I was speaking in generalities. Too often things truely are overlooked.
llama said:
Well his code was not ignored.. it was just established there are several ways for fixing this problem. It has to do with multithreading, and not with this thread.
Which is why I did not create a issue track. When I say I was ignored I mean that I felt that there was no attempt to select a solution.
OK, sorry Badmi for misunderstanding you (and Mojo too).
I'm still commited to making multithreading work, so whatever solution it will be (Badmi or mine or something else) we'll see. It's just that up to recently some related code was being worked on (just checked in this fix myself for example) so I'm waiting till these things settle down a bit first.
any update on this issue?
i don't have the time/insight to fix this. i barely find some time every week between school and work to write a bit of code for my own project.
imho this is a major bug, as textures are reloaded again and again, and therefore memory is used up at the same pace (not to mention the slow startup caused by reloading).
i just want to know if any of the dev people see this the same way. 
I'll have a look at this today, sfera.
thanks a lot
Ok, I've modified TextureManager's code to re-enable texture reuse. In the process, I've also removed the constant recreation of a TextureState in the loading code as well as removed the immediate application of the texture. A valid (non-dummy) Renderer needs to have been created prior to using TextureManager if you wish to use S3TC compressed textures, but otherwise it should be cool. I'll have this checked in later today if possible. (Our commit process here is wonky.)
don't worry, i'm on the cvs mailing list 
thanks a lot for checking in the changes to texture manager.
Sure thing.