Time Nick Message 01:34 zorman2000 paramat: Thanks for the biomes info. I didn't knew that mgv5 also had biomes 01:35 zorman2000 paramat sfan5 est31 #4774 is almost there!! 01:35 ShadowBot https://github.com/minetest/minetest/issues/4774 -- Adding particle blend, glow and animation by sfan5 02:44 paramat http://i.imgur.com/4JhWkQG.png dangerous stairs 07:14 paramat hi hmmmm, saw your message, don't concern yourself with #4061 that is reviewed and approved now. thought i'd notify you of #4829 in case you want to add input 07:14 ShadowBot https://github.com/minetest/minetest/issues/4061 -- Add loaded blocks to transforming_liquid-queue similar to updateLiquid by MillersMan 07:14 ShadowBot https://github.com/minetest/minetest/issues/4829 -- Mapgen: Make mgv7 the default mapgen by paramat 07:15 hmmmm oh no it's fine I actually did look at it a little bit 07:15 hmmmm didn't get too in depth but it seemed alright 07:15 hmmmm it's a lot of work to do a real review above what a static code analyzer could do 07:16 paramat i guess i'll merge 4061 now, guys shall i merge #4827 too? 07:16 ShadowBot https://github.com/minetest/minetest/issues/4827 -- Update credits by Zeno- 07:17 paramat that could perhaps wait until closer to release 07:17 hmmmm i mean it's one thing to do a 'code review' and it's another to actually understand what the code does, how, and why 07:18 hmmmm if i know i can't add value i'd rather not pretend to 07:19 paramat i'll merge 4061 4827 in a moment 07:20 hmmmm well hold off on 4827 for a minute 07:20 hmmmm this is not some kind of code that needs code review, it's a policy change 07:20 hmmmm how many people agree with you? 07:21 hmmmm i hoped to see a lot of "wow great +1" in the comments but.. not really 07:22 paramat 4827 is update credits 07:22 hmmmm whoops 07:22 hmmmm i confused it to 4829 07:22 paramat ah thought so 07:24 hmmmm i don't want to hold up 4061, but 07:24 hmmmm there are minor things like two identical access specifiers next to each other in a class definition 07:24 hmmmm the bracket is on a single line for a class which it shouldn't be 07:25 hmmmm and the entire ReflowScan class is really an interesting choice to be a class 07:25 paramat no problem i won't merge then 07:25 hmmmm actually nevermind on that one, it makes sense because of how much state is being shared between private helper methods 07:25 paramat credits can wait too 07:26 hmmmm but these are dumb style errors 07:27 hmmmm it's not worth holding up a PR for it that's been waiting a while 07:27 paramat ok cool 07:27 hmmmm i can't do a meaningful review without putting some serious time in it 07:27 hmmmm sorry 07:27 paramat that's ok 07:27 hmmmm can you just fix it up right before you merge? 07:27 paramat yeah 07:28 hmmmm errr wait actually nevermind you can't do that anymore unless you were to do the old way of merging PRs 07:28 paramat i'm able to alter a PR on merge 07:28 hmmmm ? 07:29 hmmmm you'd have to pull it to your local repo, cherry pick, modify, amend, then push 07:29 hmmmm maybe i'm just lazy but i wouldn't say it's worth it 07:30 paramat could you add line comments to explain the needed style changes? 07:31 hmmmm mmm 07:32 hmmmm it makes me look petty if those are my only comments on a PR 07:33 paramat to alter on merge i do: git clone, wget patch, git am patch, alter, git add, git commit -amend, git push 07:33 paramat this keeps the original author 07:33 hmmmm git fetch 07:34 hmmmm git cherry-pick commit-ID 07:34 hmmmm modify 07:34 hmmmm git commit -a --amend, git push 07:35 paramat so, it should be 'class ReflowScan {' 07:36 paramat i'm just unclear about 'two identical access specifiers next to each other in a class definition' 07:36 paramat i can alter on merge no problem 07:36 hmmmm i did it 07:36 paramat ok 07:37 paramat i see 11:24 Zeno` I suggest reverting https://github.com/minetest/minetest/commit/f8fd43 because people are complaining that it causes FPS drops (yeah, yeah, I know...) 11:24 Wuzzy hi. any comments on this?: https://github.com/minetest/minetest/issues/4844 11:24 Wuzzy oops I mean https://github.com/minetest/minetest/issues/4843 11:25 Wuzzy Zeno`: WTF? how is this even possible 11:26 Zeno` I don't know. I don't even know why it was necessary to merge 11:26 Zeno` anyone who is doing rendering work already knows how to enable wireframe mode 11:26 Wuzzy does it cause fps drop for you? 11:26 Zeno` Wuzzy, my *guess* is that it's less aggressive at culling things 11:27 Zeno` not really, but lots of things that are performance issues don't affect me 11:27 Zeno` this is a debug feature and people expect it to be optimised. I disagree with that 11:28 Wuzzy and i cant reproduce the claim that it causes fps drops ... hmmmmmmmmmmmmmmm 11:28 Wuzzy do you know when this is supposed to cause an fps drop? 11:28 Zeno` I can't reporoduce either but I am suspecting culling 11:29 Zeno` looking down also has issues apparently... (which I can see)... things go black 11:29 Zeno` but this is a DEBUG thing 11:29 Zeno` not for a user 11:29 Wuzzy any way, i dont care if the wireframe gets removed again. 11:29 Wuzzy but it would be silly if a feature gets nuked based on false or obscure claims 11:29 Zeno` as a dev I don't give a single wit whether it's optimised or imperfect because it does not affect a normal player 11:30 Wuzzy :) 11:30 Wuzzy oh, i guess i just reproducet the claim 11:31 Wuzzy its simple: Just make sure you force MT to load many MapChunks (or how they are called), then go into far view mode to make fps drop to 30 or so. if you then go to wireframe mode you will see a second very notabl fps drop to 20 or so 11:33 Zeno` Wuzzy, but does it matter? 11:33 Wuzzy but honestly, i couldn't care less. :P might be a bit inconvenient to be forced to press f5 another time for players who just use debug mode to see coords 11:33 Zeno` As a dev I, personally, am kind of used to suboptimal stuff 11:33 Wuzzy but at the end of the day, no. it does not really matter IMO 11:33 Zeno` if I am debugging then I expect things to be slower 11:34 Wuzzy so, any comments on ? 11:36 Wuzzy Zeno`: I might post an issue about wireframe making things slower, though. you know, for "documentation purposes" (aka gets ignored eternally ;-) ) 11:37 Fixer Zeno`: additional wireframe mode drops fps? how much slower? 11:37 Wuzzy Fixer: For me, it's from 30 to 20 in far view mode 11:38 Wuzzy Fixer: But at 60 FPS in non-far view mode, i see no drop 11:38 Fixer Wuzzy: when it is enabled? or when it is disabled? you compared builds with and without it? 11:38 Wuzzy uhh what? 11:39 Wuzzy wireframe mode is enabled with F5 11:39 Wuzzy I use compiled version... 11:39 Wuzzy 8a7dc838a8c1f8f2a5e9b710a8ee27d4d00715f9 to be precise 11:39 Zeno` honestly, IMO if a dev has a problem with the performance of the wireframe mode then *that* dev should fix it 11:39 Fixer if i understand correctly if you just compile with this mode available, it will be slower, no? 11:39 Fixer or it is performance problem of this wireframe mode? 11:39 Wuzzy what what whaaaaat 11:39 jin_xi ugh. whole point of initial wireframe suggestion was to make overdraw visible and do something about it 11:39 Zeno` I'm certainly not even bothered enough to look into it 11:40 Zeno` fps drop? Yeah you're in a debug mode. Accept it, fix it, or move on 11:40 Wuzzy if merely compiling wireframe mode would make MT slower, that would be indeed very bad 11:40 jin_xi that is clearly not the case. 11:40 Fixer wireframe mode is useful for debugging culling and other rendering stuff, it maybe slower, but if normal mode is OK, i don't think it is a problem 11:41 Fixer Zeno`: but fps in normal mode is good? 11:41 Fixer people are not meant to play in wireframe mode %) 11:41 Zeno` Fixer, no... I'm saying that comparing the performance of wireframe mode to how the game is meant to be played is... well, it's not an issue for me 11:42 Fixer Zeno`: same for me, it is for debug, fps is not critical in this mode and tbh it was very playable in this mode 11:43 Zeno` response should be: don't play in debug mode -- it's for debugging 11:43 Fixer yes 11:43 Fixer people who need it - will turn it with F5 11:43 Fixer on servers wireframe mode is dissalowed 11:43 Fixer problem solved 11:44 Fixer there is no problem in the first place why i'm even texting this 11:45 Wuzzy there, issue posted: https://github.com/minetest/minetest/issues/4850 11:45 Wuzzy have fun mauling it xD 11:51 Zeno` celeron55 what is your view? 11:52 Zeno` is an fps drop when using wireframe mode somehow a performance regression or not? 11:52 Zeno` I don't know why, btw, there is an fps drop in wireframe mode but that's only because I don't care enough to look into it 11:53 celeron55 wtf are you even talking about 11:53 Zeno` read up. People are experiencing FPS drops when using the debug wireframe mode 11:54 Zeno` personally I don't think this an issue at all but *shrug* 11:54 celeron55 why are they using debug wireframe mode 11:54 Zeno` fuck knows 11:55 Zeno` and that they experience FPS drops in a dev-only mode means little to me (I just don't care) 11:55 celeron55 well fuck cares then 11:55 Zeno` but maybe it can be explained 11:55 Fixer Wuzzy: what is far view mode? R? 11:55 Wuzzy yes 11:55 Fixer I don't have such fps drop at all 11:56 Zeno` they should try playing the game using valgrind :D 11:56 Wuzzy as i already said, i actually couldnt care less about wireframe mode either. :P 11:57 Zeno` I really don't care but I could not explain to the person why there is an fps drop. I guessed that it was because there was less culling but I cbf looking at the code 11:57 celeron55 it's like giving kids ice cream and then they complain it tastes bad 11:57 Zeno` well, bad ice cream is a whole different issue! 11:58 Fixer can't reproduce any fps drop in that mode 11:58 celeron55 eat the real food instead if you don't like the ice cream 11:58 Zeno` yes. bacon 11:59 Fixer Zeno`: culling is the same in both modes i think 11:59 Zeno` Fixer, I dunno. I haven't looked at the code because I don't think it's an issue 12:00 Zeno` honestly, IMO if a dev has a problem with the performance of the wireframe mode then *that* dev should fix it 12:00 Zeno` :) 12:01 Zeno` most (nearly all?) debug stuff is not optimised in any way at all 12:01 Zeno` why should it be? 12:02 Zeno` anyway, it seems we're all in agreement 12:14 est31 wireframe mode is not made for fast fps 12:14 est31 its made for developers 12:14 est31 its also not a "regression" as the commit only added the option to enable wireframe mode 12:23 Zeno` est31, that's why I'm writing a PR now to add a flux capacitor 13:59 Wuzzy 3rd try... 13:59 Wuzzy any comments on ? 17:17 Fixer_ found new lighting bug 17:18 Fixer_ https://i.imgur.com/eHzVWLU.jpg 17:19 Fixer_ https://i.imgur.com/UV7O8Rh.jpg 17:49 Fixer_ https://github.com/minetest/minetest/issues/4852 18:24 sfan5 sofar: are you saying the hunger mod writes to the *mod* directory to store world-specific values? 18:36 sofar sfan5: no, hunger actually uses a hidden player inventory, so that ends up living in the /players/ folder 18:36 sofar but for instance, the `stamina` mod I made puts data in to /datastorage/ 18:36 sfan5 in minetest/mods/? 18:37 * sofar looks and actually sees he hasnt used datastorage yet in stamina, either, lol 18:38 sofar sfan5: no, not in minetest/mods/ 18:38 sfan5 then that PR doesn't change the behaviour 18:39 sofar so currently mods can read mod folder data? 18:40 sfan5 can't with security enabled 18:40 sfan5 mods can read everything in the world folder anyway 18:41 sofar so my criticism of some mod stealing e.g. an IRC mod credentials coded in a lua file would still be valid 18:41 sfan5 yeah 18:43 sofar but yes, I think I misunderstood that part about *what* is readable :) 18:44 sfan5 the world folder was always read+write for mods 18:44 sofar yes, and that's likely unavoidable 18:45 sofar the attack vector of "player gets command block" is likely going to remain a risk at all times, though 18:46 sofar any server operator that has one of those on their server should reconsider running a server :) 18:49 sfan5 why? 18:49 sfan5 command blocks are "player-owned" which means they execute commands with the permissions of their placer 18:49 sfan5 at least the ones i've seen 18:50 sofar because if the player can execute lua, he can read mod code? 18:50 sofar maybe it would require //lua from worldedit 18:50 sfan5 command blocks are usually for commands not lua 18:51 sfan5 and if you grant any player the server priv, good luck 18:51 sfan5 ("server" is needed for //lua) 18:51 sofar defense in depth 18:52 sofar no matter how far fetched an approach may seem, people will figure out a way past unlikely scenario #1, #2, #3 etc. 18:52 ShadowBot https://github.com/minetest/minetest/issues/1 -- GlowStone code by anonymousAwesome 18:52 ShadowBot https://github.com/minetest/minetest/issues/2 -- Burned wood 18:52 ShadowBot https://github.com/minetest/minetest/issues/3 -- Furnace segfault 18:52 Krock heh 18:52 sfan5 good job ShadowBot 18:53 * sofar smacks ShadowBot 18:53 Krock at least it works :3 18:53 sofar and quick too 18:53 Amaz Unlikely scenario glowstone, burned wood, furnace segfaults :P 18:53 sofar does it kick itself if I make it quote 5+ bugs? :) 18:56 sofar a simple way around schematics inside mod folders would be 18:56 sofar copy them to the worldfolder on init 18:56 sfan5 sofar: actually you are wrong 18:56 sofar it doesn't kick itself? :) 18:56 sfan5 if you grant a player the "server" priv it's game over 18:56 sfan5 lua doesn't matter 18:56 sfan5 you can do /set secure.security_enabled false with that priv 18:57 sfan5 that is explicitly allowed since builtin doesn't run in a secure env: https://github.com/minetest/minetest/blob/master/src/script/lua_api/l_util.cpp#L81 -- https://github.com/minetest/minetest/blob/master/src/script/cpp_api/s_security.cpp#L244 18:57 sofar is there a (hypothetical) situation where a player may be capable of getting a commandblock and then change the owner of the block to an admin? 18:58 sofar e.g. I trick the admin to taking over my house? 18:58 sfan5 uh no 18:58 sofar social engineering 18:58 sofar "please add my extension to your structure" 18:58 sfan5 that is no command for changing owner 18:58 sfan5 you'd have to manually come up with something 18:58 sofar well remove area, add new area 18:58 sfan5 no 18:59 sfan5 command blocks are like locked chests 18:59 sfan5 area protection (if any) doesn't matter 18:59 sofar a mod could possibly do that 18:59 sfan5 indeed 18:59 sfan5 a mod could also have a bug that changes the owner of a different node 18:59 sfan5 you cannot protect against every scenario 18:59 sofar you can try 19:00 sfan5 it will be tireless work and unnecessary in 99.99% cases 19:00 sfan5 okay i'll wait while you do that and we stop updating minetest_game 19:00 sofar that's a cop-out 19:00 sofar at least you should acknowledge the problem 19:00 sofar I'm not saying "fix all the bugs now" 19:00 sfan5 i'm not saying that it isn't a problem 19:01 sfan5 i'm saying that it is unnecessary to try to fix all ways despite their unlikelyness 19:01 Krock some problems have priority over others 19:01 sofar well some things are making the problem worse 19:01 sofar allowing every mod to read all other mod code, for instance 19:01 sofar that's really bad 19:02 sofar I don't care how unlikely it would be 19:02 sofar I'd want that plugged 19:02 sofar that's just my experience and perspective, though, I get what your perspective is too 19:03 Amaz I'd like a fix that means I don't have to keep on telling people to disable security in order to run LOTT... It seems stupid that a mod can't read data in its *own* folder. 19:03 sofar Amaz: if the mod copies it's data from the mod folder to the world folder on init, the problem is gone 19:03 sfan5 the PR we're discussing is a workaround for that 19:03 sfan5 sofar: no that's an even worse workaround 19:04 sfan5 what if the mod includes the 80MB collection of some text 19:04 Amaz sofar: That seems very messy. 19:04 sfan5 s/the/an/ 19:04 sofar yawn? 19:04 Wuzzy OK, my 4th and final try for today: 19:04 Wuzzy any comments on ? 19:04 sfan5 ideally it could be fixed the proper way but that doesn't seem possible 19:04 sofar Wuzzy: I euh... looks ok? 19:05 sfan5 Wuzzy: i agree that it's inconsistent, not sure about the proposed syntax 19:05 sfan5 some commands may be more complicated than that propoal permits 19:05 sofar I don't think a mod should even be able to read its own assets after init 19:05 Amaz Why should it be necessary to have 20 different copies of some schematics?? 19:05 Wuzzy the proposed syntax is closely based on what is already used 19:05 Wuzzy sfan5: example? 19:05 sfan5 dunno 19:05 sfan5 i said "may" 19:06 sfan5 sofar: say goodbye to either 1) your ram or 2) your storage space 19:06 Wuzzy sfan5: Even if, I would strongly recommend any modder to even dare to introduce such a command and thus unleash True Evil unto the Minetest ;) 19:07 sofar sfan5: it's a point, but at the benefit of entirely removing the entire problem 19:07 sfan5 but that doesn't work 19:07 sofar how many 80mb mods exists that would need copying? 19:07 Wuzzy any command which is more complex than this syntax is probably way too overcomplex and it might be considered to be rethought 19:07 Wuzzy There are 80mb mods? O_O 19:07 sfan5 i hope not 19:08 sfan5 but it doesn't eliminate the problem 19:08 Fixer my mod folder is 800 mb 19:08 Fixer with .git 19:08 Wuzzy *check calender*. nope. its not april 1 19:08 sofar biggest I think is ambiance, but that's all sounds (so not shared stuff) 19:08 sfan5 what if i need 4gb of things at runtime? 19:08 sfan5 just put it into secure_mods but why? 19:08 sfan5 it doens't need to access anything else 19:09 sfan5 these are all hypotethical 19:09 Wuzzy argh crap I guess even if I start a PR about this command issue right now it probably wont get merged before feature freeze, am I right? 19:09 sfan5 Wuzzy: no 19:09 sfan5 the feature freeze is tomorrow 19:09 Wuzzy in which time zone? 19:09 sfan5 UTC i guess 19:10 sofar sfan5: I wouldn't want a mod with 4gb of assets, ever (and I'd prefer it if I had to manually get the content into the worldfolder, too) 19:10 Wuzzy cuz if its utc i only have 5h left PLUS i have to get devs to look at it... hmmm 19:10 sfan5 sofar: yeah these examples are not realistic 19:10 sfan5 but removing the ability to read at runtime is also just a workaround and can't solve the problem 100$ 19:10 sfan5 s/\$/%/ 19:11 sfan5 just as the PR is only a workaround too 19:11 sfan5 s/as/like/ 19:11 Wuzzy guess what? screw it. my feature request is not that important to me anyway. i am not going to do a quick and dirty cobbled pr only to get it in before freeze 19:12 sofar Wuzzy: right, it would be terribly short 19:12 Fixer https://github.com/minetest/minetest/issues/4853 19:12 Wuzzy woah these screenshots look CrAzY 19:14 sofar standard TNT :D 19:33 Fixer https://github.com/minetest/minetest/pull/4682 gives good results to me 19:38 nrzkt est31, interesting, android studio permit to link a cmake project with gradle 19:39 nrzkt it's new 23:16 VanessaE https://github.com/minetest/minetest/blob/master/doc/lua_api.txt#L620 some clarification is needed... why are the bits described at line 630 different from the ones immediately prior? 23:18 VanessaE hm, nevermind, I see the issue. the values 0-7 should be 0x0 - 0x07. 23:18 VanessaE otherwise the text looks like it's contradicting itself. 23:50 red-001 #4472 23:50 ShadowBot https://github.com/minetest/minetest/issues/4472 -- Make serverlist searchable. by red-001 23:52 red-001 sofar, it would only be around 20 - 30 lines shorter if the search didn't use keywords 23:52 sofar red-001: I still hate it, sorry 23:53 red-001 why is that? 23:53 sofar the keywords make it really complex, and you put 2 features into 1 commit 23:54 sofar I don't mind people making an API, and using it in 1 commit 23:54 sofar since adding an unused API is weird 23:54 red-001 unused api? 23:54 sofar but I really don't like "add feature X and on top of that, feature Y" 23:54 sofar well essentially the search bar is an interface 23:54 sofar maybe not an API, but certainly a documented interface 23:55 sofar did you add docs for the keyword stuff? 23:56 red-001 wait do you mean the stuff like "pvp:on" or "damage:false" when you say keyword? 23:57 sofar yes 23:57 sofar everything except just text search 23:57 red-001 that gets added in https://github.com/minetest/minetest/pull/4472/commits/1e2777925205c144934570aefb889d770b96e955 23:58 sofar ohh, so you did split it out? 23:58 red-001 yeah 23:58 sofar I must have missed that, sorry 23:58 red-001 into three commits