Time  Nick      Message
00:10 lhofhansl Hello. Please consider my comments om #13126. Displaying 1.234e+04 instead of 12345 for something like "blocks loaded" isn't all that helpful.
00:10 ShadowBot https://github.com/minetest/minetest/issues/13126 -- Revisit F6 statistics formatting by SmallJoker
18:10 sfan5     hmm yeah that's not optimal
18:20 pgimeno   about #13151, isn't the code likely to be optimized away by the shader compiler once all uniforms have been sent? it only depends on constants and uniforms
18:20 ShadowBot https://github.com/minetest/minetest/issues/13151 -- Detect NaN and Inf initial exposure by x2048
18:31 sfan5     I don't think any graphics driver would inline the value(s) of uniforms, but if that happens what's the problem?
18:36 pgimeno   no problem at all, I was implying that x2048's solution may not incur any performance penalty
18:37 pgimeno   I'm a little nervous about the NaN detection though, graphics cards don't respect the IEEE standard semantics because that goes against performance
18:40 pgimeno   also, x-1 == x is true for big values that aren't necessarily infinite (for single-precision floats it's true for x=16777218 on)
18:43 pgimeno   sorry, for multiples of 4 starting with 16777220