Use RavenHydroFramework from official GitHub source

Zeitsperre commented 8 months ago

Pull Request Checklist:

[x] This PR addresses an already opened issue (for bug fixes / features)
- This PR fixes #23
[x] Tests for the changes have been added (for bug fixes / features)
- [ ] (If applicable) Documentation has been added / updated (for bug fixes / features)
[x] CHANGES.rst has been updated (with summary of main changes)

What kind of change does this PR introduce?

Points the RavenHydroFramework source files to the official GitHub project
Updates the version of Raven to 3.8
Updates the CMake configuration to use git-based clones
Both the version number and the description are now dynamic (using hatch-fancy-pypi-readme and setuptools_scm).
Support has been extended to Python3.12
Now uses scikit-build-core>=0.8.0

Does this PR introduce a breaking change?

Yes. The copied source files have been removed. There is more integration with the actual project source files.

Zeitsperre commented 7 months ago

@huard

I've spent way too much time on this today. Here's what I've figured out:

Running the cmake commands directly on the upstream codebase RavenHydroFramework works fine (so long as the STANDALONE and COMPILE_EXE flags are there).
When compiling things here, it's clear that the functions defined in RavenMain.cpp are not being passed to other files. ExitGracefully() is only found in a .cpp file and not in a header, and this file isn't included in a bunch of .cpp files that make references to that function.

Unsurprisingly, if you copy and paste the following into the top-matter of all .cpp files, the ExitGracefully() error disappears:

  #ifdef STANDALONE
      #include "GracefulEndStandalone.h"
  #elif BMI_LIBRARY
      #include "GracefulEndBMI.h"
  #endif

I've been trying to read up on ways of passing the objects from RavenMain.cpp to all other files, but I have no idea if that's even what we want. I get the impression that the problem has to do with scopes (perhaps vanilla cmake is a lot less strict about ensuring that files are included explicitly?).

@analytophile am I on the right track here?

huard commented 7 months ago

@Zeitsperre

I suspect this is a false lead. Your installation setup works on master, and the code compiles fine using the original cmakelist. I believe there is a quirk related to the source files location introduced in the git fetch mechanism that we need to identify. I'll take a look later this week.

Zeitsperre commented 7 months ago

Seems to be working! The only issue I can see is that the model reports a non-fatal error:

 3.8.0 w/ netCDF
============== Early Gracefully Exit ==============
Error Statement: Version check
Error Code: 7
---------------------------------------------------

@huard any idea what that could be?

EDIT: It's nothing. Looks like it's simply the exit code for SIMULATION_DONE:


enum exitcode
{
  BAD_DATA,       ///< For bad input provided by user (requires immediate exit from program)
  BAD_DATA_WARN,  ///< For bad input provided by user (requires shutdown prior to simulation)
  RUNTIME_ERR,    ///< For runtime error (bad programming)
  FILE_OPEN_ERR,  ///< For bad file open (requires immediate exit)
  RAVEN_OPEN_ERR, ///< for bad RavenErrors.txt file open
  STUB,           ///< Stub function
  OUT_OF_MEMORY,  ///< When out of memory
  SIMULATION_DONE ///< Upon completion of the simulation
};

Zeitsperre commented 7 months ago

Needed to set the fetch-depth for setuptools_scm. See: https://github.com/pypa/setuptools_scm/issues/414

Ouranosinc / raven-hydro