Unknown W. Brackets

It's happening in D3D11 as well. So far we have a report, and it's using flags=7, which means BINDFBCOLOR_MAY_COPY (1) was set. So that narrows it down: this happens when...

Is it really a bug, though? In this case, it's probably trying to use a depal shader, but because slow effects are disabled, it's not even doing that. I think...

Oh, I guess it only doesn't set it when binding, it still sets `reallyDirtyAfterDisplay` after the draw, right. So then SKIPDRAW_BAD_FB_TEXTURE does nothing, I guess. -[Unknown]

Yeah, I think we can close this now - looks like this shouldn't happen anymore. -[Unknown]

It's a shame [GL_AMD_depth_clamp_separate](https://www.khronos.org/registry/OpenGL/extensions/AMD/AMD_depth_clamp_separate.txt) is not more widely supported outside AMD. It seems like that would allow this to work perfectly. It seems like [desktop GL 3.2+ clamping](https://www.khronos.org/registry/OpenGL/extensions/ARB/ARB_depth_clamp.txt) works the...

Added note above that Driver 76 is affected by this: it uses minz=00000, maxz=0012c, depth clamp=on. We could change the heuristic to `gstate.getDepthRangeMin() == 0 && gstate.getDepthRangeMax() == 65535` which...

I guess the Right Way to implement this is using a geometry shader. It should happen before clipping, so we can clamp and let the rest be clipped. This of...

Well, `&&` and `||` are both wrong, it's just a question of which is wrong in more cases. Since you're trying to fix a specific issue, it's obvious there are...

> It would in fact not deform triangles I believe, though yes of course, early z test breaks. Depending on the values of z and w, for a non-flat triangle...

I think that might work... just worried about the number of games that use depth clamp kinda all the time, and the early depth tests. That said, I had another...