GitLab

ENH: downgrade error on type-mismatch to a warning

- Ensight data files generated by OpenFOAM will start with the name of
  the data type (Eg, "scalar", "vector" etc), but this convention may
  fail for data files that have been generated elsewhere.

- relocate BuildIssues into doc/

- foamExec was removed as part of commit f308aa61 but
  overlooked the fact that it is partly still needed by foamJob.
  When running in parallel, it is used to source the etc/bashrc env.

  Reinstated a stripped down form, which has been relocated to
  bin/tools directory (it should not be directly called by the user).

  The previous version switching functionality has nonetheless be
  dropped since it only worked only when a rigid naming convention for
  OpenFOAM installations was followed.

ENH: add foamJob log=, -log-app, -no-log options, improved coding quality

edge it sets the globalEdgeOrientation

See #1367

- synchronization, reduction only makes sense on processor-coupled
  patches. Since cyclic baffles are within a single processor domain,
  they are not reduced. So need to skip the sanity test for these.

- the use of 'paraview --version' can be fail if the build host
  doesn't have the necessary graphics. For this case, try to obtain
  the ParaView API number from the associated include directory.

- the read offset missed the tag_ member entirely and thus the entire
  particle information would be corrupt (incorrectly interpreted) as
  well as potential violation of adjacent (trailing) memory locations.

- now suppress any '#line' if the input number number is invalid
  (ie, an empty set of tokens)

- need to be more generous when prefiltering the cell selection

- now catch these and emit a warning.
  Still need to investigate the root cause in the caller(s) or regionSplit.

BUG: DimensionedField - register new field if it does not have the same name as the copy - see #1348

- was missing range setting for the colour lookup table

- the emissivity was ignored for 'hpTa' when there is no solid resistance

STYLE: adjust debug statement in externalWallHeatFlux

Fix provided by Gavin Ridley.

- used fallback of 0 instead of the results time.
  This discrepancy caused the case file to have two timesets that
  only differed by the first (incorrect) entry.

- when running in serial but within a processor directory,
  argList::globalPath() is to be used instead of Time.globalPath()
  For other cases there is no difference.

- still had old WM_PROJECT_VERSION settings instead of FOAM_API
  for FOAM_SITE_APPBIN and FOAM_SITE_LIBBIN locations.