Skip to content

feat: added LGPL as a valid license#1050

Open
jhdub23 wants to merge 5 commits intomainfrom
feat/license
Open

feat: added LGPL as a valid license#1050
jhdub23 wants to merge 5 commits intomainfrom
feat/license

Conversation

@jhdub23
Copy link

@jhdub23 jhdub23 commented Oct 31, 2025

No description provided.

@jhdub23 jhdub23 requested a review from a team as a code owner October 31, 2025 22:01
@github-actions github-actions bot added the enhancement General improvements to existing features label Oct 31, 2025
@jhdub23 jhdub23 force-pushed the feat/license branch 2 times, most recently from f4c199e to 41384c3 Compare October 31, 2025 22:08
@jhdub23 jhdub23 changed the title feat (check-licenses): Added LGPL as a valid license. feat: Added LGPL as a valid license. Oct 31, 2025
Copy link
Member

@RobPasMue RobPasMue left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGPL cannot be accepted as a license by default - investigations need to be performed on the way the impacted package is used on our side.

@jgd10
Copy link

jgd10 commented Nov 3, 2025

The ability to release Python libraries freely but NOT as fully OSS (aka Freeware) is desirable and there is at least one library that falls into this category at the moment within ansys and there could be more in the future.

@jhdub23
Copy link
Author

jhdub23 commented Nov 5, 2025

Under what circumstances would an LGPL package X be permissible while an LGPL package Y be denied? For example: GPL - internal use only, allowed. Shipped within a closed-source product: not allowed.

I'm trying to understand the decision factors that would apply for us to deny the use of an LGPL package. I can't think of any, assuming the product team has ok'd the use of the package.

@jorgepiloto
Copy link
Member

As far as I am concerned:

  • We allow the use of free software for development
  • We don't allow free software dependencies in our projects

The reason for the last point is that we would be force to distribute our software under a free software license.

@ansys-cla-bot
Copy link

ansys-cla-bot bot commented Nov 18, 2025

All Contributor License Agreement (CLA) signatures have been captured successfully. Thanks for contributing!

@jhdub23
Copy link
Author

jhdub23 commented Nov 18, 2025

As far as I am concerned:

  • We allow the use of free software for development
  • We don't allow free software dependencies in our projects

The reason for the last point is that we would be force to distribute our software under a free software license.

LGPL packages don't force us to distribute our software under a free software license. Also, software in the ansys org is already distributed as free; every repository has an MIT or equivalent license. That does NOT extent to components it uses (e.g. our proprietary code).

@jorgepiloto
Copy link
Member

As discussed with @jhdub23, this should be fine. We are just using them as a dependencies. We are not modifying or copy-pasting their source code.

@RobPasMue
Copy link
Member

As discussed with @jhdub23, this should be fine. We are just using them as a dependencies. We are not modifying or copy-pasting their source code.

If we end up allowing it, please revisit the https://github.com/ansys/actions/blob/main/check-licenses/ignored-packages.txt packages so that if any of them are LGPL, we can accept them

@da1910
Copy link

da1910 commented Jan 8, 2026

The main reason for my reservations here remains in place - we have downstream consumers of these packages who bundle them into executables using pyInstaller. In that context LGPL dependencies are almost certainly not going to be acceptable, and we've previously discussed at length the need to work with those teams.

Permitting LGPL components as strictly parts of the build or documentation process seems reasonable, but for any of the flagship libraries, hard dependencies on LGPL libraries will prevent this bundling, and impact those workflows.

This discriminated permission process is not supported by pip-licenses or by the existing workflows, and will require manual approval and tracking (which I remain convinced we should do)

@da1910
Copy link

da1910 commented Jan 9, 2026

Per yesterdays discussion this is reasonable - I would like to see the support policy written down somewhere, but we seem comfortable drawing a line around open-source.

We also discussed improving the reporting process for SBOMs which I'll take a swing at today.

@jorgepiloto jorgepiloto requested a review from RobPasMue January 13, 2026 13:37
@jorgepiloto jorgepiloto changed the title feat: Added LGPL as a valid license. feat: added LGPL as a valid license. Jan 14, 2026
@jorgepiloto jorgepiloto changed the title feat: added LGPL as a valid license. feat: added LGPL as a valid license Jan 14, 2026
@jorgepiloto jorgepiloto added this to the ansys/actions@v10.2 milestone Jan 14, 2026
done
fi

- name: Force review and acceptance of whitelisted licenses
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

After some thoughs, I would remove this checking tbh.
While I would like to ensure that every developer pays attention to direct/transitive dependencies' licenses, I don't think adding a step in our action is the right way. If one updates the whitelist-license-check input, then one should already take responsibility for this.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Everything works on my side. I thought to add this extra-step so that it forces developers to think twice about what they are doing.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes, that is a good point @SMoraisAnsys. You convinced me 😄

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah the check is not needed as an action input. I thought we agreed on this yesterday. Modifying the whitelist acknowledges that you know what you are doing.

The bot comment as discussed is a good double-check though, but it has to be separate from the action itself

@jorgepiloto jorgepiloto mentioned this pull request Jan 14, 2026
@jorgepiloto jorgepiloto force-pushed the feat/license branch 2 times, most recently from 813351e to 83fbc43 Compare January 14, 2026 13:05
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

enhancement General improvements to existing features

Projects

None yet

Development

Successfully merging this pull request may close these issues.

9 participants