Support for optional arguments to CCPP-ized physics schemes

[davegill]

The scheme which handles vertical mixing has individual constituent fields (vapor, cloud, and ice). These are used as both generic fields to mix and also as thermodynamically separate fields with different latent heat effects. Perhaps all three fields exist, perhaps none exist in dry dynamics testing. Additionally, there are other positive definite scalars which are eligible to be sent into the vertical mixing scheme: passive tracers, chemistry variables, number concentrations, etc.

The typical Fortran method of providing this capability is with "OPTIONAL" arguments.

[davegill]

@climbfuji @grantfirl @gold2718 @cacraigucar

There is already a keyword in the metadata for OPTIONAL. Here is a snippet of the YSU metadata file:

[nmix]
  standard_name = number_of_scalar_tracers
  long_name = number of scalar tracers
  units = none
  dimensions = ()
  type = integer
  intent = in
  optional = T
[qmix]
  standard_name = scalars
  long_name = scalars
  units = kg kg-1
  dimensions = (horizontal_loop_start:horizontal_loop_end,vertical_layer_dimension,number_of_scalar_tracers)
  type = real | kind = kind_phys
  intent = in
  optional = T

[gold2718]

@climbfuji, I believe @davegill has reviewed the meeting notes on this issue. We discussed this yesterday afternoon and this proposal would apply the optional metadata attribute only at capgen time.
Currently, if capgen (or prebuild) encounters an intent(in) or intent(inout) scheme variable for which no source is found (from the host model or from a previous intent(out) variable), an error is thrown.

The proposed functionality would be to not consider it an error condition if the scheme variable is marked optional. In that case, the variable would be ignored so that it would not be present inside the scheme.

[gold2718]

@davegill,
Is this still an issue? There are a couple of new options which may do what you need, active, and default_value.
We would like to remove optional from the metadata and from CCPP physics scheme Fortran files.

[climbfuji]

Closed via #347.

[gold2718]

This was accidentally closed by #347 .

[gold2718]

Replaced with #407