Skip to content

Conversation

@traversaro
Copy link
Contributor

@traversaro traversaro commented Jun 2, 2025

Add support to build against osqp 1.0.0 . As osqp-matlab 1.0.0 was not tagged, we build on top of the latest commit of the develop-1.0 branch, see osqp/osqp-matlab#55 . Furthermore, we also use the fix osqp/osqp-matlab#61 .

@traversaro traversaro changed the title Bump revision to 5 Refresh CI Jun 2, 2025
@traversaro traversaro changed the title Refresh CI Support osqp 1.0.0 Jun 2, 2025
@traversaro traversaro changed the base branch from main to v1 June 2, 2025 14:24
@traversaro traversaro closed this Jun 2, 2025
@traversaro traversaro reopened this Jun 2, 2025
@traversaro traversaro closed this Jun 2, 2025
@traversaro traversaro reopened this Jun 2, 2025
@traversaro
Copy link
Contributor Author

The Windows build are failing due to osqp/osqp#744 .

@imciner2
Copy link

imciner2 commented Jun 2, 2025

osqp-matlab doesn't have a 1.0 tag because it hasn't been fully updated to support the 1.0 release yet. It still is actually missing some new fields we added in the various structs to support the duality gap criteria, so you won't be able to use the develop-1.0 branch there until I update it (and Mathworks was down for a while recently, so I hadn't been able to open my MATLAB instance).

@traversaro
Copy link
Contributor Author

traversaro commented Jun 2, 2025

osqp-matlab doesn't have a 1.0 tag because it hasn't been fully updated to support the 1.0 release yet. It still is actually missing some new fields we added in the various structs to support the duality gap criteria, so you won't be able to use the develop-1.0 branch there until I update it (and Mathworks was down for a while recently, so I hadn't been able to open my MATLAB instance).

Thanks for chiming in! Interestingly, the test seems to pass with the latest develop-1.0 branch, is there something that I am missing?

@traversaro
Copy link
Contributor Author

Windows failures will be solved by conda-forge/libosqp-feedstock#17 .

@traversaro traversaro merged commit a9698d5 into v1 Jun 4, 2025
6 checks passed
@imciner2
Copy link

imciner2 commented Jun 4, 2025

Interestingly, the test seems to pass with the latest develop-1.0 branch, is there something that I am missing?

It won't catastrophically fail anymore (we made the v1 MATLAB interface more resilient to changes in the solver structs, so it can still at least do basic things if they change), but you won't see any of the new solver info or be able to modify the duality gap settings until it gets updated.

@traversaro
Copy link
Contributor Author

Interestingly, the test seems to pass with the latest develop-1.0 branch, is there something that I am missing?

It won't catastrophically fail anymore (we made the v1 MATLAB interface more resilient to changes in the solver structs, so it can still at least do basic things if they change), but you won't see any of the new solver info or beable to modify the duality gap settings until it gets updated.

Clear, thanks!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants