- findZone(), cfindZone() to return pointer to existing or nullptr if
  not found. This fits with methods such as findObject() etc for other
  classes and can simplify code without checks for '-1' as not found.

- use simpler constructors for empty cell/face/point zones

- change to a templated implementation instead of relying on
  the container's writeList() method.

  This inlines the generation while also adding the flexibility to
  define different delimiters (at compile time) without the
  performance penalty of passing run-time parameters.

- deprecate get(key, deflt) in favour of lookup(key, deflt).
  Method name compatibility with HashTable.

- deprecate operator().
  The meaning is too opaque and equally served by other means:

  - use get(key) instead of operator()(key).
    Const access whereas HashTable::operator()(key)
    creates missing entry.

  - lookup(key, deflt) - instead of operator()(key, deflt).
    Const access whereas HashTable::operator()(key, deflt)
    creates a missing entry.

- make Enum iterable to allow participation in range-for etc.

- were deprecated Aug-2018, but not marked as such

mattijs authored 4 years ago and Mark OLESEN committed 4 years ago

- non-uniform offsets are generated due to truncation errors,
  which can lead to problems later on (e.g. redistributePar).

  Detect if the offsets are close to being uniform.

Adds distance-to-surface as a pre-selection
to detect cells-in-gaps. Before it could only
use inside or outside but not distance.

This also is to do with redistributePar:
this uses subsetMesh to generate parts to
send to different processors.
2) related to 1558: make sure not to choose 'mapped'
patches to move the processor patches into so
we can use the mapper cloning and correctly
size additional data (e.g. offsets). This should
be generalised to hold for any patch type
holding local data ...

Potential problem with multiple faces. Rewritten to
use DynamicList.

For convenience - build finiteVolume first

corrections and improvements for Function1

See merge request !391

- For slow oscillations it can be more intuitive to specify the
  period.

ENH: separate mark/space for Square

- makes it easier to tailor the desired intervals.

BUG: incorrect square wave fraction with negative phase shifts

ENH: additional cosine Function1

STYLE: avoid code duplication by inheriting Cosine/Square from Sine.

- deprecated Feb-2018, but not marked as such.

  The set() method originally enforce an additional run-time check
  (Fatal if pointer was already set), but this was rarely used.
  In fact, the set() method was invariably used in constructors
  where the pointer by definition was unset.

  Can now mark as deprecated to catch the last of these.
  We prefer reset() for similarity with std::unique_ptr

  Eg,
  FOAM_EXTRA_CXXFLAGS="-DFoam_autoPtr_deprecate_setMethod"  wmake

- easier support for non-mandatory functions.

  In some boundary conditions it can be desirable to support
  additional functions, but not necessarily require them. Make this
  easier to support with a Function1, PatchFunction1 NewIfPresent()
  selector.

- support for compatibility lookups

- harmonize branching logic and error handling between Function1 and
  PatchFunction1.

ENH: refactor a base class for Function1, PatchFunction1

- includes base characteristics, patch or scalar information

ENH: additional creation macros

- makeConcreteFunction1, makeConcretePatchFunction1Type for adding a
  non-templated function into the correct templated selection table.
  makeScalarPatchFunction1 for similarity with makeScalarFunction1

- makes it clearer that these are constructors and not function calls

ENH: support construction of zero-sized IndirectList

- useful when addressing is to be generated in-place after construction.
  Eg,

      indirectPrimitivePatch myPatches
      (
          IndirectList<face>(mesh.faces(), Zero),
          mesh.points()
      );
      labelList& patchFaces = myPatches.addressing();

      patchFaces.resize(...);
      // populate patchFaces

STYLE: add noexcept for zero/one fields and remove old dependency files

COMP: correct typedefs for geometricOneField, geometricZeroField

COMP: add missing clone() for fanCurve member

STYLE: reduce padding in data members

- uses ocountstream for the output, which swallows all output.
  Improves portability

ENH: improved efficiency in countstreambuf

- xsputn() instead of overflow
- more consistent seek* methods

- use bitSet instead of boolList/HashSet. Simplify looping

ENH: add construct from components

STYLE: adjust action variable name for consistency

- add intermediate class topoSetFaceZoneSource
  and corresponding New() factories

ENH: PatchParticleHistogram: add a new cloud FO

See merge request !386

  Computes a histogram for the distribution of particle diameters
  and corresponding number of particles hitting on a given list of patches.

  A minimal example by using `constant/reactingCloud1Properties.cloudFunctions`:

  ```
  patchParticleHistogram1
  {
      // Mandatory entries (unmodifiable)
      type             patchParticleHistogram;
      patches          (<patch1> <patch2> ... <patchN>);
      nBins            10;
      min              0.1;
      max              10.0;
      maxStoredParcels 20;
  }
  ```

Miscellaneous code changes - Nov 20

See merge request !387

  DOC: heatTransferCoeff models: complete remaining header docs
  STYLE: heatTransferCoeff models: use auto specifier when appropriate

  Optionally, the Nusselt number (i.e. the ratio of convective to conductive
  heat transfer at a boundary in a fluid) can be output:

  ```math
      Nu = \frac{h L}{\kappa}
  ```

  where
  ```
    Nu | Nusselt number
    h  | Convective heat transfer coefficient of the flow
    L  | Characteristic length that defines the scale of the physical system
    \kappa | Thermal conductivity of the fluid
  ```

It was observed in a MPPICDyMFoam simulation involving a single particle
in a moving mesh that the barocentric trajectory of the particle follows
an unexpected path at some arbitrary instant in time.

The issue was tracked to "hitEqn()" where cubicEqn/quadraticEqn computes
one of the roots wrongly due to the discriminant limit we set, e.g. for:

    0x^3 + 1.4334549e-33 x^2 - 9.0869006e-10 x + 0.0027666538

Although the discriminant limit was carefully selected to avoid various
problems at the time, the new change is required more due to its exposition
to a wider spectrum of applications.

Darrin Stephens authored 4 years ago and Kutalmış Berçin committed 4 years ago

From OpenFOAM Foundation https://github.com/OpenFOAM/OpenFOAM-dev/commit/e4d89daf5de85f31c98012102c7dea2e29351ff2



The main issue here was that reconstructPar is serial but coupled() in
cyclicAMIFvPatch.C could return true if both sides of the patch was present
(this->size() && neighbFvPatch().size()). However, this would result in an
evaluate call in cyclicAMIFvPatchField. This would only work if both sides
were completely contained on the same processor. The change in logic prevents
coupled() from returning true when called in serial for a decomposed case.

Signed-off-by: Kutalmis Bercin <kutalmis.bercin@esi-group.com>

  ENH: actuationDiskSource: change two FatalError to FatalIOError

  The change helps to carry out and monitor
  parallelisation tests for FOs and topoSets.

Feature rationalize mpi configs

See merge request !389

- prefix FOAM_MPI and library directories with 'sys-' for system
  versions for uniform identication.

  WM_MPLIB      | libdir (FOAM_MPI)  | old naming |
  SYSTEMMPI     | sys-mpi            | mpi        |
  SYSTEMOPENMPI | sys-openmpi        | openmpi-system |

- prefix preferences with 'prefs.' to make them more easily
  identifiable, and update bin/tools/create-mpi-config accordingly

      Old name: config.{csh,sh}/openmpi
      New name: config.{csh,sh}/prefs.openmpi

- additional mpi preferences now available:
    * prefs.intelmpi
    * prefs.mpich
    ...

CONFIG: added hook for EASYBUILDMPI (eb-mpi), somewhat like USERMPI

- EasyBuild uses mpicc when compiling, so no explicit wmake rules are
  used

ENH: support different major versions for system openmpi

- for example, with

     WM_MPLIB=SYSTEMOPENMPI2

  defines FOAM_MPI=sys-openmpi2 and thus creates lib/sys-openmpi2

ENH: centralize handling of mpi as 'mpi-rules'

    Before:
        sinclude $(GENERAL_RULES)/mplib$(WM_MPLIB)
        sinclude $(DEFAULT_RULES)/mplib$(WM_MPLIB)

        ifeq (,$(FOAM_MPI_LIBBIN))
            FOAM_MPI_LIBBIN := $(FOAM_LIBBIN)/$(FOAM_MPI)
        endif

    After:
        include $(GENERAL_RULES)/mpi-rules

- also allows variants such as SYSTEMOPENMPI2 to be handled separately

- support prefs.compiler with better possibilities to provide
  overload values

STYLE: reduce code for WM_ARCH_OPTION (seldom-used)