Time Nick Message 07:59 red-001 nerzhul, since you're here what's your view on #6932? 07:59 ShadowBot https://github.com/minetest/minetest/issues/6932 -- Remove CSM screenshot API by red-001 08:09 nerzhul red-001, i prefer implementing a rate limit and a client setting 08:09 nerzhul if user wants to use a mod to automagically generate screens to show its life on game why not 08:10 red-001 it's just that no-one wrote rate-limiting for it and even with rate limiting I would prefer if it required user interaction to enable 08:10 nerzhul yes, 08:10 red-001 kinda like the way browsers ask the users before letting websites use certain features 08:10 nerzhul rate limit + client setting by default to false 08:11 red-001 s/users/user 08:14 red-001 nerzhul, how hard would it be to isolate mods from each other in thier own lua_state and do we even want to do that? 08:15 nerzhul it's not difficult except for the callback functions 08:16 nerzhul it needs to be handled client side 08:16 nerzhul instead of lua side 08:17 red-001 is it only me or does travis take a lot longer to build stuff now? 08:18 red-001 updated 6067 yesterday and travis is still waiting with the xcode build 08:19 nerzhul macosx build is long to start since the begininng of the week and some builds randomly failed 08:19 nerzhul maybe it's due to the CPU patching on AWS cloud, some performance dropdown cna happen too 08:20 red-001 maybe 08:20 red-001 I would guess OS X uses more resources then thier linux setup 08:21 nerzhul yeah 08:21 red-001 oh I see looks like they can/are allowed only 180 active OS X VM's at a time 08:21 nerzhul it's not just a docker container 08:23 red-001 and they have a backlog in the thousands of builds 08:24 red-001 nerzhul, do we use the container for the linux or a full vm? 08:27 nerzhul it's always container on linux on travis if i remember 08:27 nerzhul but you cannot choose which 08:28 nerzhul on gitlab CI we use various different container 08:34 red-001 looks like you can choose for Ubuntu Trusty but not for Precise or OS X 08:35 nerzhul yes 08:37 red-001 looks like we use the full vm for all the builds 10:46 red-001 #6072 #6067 10:46 ShadowBot https://github.com/minetest/minetest/issues/6072 -- [CSM] Add functions to create particles and particlespawners. by red-001 10:46 ShadowBot https://github.com/minetest/minetest/issues/6067 -- [CSM] Add basic HUD manipulation. by red-001 11:09 nerzhul red-001, the chat6 submodule is an error no ? 11:11 red-001 yes 11:11 red-001 which one did I add that too? 11:12 nerzhul you also added your cmake configuration files :) 11:18 red-001 fixed 11:18 red-001 added the files to gitignore 11:51 nerzhul red-001, approved both PR 11:51 nerzhul we need a second approval and we will have those useful features 12:05 Krock red-001, hud.cpp is missing in the android makefile 12:06 * red-001 grumbles something about auto-generating files 12:11 red-001 Krock, fixed 12:47 red-001 Krock, fixed your remaining issue with #6067 12:47 ShadowBot https://github.com/minetest/minetest/issues/6067 -- [CSM] Add basic HUD manipulation. by red-001 12:50 Krock 200 12:53 red-001 505 13:09 nerzhul merging #6067 13:09 ShadowBot https://github.com/minetest/minetest/issues/6067 -- [CSM] Add basic HUD manipulation. by red-001 13:09 nerzhul thanks red-001 13:10 nerzhul macosx build is stuck 13:11 red-001 there is a backlog of close to 3000 builds last I checked 13:12 red-001 so I imagine it will be like that for a while 13:18 nerzhul Krock, #6884 is now fixed (we will see if it's okay on travis) 13:18 ShadowBot https://github.com/minetest/minetest/issues/6884 -- Game refactor: reduce game.cpp size by moving objects in the right places by nerzhul 13:48 red-001 nerzhul, can you cancel https://travis-ci.org/minetest/minetest/builds/331191378?utm_source=github_status&utm_medium=notification ? 13:48 red-001 oh nvm looks like they fixed builds not being canceled on force push 13:48 red-001 about time they did imo 15:22 paramat sfan5 sofar ShadowNinja are you ok with Shara as a new MTG dev? 15:25 nore paramat: fyi, sofar is not on this channel 15:25 red-001 who are the MTG devs anyways? 15:25 paramat yeah 15:25 red-001 I'm never quite sure of who is a core dev and who is a MTG dev 15:26 paramat it's .. 15:26 sfan5 paramat: I'm fine with that 15:26 paramat ok 15:26 sfan5 hmm maybe we should have a separate group for mtg-only devs: https://github.com/orgs/minetest/teams 15:26 paramat ^ yes 15:26 paramat i almost suggested that yesterday 15:27 paramat even if only for documentation 15:27 red-001 would be useful if you split collborators into two groups and have one for the minetest repo and the other group for the minetest_game repo 15:28 red-001 seems there is often a good bit of confusion due to this 15:28 paramat Pilz (absent), sfan, nor-e, Kroc-k, ruben-, me, so-far, shadow-ninj 15:28 red-001 and it's easily avoidiable 15:28 paramat so-far shadow-ninj are inactive 15:29 red-001 Pilz isn't around at all as far as I see 15:29 red-001 not for quite a while now 15:29 paramat in fact i think pilz is somewhat fired for absence :) 15:30 red-001 README.MD sure needs some work 15:30 red-001 it still calls itself readme.txt.... 15:30 red-001 oh the readme file in MTG wasn't renamed 15:31 red-001 paramat, so 11 isn't an MTG dev? 15:32 paramat correct 15:32 Krock red-001, all devs are in the minetest group 15:34 red-001 Krock, that would seem logical but wanted to confirm 15:35 red-001 I was almost sure there was talk of a new dev a while back but I saw no-one new on that page 15:37 nerzhul merging #6884 15:37 ShadowBot https://github.com/minetest/minetest/issues/6884 -- Game refactor: reduce game.cpp size by moving objects in the right places by nerzhul 15:37 nerzhul red-001, i think README.md is too huge, maybe it should be splited to have only a description and links to specific sections in dedicated files 15:38 nerzhul paramat, i never see pilzadam since 2 years on this channel 15:39 paramat but then, some devs are absent long term and we don't kick them, so i guess pilz remains a core dev but can be assumed not a MTG dev 15:40 red-001 also with all these CSM security related issues being marked as high priority and blocker then so should #5989 15:40 ShadowBot https://github.com/minetest/minetest/issues/5989 -- [CSM] Inventory changes are sent to server instead of being handled by a local callback. 15:40 red-001 it's a much larger issue then just reading nodes or sending text 15:40 nerzhul paramat, i don't think a very long time inactive coredev can be a coredev as in 2 years the project technic, politic can change many 15:41 nerzhul red-001, it's completely relalted to the whole formspec stuff :( 15:44 red-001 if hashing built-in at compile time is a blocker then that almost certainly should be too 15:44 nerzhul yes 15:44 nerzhul paramat, a 0.5.0/0.4.17 release date has been defined or not ? 15:46 red-001 as for hashing builtin just move the death formspec back to core or something if it's that important, hashing builtin is the most pointless security measure I heard of in a long time 15:48 red-001 the issue should be marked "can't-fix" and "non-trivial" judging by the precident set by #6114, but that's just my opinion 15:48 ShadowBot https://github.com/minetest/minetest/issues/6114 -- Detection of dirty clients 15:57 nerzhul please look at #6936 :) could be nice to have 15:57 ShadowBot https://github.com/minetest/minetest/issues/6936 -- Add minetest issue template by nerzhul 15:59 paramat so moving death formspec back will avoid having to verify builtin? if moved would we then revert part of https://github.com/minetest/minetest/commit/308bb69eefdfb2ff3b893c124ccf0321d99a1bfd ? 16:00 red-001 it's as simple as adding a fallback formspec or just remove that little piece of lua code and adding it to c++ 16:03 red-001 there might be some other stuff you might need to change but all of it si trival compare to creating a cross platform way to get the hash for all the files in builtin at compile time and then refuse to run if they change by runtime 16:03 red-001 s/si/is 16:04 red-001 I mean at that stage you might as well just implement #6114, chances are that would be simpler to do 16:04 ShadowBot https://github.com/minetest/minetest/issues/6114 -- Detection of dirty clients 16:04 red-001 (getting the hash of a file at runtime vs at compile time) 16:12 paramat !tell nerzhul sorry, i didn't reply about release dates. nothing defined yet, we're hoping for lord fingle and others to work on android and get it reasonable before release, so maybe 2+ months? 16:12 ShadowBot paramat: O.K. 16:12 paramat hmm 16:14 red-001 has anyone sent him an email or something, he seems to have disappeared after that competition he was in ended. 16:14 paramat yes i'm in touch 16:14 paramat by PM on forum 16:14 paramat a few days ago he said 'soon' 17:46 lisac paramat: maybe MT could have a small proprietary anti-cheat part 17:46 lisac which server needs to somehow detect in order to approve it as the official client 17:47 lisac for example, maybe it could be a function that takes in a number, then calculates and returns another number 17:47 lisac server generates a number, sends it to the client, which returns the result 17:47 lisac if they match, it's the real client 17:48 lisac if not, just kick the client. 17:48 lisac if you build your own Minetest, you can download the compiled anticheat library part 17:48 lisac and set your client to use it 17:50 CBugDCoder sounds like a good idea but hacked clients could set themselves to use it too so how would it do anything? 17:50 lisac oh 17:50 lisac lol right 17:50 lisac fml 17:51 lisac you could just build it in the official mt clients 17:51 lisac which prevents custom builds now. 17:51 lisac nah 17:51 lisac hey, why not render everything on server and just receive input from client? Make client just play a video server sends. 17:52 lisac cheating would be impossible. :P 17:52 CBugDCoder but that would be sooooooooo laggy especialy on popular servers 17:53 pgimeno not to mention the bandwidth requirements 17:53 nore paramat: maybe MT could have a small proprietary anti-cheat part <-- this is not happening 17:53 nore and won't be 17:54 lisac nore: yes, I'm just thinking out loud 17:54 nore minetest is an open-source project, and we won't force to include proprietary code in it 17:54 lisac or maybe we could have a better server side anticheat 17:54 nore this is the correct solution 17:55 lisac what are the most used cheats currently? 17:55 lisac Probably flying and noclip. Easiest to implement 17:55 lisac Flying is already really slowed down by the anticheat: very annoying to use. 17:55 lisac same for running. 17:56 lisac Some servers damage you while you are in walls without noclip priv. 17:58 Hijiri #6688 17:58 ShadowBot https://github.com/minetest/minetest/issues/6688 -- Custom particle generators for particle spawners by raymoo 18:01 lisac nore: maybe check if there is a node under player in PlayerSAO::checkMovementCheat(), then if the player is falling, or climbing, or has fly priv, and that way find out if he is fly-cheating? 18:03 Shara lisac: anticheat does nothign against flying from what I know. It just interferes with fast, which is not the same thing. 18:03 lisac Shara: try flying around without fast. It's annoying. :P 18:03 lisac but yes, still possible to cheat. 18:03 Shara That's your (and my) personal preference, but when you don't have fast anyway, anying already grants a huge advantage 18:04 Shara flying* 18:04 lisac I agree. We need a server side anti-fly 18:04 lisac or anti-fly-cheat 18:07 Shara Noclip is probably the biggest issue when it comes to cheating, at least from my experience (breaking into other players homes or safe areas when PVP is enabled) 18:07 Shara And you only need to fly for a moment, at any speed at all, to abuse that 18:07 lisac hey nerzhul 18:07 nerzhul hi lisac 18:08 lisac has anyone sent him an email or something, he seems to have disappeared after that competition he was in ended. 18:08 lisac red-001 seems to have been looking for you, nerzhul 18:08 nerzhul paramat, 2 months sounds reasonable, i need to test android build at a time 18:08 nerzhul i'm there for a time 18:08 lisac Shara: you could make a mod that checks if player is inside nodes 18:08 Raven262 NO, he wasn't looking for nerzhul. 18:08 Raven262 Read the logs again. 18:08 lisac if he is inside nodes a lot of times in a short time, kick him 18:08 Shara Yes, but think how frequently that has to check to actually be useful 18:08 lisac Raven262: thanks I'll check 18:09 nerzhul Raven262, if he needs me he will talk to me, then no real discussion there: p 18:09 Raven262 Just saying, sorry. 18:09 Shara And there's enough legitimate reasons (falling nodes dropping on players) for that 18:10 lisac Shara: which is why you check for a lot of nodes in a short period, like 5 nodes in 10 seconds 18:10 lisac either the player is drowning in bugged sand, or he is hacking. 18:10 lisac bonus if you check if the positions of the five nodes are different 18:10 Shara Still lots of checks. Too many on a busy server 18:11 pgimeno anti-fly might be as simple as: check how many nodes under the player are air (stop at first non-air node) 18:11 Shara And it won't even detect the issue I mainly have 18:11 nerzhul paramat, Krock sfan5 rubenwardy nore , can you all look at #6936, it's an issue template to help our users to fill issues, it's basic and can be improved before merge or after 18:11 ShadowBot https://github.com/minetest/minetest/issues/6936 -- Add minetest issue template by nerzhul 18:13 lisac pgimeno_: what about ladders and jumping? 18:13 lisac and falling? 18:14 Shara And trampolines that bounce players more than normal, and gravity changes and so on. Lots of things make it not so simple 18:15 lisac Shara: what about kick/ban votes? 18:15 lisac have a mod that allows players to call a vote to kick or ban another player 18:15 lisac would work for the time being 18:15 lisac until proper anti cheat 18:15 Shara That's not really relevant to the problem, and introduces lots of issues of its own 18:16 lisac it is the most basic form of anti cheat 18:16 lisac have other players report the issue 18:16 pgimeno_ lisac: ladders are nodes you're into. Jumping is not a problem. Falling is not a problem either, just detect constant or increasing height. Trampolines are the real problem here IMO. The mod should have a function to override the anti-cheat. 18:16 paramat ok 18:17 Shara It's not a fix for the actual issue. Andit's also something I'll never implement due to all the other problems (also I think not relelvant for -dev) 18:17 lisac pgimeno_: Don't trampolines just set the players gravity to another value? 18:18 paramat #6217 is server validation of client movement and is progressing 18:18 Shara I'd love a solution for fly/noclip cheating, but a mod won't cut it 18:18 ShadowBot https://github.com/minetest/minetest/issues/6217 -- Step height: Add as a player object property by paramat 18:18 paramat agh 18:18 paramat #6219 18:18 ShadowBot https://github.com/minetest/minetest/issues/6219 -- WIP: Server side movement by bendeutsch 18:19 Shara Glad he is still working on that 18:20 nerzhul paramat, i doesn't want to hurry up for 6219 it's a very sensible thing to code server side and network latency doesn't help 18:23 pgimeno_ lisac: I don't know. If that's the case, then it would not be a problem, as the number of nodes under the played can easily take into account gravity. I assumed a trampoline was a one-time impulse, which breaks my assumptions. It can still be detected as a parabolic movement, but network latency can screw that up as nerzul notes. 18:23 lisac IIRC, Raven262's Warzone mod jumppads use physics overrides https://forum.minetest.net/viewtopic.php?t=15199 18:25 Raven262 There are two ways to do the trampoline effect. One is a grav override other is a node group. 18:26 Raven262 I have used the override, other way was unknown to me and it would limit me to a much weaker effect. 18:28 pgimeno_ hm, 6219 looks like reimplementing client physics server side, which is going to be slow - you could as well do all physics on the server and just send the result to the client, which is what SL does 18:28 lisac SL==Second Life? 18:29 pgimeno_ yes 18:29 lisac I found the servers I tried a bit laggy 18:29 lisac not that I played more than 20 minutes of it 18:29 rubenwardy pgimeno_, client-side prediction, server-side reconcilliation 18:30 rubenwardy both sides should do player physics, but the server should take priority and correct and miscalculations client-side 18:39 Krock nerzhul, the issue template should also say something like "I have checked whether there are duplicate issues" 18:52 lisac and probably 'I checked if my build is clean' 18:52 nerzhul Krock, please add comment on the PR i will improve it :p 18:59 paramat sfan5 does #6072 have your approval now? 18:59 ShadowBot https://github.com/minetest/minetest/issues/6072 -- [CSM] Add functions to create particles and particlespawners. by red-001 18:59 sfan5 I haven't tested, but the code is fine 21:31 red-001 #6940 21:31 ShadowBot https://github.com/minetest/minetest/issues/6940 -- Fix issues with earlier CSM HUD commit by red-001 22:30 nerzhul merging #6936 22:30 ShadowBot https://github.com/minetest/minetest/issues/6936 -- Add minetest issue template by nerzhul 22:30 nerzhul and #6072 22:31 ShadowBot https://github.com/minetest/minetest/issues/6072 -- [CSM] Add functions to create particles and particlespawners. by red-001 22:55 red-001 #6939 22:55 ShadowBot https://github.com/minetest/minetest/issues/6939 -- [CSM] on_connect callback is broken. 22:55 nerzhul red-001, #6941 22:55 nerzhul :p 22:55 ShadowBot https://github.com/minetest/minetest/issues/6941 -- [CSM] Remove `on_connect` callback by red-001 22:56 nerzhul sfan5, if you are okay i merge this trivial removal 22:56 nerzhul as this callback is useless since flavours as it cannot be triggered 22:59 sofar is core.localplayer then always valid? 23:00 nerzhul yes 23:00 nerzhul if i remember it's set when the CSM is inited 23:01 nerzhul before any mod loading 23:01 red-001 yep it was set just before the on_connect callback was called 23:01 red-001 https://github.com/minetest/minetest/pull/6941/files#diff-34f48ad91ac6c202ac60b0348ae90e30L1700