Add WaveTank module + additional C binding interface improvements

[rafmudaf]

Add WaveTank module and C binding interface utilities and improvements

This pull request primarily focuses on the development of a module for driving a wave tank experimental setup coupling AeroDyn, InflowWind, MoorDyn, and SeaState to model the loads due to waves and current on a MHK turbine floating structure. To that end, the primary addition is the WaveTank module in the glue-codes/labview directory. A SeaState C-bindings interface is extended (starting from #2720) to include the full initialization sequence and the time marching simulation driven by the experiment controller (LabVIEW). To support the C binding development, I've also extended the new NWTC_C_Binding module (see #2720) with utilities for working with C and Fortran strings plus error handling.

This pull request depends on #2719 and #2720. Note that this pull request is a Draft. I will continue to update the PR description as I make more changes. I'll remove this line and change the status to "Ready for Review" when it's at a stable point.

C Bindings

Dealing with strings through the C interfaces is nuances, so I've introduced some tooling and patterns to make this more clear and consistent.

For context, a C-style string is represented in Fortran as CHARACTER(KIND=C_CHAR) :: ErrMsg_c(IntfStrLen). This is a character array of kind C_CHAR from the ISO_C_BINDING intrinsic module with one dimension of length IntfStrLen. This representation is different from the Fortran string, CHARACTER(IntfStrLen) :: ErrMsg, which is a string of length IntfStrLen (not an array). Importantly, C strings in Fortran must be terminated with C_NULL_CHAR from ISO_C_BINDING before being used by a function or subroutine expecting C-style strings: ErrMsg = " "//C_NULL_CHAR.

Since the C binding interfaces operate on both Fortran and C style strings, I've introduced the _C and _F suffix to the error message buffer variables and most other strings crossing the interface for any subroutines that have C bindings. The variable mapping and function is as follows:

Old var	New var	Description
ErrStat2	ErrStat_F2	Subroutine-local error
ErrStat	ErrStat_F	Module-global error
ErrStat_C	ErrStat_C	Error to be passed to the calling code

ErrStat_F2 should hold errors from individual commands, and then it should be incorporated into the module-global error with CALL SetErrStat( ErrStat_F2, ErrMsg_F2, ErrStat_F, ErrMsg_F, RoutineName ). Prior to returning control to the calling code, the _F error variables should be transferred to the _C variables with CALL SetErrStat_F2C(ErrStat_F,ErrMsg_F,ErrStat_C,ErrMsg_C).

An additional error handing utility is added, SetErrStat_C2C, to support glue code modules written in Fortran and combining C binding interfaces such as the WaveTank module. This subroutine expects a C-style error message string for both the local and global error message.

[andrew-platt]

should this be included?

[rafmudaf]

Yes, these files are what make the files within a directory importable by the Python interpreter

[andrew-platt]

Should this go in r-test?

[rafmudaf]

Yep it's coming soon. Just a heads up that this is very much in a draft stage. I just wanted to be able to write notes as I make changes. I've made note of this in the PR description, and I'll update the status when it's mostly complete.

[andrew-platt]

I'll hold off on reviewing further until the construction dust settles a bit more :)

[rafmudaf]

@andrew-platt Is it correct that this is checking whether errStat is 0? Where it's called above (line 342), the command is setting the error to ErrStat2. Since ErrStat is initialized to 0, I think this function would only correctly error when NumTurbines is incorrect since that's the only place that Failed() is called.

If this is indeed incorrect, I'm happy to rework it, but I just wanted to double check that I'm not misreading it.

[andrew-platt]

Yes, this is correct for Failed0(). We use this on allocations of arrays where a non-zero error status returned indicates failure to allocate.
For the Failed() routine, we use SetErrStat.

[RyanDavies19]

We should add error checks here that ensure MD is only being run w/ one coupled body, unless we change MD c-bind to be more flexible

[RyanDavies19]

Need to check the following are true:

p%nCpldBodies(1) = 1
p%nCpldRods(1) = 0
p%nCpldPoints(1) = 0

[deslaughter]

@rafmudaf Could you update the documentation for running the regression tests that use the pyOpenFAST module? It took me a bit to figure out that GitHub actions does pip install glue-codes/python/. and then executePythonRegressionCase.py imports it. Perhaps a better alternative would be to have executePythonRegressionCase.py add glue-codes/python/. to sys.path so it's always using the version directly from the repo. Sorry if this is the wrong place for this feedback, I figured this PR was related to #2720.