Ewald surface-epsilon is incorrect - Redmine #3441
The Ewald surface-epsilon correction term is incorrect when molecules
are broken over PBC.
It should be disabled in the 2020 release.
*(from redmine: issue id 3441, created on 2020-03-11 by berkhess)*
* Changesets:
* Revision c4878aff69823f930363566409f9f6a5bc7bcaf5 by Berk Hess on 2020-03-17T07:47:33Z:
```
Disable Ewald dipole correction without DD
Refs #3441
Change-Id: I492cd136b41e57b3e2198c1f0717177fb4d46a7c
```
* Revision 2548c9e0ce93d295e10d953f10763f7700ea7d17 by Berk Hess on 2020-03-18T22:04:44Z:
```
Use WholeMoleculeTransform for orires and epsilon-surface
Remove direct use of the graph from the orientation restraint
and Ewald epsilon-surface term code by passing in whole molecules
created by the WholeMolecules class.
Since this was the only remaining use of graph in do_force(),
it can now be removed.
Also enabled the epsilon-surface mdrun-test.
Fixes #3441
Change-Id: Idfec6508c6dfd9e1a656cf23613ede3793794901
```
issue