[Opm] Making OPM Flow Independent of LibECL

[Bård Skaflestad]

Dear fellow OPM Developers and Users,

At midnight tonight CEST-or possibly a little later-I will merge Pull Request 1107 in OPM-Common (https://github.com/OPM/opm-common/pull/1107) along with its downstream follow-up cousins


  *   OPM-Material #350 (https://github.com/OPM/opm-material/pull/350)
  *   OPM-Grid #397 (https://github.com/OPM/opm-grid/pull/397)
  *   OPM-Models #559 (https://github.com/OPM/opm-models/pull/559)
  *   OPM-Simulators #2060 (https://github.com/OPM/opm-simulators/pull/2060)
  *   OPM-Upscaling #281 (https://github.com/OPM/opm-upscaling/pull/281)

Collectively, these PRs will make OPM independent of the "libecl" (https://github.com/equinor/libecl) component that we've so far been using to great effect for reading and writing ECLIPSE-compatible result files (e.g., CASE.EGRID, CASE.UNRST or CASE.S0123).

Torbjørn Skille at Equinor has implemented a set of classes in OPM-Common that perform the same reading/writing task and PR 1107 is the final step to convert OPM's ECLIPSE-compatible I/O into using those classes. Note that these PRs change OPM's build system to no longer look for or link to libecl. Once everything is merged I recommend a full build of all modules from scratch, but you do not have to uninstall libecl to continue using OPM.

At the Release Manager's discretion, these changes will also be a part of the upcoming October release (designation 2019.10) of OPM Flow.

Thank you everyone for the effort in getting us to this point, and especially to Arne Morten Kvarving for the build system assistance.


Best Regards,

Bård Skaflestad
SINTEF Digital, Mathematics & Cybernetics
Computational Geosciences group