Time Nick Message 01:45 MTDiscord Has anyone asked an AI for suggestions, seems like we're spitballing. 02:02 MTDiscord It had some, not bothering to share. Only one was unique, the rest were "get transparent sorting right" 02:07 cheapie I turned transparency sorting off ages ago since it's so slow, I don't think I've noticed a single difference expect for one time when I put something together specifically to cause issues 10:08 sfan5 fwiw the default transparency sorting settings have been nerfed a lot, so it shouldn't cause issues by default anymore 10:09 sfan5 @luatic isn't that what we do? or do you mean with a trick like rendering depth-only first and then the whole thing again? 12:06 MTDiscord @luatic 13:36 MTDiscord sfan5: what we currently do is we have two separate passes for alpha="clip" and alpha="blend" meshbuffers basically 13:36 MTDiscord but if a mesh buffer with opaque fragments ends up in the latter pass, it is possible that these opaque fragments become invisible if another "blend" buffer with a smaller depth is rendered first 13:36 sfan5 solid and transparent pass, yes 13:37 MTDiscord if we rendered opaque fragments of transparent things in the solid pass already (rendering transparent things twice), we could avoid this 13:37 MTDiscord but i figure it isn't really worth the cost of doubling drawcalls for (semi)transparent things 13:38 sfan5 hmm 16:47 MTDiscord I'll admit, I tried to use transparent nodes this game jam, and had a cool effect but our defaults meant I had to switch to using particles for the effect. Ironically it came out better, but still another data point 20:08 sfan5 merging #16655, #16694 soon 20:08 ShadowBot https://github.com/luanti-org/luanti/issues/16655 -- Refactor skinned mesh weight data structure by appgurueu 20:08 ShadowBot https://github.com/luanti-org/luanti/issues/16694 -- Fix fog disable in OpenGL3 renderer by sfan5