The POS Avatax tour sporadically failed after returning from the payment screen. The race was reproduced locally by delaying the mocked Avatax response in mocked_request().

There's three somewhat related fixes.

Firstly, get_order_tax_details() calls sync_from_ui(), which emits a SYNCHRONISATION notification. Unlike the normal POS sync path, the Avatax RPC did not pass the device context. The browser therefore treated its own notification as coming from another device and started an independent reload of open orders.

That reload could replace the current order state after the tour returned to the product screen, causing the selected order line to disappear.

We now pass the normal sync context so the browser can properly ignore its own notification.

Secondly, we'll keep the complete sync_from_ui() response and replace its order, line, tax, and tax group data after the AvaTax calculation. We then simplify the processing client-side by moving towards the established pattern in the POS: missingRecursive() to load any other referenced records, and then pass that through loadConnectedData().

Lastly, clickPayButton() only waits for the payment screen element to be displayed. The AvaTax request starts from the screen's onMounted() callback, leaving a short window where the screen and its buttons exist but the request and UI blocker have not started yet. The next tour step can probably run during that window. To make sure this can't happen we explicitly waitRequest(). This first waits for requests to appear and then waits for them to complete.

runbot-error-944281

Forward-Port-Of: odoo/enterprise#125944

Bug
===
When notifying by email a tracking change, the arrow and parentheses are not rendered in the email body.

Technical and Constraints
=========================
The class `o_track` is only set in the web client template (`mail.Message`).

There's no class in the body of the email that is sent. It can be rendered with "notification templates" that we cannot change either (and they just do `t-out="message.body"`, so the body field of the mail message has to be properly rendered).

We also need existing mail messages to be rendered correctly, and so we need a way to differentiate mail messages created before the fix from those created after it, to know when to disable the arrow and parentheses.

Alternatives
============
We have tough about many solutions, this one is the best we found based on the constraints we have

1. Add a class in 19.3, use that class to not remove the arrow on previous mail message. That solution required a migration script that will change all tracking messages. Because the initial migration of the tracking was really slow, we wanted to avoid that.
2. Add a class, and keep it forever. But that solution makes the body of the mail messages bigger, which defeat one of the purpose of the initial refactoring
3. Change the outgoing email without changing the body of the mail message. That solution was really not reliable (regex change to add the arrow, and we have no clean way to target the tracking rows)
4. During the migration create a system parameter with the date, and compare with the create_date of the mail message to know if we should add the arrows or not (but we will need to keep that system parameter forever, and the code to support both to)

Task-6424104

Forward-Port-Of: odoo/odoo#282210

**Problem:**
Duplicating a task, creating a task from a task template, or generating the next occurrence of a recurring task scrambles the dependencies between its sub-tasks: each copied sub-task carries the dependencies of a different sub-task instead of its own.

**Steps to reproduce:**
1. Enable Task Dependencies on a project.
2. Create a task with three sub-tasks and chain them: the second depends on the first, the third depends on the second.
3. Duplicate the task, or use "Create from template" if the task is a template.
4. Open the sub-tasks of the new task and look at their dependencies.

**Current behavior:**
The dependencies of the copied sub-tasks are shifted: the chain runs in the reverse order of the original one.

**Expected behavior:**
Each copied sub-task depends on the copy of the sub-task its original depended on, so the new task reproduces the original chain.

**Cause of the issue:**
`_create_task_mapping` builds the original to copy mapping by pairing `original_task.child_ids` with `copied_task.child_ids` positionally, on the assumption stated in its docstring that both recordsets share the same index order. They do not. `project.task._order` ends with `id desc`, so `child_ids` is read newest-first, while the copies are created by iterating the original `child_ids` in that same order. The copies' ids therefore ascend along the original list, and reading them back through `child_ids` returns them in the exact reverse order. `zip` then pairs each original with the copy of the sub-task at the mirrored position, and `_resolve_copied_dependencies` writes every `depend_on_ids` and `dependent_ids` onto the wrong copy. This affects every caller of that method: `copy`, the task template action, and the creation of the next occurrences of a recurring task.

**Fix:**
Sorting the copied children by id restores the correspondence because id order is the order in which the copies were created from the original list, an invariant that holds whatever `_order` does, whereas the previous code silently depended on `_order` producing the same sequence on both sides.

`test_duplicate_project_with_subtask_dependencies` and `test_recurrence_copy_task_dependency` were reading the copies by `child_ids` index too, which the mirrored mapping happened to satisfy, so they passed on a wrong result; they now index them in creation order as well.

opw-6386578

Forward-Port-Of: odoo/odoo#284211
Forward-Port-Of: odoo/odoo#280893

### Steps to Reproduce:
1. Create a sign request and have 3 total signers (User, Customer, Employee)
2. Enable Signing Order and make the order as follows: (1) User, (2) Customer, (3) Employee
But make the User and Employee the same contact
3. Send and sign the request > Notice that (1) is able to sign for (3) immediately after, (2) has not signed yet.

### Description of the issue/feature this PR addresses: 
**Issue:** The signing order is ignored when the same user has to sign multiple times on a document, even if it is configured for a different person to sign in between. This happens because all signature request items are initialized in the 'sent' state upon creation, rather than strictly advancing based on the order. As a result, the system prematurely allows users to sign out of order and prompts them with their next turn too early.

**Solution:** To resolve this,  the controller was updated to include an `is_mail_sent = True` domain filter. This ensures that the UI's post-sign popup only displays documents where it is explicitly the user's active turn, rather than prompting a premature sign.

### Current behavior before PR: 
Users are able to sign prematurely, and the system will disregard the configured signing order.

### Desired behavior after PR: 
Users will only be prompted and able to sign a document when it is explicitly their turn, per the `mail_sent_order`. This way, documents are signed in order.

opw-6417327

Forward-Port-Of: odoo/enterprise#128487
Forward-Port-Of: odoo/enterprise#125573

Currently, an error occurs when the user removes the payslip period.

**Steps to Reproduce:**
 - Install the `hr_payroll` module.
 - Go to `Employees` and create an `employee`, or use an existing one.
 - Make sure the `employee's version` has a `contract start date`.
 - Go to `Payroll` > `Payslips` > `Payslips` and create a payslip.
 - Select that `employee` on the payslip, then remove the `payslip start date`.

`TypeError: '<=' not supported between instances of 'datetime.date' and 'bool'`

After the [recent commit], which computes the version from the payslip period without allowing it to 
be overridden, when the user selects an employee whose version has a contract start date, it 
checks whether the version overlaps with the payslip period [1]. However, since the payslip 
dates have not yet been set, it raises an error.

This commit ensures that the check for the version overlapping with the payslip period is 
skipped if the payslip does not have both a start and end date. It also makes the method 
depend on date_to, because if the user changes the payslip end date, it should recheck 
whether the version overlaps with the payslip period.

[recent commit]: https://github.com/odoo/enterprise/commit/569ce2af32477d410b79e98ca72729385042619b

[1]- https://github.com/odoo/enterprise/blob/8a66d7beabe6f9b28000ef12725f3a9937d5d1ee/hr_payroll/models/hr_payslip.py#L1716-L1725

sentry-7632216317

Issue:
---
Authorize payment tokenization doesn't work.

Steps:
1- Setup authorize payment provider.
2- Using portal page, add a new payment method for the user.

The created payment method is not saved.

Cause:
---
The issue was introduced in efc2788dfccd13ee6feb309430ff57e49664ff97. Before that, we were calling `_tokenize` before voiding the tx. In that PR, the `_tokenize` call was moved to `_process()`, after `_apply_updates()`.
So now what happens is that we void the tx, then call `_tokenize()`. Inside tokenize we try to create a customer profile, which fails because the tx is already voided.

Fix:
---
We can fix it by calling `_tokenize()` once before voiding the tx.

The redundant tokenize call inside the general payment tx `_process` is rendered ineffective by two safeguards:
1- There is a check for `tx.tokenize`, which neutralizes double tokenization:

https://github.com/odoo/odoo/blob/fffd987cc98d1ea0cd04e24dda2ed8b64a219cdc/addons/payment/models/payment_transaction.py#L754-L755 https://github.com/odoo/odoo/blob/fffd987cc98d1ea0cd04e24dda2ed8b64a219cdc/addons/payment/models/payment_transaction.py#L893-L896

2- If `token_id` is already set, no token value is returned:

https://github.com/odoo/odoo/blob/fffd987cc98d1ea0cd04e24dda2ed8b64a219cdc/addons/payment_authorize/models/payment_transaction.py#L237-L243

opw-6426847

Forward-Port-Of: odoo/odoo#283134
Forward-Port-Of: odoo/odoo#281014

When paying with an eWallet or gift card in POS, the reward line could end up 0.01 short of the actual card balance if the discount product's tax is configured as tax-excluded through a per-tax override, regardless of the tax's own default configuration. The card is still debited for the full balance, but the order is only discounted by one cent less, so the amount charged to the customer no longer matches the amount consumed from the card.

Steps to reproduce:
-------------------
* Top up an eWallet (or gift card) with a balance of 10.00
* On the eWallet/gift card program's discount product, set an 18% tax whose Tax Computation is overridden to "Excluded" (price_include_override = tax_excluded), independently of the company's default tax configuration
* In POS, add a product to an order and pay (partly) with that eWallet/gift card

> Observation:
Only 9.99 is deducted from the order total, while the backend correctly shows 10 consumed on the wallet/gift card.

Why the fix:
------------
The reward line's price_unit was reconstructed from a one-time backward tax computation, then kept only the tax amount for taxes whose price_include field was true, dropping it for any tax forced excluded. That price_unit was later re-taxed forward using the tax's real (excluded) configuration, and the two roundings don't agree for rates like 18%, losing a cent.

We now force special_mode "total_included" whenever an eWallet/gift card reward line's taxes are computed, not just at creation, so its tax-included total always equals the exact redeemed amount regardless of how the tax is configured, and store price_unit as that target amount directly.

opw-5819389

Forward-Port-Of: odoo/odoo#284397
Forward-Port-Of: odoo/odoo#278568

Step to reproduce:
 - create a faulty token that won't work and link it to a subscription
 - launch the recurring invoice cron
 - the following traceback occurs

```
last_tx_sudo = (self.transaction_ids - existing_transactions).sudo()
```

When the payment fails, the system rollback and we store the last_tx_sudo value in a dedicated variable. After rollback, the record does not exists anymore. Therefore, accessing the value fails.

```
  File "/home/odoo/src/enterprise/saas-18.2/sale_subscription/models/sale_order.py", line 1703, in _handle_automatic_invoices
    if not last_tx_sudo or last_tx_sudo.renewal_state in ['pending', 'authorized']:
                           ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/odoo/src/odoo/saas-18.2/odoo/orm/fields.py", line 1439, in __get__
    self.compute_value(record)
  File "/home/odoo/src/odoo/saas-18.2/odoo/orm/fields.py", line 1603, in compute_value
    records._compute_field_value(self)
  File "/home/odoo/src/odoo/saas-18.2/odoo/orm/models.py", line 4575, in _compute_field_value
    determine(field.compute, self)
  File "/home/odoo/src/odoo/saas-18.2/odoo/orm/fields.py", line 69, in determine
    return needle(*args)
           ^^^^^^^^^^^^^
  File "/home/odoo/src/enterprise/saas-18.2/sale_subscription/models/payment_transaction.py", line 25, in _compute_renewal_state
    if tx.state in ['draft', 'pending']:
       ^^^^^^^^
  File "/home/odoo/src/odoo/saas-18.2/odoo/orm/fields.py", line 1406, in __get__
    raise MissingError("\n".join([
odoo.exceptions.MissingError: Record does not exist or has been deleted.
```

Moreover, since https://github.com/odoo/enterprise/pull/45236/files#diff-c36fd7952cc2bef40716419a668de41963d49e1aa4177d9319d503fc260da588R1678-R1682

```
if not last_tx_sudo or not last_tx_sudo.renewal_state not in ['pending', 'authorized']:
```

has become

```
if not last_tx_sudo or last_tx_sudo.renewal_state in ['pending', 'authorized']:
```

But it feels strange to unlink the invoice when the payment succeed.


This PR fixes it.

Forward-Port-Of: odoo/enterprise#83913

**Steps to reproduce:**
1. Open a Sale Order report in Studio
2. Delete the Taxes column
3. Save
4. Create a sale order with at least one section line and products that have taxes
5. Print the sale order PDF

**Issue:**
- A blank column is rendered in the PDF report on section (and combo) rows whenever a column such as Taxes or Discount is removed via Studio.

**Why this happens:**
- The section row's `colspan` and the combo row's `colspan` were computed using `3 + (1 if display_discount else 0) + (1 if display_taxes else 0)`.
- `display_taxes` and `display_discount` are derived from order data (i.e. whether any line has taxes/discounts), not from which columns are actually rendered in the table.
- When Studio removes a column it deletes the `<th>` and matching `<td>` elements via XPath, but these Python variables remain `True`. As a result, section/combo rows still accounted for the removed column in their `colspan`, producing one extra cell and a visible blank column.

**Fix:**
- Introduce a `colspan_count` variable  which is incremented  inside each `<th>` body
- Use that counter for `td_section_name` and `td_combo_name` instead of the previous formula.
- Because the increment occurs inside the `<th>` element, it is skipped whenever the element is not rendered, whether because `display_taxes`/`display_discount` is `False` or because Studio's XPath removed the element entirely.

opw-6433679

Forward-Port-Of: odoo/odoo#283657
Forward-Port-Of: odoo/odoo#280719

### Issue before this commit:
Before this commit, users were completely blocked from creating an invoice for a foreign partner (e.g., "Cliente del Exterior") if all exportation journals were archived or unavailable, as the system would immediately trigger a RedirectWarning error.

### Steps to reproduce the issue:
1. Download Accounting and l10n_ar
2. Go to contacts and create a new one with:
	1. Country as United States
	2. VAT number ex. 55000002126
	3. AFIP Responsibility Type as Cliente del Exterior
3. Go to Journals, filter for sales journals and archive:
	1. Electronic Exportation Invoice (FEX)
	2. Expo Sales Journal
4. Go to invoices and create a new one for the client you just created
5. As soon as you insert the client you will receive the error: You are trying to create an invoice for foreign partner but you don't have an exportation journal

### Cause of the issue:
https://github.com/odoo/odoo/blob/014d58e3204d17db6dcba3c8ab7d8ad35003300e/addons/l10n_ar/models/account_move.py#L186-L189 The _onchange_partner_journal method rigidly enforced the use of an exportation journal for foreign AFIP responsibility types (codes 8, 9, and 10). If the query failed to find an active export journal, the code intentionally threw a hard error instead of providing a fallback mechanism.

### Reason to introduce the fix:
This fix is introduced to prevent unnecessary workflow blocks. By catching the missing journal and defaulting the document type to "Invoice B" (code 6), the user can now successfully generate the invoice using a standard domestic sales journal without being forced to configure an exportation journal.

opw-6442501


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

Forward-Port-Of: odoo/odoo#282971

Steps to reproduce
---
1. Create a quality control point on the Receipts operation type with Control per set to Operation.
2. Confirm a receipt of 2 units of the product: one pending operation quality check is created.
3. In the Barcode app, receive 1 unit and go back to the transfer with the back button.
4. The pending operation quality check is gone.

Issue
---
Going back from the Barcode app calls `post_barcode_process`, which on a partial reception splits the picked move into a done move and a remaining move, then merges the transient duplicate back with `_merge_moves`. https://github.com/odoo/enterprise/blob/b89614661682ecbd131d940539aac8afdd9d7289/stock_barcode/models/stock_move.py#L57-L60

`_merge_moves` cancels that transient duplicate through `_action_cancel` before unlinking it. https://github.com/odoo/odoo/blob/8f3100ca597559945cc42d9ef9517edbb40a900b/addons/stock/models/stock_move.py#L1400-L1401

The `quality_control` override of `_action_cancel`, picks the pending checks to drop from `is_product_canceled`, a `defaultdict(lambda: True)` keyed by `(picking, product_id)`. An operation check has no `product_id`, so its key is never computed by the loop and reads back the `True` default, so it is deleted even though the transfer still has a live move. Since an operation check covers the whole transfer, it must be dropped only when every move of its picking is cancelled. https://github.com/odoo/enterprise/blob/b89614661682ecbd131d940539aac8afdd9d7289/quality_control/models/stock_move.py#L68-L76

opw-6439179

Forward-Port-Of: odoo/enterprise#129051
Forward-Port-Of: odoo/enterprise#127427

## Steps to reproduce:

- Create another warehouse
- Create a product with a variant, like Color, values black and white
- Track the product, add a qty on hand of 50 on the black product
- Go to the sales app, make a quotation of 50 for the black product
- Confirm the quotation
- Go to the PoS, click on the product, check the available qty in the popup
- It is still 50, even though the forecasted is correct at 0

## Why the fix:

Having the actual free qty was added in this commit 682bc82 to be able to check the qty that was really free instead of the available qty. This means that we subtract the reserved_qty from the qty_available to get the free_qty.

The variant popup was forgotten in this commit, so it was still displaying the qty_available. This is why there was a difference in the qty if we press the product normally or if we long press it, because the variant popup was forgotten in said commit.

opw-6382845

Forward-Port-Of: odoo/odoo#284321
Forward-Port-Of: odoo/odoo#280330

**Steps to reproduce:**

- Create a contact with a phone number that has 9 characters
- Setup an UPS carrier, a configuration that works is UPS Saver as Service Type and UPS Package/customer supplied as a package type
- Create a quotation, put the created contact as a client
- Try adding a shipping and getting the rates
- An User Error appears, the phone number is too short

**Why the fix:**

Before this commit, any phone number that was less than 10 characters would raise an User Error, but some countries, such as Luxembourg, use phone numbers that are nine characters long or even less.

If we check the official UPS documentation (https://developer.ups.com/tag/Shipping?loc=en_EN#operation/Shipment), we can see in the Ship_to/Phone section, that the phone number should be a number between 1 and 15, not saying it should be 10 characters or more.
<img width="495" height="473" alt="image" src="https://github.com/user-attachments/assets/fed82987-ffb8-4b84-b282-6c3d3b4f304e" />


After this commit, we adapt the way we prevent the user from inputing phone numbers to fit the official UPS documentation.

opw-6307577

Forward-Port-Of: odoo/enterprise#128632
Forward-Port-Of: odoo/enterprise#122831

Deleting a draft customer invoice linked to timesheets resets their timesheet_invoice_id so the hours become invoiceable again. This write also marks the timesheets' so_line for recompute, and the re-derivation runs while the lines are no longer protected by the invoice link. When the task or project no longer resolves to a sale order item (e.g. it was unlinked after invoicing), the timesheets lose their sale order item or get reassigned to another one, so the delivered hours silently disappear from the original order line.

Protect so_line during the write and drop the pending recompute: deleting an invoice must only make the hours invoiceable again, not change their allocation.

Steps to reproduce:
- Install Sales and Timesheets
- Create a service product with invoice policy "Based on Timesheets" and "Create a task in a new project"
- Create and confirm a sale order with this product
- Log a timesheet on the generated task
- Create the invoice (keep it in draft)
- Remove the Sales Order Item from the task and from the project settings (or point them to a sale order item of another order)
- Delete the draft invoice
- Open the timesheet: its Sales Order Item is emptied (or replaced by the other order's item, whose delivered quantity now includes the hours sold on the original order), and the original line's delivered quantity is reset

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

Forward-Port-Of: odoo/odoo#283831
Forward-Port-Of: odoo/odoo#279552

When having multiple companies selected at the same time, _get_conversion_rate returns:

File "/data/build/odoo/odoo/orm/fields_misc.py", line 114, in get
raise ValueError("Expected singleton: %s" % record)

1 - Create a new company with currency EUR.
2 - Create a branch company underneath the main company.
3 - In Accounting, install fiscal localization, e.g. Belgian Companies on the company configuration settings.
4 - Select an account like 600000 Raw Materials, and enable Allow Reconciliation on this account. The exact account isn't important, only that we can make credits / debits to it to be reconciled.
5 - With only the top level company selected, make a debit of 100 USD, e.g. Vendor Bill, set in currency USD to the account 600000.
6 - Now with only the branch level company selected, make a credit of 100EUR, e.g. Customers Invoices, set in currency EUR to the same account with an amount equal to the credit in step 5. (if 1USD == 1EUR, 1-1), so that there is no residual amount, i.e. credit == debit.
7 - Now select both the top level company and the sub branch company in the company context.
8 - In Journal Items, reconcile the unreconciled journal items for the Account 600000.

With this commit we select the first company of the aml instead of every companies on the amls.

opw-6290703

Forward-Port-Of: odoo/enterprise#123774

Seven entries of the Mexican chart of accounts template carry a name belonging
to a **different** group, copied from a neighbouring entry. Each record's XML ID
still states the intended name, which is what this restores.

| Code | Field | Before | After |
|---|---|---|---|
| `6` | `name@es` | Gastos generales | Gastos |
| `252.07` | `name@es` | `account_subgroup_hipotecas_por_pagar_a_largo_plazo_nacional` | Hipotecas por pagar a largo plazo nacional |
| `602` | `name`, `name@es` | Cost of sales / Costo de venta | Selling expenses / Gastos de venta |
| `613` | `name@es` | Amortización contable | Depreciación contable |
| `614` | `name` | Accounting depreciation | Accounting amortisation |
| `701.06` | `name`, `name@es` | Interest on foreign bank charges / Intereses a cargo bancario extranjero | Interest payable by national natural persons / Intereses a cargo de personas físicas nacional |
| `702` | `name@es` | Utilidad cambiaria | Productos financieros |

### Why it is not cosmetic

The electronic accounting Chart of Accounts XML takes the `Desc` attribute of
every `<Ctas>` element from the *account group name* — `cfdicoa.xml`
(`t-att-Desc="account.get('name')"`), fed by `trial_balance.py`
`_l10n_mx_get_coa_values()`. Any `es_*` database therefore declares:

```xml
<catalogocuentas:Ctas CodAgrup="702" NumCta="702" Desc="Utilidad cambiaria" Nivel="1" Natur="A"/>
```

whereas the SAT catalogue (Anexo 24) publishes `702` as *Productos financieros*,
with `702.01 Utilidad cambiaria` … `702.10 Otros productos financieros` beneath
it. `CodAgrup` comes from `code_prefix_start` and stays correct, so the file
still validates against the XSD, but the declared description does not match the
official nomenclature. Trial Balance and Pólizas are unaffected — neither
exports group names.

### Evidence

- `252.07` contains its own XML ID as the Spanish name.
- `602` duplicates `501.01`, yet its children are `Sueldos y Salarios`,
  `Compensaciones`, `Tiempos extras`.
- `613` and `614` are swapped in one language each: `613`'s children are
  depreciations, `614`'s are amortisations.
- `701.06` duplicates `701.05` in both languages; the correct name is symmetric
  to `701.07` and to `702.06`.
- `6` is the only single-digit root group whose Spanish name does not match its
  XML ID (`account_group_gastos`).

### Notes

Introduced in d782b8b92557; correct in 15.0, where the names lived in
`account.account.tag.csv`. Still present in 18.0, 19.0 and master, hence
targeting 17.0. Template data only — existing databases are unaffected until the
chart is (re)installed, and renaming a group moves no balance.

Forward-Port-Of: odoo/odoo#277426

Forward-Port-Of: odoo/odoo#278328
Forward-Port-Of: odoo/odoo#277891

Issues
-----

1. Commercial invoice is not forwarded to the user for the return delivery.
2. US postal codes of format 12345-6789 cause the delivery to be rejected.

-----

Steps to reproduce issue 1
-----
- Set up UPS with return labels
- Create an INTL delivery & confirm

> OUT delivery has a commercial invoice in chatter, but the return doesn't

Cause
-----
The OUT and return call are not made using the same function. The OUT call is made via `ups_rest_send_shipping` which explicitly extracts the commercial invoice from the UPS response

https://github.com/odoo/enterprise/blob/1a7c8ac34348ebc1ebe2da4100bdaec57484056f/delivery_ups_rest/models/delivery_ups.py#L204-L205

We should adapt `ups_rest_get_return_label` to match.

-----

Steps to reproduce issue 2
-----
- Set up UPS
- Create an american customer with a 9 digit zip (eg 20500-0003)
- Create an delivery to the customer & confirm

> Error: Invalid sold to postal code. Valid length is 0 to 9 alphanumeric

Cause
-----
The zip code is transmitted as-is, so we should sanitise it beforehand.

https://github.com/odoo/enterprise/blob/c8c2f13b7fd17e215044fc62774f2b4a378aaf8c/delivery_ups_rest/models/ups_request.py#L368

Doc:
https://github.com/UPS-API/api-documentation/blob/69e8a3cee7f9d3bf80735ae329aed0d8be156f97/Shipping.yaml#L5410-L5420

-----
Ticket:
opw-6422500

Forward-Port-Of: odoo/enterprise#127375

Modern crawlers now send an `Accept-Language` header (for example, `en-US,en;q=0.9`), whereas historically they did not. When that language differs from the website's default language, `ir.http._match()` issues a 303 redirect from `/page` to `/<lang>/page`. Since crawlers do not retain cookies, unrecognized agents are redirected on every request and never reach the default-language page.

Customers reported that Google Search Console URL Inspection live tests only receive a redirect and that pages remain unindexed. Googlebot itself is not affected because it already matches the existing `bot` token.

`_match()` already skips language redirects for recognized bots by serving the default-language page directly. Extend the `bots` user-agent list with modern crawler identifiers, each verified against vendor documentation:

* `google-inspectiontool`: Search Console URL Inspection / Rich Results Test
* `googleother`: Google generic crawler (`GoogleOther`, `GoogleOther-Image`, `GoogleOther-Video`)
* `meta-external`: `meta-externalagent`, `meta-externalfetcher`, and `meta-externalads`, successors to the already-listed `facebookexternalhit`
* `meta-webindexer`: Meta AI search indexer
* `chatgpt-user`: OpenAI user-request fetcher (currently matched only through the `bot` substring in its info URL, which is fragile)
* `claude-user`: Anthropic user-request fetcher
* `perplexity-user`: Perplexity user-request fetcher

The redirect behavior remains unchanged for human visitors. Localized pages continue to be crawlable through their own URLs (for example, `/fr/page`) via `hreflang` alternates.

As a side effect, `link_tracker` and `mass_mailing_sms` no longer count clicks from these crawlers, and website visitor tracking skips them.

task-6213245

Forward-Port-Of: odoo/odoo#275571

step to reproduce :

1. Create a related field on `sale.order`, for example:

       x_studio_production_stage = tasks_ids.stage_id.name

2. Use this field in a filter:

       [('x_studio_production_stage', 'ilike', 'Dispatch')]

3. Applying the filter raises:

```python
   Traceback (most recent call last):
File "/home/odoo/src/odoo/19.0/odoo/http.py", line 2329, in _serve_db
return service_model.retrying(serve_func, env=self.env)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/service/model.py", line 188, in retrying
result = func()
^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/http.py", line 2384, in _serve_ir_http
response = self.dispatcher.dispatch(rule.endpoint, args)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/http.py", line 2599, in dispatch
result = self.request.registry['ir.http']._dispatch(endpoint)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/addons/base/models/ir_http.py", line 353, in _dispatch
result = endpoint(**request.params)
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/http.py", line 838, in route_wrapper
result = endpoint(self, *args, **params_ok)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/addons/web/controllers/dataset.py", line 32, in call_kw
return call_kw(request.env[model], method, args, kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/service/model.py", line 97, in call_kw
result = method(recs, *args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/addons/web/models/models.py", line 67, in web_search_read
records = self.search_fetch(domain, specification.keys(), offset=offset, limit=limit, order=order)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 1408, in search_fetch
query = self._search(domain, offset=offset, limit=limit, order=order or self._order)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 5366, in _search
domain = domain.optimize_full(self)
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/domains.py", line 446, in optimize_full
return self._optimize(model, OptimizationLevel.FULL)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/domains.py", line 460, in _optimize
previous, domain = domain, domain._optimize_step(model, next_level)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/domains.py", line 654, in _optimize_step
children = self._flatten(child._optimize(model, level) for child in self.children)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/domains.py", line 609, in _flatten
for child in children:
File "/home/odoo/src/odoo/19.0/odoo/orm/domains.py", line 654, in <genexpr>
children = self._flatten(child._optimize(model, level) for child in self.children)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/domains.py", line 460, in _optimize
previous, domain = domain, domain._optimize_step(model, next_level)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/domains.py", line 654, in _optimize_step
children = self._flatten(child._optimize(model, level) for child in self.children)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/domains.py", line 609, in _flatten
for child in children:
File "/home/odoo/src/odoo/19.0/odoo/orm/domains.py", line 654, in <genexpr>
children = self._flatten(child._optimize(model, level) for child in self.children)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/domains.py", line 460, in _optimize
previous, domain = domain, domain._optimize_step(model, next_level)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/domains.py", line 962, in _optimize_step
domain = self._optimize_field_search_method(model)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/domains.py", line 1008, in _optimize_field_search_method
computed_domain = field.determine_domain(model, operator, value)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 1928, in determine_domain
return determine(self.search, records, operator, value)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/fields.py", line 81, in determine
return needle(*args)
^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/addons/sale_project/models/sale_order.py", line 76, in _search_tasks_ids
query = self.env['project.task']._search(task_domain)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/models.py", line 5355, in _search
domain = Domain(domain)
^^^^^^^^^^^^^^
File "/home/odoo/src/odoo/19.0/odoo/orm/domains.py", line 259, in __new__
raise ValueError(f"Domain() invalid item in domain: {item!r}")
ValueError: Domain() invalid item in domain: ('id', 'any!', [('id', 'any!', <odoo.tools.query.Query object at 0x7aca184f4170>)])
```

Cause:

When searching on the related field, [_search_related()](https://github.com/odoo/odoo/blob/19.0/odoo/orm/fields.py#L768) converts the related path into an `any!` domain:

    ('tasks_ids', 'any!',
        [('stage_id', 'any!', [('name', 'ilike', 'Dispatch')])]
    )

During [Domain.optimize_full()](https://github.com/odoo/odoo/blob/19.0/odoo/orm/domains.py?utm_source=chatgpt.com#L436), [_optimize_field_search_method()](https://github.com/odoo/odoo/blob/19.0/odoo/orm/domains.py?utm_source=chatgpt.com#L1008) calls the field's search method, which invokes `_search_tasks_ids()` with `operator='any!'` and the related domain as `value`.

The existing [_search_tasks_ids()](https://github.com/odoo/odoo/blob/57c7c9938725d392a6f2cd6c89a861d2a8385c44/addons/sale_project/models/sale_order.py#L76) expects a normal search value and therefore generates an invalid nested domain.

Fix :

`_search_tasks_ids()` to directly pass the domain to `project.task._search()` when the operator is `any` or `any!`.

upg - 4584778
opw - 6475804

[here]: https://github.com/odoo/odoo/blob/19.0/odoo/orm/fields.py#L768

[here]: https://github.com/odoo/odoo/blob/19.0/odoo/orm/models.py#L5366

[here]: https://github.com/odoo/odoo/blob/19.0/odoo/orm/domains.py?utm_source=chatgpt.com#L436

[here]: https://github.com/odoo/odoo/blob/57c7c9938725d392a6f2cd6c89a861d2a8385c44/addons/sale_project/models/sale_order.py#L76

Description of the issue/feature this PR addresses:

Current behavior before PR:

Desired behavior after PR is merged:




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

Forward-Port-Of: odoo/odoo#284101

Steps to reproduce:
----------------------------------
1. Install the `timesheet_grid` module.
2. Create a project and add any timesheet to it.
3. Archive the project.
5. From the systray timer, click on the Project field.

Observation:
----------------------------------
The archived project is visible in the dropdown.

Issue:
----------------------------------
In Odoo, standard search views and `name_search` calls on `project.project` automatically respect `active_test=True`. When you open the timer, the frontend passes `{'timesheet_timer_search': True}` in the context to `name_search` with an empty query string. `name_search` overrides standard searching to retrieve recently used projects first by querying `account.analytic.line` via `_get_recently_used_records ('project_id', ...)`. `account.analytic.line` stores past timesheet logs. Even after a project is archived, historical timesheet records for that project still exist in `account.analytic.line`. Because `_get_recently_used_records` runs a `_read_group` query on `account.analytic.line` (which has no active field of its own), it fetched the `project_id` from historical timesheet entries without checking if the referenced project was active.

Solution:
----------------------------------
In `name_search`, explicitly append `[('active', '=', True)]` to the `project_domain` used when querying `_get_recently_used_records`. Standard form/list views using `_domain_project_id` already benefit from Odoo's default ORM `active_test=True` mechanism during standard `project.project` searches.

Note:
----------------------------------
Another solution was to add `active = true` in `getTimesheetTimerFieldInfo` https://github.com/odoo/enterprise/blob/22eb84cdc94ba334d42bad32fb491d35c8147c94/timesheet_grid/static/src/services/static_timesheet_timer_service.js#L322-L328

Fixing it in Python ensures that any call passing `timesheet_timer_search` in context (e.g. mobile widgets, custom RPCs, or python wizards) will benefit from the fix, rather than only patching a single OWL JS service.

opw-6445528

Forward-Port-Of: odoo/enterprise#127374

**Issue**: If a partner has a tax number that begins with a different country’s code (for instance, a Mexican RFC number that begins with “RO”, matching Romania), when a user tries to add the tax number to the partner, there will be a validation error.

**Steps to reproduce** (on fresh database with Contacts app and l10n_mx module installed):
    1. Make a new contact.
    2. Give the contact a Mexican address.
    3. Give the contact the RFC number (or `vat` field): ROS561231GR8.
    4. Try to save this change. Observe the validation error.

**Explanation**:
The `get_all_identifiers` method uses the first two characters of `partner.vat` as a heuristic to detect the issuing country, since many VAT formats start with a country code (e.g. RO1234567897). This prefix was used unconditionally whenever it matched an item from `get_tin_metadata_of_country`. without checking whether the VAT actually belongs to that country. Some countries' identifier formats begin with letters which are not country codes. In Mexico, for instance, RFC numbers start with letters derived from the partner's name, so a partner named “Sofia Rodriguez” would get an RFC starting with “RO”. Therefore, this heuristic can produce false-positive matches against unrelated countries.

**Solution**:
We no longer use a partner's vat number to detect the issuing country. Instead, we use the partner's `country_code` field as the issuing country.

opw-6471006

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