- fix overly aggressive match in the API value

- allow `INTELMPI*` generic value, this can be used to specify something
  like INTELMPI_custom and populate the corresponding wmake rule
  manually

STYLE: mention FOAM_BUILDROOT in wmake -help-full output

STYLE: adjust openfoam shell session welcome information

- adjust internal variable names to reduce collision potential

- improve handling of openfoam -etc=...

- update coded templates with qualified names

GIT: add in missing PatchFunction1 constant() method

- was missed in a previous commit

CONFIG: cleanup additional build-related env variables

- marks if the value is considered to be independent of 'x'.
  Propagate into PatchFunction1 instead ad hoc checks there.

- adjust method name in PatchFunction1 to 'whichDb()' to reflect
  final changes in Function1 method names.

ENH: add a Function1 'none' placeholder function

- This is principally useful for interfaces that expect a Function1
  but where it is not necessarily used by a particular submodel.

TUT: update Function1 creation to use objectRegistry

- previously had codeAddSup used for both incompressible and
  compressible source terms. However, it was not actually possible to
  use it for compressible sources since any references to the 'rho'
  parameter would cause a compilation error for the incompressible case.

  Added 'codeAddSupRho' to distinguish the compressible case.
  User must supply one or both of them on input.

- for compilers such as gcc and clang, may have several different
  variants installed on the computer. Use WM_COMPILER_CONTROL to
  specify the preferred variant.

  Eg,
      WM_COMPILER=Gcc
      WM_COMPILER_CONTROL="version=8"

      will compile with "gcc-8" and "g++-8"

  Good practice would be to tag output directory names with the
  version too. Eg

      WM_COMPILER=Clang110
      WM_COMPILER_CONTROL="version=11.0"

STYLE: modify message for change of gcc -> clang (darwin)

- it is now possible to include the selected mpi version in the
  top-level bashrc or prefs.sh file. For example,

      WM_MPLIB=OPENMPI-4.1.1  or  WM_MPLIB=openmpi-4.1.1

  after evaluation of the config.sh/mpi, this will define

      WM_MPLIB=OPENMPI-4.1.1  and  FOAM_MPI=openmpi-4.1.1

  During the wmake, the mpi-rules will first load the MPI 'family'
  rules (OPENMPI in this example) before trying to load
  version-specific rules if they exist.

  NOTE: the regular user-defined prefs system is unaffected by this
  change. This means it is still possible to use a file such as
  'prefs.openmpi' to define the preferred version instead or as well.
  However, it does mean inconsistent naming can be specified.
  For example, specify WM_MPLIB=OPENMPI-4.1.1 at the top-level but
  actually have FOAM_MPI=openmpi-4.0.6 in the prefs.openmpi file.
  This will make the value of WM_MPLIB misleading.

CONFIG: foamConfigurePaths support for sys-openmpi major version

CONFIG: cleanup any shadow env variables

- migrate to c++14 for most compilers *except* gcc.
  There are still many systems in use with gcc-4.8.5, which does not
  support c++14.

- initial rules for nvidia compilers (pgi is will soon be defunct).
  Not fully tested...

CONFIG: provide fallback value for the user directory name

- in containers may have an unset USER env variable.
  Default to 'user' to prevent ugly looking directory names.

- simplfies differences for OSX

- use orte-info to determine prefix/libdir for openmpi.
  This removes a run-time dependency on mpicc, which is actually
  only needed for building with MPI (not running with MPI).
  The corresponding openmpi devel package (deb/rpm) will not
  necessarily be installed on a particular system.

- retain mpicc logic if the new logic using orte-info does not
  deliver an answer. Final fallback to using 'orterun' to
  infer prefix/libdir.

- Additional logic for intel and msmpi to make it easier to
  locate these vendor packages within ThirdParty
  (ie, under ThirdParty/opt/...)

CONFIG: improve robustness

- add check for absolute path when adding PATH/LD_LIBRARY_PATH etc.

- prefix more variables with '_foam*' to prevent accidental overwrite
  of userspace shell variables when sourcing

- previously OpenFOAM library paths were set before adding in
  MPI-specific elements. However, this may mask some libraries.

  |  New library order  |  Old library order  |
  |---------------------|---------------------|
  |  user-lib (serial)  |  various 3rd-party  |
  |  site-lib (serial)  |  foam-lib (mpi)     |
  |  foam-lib (mpi)     |  ext-lib (mpi)      |
  |  foam-lib (serial)  |  system mpi-libs    |
  |  ext-lib  (mpi)     |  user-lib (serial)  |
  |  ext-lib  (serial)  |  site-lib (serial)  |
  |  various 3rd-party  |  foam-lib (serial)  |
  |  system mpi-libs    |  ext-lib (serial)   |
  |  dummy              |  dummy              |

Minor clean-up

- previously a function (unlike the csh version) but since bashrc and
  setup have been split -> replace with inline definition

STYLE: formatting/wording for openfoam starters

TUT: simplify controlDict modification, add default substitution

ENH: accept '/' for end-of-options terminator (etc/openfoam)

- makes the application or service more apparent.
  * eg.  /usr/bin/openfoam / blockMesh
  * vs.  /usr/bin/openfoam -- blockMesh

  Accept lone '-' as the end-of-options terminator, as per bash

- Adjust handling of openfoam '-c' option to flag that a command-string
  will appear, but continue with option parsing.
  Consistent with bash definition.

- adjust commented-out evaluation to avoid warnings.

  With code like this
  ```
  #if 0
  nxin    #eval{ round($nxin / 5) };
  #endif
  ```

  The handling of the "#if 0 / #endif" clause uses the plain ISstream
  parser to tokenize. This means that the "round(" is parsed as a word
  with a mismatched closing ')', whereas the "#eval" parser will slurp
  everything in until the closing brace and send it off as a string
  to the expression parser.

Andrew Heather authored 3 years ago and Mark OLESEN committed 3 years ago

    Computes a selected operation between multiple \c fieldValue function
    objects.

    The operation is applied to all results of each \c fieldValue object.

Note
    Each object must generate the same number and type of results.

Usage
    Minimal example by using \c system/controlDict.functions:

    multiFieldValue1
    {
        // Mandatory entries (unmodifiable)
        type            multiFieldValue;
        libs            (fieldFunctionObjects);

        // Mandatory entries (runtime modifiable)
        operation       subtract;

        // List of fieldValue function objects as dictionaries
        functions
        {
            region1
            {
                ...
            }
            region2
            {
                ...
            }

            ...

            regionN
            {
                ...
            }
        }

        // Optional (inherited) entries
        ...
    }

    where the entries mean:

      Property     | Description                         | Type | Req'd | Dflt
      type         | Type name: multiFieldValue          | word |  yes  | -
      libs         | Library name: fieldFunctionObjects  | word |  yes  | -
      operation    | Operation type to apply to values   | word |  yes  | -
      functions    | List of fieldValue function objects | dict |  yes  | -
    \endtable

    Options for the \c operation entry:

       add           | add
       subtract      | subtract
       min           | minimum
       max           | maximum
       average       | average

Deprecated fieldValueDelta

- The fieldValueDelta function object was originally written to compute the
difference between two fieldValue-type function objects. The multiFieldValue
object name better describes its purpose whilst being able to operate on an
arbitrary number of fieldValue-type objects.

TUT: add missing restore0Dir (finiteArea)

STYLE: doc spelling

Kutalmış Berçin authored 3 years ago and Mark OLESEN committed 3 years ago

CONFIG: update completion_cache

- additional -serial/-parallel option:
  prefer Allrun-serial or Allrun-parallel if available

- optional -output=DIR to preserve output

ENH: report missing tutorials/ directory in RunFunctions

- in some cases, additional dictionary inputs are useful for extending
  the input parameters or functionality of dynamic coded conditions.

  Typically this can be used to provide a simple set of dictionary
  inputs that are used to drive specific code, but allows changing the
  inputs without causing a recompilation.

  Accessed with this type of code:
  ```
  const dictionary& dict = this->codeContext();
  ```

boundary conditions and function objects:

* specify an additional codeContext dictionary entry:
  ```
  codeContext
  {
      ...
  }
  ```

PatchFunction1:

* The code context dictionary is simply the dictionary used to specify
  the PatchFunction1 coefficients.

  To replicated persistant data, use local member static data.
  Eg,
  ```
  code
  #{
      // Persistent (Member) Data
      static autoPtr<Function1<scalar>> baseVel;
      static autoPtr<Function1<vector>> baseDir;
      ...
  #}
  ```

fvOptions:

* currently not applicable

- meshTools include/library for many (most) coded items

- add PatchFunction1 include for coded BCs to provide ready access
  to Function1 and PatchFunction1

- use common base for handling scheme lookups.
  Extended to support local injection of schemes into the dictionaries
  during runtime.

- largely as per patch from Jong-Gwan (Jason) Do

NB: the intel-one setup adds in paths for intelmpi.
    Its mpicc version does not harmonize with the OpenFOAM
    system openmpi setup (using mpicc --showme:link).

    Needs adjustment, or use intelmpi instead.

- update name mappings for newer gcc, clang versions

Mark OLESEN authored 3 years ago and Andrew Heather committed 3 years ago

CONFIG: bad csh syntax

It was only looking for faces that were used in both
endpoints but not actually checking whether they were indeed
an edge (== consecutive vertex) in all faces. So if one
face had an additional crossing edge and another didn't it
would find more edgeFaces than the proper
'primitiveMesh::edgeFaces()' routine.
This occasionally happened inside snappyHexMesh
(e.g. motorBike tutorial)

Tomoki, Karatsu authored 3 years ago and Mark OLESEN committed 3 years ago

- previously had a very old (likely irrelevant) setting for solaris
  systems only.

- support site-specific customization.
  Eg, using etc/config.{csh,sh}/prefs.fjmpi

- remove erroneous shell redirects present in cshell files

- if the system/decomposeParDict is missing, skip check for matching
  number of processor dirs. This can make job dispatch easier.
  Does not apply if -decomposeParDict was explicitly specified.

STYLE: adjust naming of host/slaves in argList

- can use either command-line option "-load-fields" or dictionary
  entry "readFields" to specify field names to be preloaded.

  Essentially the same functionality as with a readFields function
  object but with a lot less typing.

- tutorial examples provided by Ryan Danks <ryan.danks@rwdi.com>
  illustrate using setExpr* utilities to calculate a quantity
  as a post-processing step.

- location and naming more consistent with other controls

    old: DebugSwitch  'level'
    new: InfoSwitch   'outputLevel'