Additionally:
1) Add braces wherever possible for code clarity
2) Make all input coordinates const qualified
3) Remove not needed headers, and use c++ versions of needed headers
4) Fix typos
5) Comment out unused num_nodes parameters
6) Delete not implemented functions

Adding 'mean ratio' tet metric to Cubit.  Metric created and recommended by Jay Foulk to better detect elements that become elongated in high-deformation anlayses.  Compensates for 'normalized inradius' inability to detect these elements.
In Verdict (V_TetMetric.cpp), there was already a mean ratio metric for tet4s but it wasn't hooked up to anything, save some unit tests.  This metric is formulated differently and penalizes elements with high aspect ratio more than the previous implementation.  This metric now also can be computed on tet10 elements, computing the linear version on each subtet and returning the lowest.
The author of the previous metric was Danial Ibanez.  Jay Foulk with talk with him about this different implementation of it.  Soon we'll settle on which to go forward with.

(cherry picked from commit 2d7d6dadb3840dcf63b3a62e129e46990a23009a)

(cherry picked from commit ec8fb77ccbddaa9378f57e6db3578dfc145ebd5a)

Put package CMake files into a cmake directory.

See merge request meshing/cubit!124

This fixes a failing find_package() in alegranevada.

tet_radius_ratio() is used instead.

-For higher-order tet types, computing element volume metric more accuratly by subdividing tet into multiple subtets.
-Adding unit test.

Remove files no longer needed for finding python 2 and 3.

Allowing normalized inradius to be computed on tet14s and 15s.  It just ignores mid face and body nodes.

fixing bug in quad distortion metric so it now returns a reasonable result.  Adding distortion metric for linear quads.

adding a linear version of the normalized inradius for tets.  Previously would return 0.0 for linear.

Also cleaning up some of the other tet metrics

git-svn-id: file:///scratch/cjstimp/svn/cubit/CUBIT_SOURCE/trunk@416808 d353b8dc-a10e-11dd-9ca7-d543032e456e

Increment to version 1.3.

git-svn-id: file:///scratch/cjstimp/svn/cubit/CUBIT_SOURCE/trunk@416429 d353b8dc-a10e-11dd-9ca7-d543032e456e

git-svn-id: file:///scratch/cjstimp/svn/cubit/CUBIT_SOURCE/trunk@416167 d353b8dc-a10e-11dd-9ca7-d543032e456e

Bug fix for Mesh-4075. If U-submap overlaps still exist, then call interval matching a failure, regardless of whether the prior constraints are satisfied. Fixed a 25 year old bug in the check for U-submap overlaps: in some cases the hardset intervals were not counted properly.

git-svn-id: file:///scratch/cjstimp/svn/cubit/CUBIT_SOURCE/trunk@416091 d353b8dc-a10e-11dd-9ca7-d543032e456e

Make pamgen symbols non-hidden so linking will work.

git-svn-id: file:///scratch/cjstimp/svn/cubit/CUBIT_SOURCE/trunk@415049 d353b8dc-a10e-11dd-9ca7-d543032e456e

git-svn-id: file:///scratch/cjstimp/svn/cubit/CUBIT_SOURCE/trunk@413624 d353b8dc-a10e-11dd-9ca7-d543032e456e

Adding timestep metric for hexes and tets in verdict.  Computation is now done inside verdict instead of in Cubit.  For hexes, only corner nodes are used in this calculation.  For tets10, all nodes are used.  For all other tets, only corner nodes are used.  

CubitInterface::calculate_timestep_estimate functions now exposed.  They were ifdef-ed out.

git-svn-id: file:///scratch/cjstimp/svn/cubit/CUBIT_SOURCE/trunk@413331 d353b8dc-a10e-11dd-9ca7-d543032e456e

git-svn-id: file:///scratch/cjstimp/svn/cubit/CUBIT_SOURCE/trunk@411439 d353b8dc-a10e-11dd-9ca7-d543032e456e