Wednesday, April 1, 2026
6 changes · 17.0
Enhancements to existing features
This update adds missing tax exemption reasons to the electronic invoicing system to ensure full compliance with Peppol standards. The change enables businesses to properly document all types of tax exemptions when creating electronic invoices, improving accuracy and regulatory compliance in cross-border transactions.
Original PR description
Some tax exemption reasons were missing, This commit ensures having all the tax exemption reasons introduced by Peppol task-6048561 --- I confirm I have signed the CLA and read the PR guidelines at www.odoo.com/submit-pr
The system now allows companies to register PEPPOL numbers from any country, not just those on a predefined list. This removes geographic restrictions and makes the invoicing system more flexible for businesses operating internationally or with international partners.
Original PR description
Before this commit, only numbers on the peppol list were able to be registered. Now is possible to add numbers from other countries. Task-6033336 --- I confirm I have signed the CLA and read the PR guidelines at www.odoo.com/submit-pr
This improvement makes product matching during imports more accurate and flexible. Products are now matched regardless of uppercase/lowercase differences (e.g., "Network Cable" matches "network cable"), and the system uses intelligent similarity matching instead of simple substring search to avoid incorrectly pairing unrelated products. This results in fewer import errors and better data accuracy when bringing in product information.
Original PR description
Before this commit: - Product retrieval during import relied on exact name match and substring (ilike) search. - Exact name search was case sensitive, so values like `Network Cable` would not match…
Before this commit: - Product retrieval during import relied on exact name match and substring (ilike) search. - Exact name search was case sensitive, so values like `Network Cable` would not match `Network cable`. - Substring matching could return unrelated products (e.g. `Wireless bluetooth speaker` gets matched with `Wireless bluetooth speaker battery`), leading to unrelated matches. After this commit: - Exact name search is now case insensitive, allowing matches such as `Network Cable` and `network cable`. - Substring based matching has been replaced with a similarity ratio (90%) to reduce false positives and improve matching reliability against customer database product names. Technical: - Replaced `=` with `=ilike` in the exact name search domain to make the lookup case insensitive. - Similarity ratio is computed using Python's `difflib.SequenceMatcher` on product names, with a minimum threshold of 90% to qualify as a match. - Added system parameter for configurable product name similarity threshold. task-5951469
Resolved issues and error corrections
This fix corrects an overly strict warning that appeared when generating payslips for employees whose contracts partially overlap with the payslip period. Previously, the system required the contract to fully contain the payslip dates; now it correctly allows payslip generation as long as there is at least one day of overlap between the contract and payslip period. This resolves false warnings that prevented legitimate payslip processing.
Original PR description
[FIX] hr_payroll: fix payslip warning bug Bug reproduction: 1 - Select Hong Kong (actually there is nothing about Hong Kong, you can select other companies as well) 2 - Create an employee and make…
[FIX] hr_payroll: fix payslip warning bug
Bug reproduction:
1 - Select Hong Kong (actually there is nothing about Hong Kong, you can select other companies as well)
2 - Create an employee and make its contract from 01-01-2025 to 05-03-2026 (DD/MM/YYYY) format.
3 - Generate payslip for March, the warning of "The period selected does not match the contract validity period" popups.
4 - But we do not want that, even though there is 1 overlapping day in contract with payslip we can continue.
Bug cause:
1 - In >= v.17 (not in v.19), there was a warning, when the contract dates do not fully contains the payslip dates, the warning was appearing.
2 - In v.19 it is not the case, when there is a contract that overlaps at least one dat of the payslip then we are fine, if no overlap then no contract on payslip warning should appear
Bug solution:
1 - I replaced old warning "The period selected does not match the contract validity period" with the one in v.19 "No running contract over payslip period"
Tests:
1 - There was a unit test about old warning (test_payslip_warnings), I changed that parts.
2 - I added further steps to the existing test about the new warning that should appear (No running contract over payslip period)
Note: Implemented feature: need to check what happens after v.17, should be removed in v.19 latest, maybe before as well.
task - 6006693The account reports feature had a memory leak where background data loading would continue indefinitely, preventing the system from freeing up memory. This fix ensures the loading process stops when the report is closed, allowing the system to properly reclaim memory and improve overall application performance.
Original PR description
The preloading of sections would never stop, this is an issue since this would prevent the garbage collector from collecting this big class and all it's objects. We fix this by making sure to stop the reploading when the component is destroyed. It's important to do it this way rather than clearing the timeout as the destruction could happened when the report is loading so the timeout would be unset and a new one would be started.
This fix enables the amount-in-words feature to work correctly in Czech. The underlying library had a bug using the wrong language code, which has been corrected through a temporary workaround. This ensures Czech users can now see amounts displayed as written words in their documents and reports.
Original PR description
The `num2words` library has a bug in the language code they used for Czech (`cz` instead of `cs`). This commit adds a monkey patch to map the correct language code to the existing converter class, allowing the amount in words to work in Czech. The issue was fixed in version 0.5.14 of the library, so this patch can be removed once we use Ubuntu >= 25.10 (Python >= 3.13), that contains the fixed version of the library. [opw-6088697](https://www.odoo.com/odoo/project.task/6088697)