490, in read
    return self._read_format(fnames=fields, load=load)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 3747, in _read_format
    vals[name] = convert(record[name], record, use_display_name)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 6688, in __getitem__
    return self._fields[key].__get__(self)
   File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 1744, in __get__
    self.compute_value(recs)
   File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 1915, in compute_value
    records._compute_field_value(self)
   File "/home/odoo/src/odoo/19.0/addons/mail/models/mail_thread.py", line 484, in _compute_field_value
    return super()._compute_field_value(field)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 4954, in _compute_field_value
    determine(field.compute, self)
   File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 81, in determine
    return needle(*args)
   File "/home/odoo/src/enterprise/19.0/l10n_au_hr_payroll/models/hr_employee.py", line 118, in _compute_proportion_warnings
    proportions[emp.id] * 100,
 KeyError: 4
```
Issue:
During the refactoring from `read_group()` to `_read_group()`, the `proportions` dictionary became keyed by `hr.employee` records instead of employee IDs. While the condition was updated to use `proportions.get(emp)`, the warning message still accessed the dictionary using `proportions[emp.id]`, resulting in a `KeyError`.

Fix:
Use the employee record consistently when accessing the `proportions` dictionary after switching to `_read_group()`.

Task ID - 6390373
Refrence:
https://github.com/odoo/enterprise/pull/72217/changes#diff-8ae3564d54e47eec919ef273d44f6276d8c6df60e21c5c8c8f7292b1c38cf541R460

Forward-Port-Of: odoo/enterprise#124478

In Multiwebsite settings, when the public user interactions needs email generation (appointment or event flow), the email links are generated with a base url that does not corresponds to the one from which the request started.

Case 1:
- Have website A and website B
- Create an appointment page website A
- Log in via website B
- As public user, make an appointment in Website A
- Check the generated email

Issue: button links in the email will redirect to the wrong website, so users will encounter an issue when managing the appointment. This occurs because when an user log in, the system parameter 'web.base_url' is updated with the current url.
This parameter is then used as fallback when we need to retrieve the base url without an active record

Case 2:
- Have website A and website B
- Create an event and assign it to website B
- As public user, access the event and register to it
- Check the generated email

Issue: button links in the email will redirect to the wrong website, so users will encounter an issue when managing the event.
This occurs because the record `event.registration` has no website_id field and the base url is taken from the company default website (website A)

Backport with improvements of 15bae202d8f1b5bf70bbc63b2d89025e9237e6cf

opw-4146760
opw-4336369

Forward-Port-Of: odoo/enterprise#124714
Forward-Port-Of: odoo/enterprise#122669

Steps to reproduce:
1 Install documents_hr and hr_sign on saas-19.1.
2 Go to Settings → Documents → disable Human Resources. 
3 Open the Documents app and delete the Employees - My Company folder. 
4 Upgrade the database to saas-19.2.

Issue:
- During the upgrade, the post-init hook attempts to embed the HR Sign action into predefined document folders. Since the Employees - My Company folder has been deleted (or is inactive), the folder lookup returns no record, leading to the following error:

```python3
Traceback (most recent call last):
  File "/home/odoo/src/odoo/saas-19.2/odoo/service/server.py", line 1664, in preload_registries
    registry = Registry.new(dbname, update_module=update_module, install_modules=config['init'], upgrade_modules=config['update'], reinit_modules=config['reinit'])
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/odoo/src/odoo/saas-19.2/odoo/tools/func.py", line 65, in locked
    return func(inst, *args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/odoo/src/odoo/saas-19.2/odoo/orm/registry.py", line 186, in new
    load_modules(
  File "/home/odoo/src/odoo/saas-19.2/odoo/modules/loading.py", line 465, in load_modules
    load_module_graph(
  File "/home/odoo/src/odoo/saas-19.2/odoo/modules/loading.py", line 244, in load_module_graph
    getattr(py_module, post_init)(env)
  File "/home/odoo/src/enterprise/saas-19.2/documents_hr_sign/__init__.py", line 12, in _embed_sign_post_init
    folders.with_user(SUPERUSER_ID)._embed_action(sign_action.id)
  File "/home/odoo/src/enterprise/saas-19.2/documents/models/documents_document.py", line 1643, in _embed_action
    folder.action_folder_embed_action(folder.id, action_id)
  File "/home/odoo/src/enterprise/saas-19.2/documents/models/documents_document.py", line 1614, in action_folder_embed_action
    return self.get_documents_actions(folder_id)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/odoo/src/enterprise/saas-19.2/documents/models/documents_document.py", line 1522, in get_documents_actions
    raise UserError(_('This folder does not exist or is not accessible.'))
odoo.exceptions.UserError: Esta carpeta no existe o no es accesible.
```

Root cause:
- As part of documents_sign, the documents_hr_sign action is embedded into predefined document folders (introduced in https://github.com/odoo/enterprise/pull/101890).
- The folder lookup relies on [_search()](https://github.com/odoo/odoo/blob/85992d5f5f8caeacdcc570fb9fefb4cfcec6460c/odoo/orm/models.py#L4681-L4691), which filters out inactive records 
  through the active test. Consequently, [search_fetch()](https://github.com/odoo/odoo/blob/85992d5f5f8caeacdcc570fb9fefb4cfcec6460c/odoo/orm/models.py#L1441-L1451) returns no matching

  folder when the predefined folder has been deleted or deactivated.  It will
  try to get document actions of predefined folder here, so if the folder is
https://github.com/odoo/enterprise/blob/6d334a17d10faec33d60f4a8d7d263c091bb456e/documents/models/documents_document.py#L1592
  inactive it will not be able to read the folder here.
https://github.com/odoo/enterprise/blob/6d334a17d10faec33d60f4a8d7d263c091bb456e/documents/models/documents_document.py#L1494

Fix:
- Instead of attempting to embed the Sign action into all company employee folders, filter out inactive folders before calling _embed_action().

- This ensures the post-init hook only processes active employee document folders. If a predefined folder has been deleted or deactivated, it is skipped preventing the upgrade from attempting to embed an action into an inaccessible folder and avoiding the resulting UserError.

opw-6358063
upg-4449902

Forward-Port-Of: odoo/enterprise#123846

# How to reproduce
- Have a lot of reusable & locationless packages (e.g. > 10 000)
- Go to any transfer via the barcode application

# The issue
There is a very long loading time, even in local testing. The client of the tickets experiences loadings up to 120 seconds with 50k packages

# Cause
When opening a transfer, we load barcode data by doing an API call to `_get_stock_barcode_data` : https://github.com/odoo/enterprise/blob/fe058ef501767b7ed9758fc9264f664b32c6bae8/stock_barcode/models/stock_picking.py#L85

During this we preload a lot of records, notably packages : https://github.com/odoo/enterprise/blob/fe058ef501767b7ed9758fc9264f664b32c6bae8/stock_barcode/models/stock_picking.py#L128

The issue is that in the fields we read for the packages, two of them (`location_dest_id` & `contained_quant_ids`) have a `_read_group` in their compute (or in the compute of one of the fields they depend on) :
https://github.com/odoo/odoo/blob/625e6bcbd66c45ea2f699df14e2ea12e2e28a893/addons/stock/models/stock_package.py#L65 https://github.com/odoo/odoo/blob/625e6bcbd66c45ea2f699df14e2ea12e2e28a893/addons/stock/models/stock_package.py#L146

Fortunately, this does not mean that we make a query for every records. Instead, in Odoo, we fetch records in batch of 1000.

So, for the case of the client, every time he loads the database, the backend does 50 000 / 1000 x 2 = 100 queries, which hinders performance a lot

A [PERF] commit was done to limit the number of packages that are fetched base on a config parameter. The problem is that this parameter does not have a default value, so clients still end up with the problem.

[PERF]: https://github.com/odoo/enterprise/commit/efe18bc1ea479270e42846986d7ed449b0865617

# Proposed Solution
Add a default value for that config parameter. The exact value is up to discussion

opw-6200730

Forward-Port-Of: odoo/enterprise#124538
Forward-Port-Of: odoo/enterprise#123696

**Issue:**

Envia API documentation is very bad. When trying to generate an Amazon Ground shipping with Envia, the API will sometimes reject the payload with the following error:

`Total items weight exceeds package weight. Please refer to API documentation for allowable limits. (D-703)`

Even though the [Envia documentation](https://docs.envia.com/reference/create-shipping-label) doesn't include it, it seems that Amazon requires the weight for each product. If the payload doesn't include the per-item weight, Amazon will set a default weight, which will cause the error to show up if the delivery contains a high quantity. This can be a blocking issue for some customers.

**Steps to reproduce on a fresh DB:**

- Install Inventory, Sales, and Envia Shipping
- Switch company to be based in India (need to provide an address to deliver from, I just used the Odoo India address - InfoCity Gate, Gandhinagar, Gujarat 382007)
- Configure an Amazon delivery carrier with Envia (India seems to be the only country that supports Amazon with Envia)
	- Inventory > Configuration > Delivery Methods > Envia.com method
	- Must be in a Production environment (because Envia's sandbox server does not work for some reason?)
	- Add an Envia Production Access Token (will have to use one with some funds)
	- Change 'Ship From' to India
	- Configure the `Envia.com Service Name`, set Carrier and Service to AMAZON and AMAZON - Amazon Shipping Standard
- Create a demo product, give it an arbitrary weight like 0.1 kg
- Create a sales order to an Indian customer, include the demo product with a high quantity (like 20), confirm (might have to create a warehouse and configure the customer's information)
- On the generated picking, set the Carrier to Envia.com on the Additional Info tab
- When trying to Validate, it should throw the error

**Fix:**
When generating the payload in _get_shipping_lines() in envia_request, include the item weight in the dictionary.

Related ticket: opw-6261682

Forward-Port-Of: odoo/enterprise#125017
Forward-Port-Of: odoo/enterprise#123935

Issue:
After a lock date is set on a company, certain lines on the Balance Sheet report that look back before a given date, such as "Profits (Losses) from Previous Years", can end up silently missing data.

Steps to reproduce:
1. Create a company located in Belgium.
2. Create an invoice. Give it a date in 2025.
3. Go to the Balance Sheet (Abbr Con) (BE) report. Observe that the value on the newly created invoice is captured in the line: "Profits (Losses) from Previous Years".
4. Set a lock date for Dec. 31st, 2025.
5. Return to the same Balance Sheet. Observe that the "Profits (Losses) from Previous Years" is now 0.

Note: The issue isn't caused by anything in the Belgium localization. However, not all Balance Sheets have a line that both looks back before a given date and also gets captured in snapshots. The Balance Sheet (Abbr Con) (BE) report has such a line.

Explanation:
When a lock date is set to Dec. 31st 2025, snapshots are created. The snapshots' `date` fields are all set to Dec. 31st 2025, matching the lock date. One snapshot--responsible for the line "Profits (Losses) from Previous Years"--will have a `date_scope` of `to_beginning_of_fiscalyear`. In that snapshot, the data actually aggregated only covers up to the end of the prior fiscal year, Dec. 31st 2024. So the snapshot's `date` doesn't match its own data's real cutoff.

When the report is generated, to calculate the line "Profits (Losses) from Previous Years", Odoo sums only the entries dated after the snapshot's `date`. Odoo then adds that to the snapshot's stored figure. This becomes an issue if the report is viewed as of a date in 2026. As of 2026, the line should include all 2025's entries. But, these entries will not be included. They aren't part of the snapshot, and they're skipped again since summing only resumes after Dec. 31st, 2025.

A similar problem will arise in any report lines that both get captured in a snapshot and also use a `date_scope` that cuts off before the report's as of date.

Solution:
We now set the `date` field of snapshots to the data's real cutoff, instead of the lock date. We get the data's real cutoff from `_get_date_bounds_info`, which is the same method used to build the query that searches for the account move lines to include in the snapshot. This way, the snapshot's `date` always matches the data it actually contains.

Note: snapshots created before this fix may still hold a mismatched `date`
and should be regenerated.

opw-6379956

Steps to reproduce:
===================
- Configure a product with a 12% VAT
- Create a fiscal position that maps 12% → 6%
- Open POS and apply the fiscal position to an order
- Add the product and validate the sale
- Download Receipt

Issue:
======
- The POS correctly applies and displays the mapped 6% VAT
- The payload sent to the FDM still uses the product's original 12% VAT
- As a result, the receipt displays the wrong VAT rate

Cause:
======
- VAT labels in the Blackbox payload were generated from product_id.taxes_id
- Fiscal position tax mappings were ignored when building VAT inputs
- Blackbox product prices were also computed without taking the fiscal position into account

Fix:
====
- Use taxes after fiscal position mapping when generating VAT data for the Blackbox payload
- Ensure VAT labels sent to the FDM match the taxes applied on the POS order
- Apply fiscal position tax mapping when computing Blackbox product prices

Task:6392006

FW of : https://github.com/odoo/enterprise/pull/124975

Steps to reproduce:
1 Install documents_hr and hr_sign on saas-19.1.
2 Go to Settings → Documents → disable Human Resources. 
3 Open the Documents app and delete the Employees - My Company folder. 
4 Upgrade the database to saas-19.2.

Issue:
- During the upgrade, the post-init hook attempts to embed the HR Sign action into predefined document folders. Since the Employees - My Company folder has been deleted (or is inactive), the folder lookup returns no record, leading to the following error:

```python3
Traceback (most recent call last):
  File "/home/odoo/src/odoo/saas-19.2/odoo/service/server.py", line 1664, in preload_registries
    registry = Registry.new(dbname, update_module=update_module, install_modules=config['init'], upgrade_modules=config['update'], reinit_modules=config['reinit'])
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/odoo/src/odoo/saas-19.2/odoo/tools/func.py", line 65, in locked
    return func(inst, *args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/odoo/src/odoo/saas-19.2/odoo/orm/registry.py", line 186, in new
    load_modules(
  File "/home/odoo/src/odoo/saas-19.2/odoo/modules/loading.py", line 465, in load_modules
    load_module_graph(
  File "/home/odoo/src/odoo/saas-19.2/odoo/modules/loading.py", line 244, in load_module_graph
    getattr(py_module, post_init)(env)
  File "/home/odoo/src/enterprise/saas-19.2/documents_hr_sign/__init__.py", line 12, in _embed_sign_post_init
    folders.with_user(SUPERUSER_ID)._embed_action(sign_action.id)
  File "/home/odoo/src/enterprise/saas-19.2/documents/models/documents_document.py", line 1643, in _embed_action
    folder.action_folder_embed_action(folder.id, action_id)
  File "/home/odoo/src/enterprise/saas-19.2/documents/models/documents_document.py", line 1614, in action_folder_embed_action
    return self.get_documents_actions(folder_id)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/odoo/src/enterprise/saas-19.2/documents/models/documents_document.py", line 1522, in get_documents_actions
    raise UserError(_('This folder does not exist or is not accessible.'))
odoo.exceptions.UserError: Esta carpeta no existe o no es accesible.
```

Root cause:
- As part of documents_sign, the documents_hr_sign action is embedded into predefined document folders (introduced in https://github.com/odoo/enterprise/pull/101890).
- The folder lookup relies on [_search()](https://github.com/odoo/odoo/blob/85992d5f5f8caeacdcc570fb9fefb4cfcec6460c/odoo/orm/models.py#L4681-L4691), which filters out inactive records 
  through the active test. Consequently, [search_fetch()](https://github.com/odoo/odoo/blob/85992d5f5f8caeacdcc570fb9fefb4cfcec6460c/odoo/orm/models.py#L1441-L1451) returns no matching

  folder when the predefined folder has been deleted or deactivated.  It will
  try to get document actions of predefined folder here, so if the folder is
https://github.com/odoo/enterprise/blob/6d334a17d10faec33d60f4a8d7d263c091bb456e/documents/models/documents_document.py#L1592
  inactive it will not be able to read the folder here.
https://github.com/odoo/enterprise/blob/6d334a17d10faec33d60f4a8d7d263c091bb456e/documents/models/documents_document.py#L1494

Fix:
- Instead of attempting to embed the Sign action into all company employee folders, filter out inactive folders before calling _embed_action().

- This ensures the post-init hook only processes active employee document folders. If a predefined folder has been deleted or deactivated, it is skipped preventing the upgrade from attempting to embed an action into an inaccessible folder and avoiding the resulting UserError.

opw-6358063
upg-4449902

Version - 19.0

Steps to reproduce(runbot):
- Install `l10n_au_hr_payroll`
- Go to Payroll -> Configuration -> Super Accounts
- Create a Super Account by filling all the required fields
- You will encounter the issue
**TraceBack:**
```
('hr.menu_hr_employee_payroll', 485, 'Employees > Employees', 756):
 Traceback (most recent call last):
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 346, in crawl_menu
    self.mock_action(action_vals)
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 377, in mock_action
    return self.mock_act_window(action)
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 537, in mock_act_window
    mock_method(model, view, fields_list, domain, group_by)
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 570, in mock_view_form
    [data] = record.read(fields_list)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 3490, in read
    return self._read_format(fnames=fields, load=load)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 3747, in _read_format
    vals[name] = convert(record[name], record, use_display_name)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 6688, in __getitem__
    return self._fields[key].__get__(self)
   File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 1744, in __get__
    self.compute_value(recs)
   File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 1915, in compute_value
    records._compute_field_value(self)
   File "/home/odoo/src/odoo/19.0/addons/mail/models/mail_thread.py", line 484, in _compute_field_value
    return super()._compute_field_value(field)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 4954, in _compute_field_value
    determine(field.compute, self)
   File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 81, in determine
    return needle(*args)
   File "/home/odoo/src/enterprise/19.0/l10n_au_hr_payroll/models/hr_employee.py", line 118, in _compute_proportion_warnings
    proportions[emp.id] * 100,
 KeyError: 4
```
Issue:
During the refactoring from `read_group()` to `_read_group()`, the `proportions` dictionary became keyed by `hr.employee` records instead of employee IDs. While the condition was updated to use `proportions.get(emp)`, the warning message still accessed the dictionary using `proportions[emp.id]`, resulting in a `KeyError`.

Fix:
Use the employee record consistently when accessing the `proportions` dictionary after switching to `_read_group()`.

Task ID - 6390373
Refrence:
https://github.com/odoo/enterprise/pull/72217/changes#diff-8ae3564d54e47eec919ef273d44f6276d8c6df60e21c5c8c8f7292b1c38cf541R460

Forward-Port-Of: odoo/enterprise#124478

In Multiwebsite settings, when the public user interactions needs email generation (appointment or event flow), the email links are generated with a base url that does not corresponds to the one from which the request started.

Case 1:
- Have website A and website B
- Create an appointment page website A
- Log in via website B
- As public user, make an appointment in Website A
- Check the generated email

Issue: button links in the email will redirect to the wrong website, so users will encounter an issue when managing the appointment. This occurs because when an user log in, the system parameter 'web.base_url' is updated with the current url.
This parameter is then used as fallback when we need to retrieve the base url without an active record

Case 2:
- Have website A and website B
- Create an event and assign it to website B
- As public user, access the event and register to it
- Check the generated email

Issue: button links in the email will redirect to the wrong website, so users will encounter an issue when managing the event.
This occurs because the record `event.registration` has no website_id field and the base url is taken from the company default website (website A)

Backport with improvements of 15bae202d8f1b5bf70bbc63b2d89025e9237e6cf

opw-4146760
opw-4336369

Forward-Port-Of: odoo/enterprise#124714
Forward-Port-Of: odoo/enterprise#122669

### Description of the issue/feature this PR addresses:
This PR addresses an issue where expected Mexican work hours and attendance fields fail to generate correctly in contexts where a draft payslip does not yet exist. Specifically, this was preventing the correct evaluation of work hours when an off-cycle payslip is initially generated for a given time period.

### Current behavior before PR:
The _preprocess_work_hours_data method aborts early if no hr.payslip record is found for the period. Because the method relies on an instantiated payslip to fetch calendar info, duration validation, and rule parameters, work hours cannot be properly calculated during the initial generation of off-cycle payslips. As a result, off-cycle payslips will under-calculate the total wage for a given pay period. 

### Desired behavior after PR is merged: 
The strict dependency on the payslip record is removed. If a payslip is absent, the logic will now safely fall back to the contract version's base data:
- The calendar defaults to the contract version's resource_calendar_id.
- The duration falls back to standard (is_wrong_duration = False).
- The l10n_mx_schedule_table rule parameter is fetched globally from the environment (hr.rule.parameter) instead of using the payslip helper.

This ensures that expected work hours are calculated consistently across all payroll workflows, even before the payslip is fully instantiated.

opw-6351402

---
I confirm I have signed the CLA and read the PR guidelines at www.odoo.com/submit-pr

**Steps to reproduce the issue:**

- Install the l10n_mx_edi module.
- Open the Accounting app and go to Customers → Invoices.
- From the invoice line, make the Tax Object field visible.
- Create a new invoice and add a product that has no Unit Price.
- Save the invoice.
- Edit the same invoice, set a Unit Price for the product, and save it again.
- Issue: The Tax Object field is not automatically set after the Unit Price is added.

**Description:**

- In our standard code, there is a condition that skips computing the Tax Object when the Unit Price is [0](https://github.com/odoo/enterprise/blob/615a5f756d708db95e22cdbdd6e7aca72ac769ec/l10n_mx_edi/models/account_move.py#L988). This is the expected behavior.
- However, after setting a Unit Price on the product, the Tax Object is not recomputed, so it remains unset.
- To resolve this issue, I added price_unit to the @api.depends decorator so that the Tax Object is recomputed whenever the Unit Price changes.

**Reference videos:**

Before the fix:
[screen-capture (2).webm](https://github.com/user-attachments/assets/c1162fd9-52b8-4a8c-9ac2-39cf2ca3ad69)


After the fix:
[screen-capture (1).webm](https://github.com/user-attachments/assets/1c05dc34-6778-487a-9bfc-635cf652e670)


OPW -6305080 
UPG - 4268848

Forward-Port-Of: odoo/enterprise#123009

Version - 19.0

Steps to reproduce(runbot):
- Install `l10n_au_hr_payroll`
- Go to Payroll -> Configuration -> Super Accounts
- Create a Super Account by filling all the required fields
- You will encounter the issue
**TraceBack:**
```
('hr.menu_hr_employee_payroll', 485, 'Employees > Employees', 756):
 Traceback (most recent call last):
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 346, in crawl_menu
    self.mock_action(action_vals)
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 377, in mock_action
    return self.mock_act_window(action)
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 537, in mock_act_window
    mock_method(model, view, fields_list, domain, group_by)
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 570, in mock_view_form
    [data] = record.read(fields_list)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 3490, in read
    return self._read_format(fnames=fields, load=load)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 3747, in _read_format
    vals[name] = convert(record[name], record, use_display_name)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 6688, in __getitem__
    return self._fields[key].__get__(self)
   File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 1744, in __get__
    self.compute_value(recs)
   File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 1915, in compute_value
    records._compute_field_value(self)
   File "/home/odoo/src/odoo/19.0/addons/mail/models/mail_thread.py", line 484, in _compute_field_value
    return super()._compute_field_value(field)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 4954, in _compute_field_value
    determine(field.compute, self)
   File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 81, in determine
    return needle(*args)
   File "/home/odoo/src/enterprise/19.0/l10n_au_hr_payroll/models/hr_employee.py", line 118, in _compute_proportion_warnings
    proportions[emp.id] * 100,
 KeyError: 4
```
Issue:
During the refactoring from `read_group()` to `_read_group()`, the `proportions` dictionary became keyed by `hr.employee` records instead of employee IDs. While the condition was updated to use `proportions.get(emp)`, the warning message still accessed the dictionary using `proportions[emp.id]`, resulting in a `KeyError`.

Fix:
Use the employee record consistently when accessing the `proportions` dictionary after switching to `_read_group()`.

Task ID - 6390373
Refrence:
https://github.com/odoo/enterprise/pull/72217/changes#diff-8ae3564d54e47eec919ef273d44f6276d8c6df60e21c5c8c8f7292b1c38cf541R460

Forward-Port-Of: odoo/enterprise#124478

In MPS, if a product’s Bill of Materials (BOM) is not specified at the time of addition, the system will not correctly account for batch size.

Steps to reproduce:
-------------------
* Create a product with a bom that has a batch size of 2
* Open MPS
* Add the product
  - without specifying the bom
  - Route Manufacture
* Add 1 in the Forcast Demand -> the batch size from the bom it's not taken into account.

Observation:
-------------
When updating mps, it will call get_production_schedule_view_state: https://github.com/odoo/enterprise/blob/7092dd2cc3578c3f22d1fbc82d958d57e2f93553/mrp_mps/models/mrp_mps.py#L424 this function when calculating the quantity to resplenish will call _get_resplenish_qty: https://github.com/odoo/enterprise/blob/7092dd2cc3578c3f22d1fbc82d958d57e2f93553/mrp_mps/models/mrp_mps.py#L534 to know the quantity to resplenish it will need the batch size, in mps they will only consider the batch size from the bom registered: https://github.com/odoo/enterprise/blob/7092dd2cc3578c3f22d1fbc82d958d57e2f93553/mrp_mps/models/mrp_mps.py#L863-L865 
Since there is no default value for bom_id, If there is no bom selected, there is no batch size.

opw-6259956

Forward-Port-Of: odoo/enterprise#119560

Steps to reproduce:
------------
  - Install website_sale_renting_planning.
  - Create an employee without a working calendar.
  - Create a service product and enable the planning service.
  - Configure the product:
     - type -> service
     - plan services
         - assign the resource created above.
         - enable sync shifts and rental orders.
  - open the product on the website.

Issue:
------------
   when the resource does not have a working calendar, opening the product on the website raises a singleton traceback.

Cause:
--------
   The _leave_intervals_batch method uses [ensure_one](https://github.com/odoo/odoo/blob/3364b3b38f7d3891d8c69b87fb8234eb9ac8e52b/addons/resource/models/resource_calendar.py#L531-L542),  so a calendar record is needed.

Fix:
----
   - when the resource does not have a working calendar, fall back to the resource's company working calendar
   before calling _leave_intervals_batch. 
   - I used the company calendar only when computing the resource's leave intervals. It is not used to display the company's
    non-working days.

 I found this issue while working on this task-5411076

After a preparation ticket was marked as completed, transferring one of
its items to another table through Split still kept that item in the
order's kitchen history. When a new product was later added and sent
from the destination table, the preparation display created a new
ticket containing both the new item and the already completed one.

Steps to reproduce:
-------------------
* Open a POS session and the Preparation Display
* Select Table 1, add 2 items and send them to the kitchen
* Mark the preparation ticket as Completed
* On Table 1, open Split, select one item and Transfer it to Table 2
* Open Table 2, add a new item and send it to the kitchen

> Observation:
The new preparation ticket contains the newly added item and the
previously completed transferred item.

Why the fix:
------------
Split and table transfer move preparation history to the destination
order with a new line uuid, but the preparation display still tracks
the original ticket on the source order. Without marking the moved
quantity as already sent, the server treated the transferred line as
pending and included it again in the next ticket.

Set transferredQty when moving preparation history on split/transfer,
and read it reliably in _process_preparation_changes so only items
with a real quantity increase are sent to the kitchen.

Related: https://github.com/odoo/odoo/pull/273568

opw-6146176

**Steps to reproduce:**
* Install the **l10n_pe_edi** module.
* Change the company currency to **USD**.
* Configure **Peruvian detraction** settings:
  * Set a **withholding percentage** on a product.
  * Set a **withholding code** on the same product.
* Create a customer invoice in **USD**.
* Set the **Operation Type** to **[1001] Operation Subject to Detraction**.
* Add the configured product to the invoice.
* Confirm the invoice, **send it to SUNAT** by *process now**, and generate the ZIP which has XML.

**Observed behavior:**
* The generated XML contains: `<cbc:Amount currencyID="PEN">116.82</cbc:Amount>` The `currencyID` is `PEN` but the amount is the raw USD value instead of being converted to PEN using the exchange rate.
* The PDF report also shows the withholding amount with the wrong currency symbol.

**Cause:**
* `_l10n_pe_edi_get_spot()` computed `spot['amount']` as `amount_total_signed * percent`, where `amount_total_signed` is stored in the **company currency**. When the company currency is PEN this works correctly, but when it is changed to USD the value is in USD while the XML node hardcodes `currencyID="PEN"`.
* `spot['currency']` was also set to `company_id.currency_id`, so the PDF rendered the amount with the wrong currency symbol.
* In `_add_invoice_payment_terms_nodes`, `spot_amount` was selected via `spot['amount'] if invoice.currency_id == invoice.company_id.currency_id` which relied on company currency being PEN as a proxy for the invoice being in PEN, breaking when company currency is USD.

**Fix:**
* In `_l10n_pe_edi_get_spot()`, compute `spot['amount']` by explicitly converting `amount_total` from the invoice currency to PEN using `currency_id._convert(..., env.ref('base.PEN'), ...)`, ensuring the detraction amount is always in PEN regardless of the company currency.
* Set `spot['currency']` to `env.ref('base.PEN')` so the PDF report always displays the withholding amount with the correct PEN symbol.
* In `_add_invoice_payment_terms_nodes`, always use `spot['spot_amount']` (invoice currency amount) for the `FormaPago`/`Cuota` arithmetic, since `spot['amount']` is now exclusively the PEN amount for the detraction XML node.

opw-6327571

Forward-Port-Of: odoo/enterprise#125063
Forward-Port-Of: odoo/enterprise#123782

**Problem:**
When a template is restricted to a department, appraisals for employees of that department do not use it — a generic (global) template is used instead.

**Steps to reproduce:**
1. Create an appraisal template and restrict it to a department.
2. Make sure at least one global (no department) template also exists.
3. Open a new appraisal and select an employee of that department.
4. Observe the appraisal template stays on the global one.

**Current behavior:**
The global template that was auto-selected before an employee was chosen is kept, ignoring the department-restricted template.

**Expected behavior:**
The template restricted to the employee's department is selected.

**Cause of the issue:**
On a new appraisal the template is computed before an employee (and therefore a department) is set, so a global template is pre-selected. When the employee is then chosen, `_compute_appraisal_template` reruns but the `appraisal.appraisal_template_id or ...` clause short-circuits on the already set global template and never falls through to the department's template.

**Fix:**
Evaluating the department's template before the existing value ensures a department-restricted template takes precedence over a template that was only auto-selected as a default. The existing value is still preserved when the department has no template of its own, so deliberate choices are untouched.

opw-6293936

Forward-Port-Of: odoo/enterprise#121377

Previously, when a report was loading if a variant was selected, it would display the first one when it loaded and display the second one when it loaded.

With this, we wont show the first one as we are waiting for the new one.

To reproduce:
- load the Demo data on the demo company
- Add time.sleep(5) in _get_lines
- load the Generic Tax report and wait for it to load
- click on the Group by: Account > Tax and wait for 3s
- click on the Group by: Tax > Account
- Watch the Account > Tax load and still being displayed for 3s while the Tax > Account variant is loading.

The loading of the report contains 2 steps:
- Loading the options of the report
- Loading the content of the report itself

The options are used to update the bar at the top of the screen with the filters, such as date or horizontal_split.
Since some of those filters dont need to call again for the options, we will store them in a map loadingOptionsFunctions, for them to be called when the options have been fetched.
If they are clicked when the data of the report is loading, we can directly updated the options since they dont need to update the value of the lines.

Some need to update the lines like rounding_unit, however since the call to format_column_values_from_client is faster than querying again the line,
We check if the currency changed while loading the report and if it did, we format the column values.

Following those changes, a lot of async/await were rendered useless since we are not waiting for the click of the filter before the user can click on another one.
We are able to remove them and when changing the value of an option with _updateOption or updateOption, ... they will return the promise of the reloading of the report which can be awaited when necessary.

Issue Before This Commit:
=============================
The workorder planning Gantt view only declared some fields inside the popover element. As a result, they were not fetched with the Gantt records, even though Gantt decorations also used them.

Steps to Reproduce:
===================
- Install the **mrp_workorder** module.
- Enable the **Work Orders** option in **Manufacturing → Configuration → Settings**.
- Create and plan a MO with at least one work order.
- Open the work order planning view from the **Planning** menu.
- Observe the following error: 
An `EvalError` is raised: `Name 'has_conflicts' is not defined.`

Cause of the Issue:
=============================
In this [PR](https://github.com/odoo/enterprise/pull/123693), the variables were moved inside the popover, which caused the Gantt view to lose access to them.

After This Commit:
=============================
These fields have been restored as top-level Gantt fields, ensuring they are loaded with the record data and accessible wherever they are needed.

***Issue:***
Since https://github.com/odoo/enterprise/commit/8baaad621a1555680cd77ab514f626137b5b584d the div holding the selection box and the Actions menu carries t-key="targetRecords.length" in the documents list and kanban views, so every change of the selection count destroys and recreates everything inside it. Ticking a document checkbox updates both the selection and the focused record, and with these two updates the recreation goes wrong: the new ActionMenus component ends up destroyed while its button is still in the page. The dropdown click listeners are removed with the component, so clicking Actions does nothing until the selection changes again.

That t-key was added because the selection box shows a stale count after a file upload. The real problem is that SelectionBox caches this.root once in setup. The upload flow reloads the model, which replaces the root record list, and the box keeps counting the selection of the old one.

***Fix:***
Remove the t-key from documents_list_controller.xml and documents_kanban_controller.xml and make DocumentsSelectionBox read the current props.root instead of the value cached by the parent setup. The count follows the new record list after an upload without remounting anything, and the Actions menu is not recreated so its dropdown keeps working. The upload scenario stays covered by the existing multi_view.test.js test, a new test checks that the Actions menu opens with two selected documents.

***Steps to reproduce:***
1. Open the Documents app
2. Switch to the list view
3. Tick the checkboxes of two documents
4. Click the Actions button next to the "2 selected" box

=> the Actions dropdown does not open

Ticket [link](https://www.odoo.com/odoo/project.task/6365618)

opw-6365618

Forward-Port-Of: odoo/enterprise#124238
Forward-Port-Of: odoo/enterprise#123575

Version - 19.0

Steps to reproduce(runbot):
- Install `l10n_au_hr_payroll`
- Go to Payroll -> Configuration -> Super Accounts
- Create a Super Account by filling all the required fields
- You will encounter the issue
**TraceBack:**
```
('hr.menu_hr_employee_payroll', 485, 'Employees > Employees', 756):
 Traceback (most recent call last):
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 346, in crawl_menu
    self.mock_action(action_vals)
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 377, in mock_action
    return self.mock_act_window(action)
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 537, in mock_act_window
    mock_method(model, view, fields_list, domain, group_by)
   File "/tmp/tmpao31_hxw/migrations/base/tests/test_mock_crawl.py", line 570, in mock_view_form
    [data] = record.read(fields_list)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 3490, in read
    return self._read_format(fnames=fields, load=load)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 3747, in _read_format
    vals[name] = convert(record[name], record, use_display_name)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 6688, in __getitem__
    return self._fields[key].__get__(self)
   File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 1744, in __get__
    self.compute_value(recs)
   File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 1915, in compute_value
    records._compute_field_value(self)
   File "/home/odoo/src/odoo/19.0/addons/mail/models/mail_thread.py", line 484, in _compute_field_value
    return super()._compute_field_value(field)
   File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 4954, in _compute_field_value
    determine(field.compute, self)
   File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 81, in determine
    return needle(*args)
   File "/home/odoo/src/enterprise/19.0/l10n_au_hr_payroll/models/hr_employee.py", line 118, in _compute_proportion_warnings
    proportions[emp.id] * 100,
 KeyError: 4
```
Issue:
During the refactoring from `read_group()` to `_read_group()`, the `proportions` dictionary became keyed by `hr.employee` records instead of employee IDs. While the condition was updated to use `proportions.get(emp)`, the warning message still accessed the dictionary using `proportions[emp.id]`, resulting in a `KeyError`.

Fix:
Use the employee record consistently when accessing the `proportions` dictionary after switching to `_read_group()`.

Task ID - 6390373
Refrence:
https://github.com/odoo/enterprise/pull/72217/changes#diff-8ae3564d54e47eec919ef273d44f6276d8c6df60e21c5c8c8f7292b1c38cf541R460

Forward-Port-Of: odoo/enterprise#124478

In Multiwebsite settings, when the public user interactions needs email generation (appointment or event flow), the email links are generated with a base url that does not corresponds to the one from which the request started.

Case 1:
- Have website A and website B
- Create an appointment page website A
- Log in via website B
- As public user, make an appointment in Website A
- Check the generated email

Issue: button links in the email will redirect to the wrong website, so users will encounter an issue when managing the appointment. This occurs because when an user log in, the system parameter 'web.base_url' is updated with the current url.
This parameter is then used as fallback when we need to retrieve the base url without an active record

Case 2:
- Have website A and website B
- Create an event and assign it to website B
- As public user, access the event and register to it
- Check the generated email

Issue: button links in the email will redirect to the wrong website, so users will encounter an issue when managing the event.
This occurs because the record `event.registration` has no website_id field and the base url is taken from the company default website (website A)

Backport with improvements of 15bae202d8f1b5bf70bbc63b2d89025e9237e6cf

opw-4146760
opw-4336369

Forward-Port-Of: odoo/enterprise#124714
Forward-Port-Of: odoo/enterprise#122669

Steps to reproduce:
- Set the user language to Dutch.
- Go to Accounting > Reporting > Annual Statements.
- Generate the XBRL export for a report other than the "company,
  abridged" (acon) balance sheet/P&L combination, e.g. an association
  (asso_a/asso_f) or "company, full"/"company, capital" report.
- Open the file: the `<met:bln1>` boolean facts are exported as
  "onwaar" instead of "false", which is not a valid XBRL boolean
  lexical value and gets rejected by the NBB validator.

Cause of the issue:
QWeb templates translate static text nodes by default. The base
module ships a generic `msgid "false" -> msgstr "onwaar"` translation,
used elsewhere in the UI, which silently hijacks the literal
"false"/"true" and unit tokens ("iso4217:EUR", "pure") in the XBRL
data templates whenever the file is generated in Dutch.

Solution:
Add `t-translation="off"` on the `<met:bln1>` boolean facts and the
`<measure>` unit tokens in the 5 remaining XBRL templates,
so these fixed-vocabulary XBRL values are never subject to translation

opw-6395785

### Issue before this commit:
When a company generates an EC Sales List for a foreign country (e.g., a Luxembourgish company running a Belgian report), the exported XML and PDF files incorrectly display the company's primary domestic VAT number instead of the foreign VAT number defined in the fiscal position in the tag DeclarantReference.

### Steps to reproduce the issue:
1. Download Accounting and l10n_lu
2. Switch to LU company
3. Go to Fiscal Positions in settings and create the Belgian position (insert country as Belgium and Foreign Tax ID as BE0477472701)
4. A pop up will appear saying: Click  to create the taxes for this country. so click there to create the taxes
5. Go to Invoices and create a new invoice and be sure that:
	1. be sure the customer has a VAT number in their profile
	2. in tab Other Info the Fiscal Position is set to Belgium
	3. the tax applied is 0% EU S (BE)
	4. date of invoice is in June
6. Open the tax return in 1 July
7. Open EC Sales List June 2026 (BE) and mark as reviewed all the lines
8. Click on Validate
9. Open the XML and PDF file created and see that the tag DeclarantReference is wrong because it reports the data of LU company instead of BE company

### Cause of the issue:
The `export_to_xml_sales_report` method relied on legacy code (company.partner_id.vat) to fetch the VAT number. It failed to use the centralized `get_vat_for_export(options)` method, thereby completely bypassing the foreign VAT logic correctly implemented in other tax reports.

### Reason to introduce the fix:
To ensure tax compliance by appling the correct VAT number from the foreign fiscal position.

opw-6170447

Steps to reproduce
---
1. Enable Packages and turn on "Move Entire Packages" on the delivery operation type.
2. Create a storable product P and add 2 units in different package in stock: 1 in PACK001, 1 in PACK002 
3. In the Barcode app > Operations > Delivery > New
4. Scan a first package PACK001, then a second different package PACK002
5. Scan the first package PACK001 again.

Issue
---
Re-scanning an already scanned package is meant to be rejected with a "This package is already scanned." warning, but the rejection stops working as soon as a second package is present in the transfer, so the package content gets added a second time and, once validated, the source quant goes negative (the package ends up holding a negative and a positive quant of the same product).

Commit 23613c63947 added a canPackSomeLines flag that is set to true for every package line that is not the scanned one, so any other package in the transfer makes the alreadyDonePackId && !canPackSomeLines guard false and silently skips the warning. The scanned package already exposes whether it had something left to pack through scannedPackages, so gating the warning on that flag instead keeps the check working regardless of how many other packages are in the transfer.

https://github.com/odoo/enterprise/blob/3cc1a162e61662814b0e52c0c720831952d208a8/stock_barcode/static/src/models/barcode_picking_model.js#L1976-L2006

opw-6279105

## Issue
When creating a Contract Template for a Swiss company, the template does not match the version shown in the Employee's view. Also, some fields are not correctly applied when loading a contract template on an employee (e.g. `hourly_wage`, `wage`, ...).

## Steps to reproduce
1. Install *Switzerland - Swissdec Certified ELM 5.0 - Payroll* (`l10n_ch_hr_payroll`)
2. (Create and) Use a Swiss company
3. In Employees > Configuration > Contract Templates, create a Contract Template
    - Wage Type: Hourly Wage
    - Hourly Wage: Any value > 0
    - **(Notice how the aforementionned fields are missing from the template)**
4. In Employees > Employees, create an Employee
5. On the new employee's view, on the Payroll tab, click "Load Template"
and load the template created in step 3
6. **The data from the template is not applied to the employee's contract**

## Cause
The fields loaded from a contract template are listed in the `whitelist` variable of the `hr.version.wizard`:

https://github.com/odoo/odoo/blob/5c3deb11627f4d6762c4994207bd582afb96f064/addons/hr/wizard/hr_contract_template_wizard.py#L15-L30

Multiple fields were missing from the whitelist (e.g. `hourly_wage`, `l10n_ch_has_{hourly|monthly|lesson}`, ...). These fields would not be loaded from the template when applying a template on an employee.

**This commit replicates the employee's version view on the contract template and adds the related fields to the whitelist for them to be correctly applied when loading a contract template.**

opw-5966664
opw-6128467

Steps to reproduce:
1. Install documents and sales
2. Check product document centralization from settings, and no alias should be set
3. Now go to setting>techinical>emial>aliases>remove default filter and  check total count
4. Go to products and add a document from the smart button
5. Now again check the aliases total count (1 increased)
6. Delete the document on the product.
7. Alias is not deleted
8. Try to add an alias from the settings

Issue:
- When `documents_product` is installed and product document centralization is enabled, deleting a `product.document` can leave an orphan `mail.alias` resulting in an error while creating a new alias.

Cause:
- A product document owns an `ir.attachment`, and that attachment is mirrored as a `documents.document` for `product.product` / `product.template`. In 18.0, `documents.document` inherits `mail.alias.mixin`, so the mirrored document also owns a `mail.alias`.

- Deleting `product.document` deletes its `ir.attachment`. The linked
  `documents.document` is then removed by SQL `ondelete='cascade'` on
  `attachment_id`, not through ORM `unlink()`. Because the document unlink logic
  does not run, the alias cleanup from the mail alias mixin is skipped.

Why not reproducible in 19.0:
- In 19.0, `documents.document` uses `mail.alias.mixin.optional` instead of `mail.alias.mixin`. Binary mirrored documents no longer create a `mail.alias` unless an `alias_name` is explicitly set, so product mirrored documents do not create aliases in the first place.

Solution:
- Override `ir.attachment.unlink()` in `documents`. before deleting attachment,
  collect linked document aliases then delete attachment. SQL cascade deletes
  linked documents.document rows and now we delete collected aliases.


opw-6019294

**Steps to reproduce:**

- Enable "Move Entire Packages" setting on deliveries
- Make a product A, that has a package P1, on hand qty of 1
- Make product B that don't have a package, but on hand qty of 1
- Make a delivery that has both of those products, requested qty of 1 for both
- Mark it as todo
- Go to the barcode app, select the delivery
- Select the line with product B
- Select the line with product A
--> The line with product B is not unselected

**Why the fix:**

When we have a mix of packaged products and products without a package on the same operation, they are handled separately.
The products without a package are handled in https://github.com/odoo/enterprise/blob/98c79af3fb6cb354f46fd2f58e642a72a9271443/stock_barcode/static/src/models/barcode_model.js#L388-L392 
that calls https://github.com/odoo/enterprise/blob/98c79af3fb6cb354f46fd2f58e642a72a9271443/stock_barcode/static/src/models/barcode_model.js#L1277-L1284 
But as you can see, there are no mention of the selected package line, which is stored in **this.lastScanned.packageId**.

As we do not touch this variable, the selected package line stays selected.

The same is true for the other way around, when we select a package line we call https://github.com/odoo/enterprise/blob/98c79af3fb6cb354f46fd2f58e642a72a9271443/stock_barcode/static/src/models/barcode_model.js#L394-L398 
This function does not care for the **selectedLineVirtualId** which represents the selected line without a package.

To avoid this and make it so that only one line is selected even if they have different package, we now set the corresponding value to false to unselect the other line in all situation.

This is basically how it's done in https://github.com/odoo/enterprise/blob/98c79af3fb6cb354f46fd2f58e642a72a9271443/stock_barcode/static/src/models/barcode_model.js#L1202-L1208 
to unselect every line regardless of packages.

opw-6266203

Forward-Port-Of: odoo/enterprise#122038

#### Description of the issue/feature this PR addresses:
Apple Pay (and Google Pay) express checkout via Stripe fails on the cart page with an "update postal address" error when a Starshipit delivery carrier is configured, while normal checkout with a full address works fine. No request even reaches Starshipit's own API logs, because the request is rejected on Odoo's side.

#### Current behavior before PR:
Express checkout wallets only disclose a partial shipping address before the payment is authorized (city, zip, state, country) — the street is withheld by the wallet for privacy. website_sale rates the delivery carriers against this partial address, setting the express_checkout_partial_delivery_address context key so connectors only require the fields available at that stage. The UPS/USPS/DHL/bpost connectors honor this key and skip the street requirement, but delivery_starshipit ignores it: Starshipit._rate_shipment always sends the (empty) destination street to the /api/rates endpoint, which requires it and returns success:false with "street parameter value is required". _starshipit_send_request raises this as a UserError, which propagates out of starshipit_rate_shipment and aborts the entire express checkout rating RPC (_get_delivery_methods_express_checkout), so no carrier at all is returned to the wallet — hence the endless "update postal address" prompt.

#### Desired behavior after PR is merged:
During express checkout, Starshipit is skipped when the partial address has no street (it returns an unsuccessful rate without calling the API), so the other published carriers are still rated and the wallet shows the available shipping options. Starshipit is rated normally at the regular checkout step once the full address is known. Normal (non-express) checkout is unchanged.

opw-6393393