Reapply "Compositional well"

[GitPaean]

This reverts commit c782a9f.

[GitPaean]

jenkins build this serial please

[GitPaean]

The failure is due to the invalid option access from the following code,

            this->evalSummary(reportStepNum,
                              curTime,
                              localWellData,
                              localWBP,
                              localGroupAndNetworkData,
                              localAquiferData,
                              blockData,
                              miscSummaryData,
                              regionData,
                              inplace,
                              this->outputModule_->initialInplace(),
                              interRegFlows,
                              this->summaryState(),
                              this->udqState());
        }

    const Inplace& initialInplace() const
    {
        return this->initialInplace_.value();
    }

[GitPaean]

Since FlowProblemComp does not calculate the initialInplace_ yet, so maybe the easier solution is to let evalSummary function either use std::optional or pointer as argument instead of const Inplace& initialInPlace,.

I would like some comments from @akva2 before I go ahead.

[akva2]

right. yes, I think initialInPlace should return the optional, and the function should take an optional, and deal accordingly on the inside.

[GitPaean]

this function assume that this->initialInplace() always have value, so it is moved to outputBlackoilModule.hpp.

[GitPaean]

jenkins build this please

[GitPaean]

I am not deep in understanding the usage of the reference to optional std::optinal<T>& , any comments regarding this.

In the middle of fixing, was thinking to turn to pointer.

[bska]

I am not deep in understanding the usage of the reference to optional std::optinal<T>&

Any chance you could split that part out to a separate PR? That would be easier to review/test than when it's embedded in the work that introduces compositional well models.

[GitPaean]

I am not deep in understanding the usage of the reference to optional std::optinal<T>&

Any chance you could split that part out to a separate PR? That would be easier to review/test than when it's embedded in the work that introduces compositional well models.

I can do that.

[GitPaean]

I am not deep in understanding the usage of the reference to optional std::optinal<T>&

Any chance you could split that part out to a separate PR? That would be easier to review/test than when it's embedded in the work that introduces compositional well models.

see #6137 .

[GitPaean]

This PR requires OPM/opm-common#4552 and #6137 to be able to run. So I am making as draft for now.

[GitPaean]

The dependent PRs have been merged. I am marking this PR as ready for review.

[GitPaean]

jenkins build this please

[bska]

I notice that none of the new header files are listed in CMakeLists_files.cmake. Is that intentional?

[GitPaean]

I notice that none of the new header files are listed in CMakeLists_files.cmake. Is that intentional?

I think it is just something we did not do it. I am adding all the .cpp and .hpp under flowexperimental/ to CMakeLists_files.cmake now.

[GitPaean]

I notice that none of the new header files are listed in CMakeLists_files.cmake. Is that intentional?

I just added them, let me know if you want some different organization.

[GitPaean]

jenkins build this please

[bska]

I notice that none of the new header files are listed in CMakeLists_files.cmake. Is that intentional?

I just added them, let me know if you want some different organization.

Thanks! Maybe we could add these files to the MAIN_SOURCE_FILES list instead? We might not want to install them at this time, but it's nevertheless a good idea, I think, to have them in the file list as that makes it easier for those who use IDEs.