Benjamin Uekermann

> I would go the "safe" path here and make the function throw in case it is called on a received-mesh, unless the flag enabled access to it. Sounds good...

> What's then the expected return value in this case? What we had so far: the number of vertices of this mesh partition.

Why not creating a separate `Aggregator` class, which we then could use in `PointToPointCom` and `GatherScatterCom`? > A bit of an open question how to shape this in terms of...

Related to https://github.com/precice/preeco-orga/issues/36

> Is there a way that ask preCICE if we are at the end of the simulation? That's exactly the purpose of `isCouplingOngoing`, [step-by-step docs](https://precice.org/couple-your-code-time-step-sizes.html#steering-the-end-of-the-simulation).

Open points from discussion in #1952 - Bounds at the data declaration: Users might wonder what we do with these. Incomplete list: check after writing, some transformation when reading, acceleration,...

Yes, this could be a solution to make things more explicit. I find `lower-bound` and `upper-bound` easier to understand than `valid-min/max`. An the "bound" keyword would allow us to associate...

Additional complexity: how should we treat the case when a mesh is associated to several mappings and thus to several targets or sources? I have the feeling that this is...

> We could add another parameter for the error code makes sense to me. > Do we also want to be able to test for warnings? I think this would...

> This intermediate buffer and the copy from the buffer to preCICE is pure necessary overhead. I guess you mean "unnecessary".