3.0.0 (32-bit labels) vs dev (64-bit labels):

compressible/rhoPimpleDyMFoam/annularThermalMixer
  - Same mesh features, but two files stand out as very strange:
    - constant/polyMesh/owner
    - constant/polyMesh/faces
  
  - Oddly enough, the "constant/polyMesh/faceZones" is identical, including the
    flipMaps.
  
  - But according the "log.snappyHexMesh", the meshing had several issues on
    both versions... so it's natural that there are some issues.

  - The critical difference is revealed by a full checkMesh:

        --- 3.0.0
        +++ dev
        -             rotorBlades      540      682  ok (non-closed singly connected) (-0.0375 -0.0375 0.04) (0.0375 0.0375 0.08)
        -       rotorBlades_slave      540      682  ok (non-closed singly connected) (-0.0375 -0.0375 0.04) (0.0375 0.0375 0.08)
        +             rotorBlades      540      640  ok (non-closed singly connected) (-0.0375 -0.0375 0.04) (0.0375 0.0375 0.08)
        +       rotorBlades_slave      540      640  ok (non-closed singly connected) (-0.0375 -0.0375 0.04) (0.0375 0.0375 0.08)

  - Using the filter "Clean To Grid" in ParaView, which removes duplicate
    points, revealed that 640 is the correct number of unique points.
    Therefore, the excess points is actually a bug in 3.0.0!

  - Furthermore, when using the filter "Normal Glyphs" on the surface mesh for
    the patch "rotorBlades", the normals are all well represented when the
    mesh was generated with the v4 patch on OpenFOAM-dev, which didn't happen
    with 3.0.0.


heatTransfer/buoyantBoussinesqSimpleFoam/iglooWithFridges
  - Identical meshes, including points on all digits (ascii, precision 6).


heatTransfer/chtMultiRegionFoam/snappyMultiRegionHeater
  - The master mesh has a virtually identical mesh, although:
    - faces, owner, neighour are identical;
    - points are similar (numerical precision related issues)
    - cellZones are identical;
    - the faceZones have different flipmaps... which is what we changed with v4.

  - "bottomAir" has a nearly identical mesh, only the points differ very little
    between them.
  
  - Other regions have a different arrangement of patches, making it a bit
    difficult to assess directly from comparing the files.

  - "constant/cellToRegion" files are identical.
  
  - Full checkMesh for the main mesh and for each region only reveals that the
    point map is different enough to give some small differences in checks.
    For example:
  
        --- 3.0.0
        +++ dev
            All angles in faces OK.
        -    Face flatness (1 = flat, 0 = butterfly) : min = 0.9799881  average = 0.9999923
        +    Face flatness (1 = flat, 0 = butterfly) : min = 0.9797262  average = 0.9999923
            All face flatness OK.
        -    Cell determinant (wellposedness) : minimum: 0.9215865 average: 10.2842
        +    Cell determinant (wellposedness) : minimum: 0.9215865 average: 10.28414
            Cell determinant check OK.
        - ***Concave cells (using face planes) found, number of cells: 944
        -  <<Writing 944 concave cells to set concaveCells
        -    Face interpolation weight : minimum: 0.3264674 average: 0.4792975
        + ***Concave cells (using face planes) found, number of cells: 933
        +  <<Writing 933 concave cells to set concaveCells
        +    Face interpolation weight : minimum: 0.3263706 average: 0.4792962
            Face interpolation weight check OK.
        -    Face volume ratio : minimum: 0.1168785 average: 0.8916654
        +    Face volume ratio : minimum: 0.1167739 average: 0.8916526
            Face volume ratio check OK.

  - Residuals and Courant Numbers all seemed within the same scale of
    numerical differences.


incompressible/pimpleDyMFoam/propeller
  - Very hard to compare based on files, since it ran in binary mode by default.
  - Using "foamFormatConvert -constant" revealed the meshes to be identical
    (ascii, precision 12).


incompressible/simpleFoam/motorBike
  - Very hard to compare based on files, since it ran in binary mode by default.
  - Using "foamFormatConvert -constant" revealed the meshes to be identical
    (ascii, precision 12).

  
incompressible/simpleFoam/rotorDisk
  - Identical meshes, including points on all digits (ascii, precision 6).


incompressible/simpleFoam/turbineSiting
  - Very hard to compare based on files, since it ran in binary mode by default.
  - Using "foamFormatConvert -constant" revealed the meshes to be identical
    (ascii, precision 12).


incompressible/simpleFoam/windAroundBuildings
  - Identical meshes, including points on all digits (ascii, precision 6).


lagrangian/MPPICFoam/cyclone/
  - Very hard to compare based on files, since it ran in binary mode by default.
  - Using "foamFormatConvert -constant" revealed the meshes to be identical
    (ascii, precision 12).


mesh/snappyHexMesh/flange
  - Identical meshes, including points on all digits (ascii, precision 6).


multiphase/interDyMFoam/ras/DTCHull
  - Very hard to compare based on files, since it ran in binary mode by default.
  - Using "foamFormatConvert -constant" revealed the meshes to be identical
    (ascii, precision 12).


multiphase/interDyMFoam/ras/mixerVesselAMI
  - Very hard to compare based on files, since it ran in binary mode by default.
  - Using "foamFormatConvert -constant" revealed the meshes to be identical
    (ascii, precision 12).

multiphase/interFoam/ras/DTCHull
  - Very hard to compare based on files, since it ran in binary mode by default.
  - Using "foamFormatConvert -constant" revealed the meshes to be identical
    (ascii, precision 12).


multiphase/interPhaseChangeDyMFoam/propeller
  - Very hard to compare based on files, since it ran in binary mode by default.
  - Using "foamFormatConvert -constant" revealed the meshes to be identical
    (ascii, precision 12).


multiphase/interPhaseChangeFoam/cavitatingBullet
  - Identical meshes, including points on all digits (ascii, precision 6).