[T/RM] Relax tolerances for failing CI job.

See merge request ogs/ogs!3495

frontend

RM fixes and cleanups

See merge request ogs/ogs!3477

From the equations we have
eps = eps_{m} + eps_{sw},
with eps_{sw} = -C_el^{-1} sigma_{sw}.

All other places are using correct sign.

solid_grain_pressure used in some density models was
not set and caused an access error in the ExponentialProperty.

Using central differences for Jacobian computation.
Still the HydraulicFlow and Nodal forces are not correct
and don't include the 'M \dot{x} + K x' part.

Only happened in the assemble function, which does not
have a ctest with swelling.

More Conan packages moved to CPM (2nd try)

See merge request ogs/ogs!3494

Does not interfere with CPM modules.

- Removed install COMPONENTs
- Made ogs the default component name
- Install only the default ogs component with CPack

Some web doc additions

See merge request ogs/ogs!3493

GMSH2OGS; extract boundary meshes

See merge request ogs/ogs!3482

Thanks to Shuang for reporting.

Testing conversion of the gmsh file removing the line elements from the
boundary and extracting them as additional boundary files.

Dmitri Naumov authored 4 years ago and Dmitri Naumov committed 4 years ago

Specifying strictly smaller criterion, does not allow
eps=0 to be used.

Dmitri Naumov authored 4 years ago and Dmitri Naumov committed 4 years ago

Using exact matching.

If the utility would be called gmsh2vtu, then this identification
would be outsourced (identifySubdomains tool). Now the user
expects proper OGS-usable boundary meshes to be generated.

Dmitri Naumov authored 4 years ago and Dmitri Naumov committed 4 years ago

Adding a new command line flag '-b'. With the flag being set, the line
elements (same as those removed by the '-e' flag) are extracted as
individual meshes. The new meshes are stored in files corresponding to
the material ids.

Dmitri Naumov authored 4 years ago and Dmitri Naumov committed 4 years ago

For external usage.