Time Nick Message 00:25 erlehmann josiah_wi show and tell 00:26 erlehmann pls 00:30 MTDiscord Mac provides you with a snake variable I'm certain, do an if statement on that variable 00:30 MTDiscord CMake* 00:31 erlehmann CMake & Cheese 00:31 MTDiscord eclehmann, I will be putting a minimal reproducible example and a solution in the bug report hopefully by tomorrow evening. 00:32 MTDiscord If you want to see the real deal the fix is in irr 79 00:32 erlehmann cool 00:32 erlehmann i am curious about the minimal reproducible example 00:33 erlehmann and about how you got to it 00:34 MTDiscord So am I. I have a good hypothesis but I haven't proven it. I have a plan to reproduce it in < 10 lines but I still need to see whether it will work. 00:49 MTDiscord So much for that. I'll need to copy the Irrlicht example and whittle it down. 02:14 MTDiscord Hey folks, probably pyrollo/small joker/rubenwardy/sfan5: I'm working with hypertext elements, using my markdown2formspec mod. I'm running into what seems to be max size issues, where the hypertext scrollbar stops super early (heihgt of 40 is where it stops, but should stop around 400). I'm using lua_api.md as my test example 02:14 MTDiscord any ideas as to why that scrollbar may be stopping early? 02:16 MTDiscord Hurm, interestingly enough, it works if I copy a fully working example multiple times. Let me see if I can recreate the issue with known parse-able markdown 02:17 MTDiscord hurm. never mind. It must be something I'm escaping wrong or some other error, because I can use known good markdown of equal length and the scrollbar works. 02:54 kilbith https://github.com/minetest/minetest/pull/7002 02:54 kilbith please reopen 02:54 kilbith I can make a Skyrim-like time setter with this; and it has an approval already 11:09 nrz what is the purpose of this ? you have a time set for each player ? world works how ? as some things are related to time 12:29 MTDiscord Definately useful. One usecase: survival server, admin is building and wants daytime, but wants to keep everyone else in the dark. Another usecase: a special area where it is always night, or sunset, or 5 minutes till teatime :) 12:31 MTDiscord So is there any explanation of which time the server would use to calculate light levels? 12:31 MTDiscord Because remember--light levels and time of day are mod-accessible 12:33 MTDiscord Yeah the discussion says it breaks those hmm 12:34 MTDiscord An idea would be to allow to set an area to a specific time, and also allow to set one client to a specific time regardless of server time 12:35 hlqkj add timezones to minetest! :) 12:36 MTDiscord I suppose get_timeofday() would need a pos argument then 12:37 hlqkj jokes aside, area-based time (not only phase but maybe also speed and day/night ratio) isn't such a bad idea: different realms could have different day/night cycles 12:43 MTDiscord yeah, I'm not opposed to that personally. I think it resolves the issues the solution above has and creates some interesting possibilities 12:55 MTDiscord Can't you set light for a single player currently? 12:58 MTDiscord I don't believe so 13:05 MTDiscord Yes, see ptime mod 13:11 MTDiscord hm, weird guess it does exist 13:17 hlqkj playerref::override_day_night_ratio() ?? 13:21 MTDiscord Ya 13:35 sfan5 so I guess you can already do that if you manually update the time 13:44 hlqkj yup can be done, i messed with this stuff, trying to add different realm times to MT's real control mod 13:45 hlqkj a bit tricky though, I never finished the job (as usual lol) 13:49 hlqkj laters all, have a nice time! 14:27 MTDiscord Why don't we just allow controlling lighting (param1) from Lua? 14:35 sfan5 lua can write to param1, there's just no way to prevent the server from recalculating the lighting by its own standards 14:40 MTDiscord #11641 14:40 ShadowBot https://github.com/minetest/minetest/issues/11641 -- fix integer overflow in mapgen by JosiahWI 14:42 MTDiscord This is low priority, but it's also small and should be easy to review. 14:43 erlehmann i think the 1min test is not too obnoxious 14:44 erlehmann (bc “verified no change in output on my system” is too vague. which system/compiler/processor etc.?) 14:47 MTDiscord When every single other unit test is 10ms a 1min test is going to be extremely obnoxious. I'm trying to merge a fix, not quadruple the unit test time. xD 14:49 MTDiscord I am aware sfan5. I'd like it if it wasn't overwritten. 14:49 MTDiscord Particularly considering all these "lighting" mods which currently have to replace nodes with light sources, which is fairly limited 14:51 MTDiscord It's always a radial light occupying the space of a node (which gets particularly tricky if it's in water for instance and requires such mods to register light source "illuminated" variants of such nodes) 14:52 sfan5 I'm thinking that disabling the engine calculation only really makes sense if you have a mod that wants to replace the calculation entirely 14:52 rubenwardy we could use the microsoft technical for integration testing 14:52 rubenwardy make a release for every commit, use telemetry to see regression rates 14:53 sfan5 even if you could disable light calculation on a per-block basis it'd start being out-of-date as soon as the mod has done its changes 16:13 erlehmann sfan5, consider including the following XSLT as a font converter instead of the program (note that it unambiguously sets overhang and underhang, thereby simplifying parsing) https://mister-muffin.de/p/hBYi.zip 16:14 erlehmann (assuming 0 for o and u in that case) 16:14 erlehmann sfan5, if you want to keep the exact format, i can adjust it to output that. the main benefit is that it works independently of irrlicht. 16:16 erlehmann > use telemetry to see regression rates 16:16 erlehmann oh please no lol 17:38 MTDiscord There's an implicit cast from int to float in IrrlichtMt's RNG that increases rMax above its original value. 17:39 MTDiscord I don't know how to fix this. I can make the cast explicit, but that isn't really a solution unless this bug is a feature. 17:54 erlehmann josiah_wi have you looked at the doxygen call graph to see what is affected? 18:13 MTDiscord I do not know how to use the doxygen call graph. 20:03 MTDiscord Um, I get a 300% performance increase if the chunk I spawn in happens to not have trees. 20:05 erlehmann team no trees! 20:41 MTDiscord Is there documentation I should read for using the Doxygen call graph? 20:44 erlehmann josiah_wi https://doxy.minetest.net should be the thing ig 20:45 erlehmann it will show you what functions call what other functions 20:45 erlehmann and are called by other functions 20:56 MTDiscord I can't find it in the call graph. Does Minetest even use Irrlicht's RNG? 20:58 erlehmann no idea if this even transcends the boundary 20:58 erlehmann of minetest/irrlichtmt 21:00 MTDiscord I doubt it. 21:04 MTDiscord I searched on GitHub. We don't call frand anywhere. 21:04 MTDiscord I have a hunch we don't use Irrlicht's RNG. 21:07 erlehmann deleted code is debugged code ^^ 21:08 MTDiscord If a core dev could provide input I would appreciate that. 21:09 MTDiscord If this can be fixed I think that's better than deleting it. The question is how to go about fixing it properly, if that's the case. 21:11 erlehmann what constitutes “fixing” depends on how and where it is used though 21:11 sfan5 unused code should be deleted. 21:13 erlehmann i should go through /bin and look at atimes 21:16 MTDiscord What are atimes? 21:17 erlehmann time of last access of the file 21:17 erlehmann sems i haven't used the “dir” and “ntfsinfo” commands in forever 21:17 erlehmann mknod too 21:17 MTDiscord nods 21:18 erlehmann they can go 21:20 MTDiscord Nope, missed one. There's an frand() call in CIrrDeviceStub. 21:22 MTDiscord Ah, it's just a wrapper. 22:15 MTDiscord Ohh nooo I have a linker error in Minetest master with Irrlicht master. 22:15 MTDiscord nvmnd no I don't. 22:15 MTDiscord Wrong build. 22:15 MTDiscord Ah, no, I do. 22:16 MTDiscord Do we not support building Minetest with Clang? 22:18 erlehmann that's the thing 22:18 erlehmann i was able to build it with clang 22:19 erlehmann in the past 22:19 erlehmann you might be able to isolate the affected commit 22:21 sfan5 the CI literally builds with clang 22:22 erlehmann sfan5 does my offer to maintain/bugfix the pixel font code (at least until the freetype stuff is doing the crisp thing) change anything? if i took on responsibilities by that, pls enumerate. 22:22 sfan5 no 22:23 sfan5 there is nothing to maintain, I already wrote the code 22:25 MTDiscord I will rebuild everything and open an issue if necessary. 22:25 erlehmann sfan5, what do you think of my xslt? it runs without any irrlicht 22:44 sfan5 cool I guess, I won't need it 22:44 erlehmann it would be nice to include it though 22:45 erlehmann bc then the “compile with 1.9.0mt2” bit would be obsoleted 22:45 erlehmann (for the font converter) 23:01 MTDiscord If there's ever something I can do to help out the devs or save them time, I'm interested in doing that. I think I'm adding a lot of PRs that aren't of general interest and I want to show my appreciation of the other contributors and understand the project's aim better.