DS

> > (Default on, because I think this is more important for the average user than correct transparency sorting.) > > A bit weird, considering that working transparency sorting was...

> A better approach would be limit the amount of extra drawcalls caused by transparency sorting.Then you'd sort e.g. the first 10 transparent "units" closest to the camera correctly and...

@sfan5 Idk, maybe I've done a stupid mistake. But it looks like at least for complex meshes with overlay tiles and blending, it looks awful:  For simpler nodes it...

We don't store on mapblock mesh generation to which node a triangle belongs to. One could round the vertex positions, or the triangle centers, to node positions. And then you...

> would it be difficult to track what node the triangle came from? Not really. But it would blow up the PR.

> > I'd prefer to go with the simpler first version. > > do that then. Done. (Other version is at , if someone wants that for testing.) > >...

Aren't CPPFLAGS the c preprocessor flags? CFLAGS and CXXFLAGS are already used by cmake as default for `CMAKE_C_FLAGS` and `CMAKE_CXX_FLAGS` on my machine. (Haven't tried if it also already works...

Reading node meta just requires a hashtable lookup, afaik. The issue is rather that the mapblock mesh maker thread has no access to node meta. And changing this would require...

Tested a tiny bit. Marked a dropped item with branding iron, then did `/lua branded("obj:1"):set_pos(here:offset(1, 0, 0))` (`branded` is just a shorthand for `testtools.get_branded_object`.) while flying in air. (Can also...

Thanks for the explanation! If the client prediction is correct, there should be no unsmooth jumps, ideally (there are, because it's not ideal). Even if the server step time is...