Summary
Coupling (unadapted) different solvers using the 'mapped' boundary conditions. Each solver runs in its own 'world' and the boundary condition specifies the remote 'world' to communicate with. This is in addition to the remote region and patch.
Resolved bugs (If applicable)
(Links to issues)
Details of new models (If applicable)
All the solver take a new 'world' option to name the simulation. All the 'mapped' type boundary conditions take an additional 'sampleWorld' input to indicate which world to exchange data with. Default is direct exchange so every 'evaluate' needs to be replicated on the other world. Additional there is an indirect exchange which uses a functionObject to do the exchange once per time step. This add additional explicitness but allows running with e.g. different numbers of correctors.
Risks
(Possible regressions?) (Changes to user inputs?)
Only affects coupling boundary conditions ('mapped', 'mappedMixed', 'turbulentMappedXXX' etc) so should have no impact apart from that.