I looked at using Lwjgl’s framework for creating a GLES 2.0 wrapper, but failed miserably (several times) at even working out how you’d even tell it there was a new library… instead I decided to port my Lwjgl3 example to an existing GLES 2.0 wrapper…
For a while now I’ve had a slim wrapper of GLES 2.0 for Java which is available here. Its far from perfect but does basically what I need of it (contributions and feedback welcome)
There are some minor differences which are basically to do with Java convenience, and is leading me towards thinking that rather than excepting different types of buffers as parameters for GLES functions, I should consider just having a utility to return a pointer (as a java long) for a buffer, this would simplify some things and also make the functions very much 1:1 with the native library.
So which is best – well to be honest neither really, obviously Lwjgl will enjoy wider support and cross platform native libraries (I used to know how to compile a windows DLL from Linux – I really must get round to reimplementing this…!) but mainly the difference is just that they are *slightly* different!