Time  Nick      Message
16:40 Krock     Will merge #16339 #16434 #16455 #16457 in 15 minutes
16:40 ShadowBot https://github.com/luanti-org/luanti/issues/16339 -- [Take 2] Prevent MapBlocks in generation from being unloaded by Montandalar
16:40 ShadowBot https://github.com/luanti-org/luanti/issues/16434 -- IrrlichtMt: Implement mip-mapping for RTTs by SmallJoker
16:40 ShadowBot https://github.com/luanti-org/luanti/issues/16455 -- Formspec: Implement Ctrl+Shift+Left/Right text selection by SmallJoker
16:40 ShadowBot https://github.com/luanti-org/luanti/issues/16457 -- Lua API: Unify server env checks and fix missing ones by SmallJoker
16:55 Krock     merging
17:00 Krock     done
17:12 sfan5     any opinions on whether #16267 is fine with the current approach or it should change to use RTTs?
17:12 ShadowBot https://github.com/luanti-org/luanti/issues/16267 -- Allow scaling nametag by distance and to set font size by sfan5
17:40 Krock     I have to admit that direct (pixel perfect) font rendering generally looks much better than any downscaled variants (RTT, for example)
17:42 Krock     unless someone comes up with a better scaling algorithm
17:44 Krock     from a performance perspective (and also RAM) I would however greatly appreciate the use of RTTs here
18:04 sfan5     let's merge it as-is then?
18:07 sfan5     having the RTT and old code separately could also be useful for testing later
18:24 Krock     ¯\_ (ツ)_/¯
18:24 Krock     other opinions...?
18:25 Krock     To be clear - there's no objections from my side on your PR because it is A) only few LOC and B) actually looks good
18:48 celeron55 what if the RTTs were used but they were scaled at 2x intervals?
18:48 celeron55 the question is, would that look better
18:49 celeron55 then if it looks better and you want more granularity, you could make a 1.5x version and then alternate scaling at 2x intervals from that
18:52 sfan5     sounds like mipmapping with extra steps?
18:53 celeron55 kind of
18:54 celeron55 not really though
18:56 Krock     by increasing the LOD bias of the mip-mapped' textures you could get better results, similar to what I tried in the RTT PR. However, more sharpness comes at the trade-off of jittering
18:56 Krock     or rather lowering (I used -1 instead of the default 0)