Skip to content

Conversation

@kmokstad
Copy link
Contributor

Also provide spline parameters to the Vec4 object in the boundary
integration methods, to account for possible field evaluation there.
The remaining changes are cosmetics only.

This (partly) replaces #289 which has been rebased beyond its original state.
It will remove a memory leak from the mixed NonLinEl tests (tQ2-Q1 and Q2Q1).
Have also tried to make the 2D and 3D implementations more similar to ease comparison.
Guess not that well covered by unit/regression tests yet, so please take look at it before approving.

@kmokstad
Copy link
Contributor Author

jenkins build this with downstreams please

@kmokstad
Copy link
Contributor Author

I think the failure here is only due to OPM/IFEM-AdvectionDiffusion#40 which also affects the Boussinesq test. @akva2 you fix?

@akva2
Copy link
Member

akva2 commented Jan 24, 2020

sorry about that, fixed now.

@akva2
Copy link
Member

akva2 commented Jan 24, 2020

jenkins build this with downstreams please

else if (!bou[iedge])
// Store the pointer to the first basis' edge in bou, to avoid memory leak
// since it is dynamically allocated (will be deallocated by parent class).
// If invoked with basis > 1 (less likely), we still leak though..
Copy link
Member

Choose a reason for hiding this comment

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

that certainly happens in stokes applications. but it's an improvement of course.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

yes, but only when inhomo dirichlet on pressure (or when different basis for x and y-direction).
Ideally, it should have been handled inside the SplineSurface class (like for the trivariate class).

Also provide spline parameters to the Vec4 object in the boundary
integration methods, to account for possible field evaluation there.
The remaining changes are cosmetics only.
@kmokstad kmokstad merged commit 0b88dbe into OPM:master Jan 30, 2020
@kmokstad kmokstad deleted the memory-leak branch January 30, 2020 10:32
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.

2 participants