fix(GweInputData): large GWE models throwing stackoverflow error

[emorway-usgs]

The following two lines:
this%gwerhos = gwerhos
this%gwecps = gwecps

should have instead used:

this%gwerhos => gwerhos
this%gwecps => gwecps

That discovery led to a number of other fixes since the variables in question were pointers and not copies of data. While in GweInputData.f90, updated it to follow latest style guide. Will adjust other GWE src files in the future.

While running the "Barends" model locally with 200,000 nodes, GWE was crashing with a stackoverflow error while trying to copy an entire array into another array in one line.

Amended this PR with a new smoke test that asserts two neighboring models fail when the density of water is specified differently among the neighbors. Currently, there are no methods for averaging values on either side of the interface. So, the new test first runs and verifies that the simulation runs to completion when the density of water is the same in the two neighbors. Next, the same simulation is run except that the density of water is just a little different in the 'right-hand' model.

The new smoke test should be ready to bring into parallel testing once the following is addressed by @mjr-deltares (lines 512-514ish in GweGweConnection.f90 - there is a similar comment in GwtGwtConnection.f90):

    ! we cannot validate the remainder (yet) in parallel mode
    if (.not. gweEx%v_model1%is_local) return
    if (.not. gweEx%v_model2%is_local) return

• Replaced section above with description of pull request
• Referenced issue or pull request feat(GWE): Introduce Groundwater Energy Transport #1493
• Formatted new and modified Fortran source files with fprettify
• Added new test
• Formatted new test with ruff

[mjr-deltares]

Hi @emorway-usgs , that last commit looks a bit different from what we discussed last week. Maybe have a call so I can explain my thoughts?

[mjr-deltares]

Looks good. I will make a note of the validation not being done in parallel but this can go in.