Time  Nick       Message
00:18 andriyndev Regarding the size of the patch, I agree that the patch is very large. A large part of it is AI-generated tests which I haven’t reviewed yet, so I think, they should become smaller after this. But the “production” part of the code, I think, is all necessary to implement the current features. I think, if the concept is approved, the patch
00:18 andriyndev should be split into several parts, the first part of which would be implementing support for only overriding light (or another most trivial property) and applying them for nodes but without support for overriding itemstacks. I could also write a paper or make a presentation which explains all potentially non-trivial concepts and why I used them. I
00:18 andriyndev could also write a list of the nuances which might cause controversy, to discuss them before starting the code review.
00:31 ZenonSeth  left one comment on #17705 PR, I think it's somewhat important but can make a tiny follow up PR after this too, i think if this optimization goes in 5.18 the follow up should too imho
00:31 ShadowBot  https://github.com/luanti-org/luanti/issues/17705 -- [no squash] Don't retry CDB update check + check response codes by sfan5
00:56 andriyndev Regarding overriding tiles, it required a lot of changes due to current organization of pre-processing tiles. In particular, the process of pre-processing the textures depends on other parameters such as drawtype, alpha, waving. For example, for some drawtypes it might enforce the texture to be opaque depending on client settings, and it
00:56 andriyndev determines, which shader to use for the given texture, and the shader is also a part of the TileSpec. Also, the texture might be either “ordinary” or array one, and it’d also require a different shader. So, we cannot just substitute the texture without breaking all these dependencies. Another nuance - the minimap color, which is calculated by
00:56 andriyndev mixing the main and overlay tiles, and is also calculated for every node during preprocessing. It’d also require a special handling.
14:21 rubenwardy Would be handy for the client to show messages to the user if a non 2xx status is returned with json and {message: }
14:22 rubenwardy CDB dialog specifically
15:05 sfan5      ZenonSeth: you can send a PR to my PR's branch to get your changes included
15:05 sfan5      alternatively tell me from where to cherry-pick a commit
15:06 ZenonSeth  I'll make a PR real fast to your branch
15:23 ZenonSeth  pr had to be made in your repo, afaik i can't do it on luanti's main repo
15:25 sfan5      yes
15:26 ZenonSeth  it's there - pr 6
16:45 Sheriff_U3 This comment in lua_api.md is incorrect: https://github.com/luanti-org/luanti/blob/master/doc/lua_api.md?plain=1#L12201 That position is actually in the center of the screen.
16:53 rubenwardy The element, not the screen
16:53 rubenwardy Not a very helpful comment though
16:54 Sheriff_U3 position is used for the screen though. https://github.com/luanti-org/luanti/blob/master/doc/lua_api.md#HUD
16:55 Sheriff_U3 I ran across this while creating a gui compass.
16:55 Sheriff_U3 (in the hud)
17:06 rubenwardy The position is where the top left of the element is
17:06 rubenwardy So (0.5,0.5) puts the top left of the element to the center of the screen
17:07 rubenwardy But, as said, unhelpful comment
17:08 ZenonSeth  i can add it as a change in my doc improvement pr
17:08 ZenonSeth  try to improve the comment
17:15 ZenonSeth  er default alignment is (0,0) though so only if alignment was 1,1 would its top-left corner be centered? Sheriff_U3: can you try that with 1,1 to check im not going crazy?
17:20 sfan5      release on thursday?
17:38 ZenonSeth  sounds good to me
17:45 Krock      I'll be around.
18:51 sfan5      merging #17699, #17705 in 10m
18:51 ShadowBot  https://github.com/luanti-org/luanti/issues/17699 -- Update credits for 5.18.0 by sfan5
18:51 ShadowBot  https://github.com/luanti-org/luanti/issues/17705 -- [no squash] Don't retry CDB update check + check response codes by sfan5
18:51 Krock      currently compiling to test the latter. but fine :)
18:51 Krock      2026-10-28 19:51:31: ERROR[CurlFetch]: HTTPFetch for https://content.luanti.org/api/updates/?type=mod&type=game&type=txp&protocol_version=54&engine_version=5.18.0-dev returned response code 503
18:52 Krock      this test was quite quick tbh
18:58 ZenonSeth  i didn't leave a comment but i also tested it in client before approving, its what i normally do anyway for approving
19:11 sfan5      milestone is empty now
19:11 sfan5      did we already bump the protocol?
19:12 ZenonSeth  looks like yes
19:12 ZenonSeth  LATEST_PROTOCOL_VERSION = 54 + the edit in misc_s.lua
19:16 Krock      yes it was. no formspec version bump, though. is that necessary?
19:16 Krock      might be helpful because there's a few new elements
19:17 Krock      but it's not strictly needed because the version is technically only required to increase when new args are added to existing elements
19:18 sfan5      we have no rule on that
19:18 sfan5      however https://github.com/luanti-org/luanti/commit/041a7a600 bumped it and it's in this release
19:24 Krock      thank you :)
19:24 Krock      seems like I missed or forgot about that