Graf Zahl wrote:Correct. But I have started work on it, the 2D part should be done. But at the moment the main vertex buffer that's being created will not work on a core profile and that may just be reason enough for failure.
If the screen quad render is still using the main vertex buffer that'd pretty much Game Over it.
I've updated the pull request with a few more fixes I noticed when testing on my Mac. In details:
Changed the macOS target to query Cocoa for the window client size. Made it never return zero because the Windows target always has a known window geometry and it seemed simpler to make sure that always remained true than add if statements elsewhere.
Made the window resizable on macOS because the new shared viewport code can handle this situation.
Fixed an issue where Begin2D would be called before SetOutputViewport had been called.
Clear the render buffers before first use. This may be the cause of the 'black screen' because blend operations might not like garbage input when drawing 2D.
Added a check if invalid screen dimensions are requested to error out early rather than generating GL errors or invalid frame buffer messages.
Graf Zahl wrote:Correct. But I have started work on it, the 2D part should be done. But at the moment the main vertex buffer that's being created will not work on a core profile and that may just be reason enough for failure.
If the screen quad render is still using the main vertex buffer that'd pretty much Game Over it.
Correct. But as I said, it's work in progress, so no surprises here.
Okay I'm done making my changes to PR 80 (https://github.com/coelckers/gzdoom/pull/80), after doing my last cycle of testing to try make sure there are no regressions. It should be better than what we have in master now from what I can tell.
dpJudas wrote:If the screen quad render is still using the main vertex buffer that'd pretty much Game Over it.
Correct. But as I said, it's work in progress, so no surprises here.
I see, was confused by removal of immediate mode functions.
dpJudas wrote:Okay I'm done making my changes to PR 80 (https://github.com/coelckers/gzdoom/pull/80), after doing my last cycle of testing to try make sure there are no regressions. It should be better than what we have in master now from what I can tell.
dpJudas wrote:Think we can close this one. (I sleep better at night the less tickets are open due to me )
For the things you've done for GZDoom I think everyone is willing to endure a few extra bug reports. Besides you can never perfect something if you don't know what's wrong with it.
You've fostered a huge transition for GZDoom from one rendering style to another to allow more features. And that has paved the way for other neat post-processing shaders. To think there wouldn't be some little dings and dents along the way would be a little crazy at the very least.
Thanks Eruanna. Still, the black screen and viewport bugs are annoying me because they are so central to gzdoom. If it was some new feature that had some bugs in it I'd be okay with it, but this part of gzdoom just has to work. Post processing effects are useless if you're greeted by a black screen!
)