Time Nick Message 02:11 khonkhortisan PLEASE MERGE https://github.com/minetest/minetest/pull/794/files minetest will not run without this 02:13 khonkhortisan my computer does not support ipv6, disabling it still lets the test run, failing the test still crashes the game 02:14 kahrl why does this use a hardcoded port? 02:14 kahrl I don't think it can be merged that way 02:15 hmmmm proller actually pointed that problem out in the pull request comments 02:15 hmmmm ..aand nothing happened 02:25 ShadowNinja Whoever merges it can patch it. 02:29 kahrl patch how? 02:29 khonkhortisan what should it use instead of a hardcoded port? A random one? 02:29 khonkhortisan https://github.com/minetest/minetest/issues/793#issuecomment-20123773 this is all it actually needs to run 03:06 VanessaE kahrl: about the news post... #define MAX_REGISTERED_CONTENT 0xffff <-- are you sure this is actually ok to do? 03:07 VanessaE (I mean, assuming connecting clients update to the relevant commit) 03:07 kahrl VanessaE: 0xfffU, not 0xffff 03:07 kahrl or do you mean changing it to that? 03:07 VanessaE that would have been the proposed change. 03:07 VanessaE (hence the bold) 03:07 kahrl don't set it that high 03:07 VanessaE awww 03:07 VanessaE spoil my fun :P 03:08 kahrl you will leave zero room for dummy node IDs 03:09 VanessaE well at any rate, assume I changed it to some larger value, say to 0x2000 03:10 VanessaE (probably going to need to for testing, eventually) 03:10 kahrl that would work 03:10 VanessaE ok 03:12 hmmmm hrmm, what is the point of a dummy node ID exactly? 03:12 hmmmm excuse my ignorance 03:12 VanessaE hmmmm: aside from them being used for undefined nodes? 03:12 hmmmm don't we have CONTENT_UNKNOWN for that now? 03:13 kahrl hmmmm: you can't map all unknown nodes to a single ID on the server 03:13 kahrl otherwise you'd lose data 03:13 VanessaE otherwise you won't be able to tell them apart, e.g. "oh crap, I forgot deleted that node..what was it so I can alias it?" 03:13 VanessaE +I 03:13 hmmmm oh, right, it's merely CONTENT_UNKNOWN for the *client* 03:14 hmmmm vanessae, i think it'd be more troublesome if the block gets modified somehow and then is written back as unknown 03:14 VanessaE indeed so 03:15 hmmmm how many node IDs to preserve for dummies... hmm 03:16 VanessaE I would guess an even hundred is more than sufficient 03:16 hmmmm is it really? 03:16 VanessaE If' you've accumulated any more than that, the overflow would probably not be missed. 03:16 hmmmm doesn't your "same block in every possible color"-type mods easily go over 100? 03:17 VanessaE well any number you pick is gonna be somewhat arbitrary 03:17 VanessaE no. 03:17 kahrl I would have chosen MAX_REGISTERED_CONTENT at maximum 0x7fff 03:17 VanessaE firstly, I only wrote one mod (which admittedly does go way over 100) 03:17 VanessaE but the unknown nodes only apply to what's actually in the map, no? 03:17 hmmmm well 03:17 VanessaE the nodeids I mean 03:17 hmmmm hrmm 03:18 VanessaE maybe 256 then - an even 8 bits. 03:18 hmmmm yeah hivemind 03:18 VanessaE heh 03:18 hmmmm i was literally just about to say 256, but for no real reason other than it's an even number 03:18 VanessaE I like even numbers :P 03:18 kahrl 256 sounds much too small 03:18 hmmmm because this is just what's reserved for dummy node IDs in the event that you really do have over 65000 nodes defined by mods 03:18 VanessaE ok 03:18 VanessaE 534 then. 03:19 VanessaE that leaves an even 65000 for real node id's 03:19 hmmmm i guess that's a better sounding number 03:19 hmmmm well let's think about this more 03:19 kahrl perhaps make it a minetest.conf option and set it relatively conservative by default? 03:19 VanessaE if we ever hit that number in practice, well... heh 03:19 hmmmm if you're the type of person who has mods that register that many nodes 03:20 VanessaE kahrl: +1 03:20 hmmmm you're likely going to have many unknown nodes when you go to remove one of these mods 03:20 VanessaE hmmmm: THAT much is true 03:20 VanessaE so what if we just allow it to overflow? 03:20 hmmmm so 0x7fff is a better idea 03:20 VanessaE well..hrm 03:20 VanessaE no 03:20 hmmmm i don't see the purpose in having a config option for this... i mean 03:20 VanessaE that's way too big of a number 03:21 khonkhortisan error: node def overflow - approaching real-world variety 03:21 VanessaE how about a command-line switch that wipes out all unknown nodes in a map if the limit is exceeded? 03:21 hmmmm my point is that we'd have better ways to handle variations on single nodes before any person hits 31767 03:21 hmmmm too much work 03:21 khonkhortisan Right now, we have to use abms for specific node removal 03:22 VanessaE hmmmm: 32767. 03:22 hmmmm whoops 03:22 hmmmm yeah, 32767 03:22 hmmmm so i think it should be 0x7fff 03:22 VanessaE I dunno, that feels like way too big of a divide. 03:22 hmmmm it can be raised in the future if this is a mistake 03:22 VanessaE true. 03:23 VanessaE and make it configurable as kahrl suggested. 03:23 hmmmm config bloat 03:23 VanessaE naw 03:23 hmmmm there is absolutely no need for this level of configurability 03:23 hmmmm look at how bloated the config is already 03:23 khonkhortisan I vote moving fullscreen into the gui 03:24 VanessaE well sure, the *example* is bloated, but mine for example is only 22 lines. 03:24 VanessaE hardly what I'd call blaoted. 03:24 VanessaE ok, that kinda read weird. 03:25 khonkhortisan it's not the same example 03:25 VanessaE (you get the point) 03:25 hmmmm by config bloat i mean the sheer number of available knobs 03:25 hmmmm for things that people would basically never change 03:25 hmmmm hrm 03:25 VanessaE hmmmm: no, THIS is a lot of knobs: http://mrjam.typepad.com/.a/6a00d8341c00c753ef010535df5c7e970c-800wi 03:25 VanessaE :D 03:26 hmmmm way more than that 03:26 khonkhortisan real music knobbers use this http://www.123rf.com/photo_6934982_music-system-mixer-hundreds-of-buttons-rotating-knobs-generic-background-for-music-themes.html 03:26 khonkhortisan minus the butterfly 03:26 hmmmm i would imagine we have at least 300 config settings by now 03:27 hmmmm hmmm, using the leveled nodebox for gradually sloped maps? 03:27 VanessaE is THAT what that's for? 03:27 VanessaE (I missed that discussion) 03:28 hmmmm i'd imagine that'd be as simple as multiplying height perlin noise by a factor of 8, and then placing solid nodes up to height / 8 y position 03:28 hmmmm and then place a leveled nodebox of level height & 7 03:28 hmmmm it's for weather actually 03:28 VanessaE *nod* I could see that 03:28 hmmmm but this is what it could be used for 03:28 VanessaE but pilz will hate it :P 03:29 VanessaE "doesn't fit a cubic world" etc 03:31 hmmmm doesn't really matter to me, but it's definitely not a mainstream thing 03:39 VanessaE awww 03:39 VanessaE spoilsport :P 03:51 hmmmm humm 03:52 hmmmm so i'm thinking about the height_min and height_max for biomes 03:52 hmmmm that's definitely getting changed before mapgen v7 goes mainstream, but still not sure to what 03:53 hmmmm kinda wondering what minecraft does... if only the source was available for those "more biomes" mods 03:53 VanessaE well height min/max (rather, elevation) is what plants_lib uses... 03:53 hmmmm I see they have a range from -1 to 2 that correlates with height but it's obviously not 03:56 VanessaE why was it that you wanted to ditch the min/max idea? 03:59 hmmmm it really doesn't work well 04:00 hmmmm seeing a desert biome cut off at a certain point on a hill... it's weird 04:00 hmmmm i'd prefer if the desert biome was only around areas of low hilliness 04:00 VanessaE that much is true 04:00 VanessaE but, you also need to account for other stuff that *does* cut off at certain elevations e.g. snow-capped mountains 04:00 hmmmm i'd like to go back to the concept of the terrain variance parameter 04:00 hmmmm how, exactly, i have no clue 04:01 hmmmm i have a feeling that minecraft uses a terrain variance parameter for biomes, except according to the source, it's a range of possible values 04:02 hmmmm I don't think i'd really get much further in understanding this without actually analyzing their mapgen 04:03 VanessaE the only other idea I could suggest is one I rather hate, myself - dithering. 04:04 hmmmm from what i see in the pictures, it looks like what they may do is a combination of 3d noise and 2d noise 04:05 hmmmm the 2d is the thing that decides the terrain variance and 3d is just where the actual nodes are placed 04:05 hmmmm i think this is what mapgen v5 used to do, except in minecraft its' done *differently* somehow 04:07 hmmmm anyway i'd like to say that, but then i see something like http://i.imgur.com/cEXgKTl.jpg and wonder WTF 04:08 VanessaE that looks more like perlin with a layer of simple random noise added in 04:08 hmmmm honestly, you could be right 04:08 VanessaE with a threshold function to crop them to 0 at certain elevations 04:09 VanessaE or even a third layer of simple noise being used as a crop threshold 04:32 celeron55 i vote for 0x7fff 04:35 VanessaE I'd vote for something closer to 0xbfff 04:36 celeron55 it's better to raise it conservatively (if only because raising it conservatively has worked well in the past too) 04:37 VanessaE (but that particular value has meaning to me in a different field anyway :) ) 04:37 celeron55 hmmmm: i too have had much trouble figuring out how the MC mapgen works 04:37 celeron55 (which i have mentioned already though) 04:39 VanessaE hmmmm: for shits and giggles, try the method I suggested. See what it does. 04:39 celeron55 especially those almost straight but still organic walls 04:40 VanessaE perlin with a -1 to +1 range or whatever it was + plain noise in the range of, oh, +/- 0.1, and a second noise call for the crop-to-0 threshold 04:42 VanessaE actually the crop thing might be wrong.. 04:42 celeron55 the first thing that comes to mind from that picture is 3d noise cut with a very sharp and tall 2d noise, ANDed by the shallow 2D noise that goes at the ocean floor 04:42 VanessaE *shrug* 04:43 celeron55 but there are still features there that wouldn't happen by any reasonable chance with that 04:43 celeron55 especially the "cubicality" of the walls 04:43 celeron55 (which is nice for building and i think for that reason MT should aim for that too) 04:44 hmmmm erm 04:44 hmmmm i know how it does *that* much at least 04:44 hmmmm it samples the 3d noise at an 8th of the frequency and then interpolates 04:45 VanessaE wait a minute here... 04:45 VanessaE *scratches head* 04:45 hmmmm that particular screenshot i can't explain 04:45 VanessaE celeron55: actually you're not too far off from that 04:45 celeron55 interpolation doesn't really directly result in that 04:45 celeron55 but i guess it can be made to... dunno 04:46 celeron55 (also it was 8x8x4 interpolation in some version IIRC) 04:46 hmmmm notch's own description is consistent with what i see 04:46 VanessaE if you AND a simple slope function against a triangle function, you get a really complex waveform that kinda resembles the features of that map 04:46 hmmmm yeah, exactly that 04:46 hmmmm 8x8 horizontally and x4 vertically 04:47 hmmmm see, either way, i wouldn't be able to use this even if i did figure out what goes on exactly 04:47 hmmmm the terrain generation isn't just one phase but rather making a base, then adding shit on top, then subtracting shit out later and they're all disparate functions 04:49 celeron55 you might eventually get to surpass notch as the god of large-voxel terrain, but that'll still at least take a lot of time 8) 04:50 hmmmm why do you say that? 04:52 celeron55 i've always considered world generation as a "race to get to where MC got" 04:52 hmmmm if that's it, we've clearly surpassed MC at this point 05:00 hmmmm anyway 05:31 hmmmm you know something sort of funny but clever about minecraft? their chunk arrays are oriented in the Y direction and it totally makes sense for cache coherency reasons 05:31 hmmmm most often when we're iterating through a chunk, we're iterating downward 05:31 hmmmm or upward 05:38 VanessaE true 06:05 hmmmm anyway i suppose i could do a clean-room implementation of the minecraft mapgen for shits & giggles 06:05 hmmmm it produces nice terrain for what it is and some people might like that 06:06 hmmmm i understand exactly how it all works after taking like an hour looking at it 06:17 VanessaE indeed, and clean-room work like that is, afaik, safe even from US copyright laws 06:17 VanessaE (though I guess mojang doesn't actually care) 13:40 Exio4 just cheeking, http://forum.minetest.net/viewtopic.php?id=175 13:40 Exio4 Github: https://github.com/celeron55/minetest 13:40 Exio4 isn't that a bit outdated? 13:42 Calinou it is 13:42 Calinou it's minetest/minetest 13:42 Exio4 exactly, can you update it? 13:42 Exio4 (as mod?) 13:42 celeron55 i'll update it 13:42 Calinou updated, Exio4 13:42 Calinou celeron55: too late 13:42 celeron55 well there's more to update than that 14:35 proller want to commit! https://github.com/minetest/minetest/pull/830/files 14:36 proller or maybe only lua api part 14:45 PilzAdam why hardcoding the ABMs? 14:46 proller because they must be fast and process 100000+ blocks 14:48 PilzAdam have you even tried doing it in Lua? 14:49 proller no 14:49 proller and i have no 32 core xeon 14:49 proller i can now no commit these abm 14:50 proller but i want commit lua api 14:50 proller its useful for bucket, pipeworks, ... 14:51 PilzAdam do all the neighbors need to exist or is the ABM called if one of the neighbors is there? 14:53 proller one 14:53 PilzAdam the freezemelt field is not documented, and you dont need to serialize it (the client doesnt need to know about it) 14:54 Exio4 proller: 32 cores won't matter, as i said, MT isn't very multithreaded and can't be that much 14:55 PilzAdam lines 277 and the following in content_abm.cpp are bad code style 14:56 proller make for() or idents ? 14:56 proller will use astyle on my func 14:58 PilzAdam Im not happy at all with this 15:15 proller PilzAdam, all fixed 15:20 proller and _game update for it https://github.com/proller/minetest_game/compare/weather 15:22 PilzAdam what does the nodeboxtype "leveled" do? its not documented 15:22 Jordach PilzAdam, think finite water nodes in height 15:23 PilzAdam yea, but how does it work 15:23 Jordach and snow could use this to pile up layer by layer 15:23 Jordach making snow blocks 15:23 proller leveled = 0, -- Block contain level in param2. value - default level, used for snow. Dont forget 15:24 Jordach 0 = 1/16th 15:24 Jordach 1 = 2/16th 15:24 proller Dont forget use "leveled" type nodebox 15:24 Jordach "" and so on 15:24 PilzAdam does the param2 need to be set in on_place_node()? 15:24 Exio4 where is it documented 15:24 proller no, will use default value 15:25 proller leveled =3 -- will de default 15:25 PilzAdam proller, can you docuement that properly? 15:25 proller doc/lua_api.txt 15:25 PilzAdam it isnt there 15:25 proller and leveled - already in core 15:25 ShadowNinja Isn't this for internal use only? https://github.com/minetest/minetest/pull/768 15:26 Calinou proller: 3 as default? that's thick snow 15:26 Calinou in minecraft snow is 2/16 15:26 proller PilzAdam, described in my branch 15:27 proller Calinou, now default for snow =1 15:27 proller its 1/8 15:27 PilzAdam https://github.com/minetest/minetest/commit/9733dd5b5e5516e18775665db132b2446492716c <- no documentation at all 15:27 PilzAdam proller, fix that 15:27 proller PilzAdam, forgot here 15:27 Exio4 don't you check what you commit to upstream? 15:27 proller PilzAdam, https://github.com/minetest/minetest/pull/830/files 15:28 proller here ^ described already 15:29 PilzAdam I still dont like the hardcoded ABMs, for freezing interval and chance are low enough; for melting its restricted to a hot neighbor 15:30 PilzAdam feel free to benchmark Lua vs. core and proof me wrong 15:31 PilzAdam bbl 15:31 Exio4 lua, luajit and core 15:57 sapier guys color defines are still missing 15:58 sapier I know adding those isn't a challenging thing but there already were speaking color names so I expect them to return... and I won't add them for sure 16:01 sapier sfan5 this comment was primary for you ;-) 16:01 sfan5 *sigh* 16:01 sfan5 why color names? 16:02 sapier because you can't tell me without trying what color #aa551F is 16:02 sapier but you can tell instantly what color GREEN is 16:02 Calinou color names would make it easier to mod 16:02 Calinou but don't make them a replacement to color codes 16:02 sfan5 ^ 16:02 Calinou also, can you use color codes in mods and chat? 16:02 VanessaE well you have two standards to choose from really - the so-called "web" colors, or the ones in Unified Dyes. 16:02 VanessaE also, hi. 16:03 sapier no they shall only be an addition to restore old userfriendlyness 16:03 Exio4 sapier: if that is RGB 16:03 Exio4 seems like pinkish or purpleish? no idea 16:03 sapier mods for some formspec elements chat no 16:03 Exio4 didn't try that color 16:04 sfan5 Calinou: not yet 16:04 sapier don't know I haven't tried but I can only guess that it might be something in red direction 16:05 sfan5 we'd need fixed length color names 16:05 sapier you may just prefix them too e.g. mt_color_green, mt_color_blue 16:05 sapier they're plain global variables 16:05 proller VanessaE, why you disalow add player numbers in server list ? 16:06 VanessaE when did I say that? 16:06 proller sapier, !? 16:06 sapier VanessaE didn't disallow player numbers but complain about indentions 16:06 sapier player numbers will make indentions even worse 16:06 proller sapier, then add it with bad identation 16:06 VanessaE actually I just suggested that player numbers in the list would lead to some bias. 16:07 VanessaE but otherwise, you just need a better visual presentation of those flags 16:07 VanessaE they look like noise now 16:07 sapier actually I assumed VanessaE won't accept indentions even worse than now 16:07 VanessaE maybe put them in a different background/foreground color than the rest of the text. 16:07 proller its most important info in servel list 16:07 proller nobody wants to play alone 16:08 proller mak etable ? 16:08 proller make table ? 16:08 sapier VanessaE different colors within a text aren't supported 16:08 VanessaE sapier: then a vertical line between the flags and the descriptions 16:08 proller VanessaE, did you know any serverlist without player number in games? 16:09 VanessaE proller: I don't play any other online games. 16:09 proller and players was before formspec 16:09 sapier I already tried that ... terrible result 16:09 VanessaE but I do know something of human nature :) 16:09 Zeg9 indentation IS important, but player count is even more. 16:09 proller yes! 16:09 VanessaE sapier: something *has* to be done, those flags as presented now, frankly, look like crap 16:09 proller maybe more than server name 16:10 Zeg9 if you put an ID (or just the ip) instead of the name, why not ? :p 16:10 sapier For now I can only provide terrible indentions or missing flags/playernumers ... it's up to you to decide what death you wanna die 16:10 proller and! show address:port if no name and port != 30000 16:11 proller sapier, kill indentions, return flags/playernumers 16:12 sapier adress is already shown if no name is given 16:12 VanessaE what's all this about indentations anyway? 16:12 proller sapier, print :port if not default 16:12 proller it was before too 16:12 sapier no no no indentions was your concern VanessaE 16:12 VanessaE it was? 16:12 VanessaE since when? 16:12 proller :port was 16:12 sapier it was two days ago do I have to find out the exact line? 16:13 proller since serverlist 16:13 VanessaE I only care about code formatting to the point of it being easy to read 16:13 VanessaE let the other folks here obsess about code style :P 16:15 sapier if noone cares about how this will look like I'm gonna add the playernumber BUT I won't change it back and it's up to your responsibility how it's gonna look like 16:16 VanessaE sapier: otherwise, the big issue is how it's presented in the menu 16:16 VanessaE what the player sees when they turn on the public list must NOT look like garbage 16:16 VanessaE (I mean the word literally, e.g. like garbled text) 16:16 ShadowNinja sapier: http://irc.minetest.ru/minetest/2013-07-15#i_3198641 16:17 sapier thats not a bug but a programming error 16:18 sapier 5d coordinates aren't supported 16:18 sapier sorry 4d 16:19 ShadowNinja Ah s/,/;/? 16:19 sapier exactly 16:19 sapier maybe this was ignored by now and size set to 1 but I added lots of sanity checks ... and no I won't add compatibility support for bugs 16:20 sapier but I guess this would've been wrong in old version too 16:24 proller sapier, add add add !! 8) 16:24 proller and :port too please 16:24 proller i have 2 servers in one ip and its horrile 16:24 proller ble 16:27 sapier proller if you don't calm down I'm gonna add your requests last 16:28 Calinou sapier: something *has* to be done, those flags as presented now, frankly, look like crap 16:28 Calinou way better than before IMO 16:28 VanessaE they still look like crap 16:28 Calinou color them? 16:28 sapier one line one color 16:28 Calinou make them slightly darker than normal text? 16:28 Calinou or one line one color 16:29 VanessaE sapier: put a colored vertical stripe behind them, like green/white ledger paper 16:30 sapier I can only put it on top and keeping that in sync is almost impossible 16:30 VanessaE then put them in a separate list 16:30 sapier same sync issue 16:31 VanessaE ANYTHING to visually separate them from the server names 16:31 sapier nothing that isn't a text character is available 16:32 VanessaE then use something from unicode, like the "solid block" character 16:33 sapier i can use | too but it won't fix the indention issue 16:33 VanessaE what indentation issue? 16:33 sapier iii 16:33 sapier hhh 16:33 sapier lll 16:33 sapier mmm 16:33 VanessaE I see some spacing/tabular variance 16:33 VanessaE but that's about it 16:33 sapier that issue 16:34 VanessaE then switch to a monospace font in the server list 16:34 VanessaE (or for the flags anyway) 16:34 Calinou +1 for fixed width fonts in lists 16:34 VanessaE or add something to the formspec code that lets you properly align the columns 16:34 Jordach spacer chars!! 16:35 sapier THERE IS NO ALIGNMENT POSSIBILITY AS ALIGNMENT DEPENDS ON FONT .... the only possible way is using fixed width fonts 16:36 sapier once all real bugs are fixed I'm gonna try if there is a way to use fixed width fonts ... but there's no guarantee this will look better 16:36 VanessaE U+2588 16:36 VanessaE that's the one I was after. 16:37 Jordach sapier, the lua main menu freezes with no internet connection 16:37 Jordach kaeza reported it 16:37 VanessaE sapier: as I said, or add something to the formspec code that lets you properly align the columns (i.e. if the standard doesn't allow for it, modify the damn standard) 16:37 sapier I guess he try's to read public server list 16:38 sapier post your requests to 814 while I'm fixing important issues 16:40 celeron55 how about adding flags as a variable length string of lowercase letters inside [] after tha name? 16:40 celeron55 it looks generally better in this kind of situation, but it's less convenient 16:40 VanessaE celeron55: fine by me, but that won't fix the column alignment issue he's so worried about :P 16:41 celeron55 there won't be any column alignment then 16:41 sapier I'm perfecty ok with this 16:41 celeron55 not even attempted one 16:41 celeron55 -even 16:41 Calinou put them before, would be better 16:41 Calinou else you can't look at them quickly and compare them 16:41 celeron55 it won't look good if it's variable length 16:42 VanessaE er right. read that a different way) 16:42 celeron55 it will result in the terrible variable-column sadness as currently 16:43 celeron55 i guess sapier could emulate variable columns by adding a second narrow list beside the current one and scroll it automatically based on the other 8) 16:43 VanessaE didn't I already say that? 16:43 celeron55 well consider that a half-vote for it then 16:44 celeron55 that could be put on the right side which could potentially look reasonably good 16:44 celeron55 not sure though 16:44 Calinou currently, there is no variable column sadness... they are aligned 16:44 Calinou since _ is used as placeholder 16:44 sapier celeron no I can't without adding additional events to formspec 16:44 celeron55 Calinou: characters aren't fixed width 16:45 Calinou oh 16:45 celeron55 especially * is narrower than the others 16:45 Calinou is * really useful anyway? it doesn't impact gameplay 16:45 sapier I don't get mousewheel events for example 16:45 sapier and I can't set scroll position at all as irrlicht doesn't support this for listboxes 16:45 Calinou can't use mouse wheel on new server list, yeah 16:46 Calinou the favourites list also lacks a background 16:46 sapier you can you just have to select the list 16:46 celeron55 how about just adding [] to the flags at the beginning and replacing * with something that is roughly equal width with capital letters and _? 16:46 celeron55 maybe it'd be regular enough to look like a proper column 16:46 Calinou why not remove it? see what I said 16:46 sapier celeron I already tried that result is ... acceptable for flags but numbers are something different 16:47 Calinou a "password required" message could appear when the server is selected 16:47 celeron55 okay so, is using a monospace font for only certain things a possibility? 16:47 sapier I think so yes 16:48 VanessaE Calinou: ..and the client should "grey out" the connect button and ignore "enter" if there's no password supplied to such a server 16:48 sapier but I haven't tried 16:48 Calinou +1 16:48 celeron55 try if it's easy... it could be the solution 16:48 sapier I'll try ... AFTER I fixed the real bugs 16:48 celeron55 it could be useful otherwise too, as pretty much every list tries to have columns currently and can't really have even with spacing tricks 16:49 sapier it still needs to be configurable as fixed width fonts tend to look strange 16:50 sapier may I request someone to replace formspec by a more advanced format ... except JSON 16:50 celeron55 i think fixed width fonts look awesome if you can cope with the less amount of characters they allow you to fit in a given space 16:50 sapier they look good as long as you don't have words with a lot of small chars in there 16:50 celeron55 oh by the way, irrlicht does have tables 8) 16:51 celeron55 gui tables, that is 16:51 celeron55 actually is that what is used now too? 16:51 sapier no atm it's a listbox 16:52 celeron55 well anyway, there does exists an element for this use... if somebody is crazy enough to implement formspec support for it 16:53 sapier I think implementing is less crazy then using it ... except of requirement of another separator char 16:53 proller sapier, why not json ? ;) 16:53 celeron55 proller: don't start that discussion yet again 17:07 hdastwb Is there some way to obtain server settings? I've proposed ditching sneak elevators but having a setting to re-enable them for old maps and legacy games, but it seems that the settings are all client-side 17:07 Exio4 enable a bug 17:07 Exio4 that sounds weird 17:10 hdastwb I've seen lots of structures that depend on it on servers and I've built a lot of them as well; I think there should be a better way to keep those working then to not upgrade the engine 17:11 hdastwb (or having to download an old engine to explore certain maps) 17:20 VanessaE proller: fog must rely on both humidity *and* temperature, if you do it at all 17:21 proller VanessaE, fog now not ready, and i remove getFog from core for now 17:21 VanessaE proller: I mean for later if and when you decide to approach that, since you mention it in that PR. 17:21 Exio4 https://github.com/minetest/minetest/pull/829 wtf is that commit log 17:22 VanessaE Exio4: a bad need for a rebase :P 17:22 proller it wat hello from git 17:22 proller and github doesnt update branch 17:23 proller VanessaE, ok, snow, rain, fog is subject for long adjusting 17:23 VanessaE proller: yes. 17:24 proller and we can start from changing current heat from 50 +-50 to like 20+-60 17:24 proller now is veery hot on map 8) 17:24 proller except hi levels 17:24 VanessaE the standard is -25 to +81. 17:25 VanessaE (if I interpreted snow mod correctly; plants_lib uses the same scale) 17:25 proller -25 ещщ рще ащк ьшт 17:25 proller too hot 17:25 proller maybe -40 17:25 VanessaE -25 is hot? O.o 17:26 VanessaE where do you live, outer mongolia? ;) 17:26 proller minimum - is hot 17:26 proller -40 in my country 17:26 VanessaE heh 17:26 proller and ~ -50 minimum 17:26 * VanessaE shrugs 17:26 proller and +80 too hot 17:27 proller maybe +60 17:27 VanessaE not that it matters, this is a game :) 17:28 proller ok 17:32 proller now i have no idea how to make fog position-depended 17:36 proller and need to make cloud heat-humidity depended 17:38 proller size&color 17:38 proller and maybe height 17:39 VanessaE well good luck with it 17:41 celeron55 umm 17:41 celeron55 in what units are you handling temperature 17:43 hdastwb Celsius, I assume… 17:44 celeron55 hopefully so 17:44 VanessaE well I was speaking in celsius as well. 17:46 celeron55 about hdastwb's note: we actually have a system in place that would be able to handle that *very* easily with like two lines of code 17:46 celeron55 via the privilege system similarly as fly and fast 17:48 hdastwb thus, we add a "sneak elevator" privilege? 17:48 celeron55 that'd be the way to do it 17:48 Exio4 bug_sneak 17:48 proller celeron55, is something othet than celsius ? 17:48 VanessaE why is everyone so eager to kill off sneak elevators? 17:48 Calinou add a sneak_glitch setting, on by default? :P 17:48 VanessaE proller: kelvins? ;) 17:49 Calinou VanessaE: +1, it's not a bad thing to keep it 17:49 Calinou we could make it part of the game. just like quake has strafejumping 17:49 celeron55 well i propose it will be added as a privilege 17:50 celeron55 or, well, kept, but disabled by default and enabled by privilege 17:50 VanessaE mmmmh 17:50 proller kelvins bad to fill in u8 17:51 VanessaE proller: I was kidding :P 17:51 proller ups, s8 17:51 hmmmm what do i have temperature as right now, a float? 17:51 VanessaE celeron55: one has to ask, what level of realistic behavior is the aim? 17:51 proller and fahrenheit 17:51 celeron55 VanessaE: none if you ask me 17:51 celeron55 VanessaE: something else if you ask someone else 17:51 VanessaE hmmmm: it's a perlin value right now isn't it? so float I suppose. 17:52 hmmmm I do convert values you know 17:52 VanessaE celeron55: well if there's no goal, why bother with changing sneak elevators? most healthy people could climb one if one were built in real life 17:52 VanessaE (at least for some reasonable distance) 17:53 hdastwb (with some handles installed) 17:53 VanessaE hdastwb: without - never seen a rock climber before? 17:54 hdastwb those blocks look pretty smooth to me; I didn't update to a bumpmapping texture yet… 17:54 Jordach leave the sneak elevators, because there will be endless insurgency about it 17:54 Calinou +1 17:54 Calinou keep it as an "accepted glitch" 17:54 Jordach useful* 17:54 Calinou but add an opt-in fix for server admins? 17:55 Calinou VanessaE: some guy climed a lot of buildings, several hundreds of meters high, without anything 17:55 Calinou french guy 17:55 Jordach that french spiderman Calinou? 17:55 Calinou yea 17:55 Jordach ninja'd 17:55 Calinou alain something 17:55 Jordach the french are best at athletics 17:55 Jordach anyways, sneak elevators are more useful than the elevators in travelnet 17:55 Jordach ...at times 17:56 Yepoleb hello 17:57 VanessaE Calinou: yeah, I think I remember that event (though not what his name was0 17:57 VanessaE ) 17:58 VanessaE hdastwb: as for textures looking smooth, think of cobblestone. Even without bumpmap, it clearly suggests a rough material made of reasonably large pieces. 17:58 hdastwb alright 17:59 Yepoleb can shaders be installed as a mod? 18:00 Jordach Yepoleb, ask in #minetest 18:00 Jordach no here please. 18:00 Jordach not* 18:00 Yepoleb sry 18:02 celeron55 probably nobody is against climbing 18:02 celeron55 it's the speed that some people don't approve 18:03 VanessaE that much is true 18:04 VanessaE so make climbing a sneak elevator be slower than climbing a ladder (which imho is also too slow) 18:04 VanessaE (but for a game it's okay I suppose) 18:05 hdastwb and perhaps make that speed faster when fast is on (if that's possible) 18:12 VanessaE we tried that before. 18:12 VanessaE didn't work out so well because of variations in key configs 18:12 VanessaE (particularly when descending) 18:51 sokomine i do have an issue with non-floating lava in the minimal development game: http://mg.viewskew.com/mgoblin_media/media_entries/149/screenshot_3768580702.png 19:02 VanessaE that ^^^ affects my server also, btw. 19:02 VanessaE (probably an artifact of the slow-by-default liquid flowing) 19:03 sokomine it occours in other regions as well. not only this one place 19:03 thexyz celeron55: http://forum.minetest.net/viewtopic.php?id=6571 19:03 sokomine here, for example: http://mg.viewskew.com/mgoblin_media/media_entries/150/screenshot_3768695008.png 19:04 sokomine or is it because of the minimal game? 19:04 sokomine in some of the affected lava regions, i found blocks from dungeons. might just be coincidence 19:04 VanessaE nope, my server runs a heavily-modded minetest_game and also does that. 19:06 sokomine strange 19:07 proller sokomine, use liquid_finite=1 19:11 celeron55 proller could fix non-finite liquid just as easily as he has fixed the finite one 19:11 sokomine do infinite liquids not flow anymore? for me, lava is not so important. just wanted to report that bug so that you here can see if it is one and may need fixing 19:11 celeron55 just an ABM to randomly trigger flowing... 19:11 celeron55 it's so hacky though 19:11 celeron55 well whatever 19:12 sokomine as far as flowing goes, i still dream of a mill wheel that would rotate. could be done with models i think - but people who can do models are hard to come by 19:29 sokomine another issue...rollback is unfortionately really slow. rollback_check can be speeded up significantly if external grep on the rollback.txt is used. but: that way you catch more griefers. which you then want to rollback...and then it's slow again :-( 19:50 kahrl does anyone happen to have to implement ^[inventorycube using custom software rendering instead of render-to-texture? 19:50 kahrl s/have to implement/have code that implements 20:22 proller celeron55, i can remove flowing touch abm from if liquid_finite 20:24 proller and i want to merge whole https://github.com/proller/minetest/compare/weather or only lua-luaapi part 20:26 Exio4 did you benchmark those ABMs in lua? 20:26 Exio4 (lua, luajit, core) 20:27 proller no 20:27 proller have no time for it 20:30 proller but no problem to remove these abms from core if somebody write it in lua 20:31 Exio4 if the time you used "looking around for making someone agree" would have used in benchmarking that, it would be already merged, probably ;) 20:32 VanessaE kahrl: remind me again, what the "right way" is expected to be to permanently raise the nodeid limit? 20:32 VanessaE (I mean, besides just tweaking that one line) 20:34 proller Exio4, luajut will be 5-10 times slower than core 20:34 proller and bench gets 1-2 hours to write and get results 20:35 Exio4 proller: 5-10 times slower? 20:35 Exio4 how did you get those number 20:35 Exio4 s 20:36 proller i know something about speed of languages and compilators 20:36 VanessaE proller: are you sure about that? It's supposed to be rather close to the speed of C++ 20:36 proller 70% 20:37 Exio4 http://cadoth.net/~exio4/private/performance_comparison_pathfinding.pdf 20:37 Exio4 luajit is faster than -O0 in a far pathfind-ing 20:37 proller abm must be like LiquidFreeze 20:38 Exio4 and aren't you calling lua from that ABM? 20:38 proller if it faster in one test - its nothing 20:38 kahrl VanessaE: I don't understand 20:38 Exio4 for updating nodes, that is 20:38 VanessaE kahrl: remember you were saying it would require something of a "rework" to do it right? 20:38 kahrl oh, changing the node definitions to a vector was that rework 20:38 VanessaE OH 20:38 VanessaE ok 20:38 proller Exio4, its caleed for every 10-50 mathed node 20:38 proller and it can be removed 20:39 proller and falling node must be rewriten in c++ 20:39 Exio4 proller: anyway, without real benchmarks all what we are doing is just pure guessing 20:40 Exio4 "oh, because i think this can be in this way" 20:40 Exio4 that sucks, though. 20:41 proller and i want to commit working code 20:43 Exio4 working != done in a correct way though 20:44 Exio4 but well 20:44 Exio4 whatever 20:49 celeron55 organizational aspects of code are more important in that than speed, but i don't really know how those should be handled 20:50 celeron55 it was an intention that the whole content_abm.cpp file would be just kept for the transitional period from 0.3 to 0.4 20:50 celeron55 but it does turn out that some things should be implemented in C++ for speed 20:50 hmmmm [04:49 PM] organizational aspects of code are more important in that than speed top lel m80, did you forget about how slow and laggy organizational aspects made 0.4.x? 20:51 celeron55 hmmmm: they weren't slow because of organization 20:51 celeron55 they were slow because they just happened to be slow 20:52 celeron55 in any case i'm hoping someone somehow handles this finite liquid thing (hopefully someone who has been involved with checking the previous ones so that he knows the context) 20:52 hmmmm i wouldn't go as far as to say that one trumps the other 20:52 celeron55 also there should be better diagnostics for ABMs 20:53 celeron55 (the core could fairly easily tell how much each ABM takes time and what registered it) 20:59 proller i can commit without these abms 20:59 proller only lua 21:05 proller Exio4, in your tests luajit average slower than c in 5-7 times, and it beat c with something like recursion optimizing 21:08 Exio4 sapier's 21:09 Exio4 it is an algorithm benchmark, now do real use benchmarks 21:23 sapier1 yes my performance comparison is for pathfinding algorithm only NOT overall game 21:39 sapier1 https://github.com/minetest/minetest/pull/825 should improve formspec mainmenu ... again ... last thing left is background handling (unless new issues occur) 22:56 proller https://github.com/proller/minetest/compare/weather - added bit more doc 22:57 proller VanessaE wants this api, bucket code wants too 22:58 proller if nobody agreed with abm - can commit without it 22:58 VanessaE I do? 22:59 proller work with water levels 22:59 VanessaE so many commits there I don't know where to start :P 22:59 VanessaE oh yes 22:59 VanessaE it's not so much that I "want" it as much as "does it exist?", to which the answer is clearly "yes, it does now" :) 23:00 proller commits will be squashed, look down at changed 23:01 VanessaE I wonder if "leveled" is the right word to be using. maybe "graduated"? 23:01 VanessaE (as in a graduated cylinder) 23:01 VanessaE no matter either way 23:32 PilzAdam I have some stuff that would be ready to merge: 23:33 PilzAdam - finally remove common mods as announced before 0.4.7: https://github.com/minetest/minetest/pull/820 23:33 PilzAdam - fix lighting bug when lava disappears: https://github.com/minetest/minetest/pull/820 23:34 PilzAdam - and allowing multiple singleplayer games on one computer at the same time: https://github.com/PilzAdam/minetest/commit/7da83f89a27defaa446b4a992fc2d2bc1fdb7fe0 23:34 Exio4 PilzAdam: 820 and 820? 23:34 PilzAdam eh, the second one is this: https://github.com/PilzAdam/minetest/commit/faf8350f1dedb0682588043cafc42c16f366768a 23:34 Exio4 :P 23:35 Exio4 i guess the last commit is trivial enough to get merged "right now"? 23:36 * Exio4 pokes kahrl 23:38 VanessaE PilzAdam: +1 on all three of those 23:38 VanessaE (especially the lava lighting glitch) 23:45 kahrl what's up? (I'm deep in matrix equations right now) 23:45 kahrl ah, these look fine 23:46 Exio4 you can continue swimming in those matrix equations now ;P 23:46 kahrl diving, actually :P 23:46 Exio4 how expensive is getTexture? 23:47 kahrl can be very expensive if not cached yet 23:47 Exio4 oh, it is cached? kk 23:47 Exio4 didn't know that