Benjamin Rodenberg

## Main changes of this PR Breaks develop. Only merge this PR as soon as the last release before the breaking release v3.0.0 is published. Closes #1223, #859, **more?** ##...

Remove timing "ondemand"

2

## Main changes of this PR Remove `ondemand` timing. This is a follow-up for #1222 ## Author's checklist * [x] I added a changelog file with `make changelog` if there...

Remove resetWrittenData

8

## Main changes of this PR Removes the unneeded/untested function `SolverInterfaceImpl::resetWrittenData`. ## Motivation and additional information In #1029 this function has been identified as not required for (tested) functionality. It,...

ScaleByDtAction with SCALING_BY_DT is misleading

4

In preCICE the term "dt" has multiple meanings: * time window size (what we provide in the config, time between two couplings) * time step size (actual time step size...

Simplify signature of actions

4

While working on #1350 I realized that using `dt` for actions during initialization is risky, because it is not really clear what `dt` means at thsi point in time. Under...

Clearly define and enforce allowed states and transitions in SolverInterfaceImpl

3

**Please describe the problem you are trying to solve.** Our user interface currently offers many possible ways how the user can interact with it. Flexibility is, of course, nice, but...

Generally allow initialize=true for all coupling schemes and all participants

4

**Please describe the problem you are trying to solve.** Currently `initialize=true` is forbidden for data that is sent from first to second in a serial coupling scheme (because this data...

Use-case for isReadDataAvailable and isWriteDataRequired?

10

I am not sure that the functions `isReadDataAvailable` and `isWriteDataRequired` are actually needed. Removing them is clearly breaking, but would also help us to simplify the codebase and the API....

Make calling initializeData mandatory?

9

**Please describe the problem you are trying to solve.** Simplifying the implementation of coupling schemes and have a stricter control flow with less optional branches. **Describe the solution you propose.**...

Deprecate getMeshVertexIDsFromPositions and getMeshVertices?

3

Is the function `SolverInterface::getMeshVertexIDsFromPositions` needed? I have never used it and it would be a candidate for deletion for v3.0.0 from my perspective. There was some discussion about this API...