Time Nick Message 06:35 nerzhul mergng #9813 06:35 ShadowBot https://github.com/minetest/minetest/issues/9813 -- Fix small typo by oilboi 06:36 nerzhul and #9812 06:36 ShadowBot https://github.com/minetest/minetest/issues/9812 -- [NO SQUASH] Script API: Rename fields for consistency by SmallJoker 06:38 nerzhul and #9804 06:38 ShadowBot https://github.com/minetest/minetest/issues/9804 -- Fix broken client if openal cannot be opened by nerzhul 06:42 nerzhul and #9740 06:42 ShadowBot https://github.com/minetest/minetest/issues/9740 -- Add MSVC build by adrido 06:43 ANAND lol 06:43 ANAND Inputs required regarding #9662. Thanks in advance! 06:43 ShadowBot https://github.com/minetest/minetest/issues/9662 -- Low-level HUD element to allow mods to "draw" elements on the screen 06:43 nerzhul it's the breakfast merging party, when i read reviewed PR and either merge the one which are, or the one which miss approval :D 06:44 nerzhul hard to stay under 160 PR these weeks (and i'm also one who have some opened PRs haha) 06:46 ANAND Yeah, PR count is increasing faster than they can be merged (or maybe even rejected) 06:46 ANAND Minetest is gaining popularity :) 06:47 nerzhul to be honest it depends 06:47 nerzhul 2 years ago we got so many PR, and we found to lower it to ~100 06:47 nerzhul the only thing which always increase and never decrease is the issue amount 06:48 nerzhul i don't know how we can cleanup those 06:48 ANAND I'm seeing a lot of new faces. That and the increasing PR/issue count made me come to that conclusion. :) 14:57 sfan5 merging #9820 in 5m 14:57 ShadowBot https://github.com/minetest/minetest/issues/9820 -- Server: Improve some log messages by sfan5 16:08 nerzhul sfan5, thanks for your comments permitting to enhance more the server split pr #9769. I fixed the PR and answer related comments, it should be okay now 16:08 ShadowBot https://github.com/minetest/minetest/issues/9769 -- Server class code cleanups by nerzhul 16:08 nerzhul now we have a proper way to manage detached inventories, 2 layers of enhancements from you & me 16:16 sfan5 I'll look at it later 16:50 nerzhul sfan5: #9627 approved 16:50 ShadowBot https://github.com/minetest/minetest/issues/9627 -- Make Mapgen (mostly) work multithreaded again by sfan5 17:26 nerzhul merging #9627 17:26 ShadowBot https://github.com/minetest/minetest/issues/9627 -- Make Mapgen (mostly) work multithreaded again by sfan5 17:29 sfan5 thanks 17:31 nerzhul thanks for this amazing return back 17:34 nerzhul emerge multithread can now be useful 17:35 sfan5 yeah, having only a single mapgen thread did seriously annoy me 17:35 sfan5 when you have CPUs with e.g. 12 cores these days 17:36 Lone_Wolf Does this mean I can use all of my 8 cores now? 17:36 sfan5 for mapgen sure 17:39 ANAND Wow, neato :) 17:39 ANAND Would it make sense to engage multiple cores for mapgen by default? 17:40 ANAND (Or does the PR already do that?) 17:40 rubenwardy it's a setting, defaults to 1 17:40 sfan5 it doesn't and the "new" y=48 bug with mgv7 should be investigated before that is done 17:40 Krock sorry what? one process is still limited to one CPU? 17:41 ANAND Oh, ok. I thought the y=48 was fixed too. 17:41 Krock more threads can fill NOP gaps better, but not use more than one core. Please correct me if I'm wrong there 17:42 sfan5 depends on the kind of hyperthreading your cpu has (or doesn't) 17:42 Krock Intel CPUs better have theirs turned off 17:45 Krock #9177 likely won't see a nice RemotePlayer-handover from ServerEnv to ObjectRef like requested by rubenwardy 17:45 ShadowBot https://github.com/minetest/minetest/issues/9177 -- Add core.get_player_or_load by SmallJoker 17:46 Krock problem is that the pointer might be used one, or twice. its origin being unclear 17:46 Krock ServerEnv std::vector> m_players; could probably work around it, but I'm not sure whether that's actually a good approach 17:46 rubenwardy I think invalidating refs when a player joins is perfectly fine 17:46 sfan5 I just tried /emergeblocks and it apparently only runs singlethreaded 17:47 sfan5 uh oh MT is eating all my ram 17:47 rubenwardy mods shouldn't care about keeping player data in memory using their reference 17:47 Krock rubenwardy: thing is, ServerEnv does not know about ObjectRefs at all 17:47 rubenwardy which makes sense 17:47 Krock ObjectRefs are handled in Lua, ServerEnv by Minetest 17:49 Krock I could try linking them with the shared_ptr approach, but I'd say syncing those two areas could be quite dangerous 17:49 rubenwardy are you referring to updating or invalidating the objectrefs? 17:49 Krock both 17:50 Krock ensuring that the hand-over works well (i.e. all data is kept and re-used in both places) and invalidating (so that it's really deleted) 17:51 Krock former would directly conflict what you wanted to avoid in the "kick ghost players" PR: >it feels like it would be easy to end up with networking bugs. I'd prefer a kick and rejoin 17:52 Krock but kick&rejoin isn't an option here, because the player may join at any time, so data loss is not wanted 17:52 sfan5 Lone_Wolf, ANAND: here's a demonstration https://0x0.st/i2YK.mp4 17:52 Krock data loss being temporary changes from Lua, not saved yet. 17:52 rubenwardy the kick ghost players PR is a state thing - you have to make sure to resend any HUDs / set up, and clear the known chunks for a player 17:53 Krock sfan5: interesting. which CPU? 17:53 sfan5 Ryzen 5 2600 17:53 rubenwardy the update ObjectRef could be equivalent to a close and then update to point to the online object ref 17:53 rubenwardy invalidation is perfectly fine as well 17:53 ANAND Nice! 17:53 Lone_Wolf Neat 17:53 Krock so far only seen results from Intel. AMD's threading really seems to make use of all cores. lovely. 17:54 rubenwardy I just don't like the idea of mods having so much control over how long player info is in memory, and whether a player is blocked from joining 17:56 rubenwardy sfan5: the decrease in CPU usage all at the same time is suspicious to me - is there a shared lock when writing back? 17:56 sfan5 sqlite has locking yes 17:56 rubenwardy oh, each thread map gens straight to sqlite? 17:57 sfan5 I think so yes 17:57 rubenwardy fair enough 17:57 sfan5 emergethreads are also responsible for loading blocks from the map 17:57 sfan5 even though that doesn't make too much sense multi-threaded 17:58 sfan5 at least not with sqlite 18:17 Lone_Wolf For #9792 do I add docs on everythink you listed or just the ones you have to do with pure lua? @Krock 18:17 ShadowBot https://github.com/minetest/minetest/issues/9792 -- Add note to docs on how to override privileges by LoneWolfHT 18:17 Lone_Wolf Chatcommands have override_chatcommand() 18:20 sfan5 pushing http://sprunge.us/8Erg2S?diff in 5m 18:23 Krock Lone_Wolf: tbh I do neither know what's the preferred way of modifying those tables, nor the best way to document it 18:25 Krock the tables can be modified in various ways, thus I'd rather only put notes for those that cannot be modified on Load-Time and add a description about what's generally possible with the other tablse 18:25 Krock *those that can only be modified on Load-Time 18:27 sfan5 rubenwardy: minimal really needs a formspec you can go through to ensure everything looks as expected 18:27 rubenwardy sfan5: /formspec 18:27 sfan5 ... 18:27 sfan5 does that have everything? 18:27 rubenwardy not everything 18:27 rubenwardy but I've been asking for examples to be added in every formspec PR since 5.1 18:28 Lone_Wolf Would like Wuzzy's devtest to be merged 18:28 rubenwardy well, /formspec was added when I did StyleSpecs 18:28 rubenwardy since then is when it's been maintained 18:28 rubenwardy it does contain a lot though 18:28 Lone_Wolf ^ 18:29 rubenwardy sfan5: I also have 2 PRs open which will automate formspec testing by comparing before and after 18:29 Krock the devtest game is nice, but not well usable in my setup 18:29 rubenwardy Krock: how come? 18:29 Krock rubenwardy: mainly lack of "default". broken testing worlds 18:30 rubenwardy it's for engine testing 18:30 rubenwardy default is a MTG thing 18:30 sfan5 ...is this the expected rendering? https://a.uguu.se/OF6kh4iao5sK_.png 18:30 rubenwardy yes 18:30 sfan5 good to know 18:30 rubenwardy it's loads of off placement to test clipping in the scroll container 18:34 rubenwardy #9188 and #9189 18:34 ShadowBot https://github.com/minetest/minetest/issues/9188 -- Add standalone formspec renderer by rubenwardy 18:34 ShadowBot https://github.com/minetest/minetest/issues/9189 -- WIP: Add formspec rendering tests by rubenwardy 18:39 Lone_Wolf Hmmmm, I can't switch into third person anymore 18:40 Krock change key 18:40 Lone_Wolf Where is that? I can't find it 18:41 Lone_Wolf OH 18:41 rubenwardy same 18:41 Krock keymap_camera_mode = KEY_F7 18:41 rubenwardy why? 18:41 Lone_Wolf It was set to C 18:41 Lone_Wolf Which Cinematic mode overrode 18:46 nerzhul Krock, can't you just put a accessdeined to the previous session to drop it ? 18:47 Krock yes, that's easy. Originally I wanted to recycle attachments 19:08 Lone_Wolf ATM the only things I'll be adding to the docs are for entities and privileges. I feel like it'll be simpler to just say you can edit those, AFAIK the entities are the only case where there might be something semi-unexpected, and I've mentioned that 19:10 Lone_Wolf From what I can see everything else either has a function for overriding or can't be overridden during runtime 20:02 nerzhul Krock not sure it's needed 20:26 Walker hi, i compiled old versions of minetest last time and found again that large core elements of minetest itself are implemented in lua "builtin" 20:26 Walker and I just wanted to ask, are there any projects that not only create a subgame, but also reimplement builtin? 20:27 Lone_Wolf Nodecore kinda does iirc 20:28 Walker ok nice thx 20:30 Walker wait ... is nodecore not just a subgame ? 20:30 Walker i already played it and dont replased "builtin" 20:30 Lone_Wolf It modifies builtin functions 20:31 sfan5 what falls under "large core elements" in your opinion, Walker? 20:33 Walker thinks like the builtin/game folder ... 20:34 Walker with that you can influence some basic behavior or not? 20:35 Walker things like how voxel are treated 20:36 Walker I have to say: I don't know exactly what the builtin directory contains ... is there a wiki somewhere? 20:36 Walker dont find much abaout it in the /doc folder 20:36 sfan5 half of the builtin/game folder are lua parts of basic engine functionality, the other half could possibly be removed if you wanted the engine to do nothing out of the box 20:37 sfan5 there isn't because the organization of builtin isn't even relevant for modders 20:40 Walker well, I'm going to play around a bit and see what I can learn from the structure ^^ 20:43 sfan5 rubenwardy: did you just review your own PR https://a.uguu.se/Z5lrfDj5iNNh_.png 20:43 rubenwardy I added notes for later 20:43 rubenwardy :D 20:44 sfan5 ok sounded a lot like you reviewed the PR without knowing it was your own code 20:44 rubenwardy it hasn't been _that_ long 20:44 sfan5 which would be a quite an accomplishment 20:44 rubenwardy not saying that isn't something I would do 21:01 Wuzzy how is a translated represented internally, when sent over the net? 21:01 Wuzzy translated string* 21:03 sfan5 https://github.com/minetest/minetest/blob/master/builtin/common/misc_helpers.lua#L603 21:04 Wuzzy thank you