dorie issueshttps://ts-gitlab.iup.uni-heidelberg.de/dorie/dorie/-/issues2019-09-09T19:55:21+02:00https://ts-gitlab.iup.uni-heidelberg.de/dorie/dorie/-/issues/161Tests are not being configured correctly2019-09-09T19:55:21+02:00Santiago Ospina De Los Ríossospinar@gmail.comTests are not being configured correctly#### Summary
The CMake function `dorie_add_metaini_test` in dorie requests a `TARGET` argument which is later passed to the `dune-testtools` function `dune_add_system_test`. The latter is wrong because the function [`dune_add_system_test` does not request an ´TARGET´ argument](https://www.dune-project.org/sphinx/dune-testtools/) or at least not intentionally exposed to the end user (From the source code I deduce that it is internally used to dynamically assign created targets).
On the other hand, the `dorie` target is not an executable as the function `add_static_variants` requires. (`add_static_variants` is used within `dune_add_system_test`). This explains why the pipeline fails now that [`dune-testtools` allow executables to be placed anywhere](https://gitlab.dune-project.org/quality/dune-testtools/merge_requests/124).
#### Steps to reproduce
Use the current `marter` branch of `dorie` together with the current `master` branch of `dune-testtools`.
#### What is the current _bug_ behaviour?
Test fail to be configured by CMake and therefore, pipeline fails.
#### What is the expected _correct_ behaviour?
Tests are well configured by CMake.
#### Relevant logs, screenshots, files...?
Job [#61906](https://ts-gitlab.iup.uni-heidelberg.de/dorie/dorie/-/jobs/61906) failed for 48133574d68e0fe2262f66f24cd1902715c6292c:#### Summary
The CMake function `dorie_add_metaini_test` in dorie requests a `TARGET` argument which is later passed to the `dune-testtools` function `dune_add_system_test`. The latter is wrong because the function [`dune_add_system_test` does not request an ´TARGET´ argument](https://www.dune-project.org/sphinx/dune-testtools/) or at least not intentionally exposed to the end user (From the source code I deduce that it is internally used to dynamically assign created targets).
On the other hand, the `dorie` target is not an executable as the function `add_static_variants` requires. (`add_static_variants` is used within `dune_add_system_test`). This explains why the pipeline fails now that [`dune-testtools` allow executables to be placed anywhere](https://gitlab.dune-project.org/quality/dune-testtools/merge_requests/124).
#### Steps to reproduce
Use the current `marter` branch of `dorie` together with the current `master` branch of `dune-testtools`.
#### What is the current _bug_ behaviour?
Test fail to be configured by CMake and therefore, pipeline fails.
#### What is the expected _correct_ behaviour?
Tests are well configured by CMake.
#### Relevant logs, screenshots, files...?
Job [#61906](https://ts-gitlab.iup.uni-heidelberg.de/dorie/dorie/-/jobs/61906) failed for 48133574d68e0fe2262f66f24cd1902715c6292c:v2.0 ReleaseLukas Riedelmail@lukasriedel.comLukas Riedelmail@lukasriedel.comhttps://ts-gitlab.iup.uni-heidelberg.de/dorie/dorie/-/issues/147Water flux error estimator must include Neumann boundary condition2020-02-13T16:38:01+01:00Lukas Riedelmail@lukasriedel.comWater flux error estimator must include Neumann boundary condition### Description
The water flux error estimator computes an estimate of the conformity of the solution by integrating jumps in the solution and jumps in the gradient over grid faces.
The error estimator contains two terms:
1. Solution jump $`\propto \gamma [[ u ]] / h_F`$.
At the boundary, the jump is defined by the difference between the interior solution and the Dirichlet boundary condition (zero for Neumann).
2. Gradient jump $`\propto [[ K \nabla u ]]`$.
At the boundary, the jump is defined by the difference between the interior solution gradient and the Neumann boundary condition (zero for Dirichlet). _This is not considered right now._
### Proposal
Add the Neumann boundary condition to the "gradient jump" term in the error indicator.
### How to test the implementation?
Add a unit test: For a very simple 1x1 grid, a constant initial condition and prescribed boundary conditions, compute the error indicator and verify its value.
### Related issues
None### Description
The water flux error estimator computes an estimate of the conformity of the solution by integrating jumps in the solution and jumps in the gradient over grid faces.
The error estimator contains two terms:
1. Solution jump $`\propto \gamma [[ u ]] / h_F`$.
At the boundary, the jump is defined by the difference between the interior solution and the Dirichlet boundary condition (zero for Neumann).
2. Gradient jump $`\propto [[ K \nabla u ]]`$.
At the boundary, the jump is defined by the difference between the interior solution gradient and the Neumann boundary condition (zero for Dirichlet). _This is not considered right now._
### Proposal
Add the Neumann boundary condition to the "gradient jump" term in the error indicator.
### How to test the implementation?
Add a unit test: For a very simple 1x1 grid, a constant initial condition and prescribed boundary conditions, compute the error indicator and verify its value.
### Related issues
Nonev2.0 ReleaseLukas Riedelmail@lukasriedel.comLukas Riedelmail@lukasriedel.com