Re: How about an SDL_RenderFillQuad function?
Sik the hedgehog <[email protected]>
| Newsgroups | gmane.comp.lib.sdl |
|---|---|
| Message-ID | <CAEyBR+WOEX94AB=5PqTqF7Bjjojppg=T=1k63iLTL=172CW+Tw@mail.gmail.com> |
2016-09-30 12:07 GMT-03:00, SeanOConnor <[email protected]>: > I just thought that adding a quad, triangle or arbitrary number of points > function would be so easy (as rects are drawn with triangles anyway) it > might as well be added, and I was surprised those functions weren't there > already. Because of the software renderer. That one is using dedicated algorithms instead of triangles. And honestly a good reason why we need triangle routines for it, because then at worst we can just build everything else on top of that. This is a problem that has been going on for a long while though. SDL supports multiple backends for the renderer (software, different OpenGL versions, different Direct3D versions), but they don't want to add a new function that doesn't work in all backends. And then what happens is somebody contributes a new function for one backend, but then nobody wants to bother implementing the function in the other backends, and the end result is that the function goes nowhere and gets forgotten. On top of the fact most people don't want to use the renderer anyway (they either use SDL_gpu or, more likely, they just go straight with their own OpenGL code). Actually heck, adding triangle functions would probably solve it for all backends.