When using sfml with mixed rendering, as in you have your own implemented openGL rendering which happens before or/and after calls to the sfml draw() method, it might often mess things up. For instance I know many people who use sfml to handle events/creating a window and drawing text while taking care of the rest of the rendering by themselves. Most if not all of those people have run into problems with the text not displaying properly, or their own rendering breaking after trying to draw text. I myself too have had problems with this. Apparently this is due to openGL states becoming changed and not reset by the draw method.
This seems to me as a quite big limitation when you want to use sfml in any project which needs any form of rendering which is not able to do with the sfml shapes/sprites classes and this is not a rare case. Even now, knowing this problem I still don't know how to handle it properly; which states to set before and after the draw method and it is often frustrating to solve.
What is the thought behind this behaviour and is it fixable? Perhaps it is intended even? To me it would make sense that sfml would set the things it needs for rendering, and then reset it back to the states they had to not conflict with other components. If this would add too much of an overhead, perhaps the RenderWindow could have a flag for it? Or as the very least, would this be documented so that the users would properly know which states they should themselves reset after calling the sfml draw method? It is a way too common problem.
