Time Nick Message 00:05 MinetestBot 02[git] 04Sapier at GMX dot net -> 03minetest/minetest: Remove accidentally added non-ft font files 1350ce63b http://git.io/uHGT4Q (152015-01-08T01:02:24+01:00) 01:09 acerspyro http://pastie.org/9819249 01:47 acerspyro https://www.dropbox.com/s/buw09cu473dm6c3/mt_debug.txt?dl=0 01:48 acerspyro Crash while joining VE's creative server 01:48 acerspyro Public server list fails to load half of the time, too 01:55 acerspyro VanessaE: ^ 02:04 VanessaE file it as a github issue 02:04 acerspyro I suck at issues 02:05 acerspyro My issues are always something like 02:05 acerspyro "Uhh.... Well, uhh... idk what to say.... Uhhhhhhhhh.... I guess 64-bit OpenSUSE Tumbleweed? Uhh, it's linux, and uhh, I suck at reporting bugs, and uhh, just look at the damn debug log." 02:06 acerspyro 02:07 acerspyro ch 02:07 acerspyro nvm 03:47 sofar having a hard time finding the equivalent command of /gamemode ... is such a command even available without a mod? 03:49 VanessaE sofar: use the main menu 03:49 VanessaE there's a checkbox there to change between creative and survival modes. 03:50 VanessaE some mods let you assign "creative" priv to a player, and then the inventory will adapt accordingly 03:50 VanessaE unified inventory, for example. 03:50 sofar ok, that's what I thought, thanks 03:51 sofar I'm toying around ... eventually I'll end up porting my bukkit plugins to minetest 03:53 VanessaE sounds good to me :) 03:56 * sofar installs HDX 03:56 VanessaE heh have fun with that 03:56 VanessaE that texture pack takes a lot of resources 03:57 sofar not too hard, I only grabbed the x32 ones, they seem fine to me 03:57 VanessaE oh ok 03:57 VanessaE it really starts to shine at 128x 03:58 sofar ehhhh it's a blocky game. No matter how hard you try. Can only make voxels so much pretty ;) 03:58 VanessaE heh 03:58 exio4 well 03:58 VanessaE try anyway if you have the RAM for it :) 03:58 exio4 that's what people said 03:59 exio4 then people started making less voxely models... 03:59 exio4 nodeboxes and models 03:59 sofar I'm playing on a 32bit linux box, so (yeah yeah I need to fix that) 03:59 exio4 people ain't sane over here 03:59 exio4 sofar, 32bit? what's this? 1995? 03:59 exio4 :p 04:00 sofar I have my reasons, already work professionally on linux distributions, so at home I don't want to spennd any more time than I have to 04:00 sofar I still mess around with it way too much anyway 04:01 sofar (the curse of being a paid Linux developer) 04:01 exio4 sounds the good kind of curse 04:01 exio4 +like 04:01 sofar it is 04:06 sofar fwiw here's the 2 mods I'd like to port over: 04:06 sofar https://github.com/sofar/Sedimentology 04:06 sofar https://github.com/sofar/Botany 04:06 sofar hey is the in-game mod screen broken? I can't see any mod past page 14/21 04:07 VanessaE not sure, but I always install mods manually anyways 04:08 VanessaE there was a mod for minetest that does something like that sedimentology mod 04:08 VanessaE and for botany, there's the plantlife modpack 04:09 VanessaE (particularly, plants_lib therein) 04:09 sofar yeah I'll make sure to test existing mods 04:09 sofar sedimentology is pretty complex, I doubt anyone came close to what it does in minetest so far 04:09 VanessaE maybe this one? https://forum.minetest.net/viewtopic.php?id=7111 04:10 VanessaE from his description, it looks like he was inspired by you :) 04:10 sofar does some of the basics, I'll certainly open that up deeper 04:11 VanessaE fork it 04:11 VanessaE Dan's no longer around and developing 04:11 sofar my version accounted for things like vegetation, rainfall, nearby water, temperature, elevation 04:11 VanessaE so it might save some time if you use his mod as a basis for yours 04:11 sofar don't forget I have a fully working version already 04:11 sofar that I spent hours tweaking :) 04:11 sofar well, more like weeks 04:12 VanessaE true enough, but yours is in Java, not Lua :) 04:12 sofar easy enough, one programming language or another 04:12 exio4 sofar, how can you handle Java without writing your own language that compiles to it? 04:12 VanessaE I wish you luck, we can always use more mods :) 04:12 sofar as long as I can modify blocks in the game somehow the concept is entirely similar 04:13 sofar my plugins are entirely stateless anyway 04:14 sofar his plugin is certainly a great way for me to start porting and learn the plugin API 04:15 VanessaE yep, modifying blocks in-game is pretty easy 04:16 exio4 the story of my life in a graph http://i.imgur.com/gtD8Dw6.png 04:16 sofar his plugin doesn't move blocks downhill looking at the code... 04:16 exio4 moving blocks downhill? 04:17 exio4 like, falling? 04:17 sofar that's what sedimentology does, yes 04:17 exio4 you can call falling_node(?) I guess 04:17 VanessaE exio4: HE MEANS TO FOLLOW THE PATH OF A RIVER/STREAM i THINK 04:17 VanessaE fuck 04:17 VanessaE sorry, caps 04:17 * exio4 presses VanessaE's caps lock 04:17 sofar check the animated gifs out here: http://dev.bukkit.org/media/images/69/44/Sedimentology-3.gif 04:18 VanessaE I'd rip the thing off, but on this keyboard, that feels "wrong"... 04:18 sofar it moves blocks downhill under certain circumstances, either under water or not 04:18 VanessaE sofar: ahhh gotchya 04:18 sofar I have a friend who remaps it to `ctrl` for emacs.... 04:18 VanessaE the engine has a limited form of that on new maps, goes by the term "mudflow" 04:18 VanessaE but it only works at mapgen time 04:19 exio4 hah, I did something like 04:19 exio4 it basically used a super-hack :D 04:19 sofar so I studied geology a long time ago... I've tried very hard to make that plugin as realistic and balanced as possible 04:19 exio4 I suck at geology 04:20 sofar you don't suck rocks, you lick them 04:20 * exio4 goes back with algebra and programming 04:20 exio4 sofar, toads, you mean 04:20 VanessaE hH 04:20 VanessaE haha 04:20 sofar I got to throw my 2 boys in bed, be back after story reading time 04:20 VanessaE gently 04:20 VanessaE :) 04:21 sofar oh no they like it when I really throw them 04:21 sofar (they're 4 and 6) 04:21 VanessaE heh 04:24 VanessaE there, took the stupid key off. 05:08 MinetestBot 02[git] 04paramat -> 03minetest/minetest: Return step smoothing value to 23 1356a89fd http://git.io/v8mdRw (152015-01-08T15:06:28+10:00) 05:39 Zeno` Gotta say, I like to the look of VoXus 05:39 Zeno` wonder where it's headed 05:40 Zeno` I particularly like the leaves on the trees 06:26 sofar /home/sofar/src/minetest-0.4.11/src/emerge.cpp:573: virtual void* EmergeThread::Thread(): Assertion '0' failed. 06:26 sofar heh, hitting quite them asserts still 06:26 exio4 what're you doing? 06:26 exio4 uhh 06:26 sofar just moving about the map and flying around 06:26 exio4 that 're 06:27 sofar 22:25:41: ERROR[EmergeThread0]: ERROR: An unhandled exception occurred: std::bad_alloc 06:27 VanessaE sounds like you're getting bit by that memory leak 06:27 VanessaE can you try a development build instead? 06:28 sofar oh yeah, I'll gladly do that. I can trail git if needed - I compiled from src anyway 06:30 sofar went well over 1.1g ram ... then pooped out again 06:31 VanessaE yeah, there's one or two still left; most have already been fixed and what remain have pull requests waiting to be merged 06:32 Zeno` there is still one in the server unless you change a .conf setting (because it can cause glitches) 06:32 Zeno` the glitches it causes will be there as a result of the crash anyway, so if it's happening a lot it might be worth adding anyway 06:32 Zeno` but, yes, this will require at least 0.4.11 06:36 Zeno` in minetest.conf you can set liquid_queue_purge_time = 36000 which will, every 10 minutes, half empty a queue that has a tendency to cause memory leaks. 36000 is 10*60*60 (liquid_queue_purge_time is in seconds) 06:36 VanessaE I thought you fixed that/ 06:37 Zeno` I did, but the default is now 0 06:37 Zeno` 0 = disabled 06:37 Zeno` because what happens is if the value is set too low then liquids that should be reflowed do not 06:38 Zeno` it kind of depends on how often the server runs out of RAM 06:38 Zeno` which is why it's kind of tricky to fix 06:40 Zeno` another option is to set liquid_loop_max to some very large value.... e.g. liquid_loop_max = 999999999 06:40 Zeno` which helps alleviate the problem but causes other problems 06:40 Zeno` (and doesn't get rid of the problem entirely either... just slows it down) 06:41 Zeno` that whole liquid flow stuff needs to be rewritten but it's not easy :( 06:47 * VanessaE wanders off to bed agaij 06:47 VanessaE again* 06:54 gregorycu Can we tell when we are running low on memory and start purging the liquid queue 06:54 gregorycu Whatever the fuck that means 06:54 Zeno` kind of hard to check available memory unfortunately 06:54 Zeno` especially cross-platform 06:54 gregorycu Irr doesn't provide a way? 06:55 gregorycu Are you able to explain what the liquid queue is? 06:55 Zeno` probably not, but TBH I haven't looked 06:55 gregorycu Is it something I could optimise easily? 06:55 Zeno` map.cpp lines 1616- 06:55 Zeno` that function 06:55 gregorycu Thank you 06:56 Zeno` by default the number of liquids in that queue processed per call to the function is 100000 06:56 Zeno` (this is to stop DoS by carefully placed liquid blocks) 06:57 gregorycu 100k 06:57 Zeno` but unfortunately the m_transforming_liquid queue will just keep getting larger and larger and larger 06:57 Zeno` err 10000 06:57 gregorycu That's a lot 06:58 Zeno` so what purge tried to do was... after xxxx seconds, discard the the oldest items in the queue 06:58 Zeno` yeah it is 06:58 gregorycu Um 06:58 Zeno` but the queue is millions in size 06:58 Zeno` (can be millions in size) 06:58 gregorycu So, this is a container of block coordinates? 06:58 gregorycu With duplicates? 06:59 gregorycu A queue 06:59 Zeno` it shouldn't have dupes, but I didn't check 06:59 gregorycu It shouldn't have dupes... ok 06:59 Zeno` yes, no dupes 06:59 gregorycu That was my next comment 06:59 Zeno` it's a uniquequeue 07:00 Zeno` not that tranformLiquids() also adds to the queue itself (but this is not the main -- or really even any -- issue) 07:00 Zeno` note* 07:01 gregorycu If we could tell when we are getting memory constrained, that would make our solutions way more elegant 07:01 Zeno` that queue can easily consume 1GB or more on a server with just one player 07:01 gregorycu Than, you know, picking what we think are good magic values 07:01 gregorycu But I'm aware it's hard 07:01 Zeno` That's why there is no default value 07:02 Zeno` and why the feature is currently disabled (the default purge value is 0) 07:02 Zeno` 10000 is to stop the server getting stuck in that loop for a huge amount of time because a person places a liquid source in a cunningly devised location 07:03 Zeno` perhaps the 10000 should not be a value at all, but the function only iterate for a certain amount of time 07:03 Zeno` but that will not fix the increasing queue size issue 07:04 Zeno` if the 10000 is not there and there are 9 million blocks to process on a server then the server can't do anything else (if the 10000 "limit" of iterations was not in place) 07:04 Zeno` I think I said somewhere the whole function needs to be rewritten and/or thought about a lot more 07:04 gregorycu bb in 5 min, I have comments 07:04 Zeno` but maybe that was in a PM to someone 07:05 Zeno` ok, PM me when you get back so I notice? 07:18 Zeno` 10000 used to be 3000 as well I think 07:19 Zeno` I'll grab a profile of the current server without any purge 07:26 gregorycu So, UniqueQueue uses a map, and a list 07:27 gregorycu Two containers ill suited for their purpose 07:27 exio4 errr 07:28 gregorycu What is wanted, is a set, and a queue 07:28 gregorycu I'm not saying the performance characteristics are drastically different, but they may be 07:29 Zeno` I agree 07:29 Zeno` but will they increase performance enough to "keep up"? 07:30 Zeno` well, they can't 07:30 Zeno` it would mitigate the problem of course 07:30 gregorycu I agree with that 07:30 gregorycu We do need a proper solution where the liquid transformations consume a "fair" slice of available processing 07:31 Zeno` I made a mistake... liquid_loop_max = 10000 <-- 0.4.10 default liquid_loop_max = 1000000 <- 0.4.11 default 07:31 gregorycu "fair" meaning "equitable", not "large" 07:31 Zeno` the reason for me increasing that value was all the other optimisations I made to other parts of the engine was causing the queue to grow much more rapidly than in 0.4.10 07:32 kahrl gregorycu: imho the bigger problem than using wrong data structures is that UniqueQueue.push_back needlessly makes two lookups 07:32 gregorycu Which is classically how a lot of performance problems become visible 07:32 Zeno` and "fair" is kind of hard to define. It should (probably) have to take into account client lag 07:32 Calinou liquid spreading is broken since (somewhere in 0.4) 07:33 Zeno` Calinou, which is why I said the whole function (or a group of functions) need to be reconsidered 07:33 gregorycu kahrl: that is true 07:33 gregorycu If we change it to a set, we can just insert all the time 07:33 gregorycu And check to see if the insert was successful 07:33 kahrl https://github.com/kahrl/minetest/commit/17d6a4355afcc5c452e4ddd695715bcdc97cabaf#diff-e5ab4d3af8e375631903ed35b243c3b4R28 07:34 Zeno` that makes it faster... 07:34 Zeno` but how does it ensure that the queue does not "run away"? 07:34 kahrl of course I wouldn't be against replacing list with queue 07:34 kahrl Zeno`: it doesn't 07:34 Zeno` oh, I was talking to gregorycu :) 07:34 Zeno` lag 07:34 Zeno` ircd needs to be optimised 07:35 Zeno` :D 07:35 kahrl hehe 07:36 kahrl the only thing I can think of is to remove entries from m_transforming_liquid when the corresponding mapblock is unloaded 07:36 kahrl but that would be extremely slow with the current data structures 07:37 LazyJ What if liquids didn't flow? Everything else in Minetest is a cube. Instead of "flowing", the cubes spread out a limited distance and the edges exposes to air are replaced with a slope? 07:38 Calinou then you couldn't distinguish still from flowing 07:39 LazyJ No. But if that closed a potential hazard and memory leak it may be a worth-while trade-off. 07:39 kahrl whoops, line 60 (in my link) should be // Already existed 07:40 LazyJ Zeno mentioned earlier that if liquid source nodes were placed in certain locations they could cause some serious problem. 07:41 kahrl also instead of processing liquid nodes in a FIFO fashion, it would perhaps be better to prioritise them by distance from the nearest player 07:41 Zeno` LazyJ, they can cause (in effect) a DoS on the server 07:41 kahrl (I don't know what the best data structure would be for that) 07:42 Zeno` by stalling the server in the transformLiquids() loop (and that's why the number of iterations is limited in the first place... in 0.4.9 they were not) 07:42 Zeno` current memory profile for a short execution of the server (1 client): http://i.imgur.com/JHxcGXv.jpg 07:42 gregorycu (Sorry, had to go AFK, have bali belly) 07:43 gregorycu Yes, there are two problems here 07:43 gregorycu Firstly, the liquid transformation code is slow 07:43 gregorycu Secondly, it can run away 07:43 Zeno` if you trace the source code you will see the liquid queue increases because of createBlock and createBlankBlock (just trust me or trace it yourself :D) 07:43 Zeno` gregorycu, correct 07:44 gregorycu Making the first better, makes the second a little bit less of a concern 07:44 gregorycu And is easier 07:44 Zeno` instead of limiting the number of iterations *all* the time though, maybe things can go into "suspected DoS mode" 07:44 Zeno` Not sure 07:44 gregorycu I don't like it 07:45 Zeno` I don't either. But remember this is bad enough for 1 client wandering around 07:45 gregorycu I'd honestly prefer "CPU constrained mode" 07:45 Zeno` isn't that the same thing with a different name? 07:45 Zeno` :) 07:45 gregorycu Somewhat, but one implies malice 07:46 Zeno` true 07:46 Zeno` and in most cases the issue is not due to malice at all 07:46 gregorycu And also, it detects a cause, not a symptom 07:46 Zeno` that massif data I just collected involved no malice 07:46 kahrl I also agree with limiting time instead of # of nodes 07:46 kahrl because some nodes take very little time of process and others ages 07:46 Zeno` I just had 1 client connected and continuous_forward = true 07:47 Zeno` kahrl, I agree with that as well 07:47 gregorycu In general, the server should have a "target tick rate" 07:47 gregorycu Much like a FPS target 07:47 gregorycu If the server starts to lag behind that, it starts throwing away work 07:47 gregorycu Slowly at first, and then more aggressively 07:48 gregorycu Flowing liquids are not as important as sending clients entity updates 07:48 hmmmm [02:43 AM] if you trace the source code you will see the liquid queue increases because of createBlock and createBlankBlock (just trust me or trace it yourself :D) 07:48 hmmmm WTF does this mean 07:48 gregorycu I don't know, but I'm about to find out :) 07:49 hmmmm a block got loaded and it happened to have liquid so the liquid got added to the queue? 07:49 Zeno` hmmmm, those are where the liquids are coming from 07:49 hmmmm liquids don't simply come out of thin CONTENT_IGNORE 07:49 hmmmm (i wanted to say CONTENT_AIR which was punnier but less accurate) 07:50 Zeno` err yes, you're right. This was so long ago I don't remember accurately 07:50 gregorycu What's the best way to profile this Zeno` , stand at the top of a mountain and drop some water? 07:51 hmmmm hahahaha 07:51 hmmmm that's an awesome way to DoS a server 07:51 Zeno` gregorycu, set noclip to true and walk around underground making lava caves 07:51 Zeno` hmmmm, and yet you could do that in 0.4.9 :D 07:52 hmmmm @ kahrl's idea to prioritize liquids according to player proximity 07:52 hmmmm the only problem i have with that is the liquids will probably get processed after the player already left that area 07:53 hmmmm otherwise it'd be trivial as using a priority queue (heap) 07:53 kahrl hmmmm: well it's mostly intended as a fallback 07:53 kahrl ideally the queue would always be mostly empty and completely processed in a server tick 07:54 gregorycu prioritizing liquids according to player proximity is hard 07:54 hmmmm oh i know 07:54 gregorycu And may add significant processing time 07:54 hmmmm a 3 dimensional btree 07:54 hmmmm yeah, agreed there 07:56 hmmmm i can't come up with an efficient way to search, frankly 07:56 hmmmm well 07:56 gregorycu I think the best approach is to take the CPU starvation problem as an edgecase 07:56 hmmmm i guess you could have three b-trees, one for each direction 07:57 gregorycu A supported edgecase 07:57 gregorycu The player proximity queue is optimising for the edgecase, at the expense of the common case 07:57 gregorycu As kahrl suggested 07:58 hmmmm so you look up nodes in a.. 5 node range in the X direction (std::map::find and then iterate back and forward until out of items or next element is out of range), do this 2 more times for Y and Z and then take the intersection of all three results 07:58 hmmmm :D 07:58 hmmmm and of course you need to do this for each player 07:59 hmmmm dammit why is this IRQ timing out... 08:00 Zeno` I guess it's keeping up better than before: http://pastebin.com/Ca5JrsWD 08:00 Zeno` I don't like it though 08:00 kahrl how about this rough approximation: 1. have one liquid queue per mapblock instead of a global one, 2. in transformLiquids first loop over the active blocks and then over the non-active blocks 08:00 Zeno` now it's 1200000 08:01 gregorycu I like hmm's approach, except I'd just iterate forward 08:02 Zeno` I dunno. Both are probably worth nutting out in more detail 08:02 gregorycu Anyway, I still think we're optimising an edgecase 08:02 gregorycu Why can't it keep up? 08:02 gregorycu I still don't understand 08:03 Zeno` it mostly can now 08:03 Zeno` it's when lots of mapgen is happening 08:03 Zeno` say you have 4 players all exploring 08:03 Zeno` or 1 player with no clip exploring 08:04 Zeno` there is always lava being added 08:04 kahrl well, let's try doing the limiting by time instead of # of nodes 08:04 kahrl in my experience the nodes generated by mapgen are always processed very quickly 08:04 hmmmm yes 08:04 Zeno` but they're not 08:04 Zeno` http://pastebin.com/Ca5JrsWD 08:04 hmmmm i don't even understand why it isn't time limited to begin with 08:04 Zeno` I was doing nothing there 08:05 Zeno` jsut a client flying through space 08:05 hmmmm that's because you're talking about lava nodes and kahrl is talking about oceans 08:05 Zeno` that queue should be < 100000 and probably close to nothing 08:05 hmmmm bonus points if you're working on mapgen v5 08:05 Zeno` and it certainly should be always going up :) 08:06 Zeno` I might add some code and print out the queue size on my server every 10 minutes to debug.txt 08:06 hmmmm for what it's worth, mapgen v5 caves near lots of lava pockets is pretty much the worst case mapgen caused liquid transforming scenario 08:06 Zeno` even though I don't have many players these days :( 08:08 gregorycu What governs the flow rate of water? 08:08 hmmmm viscosity 08:08 Zeno` maybe that could be added anyway so we get a 24-hour view of the queue size on an average server 08:08 hmmmm there's also a config setting 08:08 Zeno` surely someone will volunteer to collect the data hehe 08:09 gregorycu If I crank the viscosity up to something high, that should better show CPU bottlenecks right? 08:09 hmmmm also, lower liquid_update 08:10 hmmmm no, cranking the viscosity up would make less of a CPU bottleneck 08:10 hmmmm it's the opposite 08:10 gregorycu Sorry, yeah 08:14 Zeno` yes it does need to be time-based rather than node-based for a start 08:15 Zeno` kahrl, are you doing that? :D 08:27 kahrl nope, I have to leave :P 08:28 kahrl I might do it on the weekend if nobody has by then 08:36 Zeno` you're not allowed to leave 08:36 Zeno` :P 11:51 luizrpgluiz hi 12:04 lordawe hi 12:24 Jordach time for some more manglement 13:03 lordawe hi guys, please, I am running small MT server (LAN) on Windows and I think it needs some tweaking little bit. In newly generated world there are some shadow issues and I need to tweak performance a little bit - it might be because of some mod incompatibility. Is anybody, who understand it, willing to help me? Please send me a PM. Thank you very much! 13:14 Jordach the shadows are a mapgen bug 13:14 Jordach plant a sapling and it should go away 13:15 gregorycu Is there an issue for this bug? 13:17 Jordach gregorycu, nothing you can do until a mod happens to run a voxelmanip lighting update 13:17 gregorycu ? 13:18 gregorycu So this is a lua mapget not telling the engine to redo the lighting? 13:18 gregorycu mapgen 13:18 Jordach gregorycu, black areas are caused by a mod placing / removing nodes not updating the light values for the area 13:19 gregorycu Shouldn't the light values be calculated by the engine and not the mapgen? 13:22 Jordach because some nodes might be a light source or allow light to pass through 13:31 lordawe well, but I've used WorldEdit and //fixlight with no luck at all 13:32 lordawe I also tried to place a light sources in the area, but it didn't fix light either 13:33 VanessaE select to encompass the dark area both significantly below and WAY above it, //replace air air 13:33 VanessaE works better than //fixlight 13:33 lordawe I tried that too 13:33 VanessaE make your selection wider and much taller then 13:34 gregorycu I think unloading and reloading the chunk should do it 13:35 gregorycu Wait 13:35 lordawe how can I do it? 13:35 VanessaE just go further out horizontally from the shadow, and further below/above when you set //pos1 and //pos2 13:35 lordawe 2VanessaE: I cannot replace all with air, because I would damage the surface 13:36 VanessaE nonono 13:36 VanessaE //replace air air 13:36 lordawe oh I see 13:36 VanessaE replace all air...with air. 13:36 VanessaE just do it blindly 13:36 VanessaE it won't damage anything. 13:39 lordawe I have selected 4.9 M nodes, replaced air with air and WE stated: 4,04M nodes updated 13:39 lordawe and 13:40 lordawe it looks the same 13:40 VanessaE strange 13:40 lordawe yes 13:40 VanessaE screenshot of the area? 13:41 lordawe w8 13:42 lordawe HA! oh no. so by doing replace air with air ENLARGED dark area actually 13:42 gregorycu hmm... 13:42 VanessaE do it again 13:42 VanessaE same selection. 13:43 gregorycu Keep doing it until everything is dark 13:43 VanessaE no, seriously. 13:43 VanessaE every so often, the //replace command gets the lighting wrong 13:47 lordawe http://i.cubeupload.com/iPF4bD.png 13:47 VanessaE yep, seen that before 13:48 VanessaE just repeat the //replace command again 13:48 gregorycu Mordor 13:48 lordawe lol 13:49 lordawe hey! second pass fixed a part of it! 13:49 lordawe http://i.cubeupload.com/kfDJpa.png 13:49 VanessaE ok good 13:49 lordawe but there are such places here and there, I wouldn't spend my life by fixing light :) 13:49 VanessaE now you know what to do :) 13:50 VanessaE why it happens in the first place, I don't know 13:51 lordawe it might be wrong mix of mods 13:51 lordawe so I need help 13:51 VanessaE it's not mods. 13:51 VanessaE it's an engine bug. 13:51 VanessaE this happens on vanilla maps too, every once in a while 13:51 gregorycu Jordach says it's mods 13:52 lordawe i disabled windmill mod because it created a HUGE cube of rotating windmills near this place 13:52 Jordach gregorycu, it's NOT MODS 13:52 VanessaE lordawe: that's a corrupt mapblock 13:52 gregorycu Sorry, mapgen 13:52 gregorycu Lualand 13:52 lordawe there was same huge cube of travelnet boxes, so many that The Matrix sux 13:53 gregorycu The mapgen shouldn't modify light values 13:53 gregorycu As they'll get recalculated by the engine from time to time 13:53 gregorycu Which should have been the case here 13:54 lordawe As I told before, I am starting with a new map, so I can create another one easily, but tell me what mapgen to use and how to configure it in a good way 13:54 gregorycu Which mapgen did you use? 13:54 VanessaE well for starters, update to 0.4.11 13:54 lordawe this was built in V7 13:54 VanessaE and stick to mapgen v6 13:55 lordawe it is 0.4.11 13:55 VanessaE no it isn't. 13:55 VanessaE you're using 0.4.10, so say you screenshots. 13:55 VanessaE and that looks like mapgen v5 13:55 lordawe that is my remote client, but server is 0.4.11 :) see? 13:55 VanessaE oh ok 13:56 lordawe okay, I really am not sure which mapgen is this, because I have tried all 13:56 lordawe I also tried paramat7 13:56 lordawe and noisetest or what is the name 13:56 VanessaE that's either mapgen v5 or v7, or maybe paragen 13:56 gregorycu I think the lighting is handled on both the client and server 13:57 lordawe definetly not paragen 13:57 lordawe because I disabled it 13:57 VanessaE I always just use mapgen v6 without any terrain mods 13:58 lordawe I want nice mountains not too steep, I do not want float islands, and I want lakes and rivers if posibble 13:58 VanessaE I seein there you're using plantlife modpack, that should be perfectly fine too 13:58 lordawe okay, I also use moretrees mod and can I reduce the amount of those huge trees? 13:58 Wayward_One What on Earth happened to the fonts?? 13:59 Wayward_One xD 13:59 VanessaE you can, the settings are in biome_defs.lua 13:59 lordawe okay, wait please 13:59 Wayward_One Just git pulled, now they're huge 13:59 VanessaE Wayward_One: blame sapier. https://github.com/minetest/minetest/issues/2077 14:01 Wayward_One *sigh* 14:03 Wayward_One Guess I'll be using a slightly older build for now lol 14:03 Wayward_One Unless... 14:06 Wayward_One Nope, too big for me lol. Is there any way to change the size in the .conf? 14:06 lordawe Hope I will not be kicked, I will pase mods I would like to use 14:06 VanessaE fiddle with gui_scaling 14:06 VanessaE lordawe: no. 14:06 VanessaE pastebinm 14:06 VanessaE -m 14:06 lordawe k 14:07 lordawe http://pastebin.com/QUHgdca3 14:08 VanessaE install them one at a time and see how they behave for you 14:08 lordawe I have them all already and it seems to be working 14:08 lordawe I want to add mesecons too 14:09 lordawe but I am not sure if boost_cart will work with it 14:09 VanessaE I'm not familiar with that mod. 14:09 lordawe https://forum.minetest.net/viewtopic.php?f=11&t=10172 14:10 Kalabasa Hello 14:10 VanessaE hi 14:10 lordawe you can turn by movement keys 14:12 lordawe I mean I needn't carts to be operated using mesecons, just wanna know if it can co-exist without bugs :) 14:12 lordawe I see you are not familiar, but maybe somebody else 14:12 lordawe okay. 14:12 VanessaE actually I have seen this mod 14:12 VanessaE now that you linked to it 14:12 VanessaE but I haven't used it yuet 14:12 VanessaE yet* 14:13 lordawe well, so, with a new map, I should stick to v6 mapgen and tweak biomes in biome_defs.lua 14:13 VanessaE I had looked at it briefly to replace the carts mod in dreambuilder but wasn't sure if it was suitable. 14:13 lordawe of plant_life mod 14:13 VanessaE yeah, for the moretrees if you want them to be less prolific. 14:14 lordawe It would be good if size of same kind of tree vary a little bit by random number 14:15 VanessaE that requires changes to the tree models. 14:15 VanessaE I never really coded them properly to vary in size 14:15 VanessaE but it's doable 14:15 VanessaE at the time, randomness didn't work quite right 14:15 lordawe I see 14:16 lordawe so, there could be i.e. 3 tree models and it would be selecting it randomly 14:17 VanessaE yes 14:17 lordawe okay 14:17 VanessaE that's easily done, just requires designing new models 14:17 kaeza greetings 14:17 VanessaE morning kaeza 14:17 lordawe I would do it, but how can I design it, I looked at models, but I do not get it :) 14:18 VanessaE meanwhile, I'm off. 14:18 VanessaE bbl. 14:19 lordawe cul 14:39 Zeno` hmm 14:40 Zeno` i got bitten my a spider! :-o 14:40 Zeno` by 14:40 Zeno` little shit 14:40 Zeno` he now dead 14:42 Jordach #aussie 14:43 Zeno` Yes I imagine it was an aussie spider 14:43 Zeno` it is now a pile of mush 14:44 gregorycu What type? 14:44 gregorycu Huntsman? 14:44 Jordach s/H/C/g 14:44 Zeno` nah not that big. Just some kind of copper coloured little thing 14:45 Zeno` bites all up my leg though :( 14:45 Zeno` about the size of a 5 cent piece (the spider) 14:45 Zeno` well it was 14:45 Zeno` not it's about the size of... I dunno 14:45 Zeno` the size of squashed stuff 14:45 * Jordach tries to get the height right of a evergreen tree mk2 14:46 gregorycu I hate spiders 14:46 gregorycu And fear them 14:47 Zeno` fortunately most of our spiders are not particularly dangerous 14:47 Zeno` although it kinda hurts lol 14:47 sfan5 MinetestBot! 14:47 MinetestBot sfan5! 14:47 Zeno` if I saw this particular spider on my desk it would not scare me 14:48 gregorycu i thought queensland spiders are quite dangerous 14:48 gregorycu funnel web for instance 14:48 Zeno` it's only the huge ugly ones that scare me 14:48 Zeno` yeah, well, if it was a funnel web I'd not be sitting here chatting hehehe 14:48 Jordach out comes the flamethrowers 14:48 * Jordach imagines an Aussie shirtless with a flamethrower 14:49 exio4 how is it going, bunch of cool people 14:49 Jordach <3 14:49 Zeno` funnel webs are a weird spider as well; they are aggressive and will actually chase you 14:49 Zeno` I don't know of any other spider that would do that 14:49 Jordach https://cdn.mediacru.sh/O/OkKNuXQLDtWP.png 14:50 Zeno` http://upload.wikimedia.org/wikipedia/commons/1/1e/Victorian_funnelweb.jpg <-- ugly bastard 14:50 Jordach Zeno`, what about gladiator funnel web spider vs venomous giant centipedes? 14:51 Zeno` centipedes? I have them here as well. Centipedes do not run towards you when threatened heh 14:51 Zeno` Seriously I do not worry about centipedes 14:51 gregorycu Let's not talk about spiders anymore 14:52 Zeno` I saw one the other day but unless you're an idiot it's pretty unlikely you could get bitten 14:52 Zeno` gregorycu, I can't stop talking about them! I have welts all up my leg heh 14:52 gregorycu It makes me scratchy too 14:53 * Jordach considers adding in a funnel web spider mob 14:53 Zeno` :-o 14:53 Jordach it is invisible but shouts Samuel M. Jackson's "Suprise mother fucker" when it hits you 14:53 Jordach (does no damage, but does make one shit themselves) 14:53 Zeno` no damage? 14:54 Zeno` it should kill you instantly lol 14:54 Zeno` de gads... legs itchy and kinda sore 14:54 Zeno` brb 14:54 gregorycu Ice bro 14:54 gregorycu Apply ice 14:55 Jordach Zeno`, in case of spiders burn the house down...with the lemons 15:01 * Jordach now has the issue of plugging in on_generated calls to spawn trees 15:03 Zeno` err I used Savlon 15:04 Jordach Zeno`, is there a way with on_generated to find a node and run a function? 15:04 Jordach oh wait 15:04 Jordach there is 15:13 Jordach Zeno`, why are you not in the other #minetest? 15:15 gregorycu What other minetest? 15:16 Zeno` because I am asleep :) 15:16 Jordach ah 15:17 Zeno` you know, gregorycu. The "other" minetest 15:17 Zeno` the one where all the action happens 15:17 gregorycu Ah... right 15:17 gregorycu I know what you mean now ;) 15:17 Zeno` inchra.net 15:18 Zeno` anyway, I' 15:18 Zeno` anyway, I'll bbiab... gotta somehow get my legs to stop itching :/ 15:18 gregorycu What is the purpose of the other minetest? 15:18 gregorycu #minetest 15:23 ThatGraemeGuy mainly to share chat between a group of servers I assume 15:24 Wayward_One can someone help me with this? http://pastebin.com/3XgZWU9n 15:29 gregorycu I think the freenode network is pretty big 15:30 gregorycu Well mint, looks like a bad mod 15:34 Wayward_One i know, i'm trying to fix it lol 15:38 gregorycu Well, it's like saying to a mechanic "there is something wrong with my car" and then showing a photo taken 500 meters away 15:38 gregorycu You'd have to look at the code in the file specified 15:38 gregorycu At the line specified 15:39 gregorycu It may be strikingly obvious, or maybe not 15:39 gregorycu In any case, paste the single line here 15:42 gregorycu Zeno`: If you're not sleeping, it looks like if a node is updated, it will attempt to update nodes next to it, even if it has a higher water level 15:42 gregorycu Which i believe is wrong 15:45 Wayward_One gregorycu: dofile(minetest.get_modpath("castle").."/pillars.lua") 15:46 gregorycu Interesting 15:46 gregorycu It appears that minetest doesn't think the mod called castle is installed 15:46 gregorycu Did you rename the mod? 15:47 gregorycu You totally did, didn't you, you cheeky devil 15:48 Zeno` I really am asleep 15:48 Zeno` will chat tomorrow 15:50 Wayward_One lol, i tried naming it "castle" and "castles", neither worked 15:51 gregorycu Um... 15:51 gregorycu That's not good 15:51 gregorycu Can you bear with me, I'm gunna check some code 15:52 Wayward_One ok 15:58 gregorycu Ok, I'm gunna get you to add a single line of code 15:58 gregorycu above dofile(minetest blah blah 15:58 gregorycu print(minetest.get_current_modname()) 15:59 gregorycu Then restart, and try to open the map again 15:59 gregorycu Actually, let me change it slightly 15:59 gregorycu print("HELLO! " .. minetest.get_current_modname()) 16:00 gregorycu What that will do is print the mod name into the console 16:02 gregorycu Let me know what is printed 16:04 Wayward_One ok 16:10 Wayward_One gregorycu: castle 16:10 gregorycu ffs 16:10 gregorycu Why couldn't you say something else 16:16 gregorycu If you replace the line I asked you to add with the following, it will print all mods, let me know if castles is there: 16:16 gregorycu for i,m in ipairs(minetest.get_modnames()) do print("MOD: " .. m) end 16:26 Wayward_One gregorycu: just got another copy to work, but thank you for your help! :) 16:27 gregorycu ? 16:27 gregorycu Cool 16:56 Krock hi 17:46 rubenwardy Hi all! 17:47 Krock olleH! 17:48 * Jordach wonders how vmanip areas are sorted 17:48 Krock ..sorted? 17:48 Jordach local p_pos = area:index(x,y,z) \n local content_id = data[p_pos] \n print (p_pos) 17:49 Jordach and turn that result into an pos.x, pos.y and pos.z like table 17:49 Krock p_pos is a number, representing the index in the data array, containing x, y and z variables 17:50 * Jordach realises 17:50 Jordach print (x..y..z) 17:50 Krock (z * size * size + y * size + x) 17:50 Jordach the three for loops solve everything 17:50 Krock exactly 17:52 Jordach Krock, http://paste.debian.net/139771/ 17:53 Jordach i'll just see if it shits itself 17:53 Krock "axp.x, 1 do", the ", 1" isn't required at all 17:53 Krock move "local data = vm:get_data()" one line down 17:53 Krock wait. that doesn't matter at all 17:55 Jordach http://paste.debian.net/139772/ 17:55 Jordach can't compare against x,y,z but that should bunny hop the question 17:56 Krock how about {x=x, y=y, z=z}? 17:56 Krock not sure how your functions work 17:58 Jordach http://paste.debian.net/139773/ 17:58 Jordach Krock, i'm a fucking idiot 17:58 Krock I see :P 17:58 Jordach that's how MT usually takes XYZ inputs 18:01 Jordach hon hon hon 18:01 Jordach it works 18:09 * Jordach is a fucking skrub 18:09 Jordach >leaves placement missing bool in one spot, randomly makes leaves snowy and not snowy 18:14 Jordach 10 trees grown after: 0.15s 3 trees grown after: 0.13s 19:50 blaise has the build process changed from 0.4.10 to 0.4.11 ? 19:51 VanessaE nope 20:03 kaeza https://www.youtube.com/watch?v=g_ULtNYRCbg 20:05 marktraceur Wow. 20:09 VanessaE impressive 20:11 VanessaE why don't we ever see stuff this complex done with mesecons? 20:11 rubenwardy Mesecons is slow 20:12 rubenwardy Well, chunks unload 20:12 VanessaE that's what world anchors are for... 20:14 VanessaE oops. wrong window. 20:40 ipv6b sfan5, Do you know whom I should contact for the VikingCrafters server? 20:40 sfan5 no 20:41 ipv6b sfan5, have you played any on that server? 20:42 sfan5 no 20:42 ipv6b its a popular one 20:42 sfan5 !server vikingcrafters 20:42 MinetestBot sfan5: VikingCrafters [=:=] Survival & PvP | 198.50.161.137 | Clients: 22/77, 25/50 | Version: 0.4.11 / minetest | Ping: 92ms 20:48 ipv6b Anybody who plays on VikingCrafters? 20:54 ipv6b That server has existed for a very long time. 20:55 ipv6b or at least one with a similar name 21:37 MinetestBot 02[git] 04ShadowNinja -> 03minetest/minetest: Remove included SQLite3 13a3db918 http://git.io/nLBUcg (152015-01-08T16:34:42-05:00) 21:37 MinetestBot 02[git] 04sfan5 -> 03minetest/minetest: Add SQLite3 libraries to buildbot 13d8cf64a http://git.io/RwGa4g (152015-01-08T16:34:42-05:00) 21:39 apdapreturns Hello 21:39 ipv6b hi apdapreturns 21:39 VanessaE hi 21:40 Brains VanessaE: Was there ever a cause found for that weird appearance of windmills and unknown blocks and the like that was mentioned in channel the other day? 21:41 VanessaE Brains: no, I can only conclude it was a corrupt mapblock or a mod that was creating weird mapblocks in general 21:43 Brains VanessaE: I had something similar in a git-build dreambuilder-game only world the other day. (I am an idiot and didn't note the commit I built...) I've since updated both, rebuilt, and started a world... Haven't seen it since so dunno if it was fixed or just not encountered again. 21:45 * Brains was thinking it was related to a crash he was having but that turned out to be the facedir > 23 thing. 21:45 VanessaE I'm still getting crashes while connected to my Creative server. when I can be bothered, I'll get a backtrace (though Wayward_One I think already did) 21:49 Brains My crashes were client-only and I never found any errors/warnings in the server log. What kind of logging options do you use for your servers? 21:50 VanessaE this is client-side also 21:51 VanessaE and my servers are just basic logging, and capturing the console output to a separate file also. enough to catch a crashdump. 21:56 lordawe 2Brains: about that windmill cubes, I do not think it was just a corrupt mapblock. Map was fresh and it appeared in two maps I've tried, so I disabled the mod and deleted it 21:57 lordawe maybe the mod is not 0.4.11 compatible, or it has to do something with v5 and v7 mapgens 21:58 Brains lordawe: I was running dreambuilder alone so I'm guessing that it isn't just a non-0.4.11 mod. 21:59 VanessaE that would be no. 22:00 VanessaE dreambuilder works fine in 0.4.11 22:00 VanessaE and has the windmill mod in it 22:00 VanessaE it's a mapgen bug or a mod generating a corrupt mapblock 22:00 iqualfragile_ who maintains pipeworks atm? 22:00 VanessaE iqualfragile_: mostly nore 22:01 VanessaE I started pipeworks, but I kinda maintain a hands-off approach to that mod these days 22:01 * Brains makes a note to look into the default settings stuff a bit later... His worlds were generated with very little in the way of settings (and no GUI at all). 22:02 VanessaE (because others generally do better than me at maintaining it) 22:03 iqualfragile_ because it seems to make minetest eat a gig of ram and then die 22:04 VanessaE that's... not right 22:04 VanessaE minetest has some known memory leaks 22:04 VanessaE couple of pretty bad ones afaik 22:04 VanessaE but nothing that should be triggered by pipeworks 22:05 carjack good evening 22:05 VanessaE bbl 22:06 carjack big bad lion? 22:06 VanessaE be back later :PO 22:06 VanessaE :P 22:10 iqualfragile_ VanessaE: its definitly caused by pipeworks, when i disable it everything works 22:11 lordawe VanessaE: pls, what lua file did you say I can tweak biomes ? 22:20 acerspyro VanessaE: You seem pretty in-demand today, but I just installed release 0.4.11 from my repos, and it crashes aswell on your server. 22:31 twoelk I just tried to tour that french server to see the cathedral but my brand new client keeps crashing as well 22:31 twoelk or rather it freezes and eventually crashes 22:31 acerspyro What distro? 22:32 twoelk uhm the krock version for win32 22:32 acerspyro ok nvm 22:33 acerspyro krock? wat? 22:33 twoelk hm the official 0.4.11 doesn't seem to like this pc much as well :-( 22:34 acerspyro idk, it only crashes in OpenSUSE, but it didn't in Fuckuntu. 22:35 * twoelk is starting up his laptop to compare 22:35 carjack she said bbl 22:36 acerspyro I wasn't there when she did 22:36 iqualfragile VanessaE: see -dev 22:36 carjack yes that is why I am say'ng 22:36 Zamack Is there an official/popular minetest mumble or teamspeak server? 22:38 carjack no official servers as far as I know 22:39 twoelk don't think there is such a thing as any official Minetest playing server. Only developing seems to producing official content. All playing is unofficial ;-P 22:40 iqualfragile btw: there is an unofficiall patch that adds mumble positional audio to minetest 22:40 Zamack Oh, cool. 22:40 Zamack Well, thanks for the info guys. 22:41 twoelk bah, got no micro hooked to this outdated box 22:41 acerspyro I do :D 22:42 Zamack that's okay, you could just listen/type 22:43 twoelk eh? mumble can say what I type? 22:43 Zamack Yes, actually...with text to speech on. 22:43 Zamack if everyone has it enabled I mean 22:44 twoelk cool feature 22:53 oreminer Hi! I seem to be too stupid to get normalmaps working, using hdx-512 and hdx-512-normalmaps with MT 0.4.10, copied all images to ~/.minetest/textures/hdx/, it just ignores the normalmaps 22:54 oreminer (shaders, bumpmapping is on, only generate normalmaps + parallax is off) 22:54 Wayward_One hmm 22:55 Wayward_One btw, latest version is 0.4.11 :P 22:55 oreminer I know, but nobody updated the gentoo portage ebuild yet, and I was too lazy to do it locally ;-) 22:55 Wayward_One lol 22:56 Wayward_One did you try with normalmaps on? 22:56 oreminer Yes, does not change anything.... 22:56 carjack gentoo :) 22:56 acerspyro lol 22:56 acerspyro I am on SuSE, and 0.4.11 is in the repos already. 22:56 Wayward_One hmm... I think VanessaE (creator of that TP) might know more 22:59 oreminer Mhhhm... that's interesting... if I ONLY use the hdx-512-normalmaps folder (which only has the normalmaps) and put that in the textures folder, I get the normalmaps 23:00 oreminer but on the non-HDX textures 23:12 Wayward_One weird 23:21 oreminer I'll give up for today... good night all! :-) 23:39 MinetestBot 02[git] 04Sapier at GMX dot net -> 03minetest/minetest: Switch default scaling factor to 0.75 as most ppl seem to have 96dpi screens instead of previous assumed 72 137ad17a2 http://git.io/tT9D6w (152015-01-09T00:37:02+01:00) 23:48 carjack yes 72dpi is not very popular