From: Mathias <Mat...@gm...> - 2006-07-03 19:13:04
|
On Monday 03 July 2006 08:23, Melchior FRANZ wrote: > You apparently missed that I had started to work on a new HUD > implementation in src/Instrumentation/HUD/. It is based on the old HUD, b= ut > heavily changed, cleaned up, extended. I spent at least 20 hours work on > that, and you didn't bother to tell me sooner what your plans were, > although I talked about this effort in several emails during the last > weeks. This was the poorest timing that one could have thought of. > > This new code still requires several more hours of work. Tapes don't > currently work for the F16, everything needs more cleanup etc. A few FIXM= Es > to fix, some TODOs to do. But now that I know that you are working on yet > another implementation, I consider every minute spent on mine a waste of > time. I don't like to work for the waste bin. No, I am not working on an other implementation. I have expressed that it would be very nice if the direct OpenGL calls coul= d=20 be put into the scenegraph. There is not wasted work here. The old code nee= ds=20 to be cleaned up anyway. I doubt that this should be implemented from scrat= ch=20 in one pass. I would more favour an incremental change, may be on top of wh= at=20 you have now. I have also told you that relying *entirely* on RenderTexture is not a good= =20 idea. Just because it does not work on *many* OpenGL cards. If you have either some code that prerenders the hud there is no big=20 conceptiual difference if you either render that to a texture and pin that= =20 texture onto a surface. Or if you set up some clipping planes and an=20 aprioriate transformation matrix and render the lines directly into the=20 framebuffer. Melchior, calm down a bit ... Greetings Mathias =2D-=20 Mathias Fr=F6hlich, email: Mat...@gm... |