Make queue_job_cron LGPL

[baimont]

[pedrobaeza]

Although I'm not a contributor of this module, my take on this: https://github.com/OCA/contract/pull/1259#issuecomment-3008159581

[rousseldenis]

You can't relicense a project without the agreement of every single contributor. Sampling other OCA repositories, it looks like AGPL is standard across all of them, so this pull request is likely DOA.

@amh-mw Acsone is the original author of the module and in particular @Cedric-Pigeon

As for queue_job, this is a technical module and serve as base for others.

@etobella @ValentinVinagre Please remove your blocking as you are not a contributor of this module. Many thanks

[etobella]

Have you asked for the people doing the last 3 migrations for example?

@QuocDuong1306 @nguyenminhchien @JoanMForgeFlow

There are other contributors, but just to add some of them.

If there is consensus on this topic from all contributors, I will unblock it, obviously :wink:

[nguyenminhchien]

This PR should be closed as requested by lmignon: https://github.com/OCA/contract/pull/1259#issuecomment-3015155659

[rousseldenis]

This PR should be closed as requested by lmignon: OCA/contract#1259 (comment)

This is not the same case.

[nguyenminhchien]

This PR should be closed as requested by lmignon: OCA/contract#1259 (comment)

This is not the same case.

I think they are the same subject, no?

[ValentinVinagre]

@rousseldenis

You can't relicense a project without the agreement of every single contributor. Sampling other OCA repositories, it looks like AGPL is standard across all of them, so this pull request is likely DOA.

@amh-mw Acsone is the original author of the module and in particular @Cedric-Pigeon

As for queue_job, this is a technical module and serve as base for others.

@etobella @ValentinVinagre Please remove your blocking as you are not a contributor of this module. Many thanks

I can't remove the block because I only have read permissions. Can you cancel my review?

[rousseldenis]

I can't remove the block because I only have read permissions. Can you cancel my review?

No, the fact you react here is sufficent :smiley:

[rousseldenis]

I think they are the same subject, no?

I checked with @lmignon. It was just about contract :sweat_smile:

[rousseldenis]

@etobella @QuocDuong1306 and @JoanMForgeFlow are no more active. I think we have the full pack :smile:

[yvaucher]

I just want to react to the claim that you cannot re-license without the approval of all contributors, it's not true due to the CLA that all contributor must have signed. It can be decided by the OCA. So it's a collective decision one person cannot put a veto. (It's nice to ask though and be clear about why)

I'm also totally in favor of keeping the default as AGPL accross OCA repo.

Here is my take, queue_job modules are LGPL and cron is a feature of the core. It would make sense to also have it as LGPL. I'm not against in this case. However, I don't see the need for it.

Let's assume the following dependency graph:

flowchart LR
    queue_job_cron["queue_job_cron"] --> queue_job["queue_job"]
    specific_addon["specific_addon"] --> queue_job

    style queue_job_cron fill:#C8E6C9,stroke:#00C853,color:#000000
    style queue_job stroke:#2962FF,fill:#BBDEFB,color:#000000
    style specific_addon fill:#E1BEE7,stroke:#AA00FF,color:#000000

The specific_addon, don't inherit on queue_job_cron, thus you can still configure your cron as a job even if it launches a proprietary method. Basically the only thing that you cannot do is to put the field run_as_queue_job in xml data inside your specific_addon.

EDIT: fixed field name

[etobella]

@yvaucher It is true that CLA allows OCA to change the license, but this has never been used in the past. Also, it is not specified how this change of license can be launched according to this document or the bylaws.

Before applying it, we should define the way that this specific change can be applied.

For this reason, we always ask for approval of all contributors to avoid this problem.

[yvaucher]

@etobella You are right I was just nuancing @amh-mw assertion. Mostly because there is on the balance contributors that might never answer or you might have 1% of contributor that veto the change. For the contributions that got no approval it should be removed legally speaking, that's why invoking the CLA could be an option.

I also want to remind that the OCA is commited to keep a FOSS license. And the license could for instance be changed (like the change from GPL to AGPL to close a loophole).

To me the current issue is just a followup of the license change on connector and queue_job which is sad but already happened in 2019.

[etobella]

Well, also changing licenses should require a complete audit of CLA signed, because we had a problem in another repo for this reason.

In any case, I would prefer to keep as many AGPL modules as possible too.

[rousseldenis]

@etobella @yvaucher @baimont Don't miss to read OCA FAQ about that:

https://odoo-community.org/resources/faq

If I read correctly, contributors that are not authors are not involved in licensing change.

[etobella]

Well, Certainly, it says only authors, but it states all, nothing about disappeared authors, so you should need the author (or their company at that time approval) approval from disappeared authors too (also, it is not defined what a disappeared author is....). Also, it is explained that if we want to change without all author approvals, we need to make a delegate vote.

Can I change the license when I port a module to a new version?

No. You can only change the license of a module if you have the copyright or if all authors of the module agree to change the license.

Can the OCA change the license of a module?

Through the signed CLA, the OCA has the right to relicense a module under any OSI certified license. But the CLA is here to protects the contributors, not to relicense their work. If that would happen, it’ll be subject to the vote of the Delegate assembly (the board alone cannot decide).

Thanks for sharing @rousseldenis