Hi David and others,

I'm (in between working on the Voxeliens Linux port) trying to get the SWIG bindings up and running again and I'm making good progress. One problem I've hit though is the deprecated constructors available in LargeVolume and SimpleVolume that take 3 uint_32s. SWIG seems to be trying to wrap the constructor since it doesn't know it's deprecated and is failing. This constructor was deprecated last March and so anyone using the last three snapshots or Git directly will not be using this constructor.

I think it's safe to say that the danger has passed and it should be safe to remove it now and so I wanted to make sure there were no objections.

In the future we're planning on doing proper PolyVox releases (with version numbers and everything!) and so at that time we'll have a proper procedure for dealing with deprecations (e.g. making sure we leave it at least one release before removing).

_________________
Matt Williams
Linux/CMake guy

Yep, you can take that out. It doesn't even do anything, it just there to ensure people see the message to update their code.

Ok, I've now removed the constructor. I've also re-enabled a bunch of the Python bindings so that it is now possible to create a SimpleVolume and extract a SurfaceMesh from it. To make sure this keeps working I've added a test in Python which should run with everything else on the CDash server tonight.

Take a look at the inside of the setUp() method in tests/TestSurfaceExtractor.py to see how it works.

_________________
Matt Williams
Linux/CMake guy

Wow, this is really cool! I think now we can start to get the bindings into a good shape and keep them there

Ok, well it seems I forgot to upload one of the files but I'll do that when I get home so that it will build correctly tonight.
Yes, I think that it would be worth keeping them compiling all the time from now on. As I go, I'll try to add some more tests and probably even an example (I was thinking a 3D Game of Life as a stress test ). Based on that I'll write a page in the documentation about how to use them in Python.

At some point we should probably also look at C♯ to make sure they compile there since I know we've had people using them in the past. It would probably require some help from the people who are using the bindings though.

_________________
Matt Williams
Linux/CMake guy