Time Nick Message 00:00 est31 Also, I agree to remove the renderer option, its more contraproductive than productive i think 00:02 paramat push complete 00:12 est31 RealBadAngel, if you code something for minimap, can you perhaps also add it to the hud, so that it can be gamified, e.g. you will have to craft a map item, which shows you the minimap? 00:13 est31 also there is further speedup potential for getPixel, the requests to the cache are done for every single pixel 00:13 est31 those can be expensive 00:13 est31 better to cache it for all heights somehow 02:26 yh1986 @est31, Minetest-0.4.12.11-android you can see sun or moon and stars ? 03:08 VanessaE can someone explain how this could happen? http://pastebin.com/XPWSyhEz 03:08 VanessaE relevant line in the farming mod: https://github.com/VanessaE/dreambuilder_game/blob/master/mods/farming/api.lua#L204 03:09 VanessaE (it an unmodified copy of the upstream minetest_game version) 03:09 VanessaE it's* 03:10 VanessaE the function it fails to call is here: https://github.com/VanessaE/dreambuilder_game/blob/master/mods/farming/api.lua#L113 03:16 cornernote does it happen every time, or just once ? 03:17 VanessaE in this case, it happened several times in a row 03:18 VanessaE I checked the online copy, compared it with the minetest_game code, it's the same. I don't understand how such an error is possible :-/ 03:21 cornernote does it happen when onplace() is called? (i assume it doesnt happen on startup) 03:22 cornernote i see the reason, on the phone 03:22 VanessaE it appears so 03:23 VanessaE it crashes even when I just hoe some ground 03:23 cornernote https://github.com/VanessaE/dreambuilder_game/blob/master/mods/farming_plus/init.lua#L1 03:23 * init looks at channel 03:23 cornernote looks like this is loaded after farming 03:23 VanessaE hm, that could do it 03:29 VanessaE looks like farming API changed or something 03:40 VanessaE hm, no.. 03:52 VanessaE I see what happened. old version of farming_plus snuck in. 10:05 nore pushing that in 5 minutes: https://github.com/Ekdohibs/minetest/commit/5dc9df8bbe3a3936489e1a9a66406ae0bead41c1 10:07 VanessaE Ekdohibs?? 10:07 nore I just learned that my previous name was a trademark :/ 10:07 nore so I changed it 10:07 Calinou that doesn't make it invalid 10:07 VanessaE seriously? O_o 10:08 nore yeah, but I don't want any problems :) 10:08 nore and that one isn't used anywhere 10:08 VanessaE I dunno, in a google search for "novatux", your github is the first result. 10:08 nore hm, not for me 10:08 VanessaE and the second..and third.... :P 10:08 Calinou thank Google's search bubble 10:08 VanessaE fourth is some linux distro 10:09 VanessaE or tool, idk what 10:09 nore well, it is a trademark selling linux computers I think 10:09 VanessaE eh 10:09 VanessaE well better safe than sorry I guess. 10:09 nore but anyway, I'm sure not to have any problems if I change it :) 10:10 nore (and I carefully checked that google didn't give any results this time) 10:10 Calinou ask all those guys who use pony names as usernames :P 10:10 jin_xi yes its a really weird nick :) 10:10 nore (and yes, that name is stupid, I used a password generator to get it) 10:10 nore :D 10:10 VanessaE that's a hell of a way to come up with a handle :"D 10:10 VanessaE "D 10:10 VanessaE ... 10:11 nore anyway, pushing it :) 10:37 TBC_x my github nick is way weirder 10:50 kilbith RealBadAngel: how do you like that ? https://lut.im/u4nXhXCi/r7yWwGZq 10:50 kilbith then we can fill the right column for screen and mouse settings 10:51 VanessaE shaders option needs to allow for bumpmapping-only mode also 10:51 VanessaE off/low/mid/high is too vague 10:51 kilbith say to Valve 10:51 kilbith say that* 10:51 VanessaE we ain't valve :) 10:52 kilbith also you feel forced to react on literally everything ? 10:52 kilbith i asked RBA 10:52 VanessaE if it's gonna affect how I use the project, or for people who use my stuff, well yeah. I'll react. 10:53 kilbith you react on everything for the sake of saying something 10:53 VanessaE *sigh* 10:53 kilbith and it's often hollowing relevant 10:55 VanessaE do whatever you want then 10:55 kilbith yeah, i don't need your input, really 10:55 * VanessaE debates deleting the HDX normalmaps repository 11:05 TBC_x separate menu screens/formspecs would be better, then you can use tabs to expand settings 11:06 TBC_x IE you would have tabs in a submenu, not in the root formspec 11:06 TBC_x will see what RBA had in mind though 11:12 kilbith what do you mean by splitting menu screens/formspecs ? 11:14 kilbith you mean move the options in a separate formspec ? 11:14 TBC_x practically everything, that don't belong together 11:14 TBC_x more traditional menu screen 11:14 TBC_x for example I kinda like the one in AoE 2 11:15 TBC_x to be more precise 11:16 TBC_x The sub-screens could look like today's main 'menu' 11:16 TBC_x but the root menu would have buttons 11:16 TBC_x but I don't know whether there was a discussion about this before 11:16 kilbith i see, more organized in sub-places 11:16 TBC_x yes 11:18 TBC_x then there will be more space for more options 11:27 TBC_x +1 if we transform formspecs into mapblock representation 11:28 TBC_x but that is very, very, very experimental task 11:28 TBC_x IE to manage inventory, you have to manipulate nodes 11:33 TBC_x though, this is mission impossible... but anyone can prove otherwise, if they have all the time in world 12:14 celeron55 what kind of policy should we have for using the minetest logo? i mean, the license is clear, but do people care for the trademark-ish concerns at all? 12:14 celeron55 someone is asking about using the logo as the logo of a let's play youtube channel 12:15 VanessaE I use the logo in a couple of textures here and there, because it just "fits"; I don't see a problem using it in a youtube video so long as it's not trying to misrepresent minetest or something. 12:15 celeron55 the concern from our standpoint probably is that it might make the channel weirdly official looking and would misrepresent this community 12:16 celeron55 no not video, but the channel header 12:16 VanessaE ohh 12:16 VanessaE that's a lot more grey 12:16 VanessaE in that case I think I'd suggest some basic courtesy 12:17 VanessaE a disclaimer or something, or changes to the logo to differentiate it from the real one 12:18 VanessaE going hard-ass on someone using the logo is probably not good for minetest's image :) 12:18 celeron55 i guess i can answer something like "you can, but you have to avoid misrepresenting yourself to be a developer of minetest or some other kind of official source" 12:19 VanessaE yeah, seems fair 12:19 celeron55 which is common sense, but i guess sometimes it has to be written down 12:20 nrzkt +1 12:39 dzho well, trademark law is not something you can just make up 12:39 dzho I mean, you *can* but it isn't going to mean much. 12:44 celeron55 well for all practical purposes, Minetest is an unregistered trademark owned by me 12:44 celeron55 how effective that actually is depends on the country 12:45 dzho sure 12:47 celeron55 it might be interesting to ask some FOSS lawyer about whether it might actually be enforceable at all 8) 12:47 celeron55 my bet would go to "nope", but at least all the good people follow it nevertheless 12:49 celeron55 and it's not even worth taking into court because it's a bad name to begin with, and that's why nobody would put any money on it anyway so none of this will ever be tested 12:49 dzho it could go the other way 13:01 TBC_x The logo IMO should be modified in such a way that anyone can tell the differences between original, for example replacing the tree with a mushroom 13:02 TBC_x only in a case, where the logo does not represent original Minetest 13:04 TBC_x I think this covers some of the licenses? 13:23 TenPlus1 hi folks 13:24 nrzkt hi 13:30 TenPlus1 how's the latest build fairing with the random crashes > 13:38 RealBadAngel celeron55, our code is free, but our logo shouldnt be imho 13:39 RealBadAngel we are known already as best possible free alternative to mc 13:39 TenPlus1 which is great if the minetest binary was stable 13:39 RealBadAngel it is made stable from time to time 13:40 TenPlus1 0.4.12 was the last release and it's not vey stable 13:40 RealBadAngel next one is coming soon 13:40 TenPlus1 glad to hear... has the memory leak problem been looked into ? 13:41 RealBadAngel which one? 13:41 TenPlus1 I ask because Xanadu server is currently unplayable since it's crashing all the time due to memory leaks and std::bad_alloc errors 13:42 TenPlus1 I know you fixed the one with the minimap.. what others are there ? 13:48 TenPlus1 sorry if I seem miffy... minetest is an amazing game and having a popular server like Xanadu running like a wounded goat doesnt help with promotion... lol 13:49 TBC_x Xanadu is not the only server in the universe 13:50 TenPlus1 I know... am talking from a personal view... 13:51 TenPlus1 we (xanadu) had 45 players at one time which was great, but now it's quiet since it keeps crashing and players get frustrated 13:54 TBC_x crashing servers are IMHO owner responsibilities 13:54 TenPlus1 tbc, usually I would agree but it's the minetest binary causing the crashes, not the mods it's running... those are running fine... 13:54 RealBadAngel and ours 13:55 TenPlus1 and the serer has 4gb of memory of which only 600mb is used on minetest 13:55 RealBadAngel TenPlus1, we need feedback with logs 13:55 TenPlus1 I've been going over every mod time and time again incase I missed something... 13:55 RealBadAngel we have to know what is the reason in order to fix it 13:56 TenPlus1 RealBadAngel, it's the same errors every time... closed due to server error... stg::bad_alloc or segfaulting 13:56 TenPlus1 nothing specific to tie it down to one thing 13:56 RealBadAngel is there a chance you can run a debug build? 13:57 TenPlus1 the launchpad site I get the daily builds from doesnt seem to have a dgb build 13:57 TenPlus1 and the server owner isnt tech savvy 13:58 TBC_x why can't you get him here? 13:58 nrzkt TenPlus1: using stable builds not master can prevent crash due to non declared stable builds. 13:58 nrzkt if you want stable you take stable. It's like tell, i'm using debian sid and some things are not working properly, it's not normal. 13:58 TenPlus1 nrzkt, I used stable builds of 0.4.12 and it still crashed due to std::bad_alloc errors... I was hoping each daily would fix previous problems 13:59 nrzkt ok 13:59 TenPlus1 and going back to 0.4.10/11 is out since most of the mods have been changed to suit 0.4.12 now 14:01 TenPlus1 am using Lubuntu 14.04 LTS which is rock solid as a base, and yeah the server may be an old pentium D but works amazingly well bringing testament to how well Minetest runs on old hardware 14:16 * VanessaE senses the pending arrival of Zeno` 14:23 * nrzkt sent VanessaE 14:23 VanessaE heh 14:24 VanessaE well guess he's not gonna sign on :P 14:27 RealBadAngel TenPlus1, without backtrace we cant help 14:27 RealBadAngel it could be anything, including storm on the Sun 14:27 RealBadAngel we just dont know 14:28 TenPlus1 am searching for a debug version to put on server 15:07 TenPlus1 bye folks 16:48 est31 I didn't find novatux as registered trademark, neither french nor us, http://bases-marques.inpi.fr/ 16:48 est31 but yea I understand, better avoid trouble 16:56 Calinou lots of people have names that are also trademarks, without any issues to be honest :P 16:59 TBC_x Is network/connection.cpp documented somewhere? 17:00 est31 network/connection.h? 17:00 TBC_x oh, missed that single comment block 17:27 TBC_x CONTROLTYPE_DISCO Is for actual game data I expect 17:28 TBC_x what purporse serves sender_peer_id? 17:28 est31 DISCO could be short for DISCONNECT 17:28 TBC_x oh 17:36 TBC_x what is field sender_peer_id for anyway? 17:40 est31 dunno 17:41 est31 it might be possible to even fake the peer id as sender, and get admin privileges, im not sure 17:41 est31 has to be investigated, nrzkt has said something along tha 17:41 est31 t 17:41 TBC_x that's why am I asking 17:42 nrzkt peer_id is bad, we should use a cookie_id which is uniq at each session, but this is another problem 17:42 TBC_x what does the _id resolve anyway? 17:42 TBC_x i mean peer_id 17:44 TBC_x do you think that cookie_id couldn't be faked too? 17:45 nrzkt in fact peer_id is used everywhere, and if i'm right it's sent to other clients in some packets (but i could be wrong). Also peer_id is a min(available_id) and can be used to spoof current connected players packet 17:47 est31 can't we just take the ip address for comparison? 17:47 TBC_x IP:port 17:47 est31 e.g. check whether the incoming packet's peer id matches with the ip 17:47 est31 and port too 17:47 nrzkt TBC_x solution is better, it's not the best but it's better 17:49 TBC_x The only thing that can be abused can be AFAIK chat 17:49 est31 well yeah 17:49 est31 /grant player all 17:49 nrzkt how can this be spoofed ? 17:49 est31 its the easiest I guess 17:49 nrzkt we don't check the incoming player name ? 17:50 TBC_x Is that a obstacle? 17:50 nrzkt we don't need to have incoming player name, only a peer_id 17:50 est31 that can be faked 17:50 est31 nrzkt, the problem is that the peer id can be faked 17:50 est31 and as you said is guessable 17:50 nrzkt and we send to other players peer_id + messages, in MT it's name + peer_id because we don't have ID mappers clients side, and i hope we can have this later 17:51 nrzkt when i will rewrite sessions before the major release we will change this u16 peer_ids by better things 17:51 nrzkt a unique session cookie eg a UUID is good 17:51 est31 good you say it 17:52 est31 because I have changed the chat packet in a pr 17:52 est31 in the utf-8 pr 17:52 nrzkt oh 17:52 nrzkt this is a good idea 17:52 nrzkt did you change ids ? 17:52 est31 no, because I didn't think of, and we dont have clientside ids 17:53 nrzkt it could be good to have a id solver for that :) 17:53 est31 but before we make ids, we have to close this vulnerability, if it indeed exists 17:53 nrzkt i think it could be good to have this thing added to client and server core for future release (not this release merge window is too light) 17:54 est31 agree 17:54 est31 we'll have to do it later on 17:54 est31 I propose we raise to protocol 26 now then 17:54 nrzkt why ? 17:55 nrzkt we have 25 for this 0.4.12 as we said 17:55 est31 no 17:55 est31 because there is a change from RBA which needs a raise 17:55 nrzkt we will not change packets since release, no ? 17:55 est31 I had the chat change planned for 26 too 17:55 nrzkt protocol v25 is experimental because it was not release 17:55 est31 but if we want ids instead of names, then we need a raise 17:56 nrzkt yes, but we will do it after the current release 17:56 est31 yes 17:56 nrzkt before launch the 0.5.0 of 0.4.13 17:56 est31 nrzkt, nobody forces us to 25 for 0.4.12. 17:56 nrzkt i have some design ideas but there is a little difficult thing with name resolution using ids 17:56 est31 we have to study what goes wrong if somebody hasn't updated 17:57 est31 if only small things go wrong, it can be changed 17:57 nrzkt we give stable release, clients should use stable release. dev releases are not our problem it's a risk owned by compilers 17:57 est31 if you can't log in anymore or chat is failing, we haveto do a new protocol 17:57 est31 no 17:57 est31 nrzkt, this way we piss off people who install nightlies 17:57 est31 thats not the way we should do it 17:58 nrzkt ok, install firefox and complain about crashes. 17:58 nrzkt firefox nightly 17:58 est31 nrzkt, crashes are involuntary 17:58 nrzkt and compat problems and new features problems neither 17:58 est31 and yes, please complain, we need the complaint info to fix bugs 17:58 TBC_x any client ID sent by client is wrong 17:58 nrzkt firefox ensure compat with stable release not with the nightly cycle itself 17:59 est31 nrzkt, they dont change basic features, they always keep backwards compat 17:59 est31 nobody just changes how http works in firefox nightlies 17:59 est31 also its not comparable 17:59 nrzkt i talk about stable release compat yes 18:00 est31 because firefox is just a client 18:00 est31 and we have client and server 18:00 nrzkt yes 18:00 nrzkt servers can have more features than client, it's normal 18:00 est31 what the fuck does this have to do with the discussion 18:00 TBC_x nobody just changes how http works in firefox nightlies because http does not uniquely identify client 18:00 nrzkt but if a client has an incomplete protocol v25 implentation because user uses nightly without updating it it's not our problem 18:01 est31 TBC_x, we discuss something else here, whether to raise protocol version for RBA's change 18:01 TBC_x oh that 18:01 est31 perhaps I havent communicated this yet 18:01 nrzkt you want to merge it before next release ? 18:01 est31 its already merged, but not activated 18:01 est31 because it was intended to be activated with chat 18:02 est31 as I said 18:02 est31 both for protocol 26 18:02 nrzkt okay, then after the current release cycle, right ? 18:02 est31 no 18:02 est31 before 0.4.13, with that 18:02 est31 chat* 18:02 TBC_x We need a mechanism to separate protocol versions apart 18:02 est31 but that has been the intention, now we have chat after the current release 18:02 nrzkt then we are in 0.4.13-dev and protocol v25 isn't present in 0.4.12, right ? 18:02 est31 now the intention is to have chat after 0.4.13 18:03 TBC_x and that's not uni-protocol 18:03 est31 nrzkt, agreed 18:03 nrzkt then protocol v25 is the 0.4.13 version :) 18:03 est31 but one raise per release isnt what we should to 18:03 est31 do 18:03 nrzkt we can update it if you want 18:03 nrzkt but please don't use backward compat for inner release cycle dev 18:03 est31 instead we should raise every time if we change something basic to the protocol 18:03 est31 small changes arent needed 18:04 nrzkt if we change 4 packets in the release and one was modified Two times before a release, don't give compat to first version of the packet. 18:04 est31 nrzkt, agreed, after we have released, we can drop backward compat for something added and removed 18:04 est31 agreed 18:04 nrzkt compat is for stable releases not inner nightly 18:04 est31 thats ok 18:04 nrzkt perfect :) 18:05 TBC_x protocol_id is the field containing protocol version? 18:05 TBC_x or minetest_protocol? 18:05 est31 nrzkt, we do give compat for the first version of the packet, but drop it after the release. 18:05 est31 this allows the switch to be smooth 18:06 TBC_x ehm... how is the protocol negotiated? 18:06 TBC_x fields defined in connection.cpp or in payload? 18:06 est31 in init packets 18:06 est31 in payloads 18:07 est31 (from a connection.cpp perspective) 18:07 est31 I think the packet is called TOSERVER_INIT 18:07 est31 there we send maximum and minimum client protocol versions 18:08 TBC_x theoretically, we can bump protocol_id and create new protocol when the old becomes obsolete 18:08 est31 then the server answers with TOCLIENT_HELLO i think, with the actual protocol 18:08 TBC_x to keep backward compatibility we can listen for old protocol_id 18:08 est31 TBC_x, currently we do it otherwise 18:10 TBC_x the only issue i see is that all compatible protocol versions are kept in a single file 18:11 TBC_x and is hard and confusing to use older version 18:12 TBC_x the transport layer could use some stripping 18:12 TBC_x connection.h 18:12 TBC_x imho 18:14 TBC_x all the protocol version specific things could sit behind an vtable 18:16 TBC_x btw, anyone done any benchmarking on priority channels? 18:19 est31 whats a vtable? 18:19 TBC_x virtual table, you get this on virtual classes 18:20 TBC_x in virtual methods 18:20 est31 TBC_x, so you want to copy the entire file only if one method changes? 18:20 TBC_x that's one way 18:20 TBC_x or you can subclass 18:20 est31 a subclass chain? 18:21 est31 its not nice, I think the whole protocol should be in one file, this makes studying it easier 18:21 TBC_x then keep the classes in a single file 18:21 est31 imagine if a packet A was changed in protocol 10, the answer packet in protocol 25. 18:22 est31 then you have them at two completely different parts of the file 18:22 TBC_x I expected that old protocol versions are supposed to be cast into concrete 18:22 est31 and in order to find out whether the packet checked 18:22 est31 whether the packet changed* 18:23 TBC_x besides, unit tests should cover that 18:23 est31 you will have to scan the entire list of changes 18:23 est31 no 18:23 est31 unit tests are the last resort, before human testing 18:23 est31 they shouldnt be an aid to help you understand the protocol 18:23 est31 they should be "oops, sorry" 18:24 TBC_x well... system tests or whatever is the jargon 18:25 est31 ok, so I push a commit which raises the protocol to 26 now, ok nrzkt? 18:25 nrzkt ok 18:32 est31 pushed 18:32 est31 https://github.com/minetest/minetest/commit/c35747bc7d8d8c6c3e21b18a40fed84cd119dd7d 18:51 TBC_x hmm... minetest_game in 0.4.12-stabe has bug in crafting formspec 18:52 est31 ? 18:52 TBC_x when you grab crafting output, there is no check whether you're holding anything 18:54 TBC_x with the cursor 18:54 est31 yes 18:54 est31 but you get both back 18:55 est31 sort of a bug yes 19:06 nrzkt est31, can you review #2952 it's a trivial merge 19:06 ShadowBot https://github.com/minetest/minetest/issues/2952 -- Remove some old dead code. Fix some Clang warnings in SRP by nerzhul 19:06 nrzkt (i noticed a tab issue, i fix it soon, just review the other parts) 19:07 est31 nrzkt, can you make a pr to csrp too? 19:07 est31 https://github.com/est31/csrp-gmp 19:07 nrzkt okay, will do it 19:08 nrzkt done 19:10 est31 I dont see it 19:13 nrzkt oops don't validate the change, done now 19:14 nrzkt okay tab fixed. I switched from eclipse to qtcreator, it's better for c++11 needs :p with eclipse it's a pure pain 19:16 nrzkt and it's faster, cmake is properly included :p 19:31 nrzkt will merge #2952 in ~20mins 19:31 ShadowBot https://github.com/minetest/minetest/issues/2952 -- Remove some old dead code. Fix some Clang warnings in SRP by nerzhul 19:44 paramat nore sfan5 any comments on game#588 ? 19:44 ShadowBot https://github.com/minetest/minetest_game/issues/588 -- Default: Thicker snow nodebox by paramat 19:44 sfan5 I'll look at it later 19:45 paramat cool 19:49 paramat aha i will fix underwater mushrooms very soon =/ 19:52 kilbith est31, sfan5 : https://github.com/minetest/minetest/pull/2953 20:37 paramat pushing soon game#590 20:37 ShadowBot https://github.com/minetest/minetest_game/issues/590 -- Flowers: Keep mushrooms above water level by paramat 20:54 paramat now pushing 590 to game 20:57 paramat push complete 21:05 sfan5 paramat: i agree with C1ffisme the thick snow on trees looks a little unrealistic 21:06 paramat well, pines can get very thick snow on them, and it looks gorgeous 21:07 kilbith http://homesinestes.com/wp-content/uploads/2009/04/loaded-pine-tree.jpg 21:08 paramat heh i was about to search for some images 21:08 kilbith you're right on this, paramat 21:09 paramat i'm going for that 'scandinavian forest' feel 21:12 est31 paramat, can you approve this: https://github.com/est31/minetest/commit/288302a1f0db8c2c48d8b3a24db9889aaaabcd51 21:12 * paramat looks 21:12 est31 the problem here is that dedicated_server_loop in server.cpp has the opposite meaning 21:16 paramat oh man, this is beyond me, and i'm super tired, best to ask another dev, sorry =) 21:16 est31 :( 21:16 est31 sfan5, can you look? 21:16 est31 np paramat 21:18 sfan5 est31: imo thats a trivial change, you can push that 21:18 paramat anyway yes, an image search for 'snowy pines' shows they can be loaded up even more than my commit. thick snow looks good, the current reminds me too much of depressingly thin english snow =( i love snow 21:20 est31 its 25 cm snow in real life. but otoh, pines are larger I think than in real life 21:20 est31 the main problem I think is that the snow doesnt cover the sides 21:21 est31 this is a problem of us being a voxel game :) 21:21 est31 sfan5, ok pushed 21:21 est31 thx