Time Nick Message 00:07 MTDiscord Hardware support meaning software support, meaning OS provided sounds. Android doesn't have .mod built-in 00:08 MTDiscord So using .mod requires us to carry our own soundfiles, and sound fonts are some of the most effective methods. 02:08 MTDiscord Sigh, libmodplug is the best library for .mod and it's linux only... fluidsynth works on android, windows, and linux but it has a decent amount of deps. wonder how many can be ignored or not.... 02:15 MTDiscord Huh, okay not as bad as I thought, It requires glib and gthread at a minimum 02:16 MTDiscord And that allows us to generate directly into raw PCM (wav minus header info), which we would pipe to openAL 02:17 MTDiscord alternatively, it supports SDL2 backend for linux/windows and oboe for android 02:18 MTDiscord Further, we could just use precompiled stuff from package managers and distribute the .dll with the windows distribution like we do now... 02:25 MTDiscord Yeah, this is not as bad as I suspected, we can retrieve latest from package managers for fluidsynth on all the .yml's, add one dep to the readme.md, and update our build pipelines for windows and android with the build artifacts from the libfluidsynth pipelines. It's actually quite doable... 04:43 ronoaldo I'm still seeing some issues when building from master, about a potential out-of-bounds warning: https://i.imgur.com/rvimCRP.png 04:43 ronoaldo Should I file an issue about it? Not sure if a false-alarm ... 05:34 Wuzzy Changelog says "Removed support for bumpmapping, generated normal maps, and parallax occlusion" 05:34 Wuzzy like, completely removed? what about node stuff? 05:35 Wuzzy DevTest still has test nodes for those things. are these test nodes useless now? 09:01 Zughy[m]1 I guess so 09:16 rubenwardy ronoaldo: #10238 09:16 ShadowBot https://github.com/minetest/minetest/issues/10238 -- Warning -Warray-bounds on H_ns in src/util/srp.cpp 11:05 pgimeno I think #6855 was probably the wrong approach: https://irc.minetest.net/minetest-dev/2019-11-27#i_5615954 11:06 ShadowBot https://github.com/minetest/minetest/issues/6855 -- Fix Wstringop-overflow warning from util/srp.cpp by HybridDog 11:07 pgimeno "I don't conceive the reason of this warning, however, when using an unsigned int instead of size_t for len_bytes, the warning doesn't appear." 11:07 pgimeno if the root cause is not identified, applying a random fix can cause problems 11:08 sfan5 yes, changing it to uint32_t doesn't fix it 11:08 sfan5 it won't break anything however 11:09 pgimeno the root cause is probably that the -1 in the default case is not handled by the callers 11:10 pgimeno this -1: https://github.com/minetest/minetest/blob/master/src/util/srp.cpp#L357 11:10 sfan5 just make it return zero 11:11 pgimeno maybe, but are all cases where the algorithm is not valid properly handled? 11:11 rubenwardy it should do FatalError 11:11 rubenwardy and not have the default 11:11 rubenwardy although, there are compile guards for the hashes for some reason 11:12 pgimeno FatalError sounds more reasonable 11:13 sfan5 I thought Minetest only ever calls srp with the same hash algo 11:14 pgimeno maybe the path to that point is too convoluted for the compiler to deduce that 11:15 pgimeno or maybe there's a code path where that's not the case, I haven't looked in that much depth 13:08 rubenwardy there are a lot of use_texture_alpha warnings 13:08 rubenwardy is it really necessary? 13:10 rubenwardy spamming just Field "use_texture_alpha": Boolean values are deprecated; use the new choices isn't good either. It should mention the mod and file, or only log once 13:13 rubenwardy Mineclone2 has 540 texture alpha warnings 13:14 rubenwardy the engine should choose the correct mode based on drawtypes and texture, and the modder can then change it if it's wrong. All these warnings are a bad API 13:19 sfan5 there is one warning per nodedef, I think it writes the stacktrace to infostream 13:19 rubenwardy Modders aren't going to spend hours fixing 540 warnings for something they don't even see the difference of 13:20 rubenwardy They'll just ignore the warnings 13:21 sfan5 and "should choose the correct mode based on drawtype" already happens, "based on texture" reintroduces the exact problem this is supposed to avoid 13:22 sfan5 is your suggestion that we keep supporting use_texture_alpha = true/false without warnings? 13:26 sfan5 that'd be doable and wouldn't hurt since the server can just pick whatever mode it used to mean 14:02 MTDiscord Yeah that's what I expect Ruben means 14:53 rubenwardy use_texture_alpha = true/false is something that can be updated pretty easily, and is only 122 of the deps 14:53 rubenwardy the fast majority are about the missing use_texture_alpha thing 14:53 rubenwardy *vast 14:55 rubenwardy eg: assuming use_texture_alpha = "clip" 15:01 sfan5 the point there is that modders have to specify if they want alpha used because you don't want that to depend on the current texture pack 15:08 rubenwardy what's wrong with it depending on the current texture pack? invisibility cheats? 15:11 rubenwardy wsor is wondering if the C/V keybinding can be reverted 15:11 sfan5 yes, the same that used to exist for drawtype="normal" 15:12 rubenwardy Normal defaulting to opaque fixes most of those, surely 15:12 sfan5 yes 15:12 sfan5 but this is about nodeboxes and meshes 15:15 MTDiscord to be specific, revert https://github.com/minetest/minetest/pull/9779. you can see a example of the complaints, https://github.com/minetest/minetest/issues/10678, and https://github.com/minetest/minetest/issues/10617. most people i know changed it back. so im asking if we could revert it for 5.4 15:16 MTDiscord also c and v is very annoy for copying and pasting 15:16 MTDiscord *annoying 15:16 MTDiscord +1 from em 15:16 MTDiscord *me 15:25 MTDiscord there PR https://github.com/minetest/minetest/pull/10906 15:26 rubenwardy I'm on the fence. The change annoyed me but I think most of the damage is done. We can't just never change key bindings, but perhaps a better way to handle it is with hints like "Minimap has changed to V" when you press F8 15:26 MTDiscord C and V is very dumb cause control c and control v 15:27 MTDiscord and a lot of people just changed them back in there client after the annoyance 16:07 Zughy[m]1 I used to critique C for the camera, but after playing with it a bit I think it's handier than before 16:07 Zughy[m]1 As you don't have to move your hand 16:13 MTDiscord ehh, how often do you change your pov 16:14 MTDiscord it should be kept away from common keys to prevent accidental strokes too imo 16:14 MTDiscord the fn row provides that isolation very well imo 16:18 celeron55 i think it's a handy key for a useless functionality and should probably be reserved for something useful 16:19 Zughy[m]1 changing camera is useful. Source: me playing 16:19 MTDiscord eh, depends on what you are playing 16:19 celeron55 i can say the opposite 16:20 MTDiscord #metoo 16:20 celeron55 so it depends on the play style 16:21 MTDiscord c55: i always had a question... when you first wrote minetest, did you think of it survival first, or creative first? 16:23 celeron55 i'd say neither 16:23 MTDiscord oh? 16:26 celeron55 it's not easy to explain, maybe something like adventure, or experience, or emergence first 16:26 celeron55 but there was no plan 16:27 MTDiscord Something he didn't add yet i guess 16:28 MTDiscord MT isn't like MC, anyone can make a mode for free and anyone can add it, so even if no creative (for example) someone can makes that creative 16:28 MTDiscord ah, i see, thanks 16:29 celeron55 in any case personally i don't really like the creative/survival division 16:30 celeron55 it's very utilitarian so it does work though 16:31 MTDiscord Me too. Adventure is better 16:31 MTDiscord Like a story the player play according to 16:32 celeron55 well i don't think a story would fit either 16:32 MTDiscord Like< 16:32 celeron55 i mean, stories always work, but you don't need a game or a sandbox for them 16:33 MTDiscord Indeed, and stories can be singleplayer or team who plays in the same progress 16:33 celeron55 imo the best way to tell a story on a computer is a visual novel, anything else just gets in the way 16:35 MTDiscord Some good real life stories with some exaggreting 16:36 MTDiscord Can be good 16:36 MTDiscord By the way, are you planning to or do you have enough time to host a server on minetest? :P 16:37 MTDiscord By the way, are you planning to or do you have enough time to host a server on minetest? :P 16:49 Wuzzy hey everyone. the "ores" mapgen flag was added by paramat, but i noticed it is not being exposed to the world creation dialog 16:49 Wuzzy is this intentional? 16:49 sfan5 no idea 16:49 sfan5 does it work as expected? 16:49 Wuzzy yes 16:50 Wuzzy you have to go all the way to the advanced settings tho 16:50 Wuzzy i just wonder if its better to add an "Ores" checkbox to the world creation dialog or if this setting is too unimportnat 16:50 Wuzzy just an usability thing, to be clear 16:50 Wuzzy its not broken, dont worry 16:51 Wuzzy i note we also dont expose the Biomes flag as well 16:51 sfan5 ¯\_(ツ)_/¯ 16:52 Wuzzy i have no strong preferrence one way or another, its just something to think about 16:53 MTDiscord anyhow, may i make a shameless plug about a project of mine that is somewhat on-topic here? 16:56 Wuzzy no. muhahaha. we have #minetest for chitchat 16:57 MTDiscord fair 18:09 rubenwardy #10907 18:09 ShadowBot https://github.com/minetest/minetest/issues/10907 -- Game closes after i open it up but before the game runned fine. 18:14 Krock it worked but not it doesn't 18:16 Krock any PCI(E) hardware analyzer software could help there to find the correct drivers 18:16 Krock just because it worked previously doesn't mean that it worked correctly 20:31 Wuzzy bugfix PR: #10742 20:31 ShadowBot https://github.com/minetest/minetest/issues/10742 -- Fix world-aligned node rendering at bottom by Wuzzy2 20:35 Krock rubenwardy: https://github.com/minetest/minetest/suites/1943672677/artifacts/38416220 results in error 404 in a private tab. indicates some sort of permission wall 20:35 rubenwardy You need to be logged into githubt 20:36 Krock ah, so just that. I was worried you might need specific repo access or similar 20:37 Wuzzy i get now a lot of errors in MCL2 about specified ore types suddenly requiring "noise_params" 20:37 Wuzzy this was never an error before 20:38 Wuzzy what changed? 20:38 rubenwardy There was a bug that meant the ores never checked for noise paramd 20:38 rubenwardy Like, they all accepted not having it 20:39 Krock > -- Needed by "sheet", "puff", "blob" and "vein" ores. 20:39 Krock but optional for others 20:39 rubenwardy Yeah, the C++ for this did polymorphism wrong and always returned "not required" 20:40 rubenwardy Do these ores still spawn? 20:40 Krock #10846 20:40 ShadowBot https://github.com/minetest/minetest/issues/10846 -- MgOre: Fix invalid field polymorphism by rubenwardy 20:40 Krock no. "ore" is literally deleted 20:41 rubenwardy I meant before the fiz 20:42 rubenwardy We should avoid printing errors whenever possible. So if this worked before, then it should probably emit a warning at most maybe 20:44 Krock yeah. lower the log level just like #10905 but in this case it must not delete the ore so that it's still generated 20:44 ShadowBot https://github.com/minetest/minetest/issues/10905 -- Reduce "Empty translation" debug log message level from error to info or action 20:44 Krock just using some dummy noise params 22:10 Wuzzy hmmmm 22:11 Wuzzy how on earth did the ores work before then, if NoiseParams was never specified?