I have no real interest in game development or OpenGL/JOGL. However, I have struggled with SWIG for wrapping a number of C libraries/DLLs. From my read of GlueGen, it looks like a much better, and higher performance product that the other alternatives. Has anyone considered splitting out this project from JOGL? I think it would garner more support if it wasn’t integrated into the JOGL project itself.
GlueGen was designed with broader scope in mind than just the JOGL project, but splitting it out from the JOGL workspace isn’t a high priority right now. I believe the current JOAL workspace generates its glue code by expecting a sibling JOGL workspace with a built GlueGen to be present. Given other priorities it’s unlikely that we will perform this split anytime soon but if you want to do it and modify the JOGL build scripts to reference a side gluegen workspace I’d be willing to consider these changes for inclusion in the workspace.
I have looked into the code a little and this appears doable. I will look into this further. Do you have any pointers to docs or notes that might help me begin?
There isn’t a lot of documentation right now for either the GlueGen tool or its use in the JOGL build process. The various build.xml files in the make/ subdirectory show how GlueGen is compiled and used, and there are some comments and printouts indicating which portion of the autogenerated code each invocation produces. It’s a little complicated, but you shouldn’t have to change much. You might also want to look at the JOAL build.xml files, which were written later than JOGL’s and which auto-detect some things like which platform you’re building on.
Incidentally, the lack of documentation in general is a problem. Even if GlueGen is split off into its own workspace, I doubt there will be enough documentation for it to be used even by interested users, and there isn’t enough time right now for the core JOGL team to write docs.