…to hours

## Issue:
'request_hour_from' and 'request_hour_to' should be computed using the default
 calendar attendance values, but instead they display 12:00 AM, meaning their
values remain 0.0.

## Steps to Reproduce:

- Open the Time Off app.
- Configure a Time Off Type with the Duration Type (request_unit) set to Hours.
- Set this Time Off Type as the default one while creating a new Time Off.
- Create a new Time Off request.
- Observe that 'request_hour_from' and 'request_hour_to' are not computed and display 12:00 AM (0.0).

## Root Cause:
While fixing the issue related to preserving leave hours when changing Time Off Types, a regression was introduced.

Related PR: https://github.com/odoo/odoo/pull/227235

During the creation of a new Time Off with the duration type set to Hours by default, the field 'request_unit_hours' is already True. This prevents the computation of 'request_hour_from' and 'request_hour_to', resulting in both values remaining 0.0.

## Solution:
Compute 'request_hour_from' and 'request_hour_to' when their values are still 0.0 (initial Time Off creation case), and skip recomputation only when switching between Time Off Types in order to preserve manually entered hours.

Steps to reproduce : [Video](https://drive.google.com/file/d/1P2MFIj8ZxFtFxv2FlX4Zdq6P-5veTxUb/view?usp=sharing)

OPW: 6209992

Forward-Port-Of: odoo/odoo#266104
Forward-Port-Of: odoo/odoo#264707

# How to reproduce
- Start from a fresh database
- Install the account module
- Import companies with a data file. These companies need to have a country set and the
data file needs to contain some bad data
- Click on the Test button

# The problem
An Odoo Server Error is displayed saying : "savepoint xxx does not exist", which prevents
the import or hides other potential error. More importantly, the date is imported even though it was a test run.

# Cause
Importing companies with a country will import their respective l10n modules using
`button_immediate_install()` :
https://github.com/odoo/odoo/blob/6de867f1c92bacedc0574b63e9e6a2a57fe805dd/odoo/addons/base/models/res_company.py#L319-L322
https://github.com/odoo/odoo/blob/661ddbb7f12e32394a3c11b5a4cd2f38a9e156f5/odoo/addons/base/models/res_company.py#L237

This import calls `cr.commit` :
https://github.com/odoo/odoo/blob/661ddbb7f12e32394a3c11b5a4cd2f38a9e156f5/odoo/addons/base/models/ir_module.py#L632-L634

The issue is that when importing a data file, we create savepoints :
https://github.com/odoo/odoo/blob/661ddbb7f12e32394a3c11b5a4cd2f38a9e156f5/odoo/orm/models.py#L971-L974

And if an error arise during our import, we rollback to the appropriate save point. Sadly,
commiting erases any existing save point, so a savepoint error is rased and the
data is not rolled back.

The problem stems from the fact that importing module is simply not transactionnal and
`button_immediate_install()` is not expected to be called with a savepoint. This
was already partly adressed by :
https://github.com/odoo/odoo/commit/66dcee9aa70dc72a332fde64dbc802266bbe4a5a

But it did not cover the file importing case

opw-6174983

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

Forward-Port-Of: odoo/odoo#265188

Currently the query to get the closing difference account move is done by searching for the reference of the move, which is not very efficient. This commit optimizes this query by adding the journal to the search criteria, which allows us to benefit from the index on the journal_id field.

Here is an example of the before after on a database with 39 million account_move records. Meanwhile only 10-20K account_move are linked to specific journals used in POS payment methods.

All measures are performed with a warmed up cache

[Explain Before](https://explain.dalibo.com/plan/h8edf56c09d7dfd7)

### Benchmark:
<table>
	<thead>
		<tr>
			<th># of am</th>
			<th>Before</th>
			<th>After</th>
		</tr>
	</thead>
	<tbody>
		<tr>
			<td>38982635</td>
			<td>~17s</td>
			<td>~22ms</td>
		</tr>
	</tbody>
</table>

[Explain After](https://explain.dalibo.com/plan/be2397f176a6b29d)

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

Forward-Port-Of: odoo/odoo#262148

**Steps to produce:**
- Install `sale_mrp_renting`.
- Enable `Rental Transfers` from settings.
- Create a rental product.
- Create two variants of the product.
- Create a BoM for one variant and set its type to `Kit`.
- Create a rental order using the other variant.
- Try to confirm the order.

**Issue:**
In versions 17 and 18, a UserError is raised-
```
The unit of measure Units defined on the order line doesn't belong to the same
category as the unit of measure False defined on the product. Please correct
the unit of measure defined on the order line or on the product, they should
belong to the same category.
```

From version 18.2 onward, a different error occurs
```
ZeroDivisionError: float division by zero
```

**Root cause:**
In versions 17 and 18:  At [1], since the BoM is created for a different variant , no BoM is found for the selected variant. As a result, when `_compute_quantity` is called at [2], the `bom.product_uom_id` is empty, which leads to the `UserError` from  `_compute_quantity` method. In version 18.2+:  At [1], as the BoM is empty. Then at [3], `_compute_kit_quantities` is called with an empty BoM, and at [4], this results in a division by zero error.

**Solution:**
Skip the computation when no BoM is found and directly return the quantity to avoid both the `UserError` and the `ZeroDivisionError`.

[1]https://github.com/odoo/enterprise/blob/eaa70dcab2b49eadca1cf36008dd80cfccbb8e4e/sale_mrp_renting/models/sale_order_line.py#L13
[2]https://github.com/odoo/enterprise/blob/eaa70dcab2b49eadca1cf36008dd80cfccbb8e4e/sale_mrp_renting/models/sale_order_line.py#L20
[3]https://github.com/odoo/enterprise/blob/eaa70dcab2b49eadca1cf36008dd80cfccbb8e4e/sale_mrp_renting/models/sale_order_line.py#L21
[4]
https://github.com/odoo/odoo/blob/91b09dbea5c8a306b5e9d2120466777f0248b360/addons/mrp/models/stock_move.py#L676

**opw-6082434**

Forward-Port-Of: odoo/enterprise#118207
Forward-Port-Of: odoo/enterprise#114176

Currently, a traceback occurs when the cron "Payroll: Update data" runs after a referenced partner record has been deleted.

Steps to reproduce the error:
- Install ``l10n_us_hr_payroll`` module with demo data
- Switch to ``My US Company``
- Go to Contacts > Delete ``Internal Revenue Service (IRS)`` contact
- Run the cron ``Payroll: Update data``

Traceback:
```py
ValueError: External ID not found in the system: l10n_us_hr_payroll.res_partner_irs
```
```py
ParseError: while parsing /home/odoo/src/enterprise/ saas-19.3/l10n_us_hr_payroll/data/hr_salary_rule_data.xml:322, somewhere inside
```

https://github.com/odoo/enterprise/blob/21477b333f7a33cb582cd806236e4f9f8346d022/l10n_us_hr_payroll/models/hr_payslip.py#L12-L21
The issue occurs because the cron ``Payroll: Update data`` calls
``_get_data_files_to_update`` method, which loads
the ``data/hr_salary_rule_data.xml`` file containing a reference to the
deleted partner record at [1]. Since the external ID no longer exists,
loading the XML file raises the traceback.

[1]: https://github.com/odoo/enterprise/blob/21477b333f7a33cb582cd806236e4f9f8346d022/l10n_us_hr_payroll/data/hr_salary_rule_data.xml#L484

sentry-7496380516

Problem:
Undo does not work correctly after inserting a table of content when no paragraph follows it.

Cause:
`SelectionPlaceholderPlugin.onSelectionChange` clears attributes from the next base container and adds a history step whenever the selection changes.

In the table of content case, this creates a loop:
- Attributes are cleared and a history step is added.
- Undo restores only the cleared attributes.
- The selection falls back into the empty paragraph after the table of content.
- `SelectionPlaceholderPlugin.onSelectionChange` runs again and adds another history step.

As a result, undo never reaches the previous user action.

Solution:
Avoid adding a history step in
`SelectionPlaceholderPlugin.onSelectionChange` when the current step is
not modified by any user interaction.

Steps to reproduce:
- Write some text.
- Insert a table of content using `/toc`.
- Press Ctrl + Z.
- Observe that nothing happens and the previously typed text cannot be undone.

task-6216910

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

Forward-Port-Of: odoo/odoo#264743

**Issue**:
When the name of a MO changes before confirmation, the picking origin
may remain incorrect after confirmation.

**Steps to reproduce**:
- Make sure that multi-step route is enabled in the settings
- Configure the manufacturing route as 2-step
- Go to Inventory > Configuration > Warehouse Management > Operations Types
- Clone the "Manufacturing" operation type and assign a different Sequence Prefix
- Create and save a MO, without confirming it
- Change and save the operation type to the cloned one (the MO name changes)
- Confirm the MO
-> The picking source document uses the previous MO name instead of the new one

**Cause**:
The source document of the picking (`origin`) comes from its move:
https://github.com/odoo/odoo/blob/95c73aa4dd7433f394799fdaaad57a84d750ec5a/addons/stock/models/stock_move.py#L1526
The move origin comes from the procurement values:
https://github.com/odoo/odoo/blob/95c73aa4dd7433f394799fdaaad57a84d750ec5a/addons/stock/models/stock_move.py#L1575C13-L1575C56
Which relies on `self.reference_ids[0].name`:
https://github.com/odoo/odoo/blob/95c73aa4dd7433f394799fdaaad57a84d750ec5a/addons/stock/models/stock_move.py#L1639
which is never updated, causing the origin to keep the previous MO name.

opw-5979778

Forward-Port-Of: odoo/odoo#255874

Issue
-----
Forced full packaging reservation setting is ignored when there is a big quant in stock.

Steps to reproduce
-----
- Enable packagings
- Create a product category "Super Category"
    - Reserve Packagings: Reserve Only Full Packagings
- Create a stored product "AAA"
    - Product Category: Super Category
    - 50 units on hand
    - Packaging: 6-Pack (6 units)
- Create a delivery for 15 units of AAA

> Reservation is made for 15 units

Cause
-----
The rounding to a multiple of the packaging quantity takes the stock quant into account. For our example case, we have 8 full 6-Packs on hand, so the `available_quantity` gets set to 48 when doing

https://github.com/odoo/odoo/blob/5e458236ca2ff2ab92c4893495e7a721be902c40/addons/stock/models/stock_quant.py#L923-L925

This leads to the reservation quantity being min(15, 48) = 15

https://github.com/odoo/odoo/blob/5e458236ca2ff2ab92c4893495e7a721be902c40/addons/stock/models/stock_quant.py#L927

-----
Ticket:
opw-5974333

Forward-Port-Of: odoo/odoo#263934
Forward-Port-Of: odoo/odoo#257342

The employment subsidy calculation was incorrect in two main scenarios:

### 1. Incorrect threshold prorating:

The system was comparing the salary against the full monthly limit even for partial periods (weekly or bi-weekly). This resulted in employees wrongly receiving the subsidy when their proportional salary actually exceeded the limit.

Example: In 2026, the 14-day threshold should be 5,292.67 (11,492.66 / 30.4 * 14). Currently, an employee earning 10,000.00 in those 14 days still gets the subsidy because it's being compared against the full 11,492.66.

### 2. Cumulative monthly cap:

When multiple payslips occur in the same month, the total subsidy sometimes exceeds the statutory monthly maximum (536.22 for 2026) because the cap wasn't enforced across all slips.

Example: The 2026 maximum monthly subsidy is 536.22. In a month with three partial payslips:
- Mar 1st - Mar 14th: The system grants 246.68.
- Mar 15th - Mar 28th: The system grants 246.68.
- Mar 29th - Apr 11th: For the 3 days belonging to March, the system grants an additional 52.86.
Total subsidy for March reaches 546.22, exceeding the legal cap.

### Changes included in this PR:
- Updated `l10n_mx_rule_parameter_uma` to include monthly and annual values. This prevents rounding discrepancies.
  Example: the 2026 annual UMA published is 42,794.64. In a rule the calculation is: l10n_mx_uma * 30.4 * 12 = 117.31 * 30.4 * 12 = 42,794.68 resulting in a ~0.04 difference.

- Create a new rule parameter `l10n_mx_rule_parameter_subsidy_salary_limit` to have the subsidy eligible threshold. Starting in 2026, the government's rounding changed from zero decimals(e.g., 9,081.00 in 2024, 10,171.00 in 2025) to two decimals (11,492.66). Storing these as explicit parameters avoids the precision errors.

- Added comprehensive unit tests covering:
  - Complete periods: validates standard payslips aligned with the month calendar (bi-monthly, monthly, bi-weekly).
  - Overlapping periods: validates split-month scenarios (14-day, 10-day, weekly) where periods cross month boundaries:
    Example of self._overlapping_period("weekly", 7, 2646.33, (35.24, 88.10), (3, 123.34), (77.53, 35.24))
    This test covers 5 weekly payslips with the following subsidy
    distribution:
    - Tuple `first_payslip` => (35.24, 88.10) means that:
      First payslip (Apr 29 - May 5), the subsidy is 35.24 for April and 88.10 for May.
    - Tuple `mid_payslips` => (3, 123.34) means that:
      For the next 3 payslips fully in May, the subsidy is 123.34 each.
      Payslip 2 (May 6 - May 12): Subsidy for May = 123.34
      Payslip 3 (May 13 - May 19): Subsidy for May = 123.34
      Payslip 4 (May 20 - May 26): Subsidy for May = 123.34
    - Tuple `last_payslip` => (77.53, 35.24) means that:
      Last payslip (May 27 - June 2), the subsidy is 77.53 for May and 35.24 for June.
  - Across years: subsidy amounts and limits are updated annually.
    Therefore, if a period overlaps two years, a salary amount might be eligible for a subsidy in January but not in the previous December, and the paid subsidy is increased in January due to the new limits.

- Cleaned up redundant tests (test_regular_payslip_subsidy) and adjusted decimal precision.

- For split-month `schedule_pay` periods, the first payslip might generate a subsidy. However, in subsequent payslips, due to commissions or a wage increase, the employee may exceed the monthly subsidy salary limit.
  In those payslips, a warning is shown to notify the user that a manual adjustment is required.
  Created tests to validate these cases.

target: 19.0
task-5419659

Forward-Port-Of: odoo/enterprise#116779
Forward-Port-Of: odoo/enterprise#107601

Issue before this commit:
=========================
For products tracked by serial/lot with packaging units, the delivery slip correctly shows the packaging unit and quantity before validating the transfer. However, after validating the transfer, the packaging unit and its corresponding quantity are no longer displayed in the delivery slip report.

Steps to Reproduce:
=========================
1. Install stock and sale_management modules.
2. Enable Units of Measure & Packagings and Display Lots & Serial Numbers on Delivery Slips from settings.
3. Create a product with tracking by lot/serial number and configure a packaging unit.
4. Create a SO using this product with a packaging unit and confirm it.
5. Open the related transfer and print the delivery slip before and after validation.

Cause of the Issue:
=========================
The delivery slip report template (stock_report_delivery_has_serial_move_line) does not display packaging unit information after validation for move lines when the packaging unit differs from the product unit of measure.

With This Commit:
=========================
This commit ensures that packaging units and their corresponding quantities are displayed on the delivery slip after validation when the packaging unit differs from the product unit of measure.

Steps To Reporduce: [Video Link](https://drive.google.com/file/d/10DmFKW1Y_Tm-AyKzPrqtFMY8orBkKbIm/view?usp=sharing)

opw-6142052

Forward-Port-Of: odoo/odoo#266028
Forward-Port-Of: odoo/odoo#265206

### Issue before this commit:
When importing an Italian electronic vendor bill using the AssoSoftware standard, pension fund taxes (Cassa Previdenziale) are not applied to the invoice lines.

### Steps to reproduce the issue:
1. Download Accounting and l10n_it_edi_witholding
2. Change VAT number of IT company with the one in the xml
3. Go to Taxes > 4%F.Pens. > Advanced Options and change Pension Fund Type with TC02
4. Import xml of the ticket in vendor bills
5. P.Fund tax is not assigned

### Cause of the issue:
The issue is caused by the following line:
https://github.com/odoo/odoo/blob/669b9b84f4d5c8765dc4b451d5da6a95dbb9ded8/addons/l10n_it_edi_withholding/models/account_move.py#L247 Currently, the parser strictly expects the optional <RiferimentoTesto> tag alongside <TipoDato>AswCassPre</TipoDato>. However, several third-party software providers generate valid XML files containing only the  AswCassPre block without any optional child tags.

### Reason to introduce the fix:
Ensure that the pension fund tax mapped to the line's VAT rate is correctly applied whenever the AswCassPre data type is present, even if the optional reference tags are omitted.

opw-6189225

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

Forward-Port-Of: odoo/odoo#265914
Forward-Port-Of: odoo/odoo#264083

# The bug

When accessing a done transfer with two lines where one line has a result package ID and the other does not, the computed field `has_lines_without_result_package` returns `True`. This field is used in the `stock_package_m2m` widget to append a `No package` tag when a move has this field set.

https://github.com/odoo/odoo/blob/eaa6c4352aec2be8519360c282f3f6504a2f263c/addons/stock/models/stock_move.py#L266-L269

https://github.com/odoo/odoo/blob/eaa6c4352aec2be8519360c282f3f6504a2f263c/addons/stock/static/src/widgets/stock_package_m2m.js#L9-L24

This works fine when package history exists, as it accesses the `package_ids` field to generate the tags. However, for recently upgraded databases, no package history is available. When the `_compute_package_ids` method runs, it attempts to access data from an undefined history record, triggering a traceback.

https://github.com/odoo/odoo/blob/eaa6c4352aec2be8519360c282f3f6504a2f263c/addons/stock/models/stock_move.py#L271-L278

# The fix

The fix is straightfoward: in `_compute_package_ids`, if a move is in the `done` or `cancel` state and has no package history, we fallback and populate `package_ids` using the same logic applied to states other than `done` or `cancel`. This behavior specifically targets and fixes issue for databases recently upgraded to v19.

task: 6070541


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

Forward-Port-Of: odoo/odoo#265032

[FIX] hr_attendance_gantt: fix gantt view with public holidays

Bug reproduction:
    1 - Select flex schedule employee (or change its schedule to 40h flex one) and make its contract before 01/01/2026
    2 - Create a new public holiday on 01/01/2026 (from 00.00 to 23.59 or 23.55 (depends on version, it does not matter))
    3 - in attendance app the cell from 00.00 to 01.00 seems white for that day and for selected employee (this cell seems like not holiday and employee can work)

Bug cause:
    1 - After a long traceback, _gantt_unavailability in hr_attendance_gantt/HrAttendance, if an employee is flexible then unavailable_intervals is calculated with the Brussel time zone
    2 - All other unavailable intervals are converted to the UTC in the function of _gantt_unavailability except in the final lines of the function.
    3 - When the employee is flexible and since the conversion is not done in the final lines, it remains 1 hour more (UTC+1), it is from 1 am to 1 am of next day instead of 0 am to 23.59.

Bug solution:
    1 - I converted the timezone to UTC to solve the problem.

task - 6067070

Forward-Port-Of: odoo/enterprise#112493

### Description of the issue/feature this PR addresses:

Some stock queries determine whether a location belongs to the subtree of a set of locations by checking the parent_path prefix against candidate parent locations. This is done using a correlated EXISTS subquery with a LIKE parent.parent_path || '%' condition.

When the list of candidate locations becomes large (for example tens or hundreds of thousands of ids), this approach causes extremely poor performance because the database must repeatedly compare hierarchical path strings for every candidate row.

This PR improves the performance of this ancestry check by replacing the string prefix comparison with a direct check on the ancestor ids contained in parent_path.

### Current behavior before PR:

The query determines whether a location belongs to the subtree of one of the provided locations using:

location.parent_path LIKE parent.parent_path || '%'

For each row, PostgreSQL must evaluate a correlated subquery against all candidate parent locations. Because this relies on string prefix comparisons on parent_path, when the location list is large, this results in extremely slow queries.

### Desired behavior after PR is merged:

Instead of performing string prefix comparisons, the query extracts the ancestor ids directly from parent_path.

The path is:

1. Trimmed to remove leading and trailing /
2. Split into an array of ancestor ids
3. Expanded using unnest
4. Checked for intersection with the provided location ids

This converts the ancestry check from repeated string comparisons into a simple integer membership check.

### Benchmarks

Comparing performance of old subquery:

```
SELECT stock_location_inner.id 
FROM stock_location AS stock_location_inner 
WHERE EXISTS (
    SELECT 1
    FROM stock_location parent
	WHERE parent.id IN (long list)
		AND stock_location_inner.parent_path LIKE parent.parent_path || '%%'
);
```

to new one:

```
SELECT stock_location_inner.id 
FROM stock_location AS stock_location_inner 
WHERE EXISTS (
    SELECT 1
    FROM unnest(
        string_to_array(trim(both '/' FROM stock_location_inner.parent_path), '/')::int[]
    ) AS path_id(id)
	WHERE path_id.id IN (long list)
);
```

Depending on the number of elements in 'long list'

| # of elements | Before | After |
| --- |---|---|
| 130,000 | 21min | 0.8sec |  
| 10,000 | 95sec | 0.5sec |  
| 1,000 | 10.5sec | 0.5sec |  

In practice, on the reference ticket this causes the "Validate" button on a stock picking to go from timing out to taking 8 seconds. 

### Reference

opw-5932436

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

Forward-Port-Of: odoo/odoo#255399
Forward-Port-Of: odoo/odoo#254245

**Steps to reproduce:**
1. Create user with role: User. Sales: Administrator, Accounting: Administrator and Documents: System Administrator.
2. Create a SO, create invoice, confirm, and send.
3. Now go back to the SO, and try to link the INV document to the chatter.

**Cause:**
When linking an existing document to the composer, the underlying attachment is copied. If the source attachment is bound to a specific field (e.g., `res_field = 'invoice_pdf_report_file'`), the `copy()` operation duplicates this field reference. Odoo's native security checks then attempt to verify access to that specific field on the target model (`mail.compose.message`). Because the composer does not have this field, the check fails and throws an AccessError, even if the user has full rights to the source document.

**Solution:**
Explicitly set `"res_field": False` during the copy operation. This strips the original field binding, cleanly converting the file into a standard, generic chatter attachment for the composer without bypassing the standard security framework.

opw-5916364

Forward-Port-Of: odoo/enterprise#107723

On /shop, when a snippet block sits above the searchbar, the search dropdown was rendered partially hidden behind that block (cropped/unreadable items).

Steps to reproduce:
===================
1. Go to /shop.
2. Add a snippet block above the searchbar.
3. Type in the searchbar.

=> Observed: search results appear cropped, with upper items
   hidden behind the snippet block above.

Root cause:
===========
the products grid column (`#products_grid`) has `overflow: auto`, https://github.com/odoo/odoo/blob/d9bb1c1dc90f97b63b87ad762fc4ab36abf7e05f/addons/website_sale/static/src/scss/website_sale.scss#L442 which clips any absolutely-positioned descendant
that extends past its bounds. The dropdown's containing block is the searchbar `<form>` (position: relative), which lives inside that column. When the dropdown grew (or flipped to dropup) and extended outside the column, the part outside was clipped, and any positioned snippet siblings above the column painted over the clipped area.

Fix:
======
while the dropdown is mounted, lift the `overflow: auto` on its ancestor `div.col` so the menu can extend past the column and paint on top of other content. Done from JS so no SCSS rule has to target the searchbar-specific column.

opw-6216317

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

Forward-Port-Of: odoo/odoo#265982
Forward-Port-Of: odoo/odoo#265558

It is possible to order as many products as we want of a rental product
that synchronizes shifts depending on a specific service

Steps to reproduce:
1. Install website_sale_renting_planning module
2. Go to Rental > Products and create a new product "test" with Sales enabled, Product Type "Service", Plan Services enabled as "Developer", in the Sales tab, enable Is Published and in the Rental prices tab, create a pricing for Daily period
3. In the General Information tab, click on the internal link to "Developer"
4. Enable Sync Shifts and Rental Orders
5. Go to the eCommerce website and search for product "test"
6. You can add as many quantity of the product to your cart

Issue:
We don't limit the maximum quantity of the product

Solution:
Look through the renting availabilities of the product and set the
maximum quantity to the minimum of the availabilities relevant to the
renting dates selected

opw-6009928

Forward-Port-Of: odoo/enterprise#113525
Forward-Port-Of: odoo/enterprise#111793

**Problem:**
Clicking the timesheet timer in the systray while viewing a subtask raises a MissingError, preventing the user from starting the timer.

**Steps to reproduce:**
1. Open a project task that has subtasks
2. Click the subtasks smart button and open a subtask
3. Click the timesheet timer icon in the systray
4. Observe the "missing record" error

**Current behavior:**
The systray RPC fails with MissingError.

**Expected behavior:**
The timer opens, with the subtask's project and task prefilled.

**Cause of the issue:**
The frontend sends `currentState.active_id` regardless of the model it refers to. When the user navigates to a subtask via the smart button, `active_id` is the parent task's id. The systray controller then does `request.env['project.project'].browse(active_id)`, which returns a truthy recordset for an id that does not exist in `project_project`. Reading `.allow_timesheets` on that recordset triggers the DB fetch and raises MissingError. The same pattern exists in the `helpdesk_timesheet` override for both the `project.task` and `helpdesk.ticket` branches.

**Fix:**
Adding `.exists()` after the browse validates the recordset before any field access, so a stray `active_id` falls through to the task's own project (or ticket's project) instead of crashing.

opw-6169548

Forward-Port-Of: odoo/enterprise#117039

Problem:
In email templates, adding a banner/info block and saving then reloading causes the horizontal padding to be lost and the icon to become misaligned.

Cause:
During save, `convert_inline` processes the content via `_normalizeStyle`, which iterates over `CSSStyleDeclaration` using index-based iteration. This only yields longhand properties (e.g. `padding-left`, `padding-top`), never shorthands like `padding`. When the shorthand contains `var()` references (e.g. `padding: var(--y) var(--x)`), the browser cannot resolve the longhands and leaves them empty, so they are silently dropped during style extraction. Adding shorthand support to the iterator was not viable, as the rest of the pipeline expects longhand-only styles, and safely converting `padding: var(--y) var(--x)` to longhands is not possible without first resolving the variables.

Solution:
Replace the `padding` shorthand in the banner template with explicit longhand properties (`padding-top`, `padding-bottom`, `padding-left`, `padding-right`).

Steps to reproduce:
1. Open an email template
2. Add a banner/info block
3. Save the template
4. Reload the page
5. Observe horizontal padding is lost and icon is misaligned

task-6230530

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

Forward-Port-Of: odoo/odoo#265756

### Issue:

The `Print Label`, `Register Production`, `Register By-products`and `Register Consumed Materials` are all available in the test types at control point creation.

### Expected behavior:

These test types are only meant for manufacturing operations and are supposed to be hidden by the field domain:
https://github.com/odoo/enterprise/blob/f56aa85b4ad32c5d9ad5593df1366d72e88da0e4/mrp_workorder/models/quality.py#L102-L104 https://github.com/odoo/enterprise/blob/00d6cccd75c402378698a6fd11ee2692f2361c7f/mrp_workorder/models/quality.py#L20-L24

### Cause of the issue:

Since saas-18.1: 5ef007a2116e528b796ebe80fb291ba5f1a94c8f domains are optimised into equivalents SQL clause with better sql performances. This optimization results in the following match for boolean fields: 
`('field', '=', True)` -> `('field', 'in', OrderedSet([True]))` 
`('field', '=', False)` -> `('field', ' not in', OrderedSet([True]))`
Because of these:
https://github.com/odoo/odoo/blob/82b16e8feb3d60a9a3855e3adb3164ae5a6c041d/odoo/orm/domains.py#L1058-L1079 https://github.com/odoo/odoo/blob/82b16e8feb3d60a9a3855e3adb3164ae5a6c041d/odoo/orm/domains.py#L1215-L1236
Now the issue is that the specific `search_method` of the `allow_registration` field is then called with this optimized domain: https://github.com/odoo/odoo/blob/82b16e8feb3d60a9a3855e3adb3164ae5a6c041d/odoo/orm/domains.py#L860-L866 https://github.com/odoo/enterprise/blob/00d6cccd75c402378698a6fd11ee2692f2361c7f/mrp_workorder/models/quality.py#L20-L24
And since `value` is defined as a non empty ordered set in both cases it the search method returns a True leaf as search domain.

opw-5915197

Forward-Port-Of: odoo/enterprise#118073
Forward-Port-Of: odoo/enterprise#117068

__
## Short functional explanation of the error
When setting attendances on several consecutive days for a flexible employee, with an overtime ruleset containing a single rule. This rule being based on week and quantity.
When regenerating overtimes for this ruleset, the overtime hours generated isn't correct.

## Reproduction Steps
1. Create an employee. In the Payroll tab, set a start date for their contract. Set Work Entry Source as Attendances. Set their Working Hours as a flexible schedule. Set their weekly hours at 40.
2. Create an Overtime Ruleset. Add a single rule, based on Quantity, if the worked hours on a `Week` differs `from the amount defined on the contract`. Check Pay Extra Hours and leave the Work Entry Type to use as Overtime Hours.
3. Go back to the employee. In Settings, set the Overtime Ruleset field as the new Overtime Ruleset you just created.
4. Create 5 attendances, each from 8 am to 6 pm, from Monday to Friday.
5. Go to the Overtime Ruleset you just created and click on Regenerate Overtimes.
6. Go back to Attendances. Search for your employee, and click on the list view.

### Expected behavior
The employee's schedule is 40 hours per week. They worked 50 hours. 10 hours should be considered as Worked Extra Hours.

### Unexpected behavior
18 hours are considered as extra hours.

## Origin of the issue
To compute the expected duration of the day, we run: https://github.com/odoo/odoo/blob/7fc5edc29f854d619dbcb5fcc3503fb18ca05335/addons/hr_attendance/models/hr_attendance_overtime_rule.py#L303-L304 where `schedule['work']` will contain intervals on 5 consecutive days, from 8 am to 4 pm.
However, the last day of the employee's attendances isn't contained in these intervals. As a result, `period_schedule` will contain 4 days (the common days between the employee's Attendance days and `schedule['work']` ) and thus, `expected_duration` will be set at 36 hours instead of 40.
In the case where overtimes are computed based on hours from the 
contract, for flexible employees, the expected hours are the ones
indicated on their schedule.

__
opw-6131543

Forward-Port-Of: odoo/odoo#265120
Forward-Port-Of: odoo/odoo#263335

## [FIX] core: lxml compatibility v2.14.0+ (HTML parsing)

In version 2.14.0, libxml2 fixed a long standing quirk in its HTML
handling where it always implies `<p>` start tags [1].

As a result, there is a difference in behavior between pre and post
2.14.0 produced HTML when no start tag is provided:
- pre: always has a `<p>` tag
- post: depending on the case, could have either a `<span>` or `<p>` tag.

This commit introduces a monkeypatch of the lxml's HTML parser when
built with libxml2 2.14.0+ to maintain a similar behavior with older
versions.

[1]: https://gitlab.gnome.org/GNOME/libxml2/-/commit/8cf6129bbd836e666e7eda8c9e61c00387ae388b


## [FIX] base,l10n_it_edi: catch TypeError/ValueError for lxml 6+ compat

Updates exception handling to account for stricter type checking
introduced in lxml 5/6 and libxml2 2.12+.

Note: Ubuntu 26.04 (Resolute) provides lxml 6.9.2/libxml2 2.15 while
Debian Trixie has lxml 5.4.0/libxml2 2.9.14. Don't be fooled by the
version `2.12.7+dfsg+really2.9.14-2.1+deb13u1` which actually means that
Debian has reverted/held back the core engine to 2.9.14 while adding
commits from 2.12.7.

Forward-Port-Of: odoo/odoo#259348

### **Description of the issue/feature this PR addresses:**

**Issue:** In Odoo 18/19, purchase move lines default to the WH's main stock location (`lot_stock_id`) as the `location_final_id`. However, when a user configures a sub-location on the Receipt Operation Type, the picking destination is correct, but the move lines are defaulted to the main warehouse. This mismatch causes the Forecasted Quantity to not increment for the intended sub-location

**Solution:** Prioritize the `default_location_dest_id` before falling back to the default stock location

opw-6032018

### **Current behavior before PR:** 
When confirming a PO, the `location_final_id` on stock moves defaults to the `lot_stock_id`, regardless of the specific destination set on the Operation Type. This causes a mismatch in 1-step receiving flows where a sub-location (e.g., WH/Stock/Test) is intended, since the move lines revert to the root warehouse location (WH/Stock). Thus, the forecasted quantity for the specific sub-location doesn't increment as expected.

### **Desired behavior after PR is merged:** 
The `_get_final_location_record` method will now evaluate if the Operation Type's `default_location_dest_id` is a child of the warehouse's main stock. If it is, the sub-location is used as the `location_final_id` for the moves and move lines. This ensures that the forecasted quantity reflects the intended destination upon PO confirmation while still maintaining the fallback to the warehouse root for standard multi-step routes.




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

Forward-Port-Of: odoo/odoo#259233
Forward-Port-Of: odoo/odoo#254527