Skip to Content
  2. Mailing Lists
  3. Contributors
  4. Re: OCA/bank-payment-alternative

Archives

Contributors

Re: OCA/bank-payment-alternative

Re: OCA/bank-payment-alternative

Re: OCA/bank-payment-alternative

Hello,

Akretion founder here and Odoo full time since 2008, co-created the OCA along with the old "Certified Training Partners" back in 2013, often still topping amongst the top 10 OCA contributors and carrying the hardest Odoo localization (Brazil, larger codebase than TinyERP 4.2, soon larger than Odoo 5, 90% test coverage).

Just a word to say, this is a very sad situation. While I trust Alexis about the quality of his incredible work and thank him for leading the French OCA localization, I should say that "skipping a few native fields" is certainly not more condemnable than skipping Odoo odoo versions.

Yes at Akretion we largely skept version 9, 11, 13, 15 and 17 and having a better focus on even versions helped us a lot to float the boat.
But hell, all that would never be possible if some people were not adopting these odd versions. There would be no working OpenUpgrade path if some major players (including Tecnativa, Dixzmit, Forgeflow) didn't adopt these odoo versions. How many modules would not have made it to v14 if these players (and others like CamptoCamp, Acsone...) didn't bit the bullet for us in v13. Cause yes it's easy to jump on even versions once the "community" did half the work on the odoo versions.

And certainly part of the mess is due to Alexis not aligning with the other banking PSCs during versions 15 or 17 and then path diverging at 16 or 18.
Again, we know why we skip these odd versions, but skipping them and then pointing the finger to the others for not adopting the "native way" is not something I can agree with sorry.
Certainly the "keep throwing shit at the wall until something sticks" roadmap from Odoo, that is sell shit early, fix it years later if they can, or "fake it until you make it", is hardly sustainable for an open source community, but we are all trying to cope with it and skipping versions is no better than not adopting what is often just "shit at the wall" early on.

Moreover, there are other cases very major cases were Pedro was the first to push important native refactors (or "innovation" if you like) in account_payment_order:

Finally, I think some people do not follow the Github OCA notifications and don't have a clear vision about the INCREDIBLE AMOUNT of WORK done in the OCA by the so called dictator Pedro Baeza, and even Enric from Dixmit or other people supporting the legacy OCA/account-banking way.

Yeah, it's easy to point the finger to Pedro and many people may not like that he did a negative review of their PR. But I think the guys from Acsone just don't accept either untested or sub-optimal code that Pedro uses to block. It's just that Pedro does reviews at a scale nobody does so it's easy to point him as "the bad guy" of the OCA.

But my important point here is: we should stop behaving as spoiled children here. Without Pedro I think we would all be trapped in bullshit jobs running after our Odoo Enterprise licenses monthly sale targets until Odoo investors kill us one by one as they did with all their serious "partners", most of them not even there anymore to warn the next lemmings.

Before making critics to Pedro and people like Enric, people should REALLY be aware of what it mean to achieve these contributions figures years after years after years:

People who do not follow the Github notifications and think work only happens during the official meetings are just not getting the big picture.
So once you understand our whole OCA thing relies on people like Pedro, I think you should adjust to make it easier, not harder for these guys.
And people saying "if Pedro were not doing it, others would do it, it would be more democratic", Sorry but I call that wishful thinking from people not used to running open source projects.

If no converging way can be achieved, at least I think people behind the fork should assume to work to minimize the impact. For instance, help extract payment_order_dependencies in other repos, or provide common abstractions they can depend on...

About skipping versions: my recommendation at Akretion is even if we keep skipping versions for our projects, we should help migrate the main modules we pretend to be responsible for the odd versions to avoid that kind of situation.

So I hope you guys will find the best way, but here are my 2 cents.


On Fri, Jun 27, 2025 at 10:32 AM Enric Tobella Alomar <notifications@odoo-community.org> wrote:
Alexis,

You removed all history from 17 and part of 16. Some of the fixes that you made were already done by other people.

I am just analyzing https://github.com/OCA/bank-payment-alternative/pull/5. The same might happen with all other modules of your alternative, but I don't have the energy or time to do it...

Just to start: you didn't rewrite the history. Commits will be lost on next migration (removing people attribution in the next migration, not now...)

We shared with you how to do it here in a non-blocking review (we didn't want to block your job): https://github.com/OCA/bank-payment-alternative/pull/5#pullrequestreview-2931804750

About removing people attribution: I will raise 2 small examples because it takes time to compare.

- https://github.com/OCA/bank-payment/commit/1d37132360fae0d8b6d3204e63007862f3325922 was lost. It improved how it was handled the tests but you ignored it...

Do you need more examples?

My recomendation would be: "check the history of branch 17"

--
Enric Tobella Alomar
CEO & Founder

_______________________________________________
Mailing-List: https://odoo-community.org/groups/contributors-15
Post to: mailto:contributors@odoo-community.org
Unsubscribe: https://odoo-community.org/groups?unsubscribe



--
Raphaël Valyi
Founder and consultant

by Raphaël Akretion - 07:20 - 27 Jun 2025

Reference

  • OCA/bank-payment-alternative
    Dear OCA friends,

    On March 26th 2025, Stéphane Bidoul sent an email on this mailing-list with the subject "The future of OCA/bank-payment". If you missed it, you can find it in the archives:

    After some long discussion/debate, it was decided to let the chance for an alternative OCA banking project that would use the native payment method object of the "account" module (account.payment.method.line). This project is named OCA/bank-payment-alternative and you can find it here: 

    We had to change the module names in OCA/bank-payment-alternative. Here are the new names that were chosen :

    OCA/bank-payment OCA/bank-payment-alternative
    account_payment_order account_payment_batch_oca
    account_banking_pain_base account_payment_sepa_base
    account_banking_sepa_credit_transfer account_payment_sepa_credit_transfer
    account_banking_mandate account_payment_mandate
    account_banking_mandate_sale account_payment_mandate_sale
    account_banking_sepa_direct_debit account_payment_sepa_direct_debit
    account_payment_mode native + account_payment_base_oca
    account_payment_partner native
    account_payment_sale account_payment_base_oca_sale

    Here are the main changes in the datamodel introduced by OCA/bank-payment-alternative :
    1. Adoption of the native object account.payment.method.line as "Payment Method" (replaces the OCA object account.payment.mode)
    2. Introduction of account.payment.lot, which represent the Payments Lots (corresponds to the PmtInf i.e. Payment Information block of the SEPA XML files). If your bank groups the debits/credits on your bank statement for the payments, a lot matches one debit or credit payment line on your bank account. This object is generated upon confirmation of the payment order and is used in the generation of the XML (and should soon be used in the bank statement reconcile interface)
    3. the datamodel of the mandate has been simplified: the field "format" also has the information of the "scheme" field, so "format" now has 3 possible values : basic, sepa_core or sepa_b2b. The field "scheme" has been removed. The field "type" has 2 possible values: recurrent or oneoff (instead of 3 possible values : generic, recurrent or oneoff). The field "recurrent_sequence_type" has been removed because we don't need to handle the "first" vs "recurring" sequence any more : since November 2016, "the requirement to use the sequence type ‘First’ in a first of a recurrent series of Collections is no longer mandatory" according to the European Payment Council (EPC), cf SDD Core Rulebook. The "final" sequence is now supported by the state field which has a new "final" state that can be activated via a button. The field "partner_id" is NOT a related field of partner_bank_id any more, which solves the bug "account_banking_mandate: Change in the filtering behavior of the "Bank Account" field" https://github.com/OCA/bank-payment/issues/1473 With all these simplifications on the mandate datamodel, the form view and list view of mandates are more user-friendly.
    4. The field generated_user_id has been removed from account.payment.order (the info is present in the chatter)
    5. On account.payment.method, the boolean field "payment_order_only" has been renamed to "payment_order_ok" (for consistency)
    6. The boolean field "convert_to_ascii" has been removed from account.payment.method : we consider that this option should be always on (there is a hook in the code in case you need to disable it).
    Here is a summary of the new features and improvements by order of importance :
    • Take into account the boolean field "allow_out_payment" of res.partner.bank on payment orders : when you try to confirm a payment order that has bank accounts on payment lines with allow_out_payment = False, you will get a blocking error message. The affected payment lines will be shown in red and you will have a smart button that gives access to the bank accounts that are not allowed to send money to. To enable "allow_out_payment", you need to be part of a specific group "Validate bank accounts" (XMLID account.group_validate_bank_account). As a consequence, we don't inherit the native ACL of res.partner.bank that give full rights to partner manager any more : the security is handled by "allow_out_payment".
    • by default, there is a sequence of payment orders and another sequence for debit orders. It is possible to configure a specific sequence for a payment mode.
    • add support for "Regulatory Reporting" in the SEPA XML structure (tag RgltryRptg). Needed in some countries for international non-SEPA credit transfers.
    • replace unstructured address by structured address in SEPA XML file (mandatory starting 11/2025)
    • add support for pain.008.01.08 (SDD) and pain.001.001.09 (SCT), which are now the recommended versions of the EPC
    • easier download of the banking file after generation
    • add field acc_number_scrambled on res.partner.bank for easy and direct use of scrambled account number
    • search on partner_id from account.payment.order
    • support currencies with decimal_places != 2 in ISO20022 XML file generation
    • on mandates, fields format, type, signature date and partner are not editable any more when state != draft
    • remove support for pain.001.001.02/04/05 (SCT) and pain.008.01.03/04 (SDD) which have never been selected by the EPC, to simplify the code
    • stop using safe_eval() in XML generation
    • replace all @api.onchange by computed fields
    • add sql unicity constraint on payment order number per company
    Many bugs have been fixed:
    You can have a complete list of the changes and bug fixes of OCA/bank-payment-alternative on these 2 issues:

    But, for me, the most important of OCA/bank-payment-alternative is the investment made on the quality of the code, not only on the SEPA XML generation but also on the lower-level modules.

    OCA/bank-payment-alternative will continue to focus on code quality and innovation (which involves datamodel changes when needed).
    Now that the modules are merged, the next priority should be on providing migration scripts for v17 to v18 migration.

    So, for v18, you now have the choice between OCA/bank-payment and OCA/bank-payment-alternative. Happy hacking !

    --
    Alexis de Lattre
    Akretion France - 27 rue Henri Rolland - 69100 Villeurbanne - France
    Mail : alexis.delattre@akretion.com

    by Alexis de Lattre - 03:34 - 21 Jun 2025