Time Nick Message 08:30 kahrl so I deleted my minetest.conf, made a new world, made myself a bucket, grabbed some water from the ocean, dropped it on top of a mountain 08:30 kahrl neither does it flow nor does the hole in the ocean get filled 08:30 kahrl this really needs to be fixed before 0.4.8, any progress on that? 08:31 kahrl (I'll mark #950 as milestone since it's essentially the same) 08:36 kahrl there needs to be a decision whether #882 or #911 or a compromise or something else will be merged 12:26 PilzAdam kahrl, I guess increasing liquid_loop_max to 10000 again should work 12:28 proller or https://github.com/minetest/minetest/pull/895 12:30 PilzAdam this seems to increase the speed of the default liquids 12:32 proller only if queue big 12:33 proller but now liquds have no constant speed 12:33 proller bigger queue -> slower 12:34 PilzAdam you increase liquid_update, that makes the liquids faster when the queue is not full 12:38 proller no, because decreased liquid_loop_max = 500 12:41 VanessaE the problem with high values of _max is risk of huge server lags 12:41 VanessaE there's basically no protection against excessive CPU usage in the water flowing code - or there wasn't at the time the default value was reduced. 15:00 celeron55 to sapier and whoever: minetest is supposed to be buildable by these: GCC of debian stable, latest clang, VC2010 and latest binary distribution of mingw 15:01 celeron55 this has been the case forever; it probably should be written down so that nobody has to ask it anymore 15:03 celeron55 (vc2010 will be dropped in favor of vc2012 some day; i have no idea when) 15:04 celeron55 (more like "some year") 15:09 BlockMen i hope not to soon because i have full version of 2010^^ 15:23 BlockMen celeron55, have you any comments on #928? 15:32 BlockMen or kahrl, or thexyz? 17:17 celeron55 there's some random guy asking by email basically about becoming a sponsor and getting some features in exchange; features which seem to roughly be about getting a largeish survival server instance with proper mobs running 17:17 celeron55 no details other than that really 17:21 celeron55 probably not worth the hassle as usual, but if someone has something to say or ask or wants this contact for some other minetest-related purpose, just tell 17:23 sapier sounds interesting ... but "proper mobs" is a little bit unclear 17:28 NakedFury proper mobs. way to be unclear 17:29 celeron55 well that was really only my interpretation 17:29 celeron55 here's the email content: http://pastebin.com/raw.php?i=vhV7H7hC 17:30 sapier hmm "rough compatibility with minecraft" I guess there will be some ppl that don't like that direction ;-) 17:31 sapier using the simplest possible interpretation "same mobs as in minecraft" this wouldn't be that difficult to be done ... but ai will still be limited by cpu performance 17:31 celeron55 whatever that means, there's no reason it would go in upstream minetest_game or whatever; it seems to me mostly like a request for customization or such 17:32 celeron55 as in "i'd like to run a server and i have some money or whatever but don't care to develop content myself" 17:32 sapier I'd assume same yes 17:33 proller all mobs logic must run in fast c++ core 17:34 sapier proller how may ai did you do recently? 17:34 celeron55 (in any case, i'm fine with throwing these kinds of customization things to any interested individuals that i deem trustworthy) 17:35 proller sapier, ~0 17:35 sapier it's not the logics but environment operations that require cpu power 17:35 sapier did you try mobf with delayed spawning enabled? almost no delay left 17:35 celeron55 (who can then discuss about engine development on behalf of the "guy who can pay for work") 17:36 proller but mobs now totally stupid 17:36 sapier everyone says mobs are stupid but once I ask why they are stupid noone answers 17:37 sapier proller could you give an example? I'd gladly fix it in next version ... but I can't fix things I don't know about 17:38 sapier and yes simple mobs are stupid they're designed to be stupid in order to avoid cpu usage 17:41 BlockMen intelligent mobs would be to hard for players :P 17:42 sapier that's one point and if their iq rises above 100 we'd reach singularity thus rendering mankind useless 17:42 sapier ;-P 17:43 Sokomine i'd really love to see mobf beeing used on servers - at least the friendly mobs 17:43 sapier no without joking are there any examples what behaviour is "stupid" 17:44 Sokomine hm :) i'd say "see a player - attack" is stupid. "see a player - RUN!" would be intelligent in most cases :) 17:44 sapier vombies once was able to kill player ... ppl didn't like that either ;-P 17:45 Sokomine perhaps those players who want mobs to fight against ought to tell what they want 17:46 Sokomine for me, having animals as decoration in the landscape and on farms is most important. younger players may get more excited about beeing able to kill something 17:48 sapier still I'd like to get an answer "WHY MOBS ARE STUPID" to me it seems like mob bashing atm as noone ever answers that question 17:49 BlockMen stupid as "not knowing any behavior of the enemy, no plan of attack" 17:50 BlockMen "no change in behavior if hit by player" and so on 17:50 sapier ok so you're talking about fighting behaviour 17:51 sapier thanks that's a point to start at ... by now I did focus primary on movement if we got to the point where movement isn't main point of annoyance I gladly improve fighting capabilities 17:52 BlockMen hmm...the way of moving itself is ok IMO, but the "random walking around" could be called stupid ;) 17:53 BlockMen but i have to say that i have not played much with mobs...so i may be wrong 17:56 sapier it's random so it should be stupid ;-) 17:56 sapier they just avoid to drop or run to areas where they shouldn't be e.g. sheep in desert 17:58 BlockMen sheep should seek for "default:grass" (1-5), move to it and remove that node 17:59 BlockMen and that wouldnt be thaaat cpu intensive 18:00 sapier they do in current dev version 18:02 * BlockMen didnt know 18:02 BlockMen nice 18:19 sapier thexyz I may have a solution for your localization issue 18:21 thexyz what's it? 18:21 sapier different gettext versions (re)evaluating environment in different orders 18:30 sapier wow those microsoft guys are really really strange yesterday vs2012 run fine on my win7 64 bit starting it now I get told it's known to be incompatible and can't be run ... the only thing that might have changed are automatic os updates 18:30 sapier I didn't even reboot :-) 18:33 BlockMen "sapier: I didn't even reboot :-)", its windows. you should do that ;) 18:33 sapier trying atm let's see if it's gonna work after that again 18:36 celeron55 inb4 it breaks more 18:37 sapier no windows still is convinced that microsoft visual studio express 2012 is incompatible to my win7 64 bit from today ... but not the one from yesterday 18:38 sapier ok they released a hofix for it ... so that damn os is yelling to microsoft what applications are started each time you start it 18:39 sapier ok maybe not each time ... but at least once 18:43 proller sapier, mobs must be intelligent as you (if you playing inside mob) 18:43 sapier that ain't possible in general ;-) 18:43 proller birds must fly fast (as birds) 18:44 sapier they do there's just a little problem once a bird flys fast it's gonna leave active area quite soon 18:44 proller various mobs must have various behavior (neutral, angry, ....) 18:44 sapier they already have 18:44 proller birds can fly by circles 18:45 proller with low hp mobs must run away 18:45 sapier why? 18:45 sapier did you ever hear of dodo's running away on mauritius? 18:46 proller maybe not all 18:46 proller and not always 18:47 sapier as of last version mobf has factions support 18:47 sapier including reputation mechanism 18:48 proller ok. 18:49 sapier used correct this will result in neutral angry behaviour 18:49 proller and need more biome-specific mobs, for snow and maybe lava 18:50 sapier true but a lot of work without ppl helping e.g. by testing 18:50 celeron55 birds don't generally fly most of the time; they just sit in tree and then sprint to an another tree and then sit there 18:50 celeron55 +a 18:51 sapier that's why mobf has only gulls those sit on ground (mobf ones don't) and fly long times 18:51 proller or sit | walk on ground and fly if you near 18:52 sapier good idea ... I think that'll be done for mobf 2.4 ... requires a little bit more work 18:59 Sokomine the gulls sailing in the sky are just fine. mobf does well there. no need for faster birds 19:01 Sokomine it just gets problematic when they sail 10 cm above ground at low velocity. as long as they're high enough it's fine 19:02 sapier Yes I know ... actually there are mechanisms to avoid this ... obviously this doesn't work everytime ... especialy if server/client get out of sync 19:03 Sokomine *nod* but in general those gulls are a very nice view 19:03 Sokomine hm. that "cyan wool" problem needs to be solved i'm afraid 19:03 sapier cyan wool problem? 19:04 Sokomine newer clients support more than 4k diffrent nodes. in order to do that, one node got a special one. unfortionately, on older worlds, this number is already occupied by a node. in case of redcrabs server, that's cyan wool 19:04 Sokomine it's not just an unkown block. it's impossible to dig with a new client 19:05 Sokomine the map needs to be rewritten once 19:05 Sokomine players switching to 0.4.8 will eventually have that problem as well... 19:05 sapier can't this be done on the fly once a map is opened? 19:06 Sokomine i don't know 19:07 Sokomine we also have that flowing liquid problem with old world vs new client. but that gets solved automaticly once the server runs a newer version. the "cyan wool" problem can't be solved that way. for servers, rewriting the map once might be the solution 19:08 proller liquid fixed in 882 19:08 PilzAdam Sokomine, I already told you that the content IDs in that are used at runtime have nothing todo with the content IDs that are used in the database 19:09 PilzAdam I dont remember any changes in this direction, so it might just be database corrpution 19:17 sapier thxyz ... grrrr same thing working on mingw doesn't work for msvc 19:17 thexyz what thing? 19:18 Sokomine no, it's not database corruption. someone (hmmm?) explained to me that it was due to allowing 4k nodes 19:18 sapier refreshing locale using current gettext does work ... using latest supported gettext version for vs doesn't seem to work 19:18 Sokomine just connect to redcrab.suret.net 30401 and try to interact with some cyan wool 19:19 Sokomine or linuxgaming.us with the asphalt with middle line 19:19 PilzAdam server client incompatibility is a different thing 19:20 Sokomine hm. so it might be due to the download+save variant? and will be sorted out once a new server is run? ok, in that case it won't be a problem 19:20 BlockMen devs, what about #928? any chance to get it into 0.4.8? 19:24 RealBadAngel squash it first 19:24 RealBadAngel and i think it could be merged 19:26 PilzAdam has someone tested it yet? 19:27 BlockMen RealBadAngel, sure, will do 19:27 celeron55 where are screenshots showing what it can do? 19:28 BlockMen PilzAdam, http://pastebin.com/5LwX8hhH have fun ;) 19:28 BlockMen celeron55, posted 1 or 2 months ago already 19:28 PilzAdam I like the "255,138,138,138;255,192,192,192;50,0,0,0;255,255,255,255;255,0,0,0" part 19:29 celeron55 uhm 19:29 celeron55 how did the discussion about this color stuff end previously? 19:29 BlockMen it ended that argb is ok (AFAIK) 19:30 BlockMen and here the old screenshot http://i.imgur.com/UnrmevT.png 19:30 sapier ok localization with latest official win32 gettext version and visual studio2012 doesn't seem to work at all 19:30 RealBadAngel argb or rgba are both fine 19:32 celeron55 RealBadAngel: the questions is about whether to use #hhhhhh or comma-separated integers 19:32 celeron55 -s 19:32 celeron55 or was 19:33 BlockMen http://irc.minetest.ru/minetest-dev/2013-08-04#i_3237919 19:33 RealBadAngel celeron55, it doesnt really matters what format we choose 19:34 celeron55 RealBadAngel: wat 19:34 celeron55 it matters by a ton 19:35 RealBadAngel why? 19:35 sapier readability 19:35 celeron55 RealBadAngel: what if it turns out to be a bad decision later and needs to be changed? 19:35 celeron55 then minetest needs to still support the bad decision too 19:35 VanessaE (I lean toward #hhhhhh, fwiw) 19:35 RealBadAngel for me hex is more readable, but that is not for most folks out there 19:36 kaeza #rrggbb looks more clean 19:36 celeron55 but anyway, it looks like for some reason i decided previously that integers are good so i'm trusting myself on that 19:36 VanessaE hex is what's used on most systems where an RGB triplet is needed, I think 19:36 celeron55 the issue is/was that #rrggbbaa is very non-standard and it's unreadable too with four octets 19:37 VanessaE how about #rrbbgg,aa then? 19:37 VanessaE er rgb* 19:37 RealBadAngel shaders already are using vec4 19:37 VanessaE (you know what I meant) 19:37 RealBadAngel in format rgba 19:37 BlockMen not that much IMO 19:37 BlockMen *better 19:38 RealBadAngel 4 components vector of floats 19:38 sapier supporting both isn't an option is it? 19:39 celeron55 sapier: it's always the worst option possible; anyway if i can, i'm locking this now to the previously made decision if at all possible 19:40 celeron55 (it's a waste of time to go over the same things again as nobody has learned anything new in this time anyway) 19:40 celeron55 BlockMen: change "inv" in your documentation to the full word "inventory" 19:41 celeron55 or alternatively remove the word 19:41 sapier it's just a color format so supporting different formats isn't as bad imho its like using 10=0xA=012=1010b 19:42 sapier but I'll be fine with any option 19:42 BlockMen kk 19:42 celeron55 so who has tested #928? 19:44 kahrl I'll have to think about formspec_table then 19:44 kahrl currently it uses #rrggbb colors with commas separating options 19:46 celeron55 that gives us a solid-ish reason to go for #rrggbb[aa] instead, in which missing aa is read as ff 19:47 VanessaE sounds good to me. 19:50 celeron55 listcolors[#8a8a8a;c0c0c0;#00000032;#ffffff;#000000] 19:50 celeron55 this does look a million times more clear 19:50 RealBadAngel indeed 19:50 kahrl I'll probably merge those separate "background" and "background_alpha" options in formspec_table 19:50 kahrl and highlight and highlight_alpha 19:50 celeron55 +# 19:51 celeron55 so it looks like BlockMen got a bit more work to do then 19:52 * BlockMen .. yay 19:52 BlockMen :P 19:52 RealBadAngel hehe 19:52 RealBadAngel not to mention code is shorter 19:53 BlockMen can someone test it anyway...then i know what else (i might have) to change 19:53 celeron55 remember to require the use of #; that way different formats can be differentiated if needed in the future 19:53 kahrl fwiw, here is my hex color parsing code: https://github.com/kahrl/minetest/commit/9f3a19468293e015786c3342c170d2076419f435#diff-cba62fd9bbe7903df6e74be690f396fbR1131 19:54 kahrl not terribly great code but it works (doesn't do #RGBA yet, though) 19:54 RealBadAngel btw, im finishing my work on shaders, last thing i need are refraction and reflection rendering to texture 19:55 RealBadAngel the question is, where to use them 19:55 BlockMen hmm...then a new function that tables use too? 19:55 BlockMen i mean not new, extended yours 19:55 RealBadAngel its impossible to get correct reflections on every height 19:55 kahrl BlockMen: yeah, would be best... I wonder where to put it 19:56 BlockMen guiFormspec? 19:56 RealBadAngel limit those effects to water level? 19:56 kahrl sounds fine 19:56 BlockMen since it is for formspecs 19:57 BlockMen ok, i dont have much time atm. it might not be finished until next weekend 20:09 BlockMen actually i just need change this function -> https://github.com/minetest/minetest/blob/master/src/guiFormSpecMenu.cpp#L2636 20:27 sapier thexyz could you try this version https://filetea.me/t1sw00pwZFISuqxQaYiKrjb4w 20:36 kahrl BlockMen: this should work: http://paste.dy.fi/3tW 20:36 kahrl (self contained program with test cases) 20:44 BlockMen kahrl, looks good. will test it asap 20:50 thexyz sapier: try what? 20:51 thexyz and please next time upload it to some sane place 20:51 thexyz I don't think it's that hard 20:53 thexyz alright 20:53 thexyz this time you didn't include libstdc++-6.dll because, dunno, you hate us or something 20:53 sapier it is as all sane places tend to go away very soon 20:53 sapier pomf.se is already offline again 20:53 sapier sfan doesn't support files > 5mb ... 20:54 thexyz pomf works for me right now 20:54 thexyz dunno what you're talking about 20:54 sapier great it didn't work 10 min ago 20:54 thexyz great indeed 20:56 sapier I wonder if there's a tool that automaticaly copys dlls 20:58 sapier did you manage to get gettext working with vs at all? I failed by now 20:58 sapier no matter what I do non ascii chars aren't displayed 20:59 sapier http://a.pomf.se/J6r0F9.7z hope i didn't miss another dll 21:04 thexyz what do you mean to "get gettext working with vs"? I didn't compile it from source; otherwise it works just fine 21:04 sapier it works as long as it doesn't require to use a different charset as system default 21:05 sapier once I set language to non ascii charset variant I get totaly messed up gui 21:09 thexyz check the current official build, it seems to work fine 21:09 sapier which is vs2010 is it? 21:10 sapier did I forget other dlls? 21:11 thexyz no, it launches now 21:11 thexyz hm.. this is odd 21:11 sapier what exactly is odd? 21:13 thexyz I can make official 0.4.7 build work with japanese translation on windows with russian locale when I rename "ja" to "ru" in "locale" directory; but if I set LANG from cmd and then launch minetest.exe it just crashes 21:13 sapier isn't odd I have similar behaviour for various combinations ... those gettext dll/libs aren't a perfect match 21:14 sapier they never are if you use different compiler versions 21:14 sapier of course they should be ... but usualle they aren't 21:14 thexyz so are you willing to provide windows builds from now on? 21:16 sapier if no one else wants to do this ... no problem still imho we should find a solution for the gettext vs issue ... we can't rely on some old vs version to use a even older gettext variant ;-) 21:16 thexyz we cannot because ? 21:17 sapier I had to install a hotfix to vs2012 to get it run again today because microsoft did a automatic update last night ;-) 21:17 sapier ok not for this exact reason but for that kind of things to happen 21:18 sapier but I don't have a good idea how to solve that issue ... anyone else? 21:21 sapier does my mingw build work for you with console set as well as minetest.conf?