Time Nick Message
00:00 ShadowNinja hmmmm, kahrl, PilzAdam: Volunteers?
00:27 VanessaE hmmmm: about weather. the values in the engine are not compatible with any worlds that used plants_lib. Can we please change this to something that won't leave e.g. palms buried in snow and ice?
00:31 proller hmmmm, make np_heat np_humidity adjustable via params like other noises
00:32 VanessaE I don't care if they can be configured, I just care if a user ends up with an existing world (read: their favorite server) that's been switched to use the new system and suddenly everything is growing in the wrong places.
00:32 VanessaE (or appears to have grown)
02:52 hmmmm https://github.com/kwolekr/minetest/commit/9d556c70782d983b9534ac425386aa0501790118
03:10 VanessaE seems okay
03:10 VanessaE (at least on a quick look at the code)
03:11 VanessaE seasonv += (f32)p.X / 3000; // you can walk to area with other season
03:11 VanessaE why no check against .Z also?
03:14 VanessaE oh wait
03:15 VanessaE hrm, no that's for humidity down there.
03:26 hmmmm man... mapgen v7 produces terrain that looks like it's from the original FarCry
03:27 hmmmm not sure if that's a good thing all the time
03:27 VanessaE hmmmm: see my commentary above?
03:27 hmmmm i totally get what celeron means by "way too vast" though
03:27 hmmmm yes, i did, that's proller's stuff though.
03:27 VanessaE (re: temperature/humidity maps being compatible with plants_lib)
03:27 hmmmm not mine
03:28 VanessaE actually he's pretty certain you won't allow any changes to the settings
03:29 hmmmm well I'm gonna have to change the default settings for heat and humidity eventually
03:29 hmmmm the challenge is finding an appropriate distribution of values
03:29 VanessaE you mean numerical distribution or geographical?
03:29 hmmmm numerical
03:29 VanessaE oh ok
03:30 hmmmm i just never really got around to doing that because a). it'd screw up everybody's biomes they have defined for their games, and b). that part is boring, i just wanna code
03:33 hmmmm grr :/
03:33 VanessaE http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/screenshot_692120453.png
03:33 VanessaE here, this'll cheer you up slightly :D
03:34 hmmmm what is it?
03:34 VanessaE those are moretrees palms....buried under half a meter of snow :D
03:35 hmmmm the snow ends at a block
03:35 hmmmm i suppose that's mod-made or something?
03:35 VanessaE don't mind that
03:35 VanessaE that's just new terrain after the weather mod was installed.
03:37 VanessaE figured the dichotomy would amuse you :)
06:57 hmmmm https://github.com/kwolekr/minetest/commit/9a7bd417cb7278021290dd5bd0da43db8ed77056
07:27 kahrl_ hmmmm: is it possible that murmur_hash_64_ua accesses invalid memory?
07:27 kahrl_ (if the input string to read_seed is not a multiple of 8 chars long and crosses a page boundary)
07:27 hmmmm nope
07:28 hmmmm as you can see, i take care of unaligned access.... i even say so
07:29 kahrl_ oh that's what the switch is for :)
07:30 kahrl_ yeah I shouldn't try to read code when tired
07:31 hmmmm alright i have a better idea
07:31 hmmmm i'll add an optional parameter onto engine.create_world which will be the seed, passed along to initializeWorld
07:31 hmmmm which will write the initial map_meta.txt
07:32 kahrl sounds good
07:32 hmmmm but that still doesn't fix the face that there's not enough space to do anything with the formspec window
07:32 hmmmm s/face/problem/
07:32 hmmmm why did i say face
07:32 kahrl maybe pass mapgen parameters as well? (only the selected mapgen currently)
07:32 hmmmm hey don't get cocky
07:33 hmmmm next thing you're going to want is for me to pass the flags, and then the noise parameters, and then the...
07:33 kahrl and everything in json please
07:34 hmmmm oh i know why i said face
07:34 hmmmm i meant to say fact but i made a typo and i'm too tired to notice it
07:35 kahrl #876 does it by reducing the height of the games textlist
07:37 hmmmm that's retarded
07:37 hmmmm i mean the fact that you need to shorten another element to fit like..... 4 fields
07:37 hmmmm i know what needs to be fixed next
07:38 kahrl well you would have to change the size[] field
07:38 kahrl which is hardcoded in mainmenu.lua:151
07:38 hmmmm oh
07:38 hmmmm i thought it was a deeper problem than that
07:39 kahrl I guess just check if tabbuilder.current_tab == "dialog_create_world" there
07:40 hmmmm i'd rather spend 15 more minutes making a more elegant solution so that i don't need to hardcode a hack
07:40 hmmmm i'll have each dialog specify a size as well
13:54 proller hmmmm, weather moving look ok (iw works ;), commit it and i must update other my pulls
16:40 hmmmm well I already did commit it last night
16:41 proller ok
16:42 proller i still want to commit weather and liquid branches
16:43 proller it make weather 80% complete
16:43 proller and liquid 95%
16:46 hmmmm i need to take a look at those
16:46 hmmmm about the waterfall thing..... not sure about that
16:47 hmmmm a lot of people don't like it, and that'd totally screw up the way liquids work
16:52 proller waterfall not ready
16:52 proller its only after fast nodeupdate
16:52 hmmmm then why make a pull request for it?
16:52 proller do not enable liquids by default
16:53 proller it was working, pull for discussion
16:53 proller and math pull now in progress, but works
16:54 proller last 4 pulls is ready and tested
16:54 hmmmm oh, you should specify that in the pull request description, that it's not ready yet and you'd just like feedback
16:55 proller [WIP] ok ?
16:55 hmmmm yea
16:57 proller edited
16:58 proller hmmmm, also lot people dont like default liquids, because it does not seem as liquid
16:58 hmmmm zz
16:58 hmmmm a lot more people don't like finite liquids
16:59 hmmmm we should have a poll and see exactly what percentage of people prefer finite liquids over old liquids
17:00 proller don't like finite liquids because it not as in minecraft
17:00 proller yes, pool can show real likes
17:00 proller poll 8)
17:24 celeron55 has someone modified the layout of the forum?
17:24 celeron55 the visual layout, that is
17:25 thexyz where's screenshot?
17:25 celeron55 it's not strictly broken, but for some reason i noticed that it uses the screen space on this laptop quite badly due to padding and margin
17:26 thexyz that's punbb, what else would you expect
17:26 celeron55 hmm, actually now i see why i noticed that
17:27 celeron55 for some reason my browser had chosen a different page scaling factor than it usually uses
17:27 celeron55 umm actually no
17:30 celeron55 http://i.imgur.com/goFKra6.png
17:31 celeron55 it looks like that, with my usual 90% scaling on this 1024px wide screen with tree style tabs like that; it roughly corresponds to an only slightly narrower view than 1024px which really should be expected to work reasonably well
17:33 thexyz yes, I remember VanessaE complaining about something like that
17:33 thexyz it's punbb
17:33 celeron55 but nothing has changed, right?
17:34 celeron55 anyway, this is a really crappy website layout by today's "responsive layout" standards; if you make the window narrower, it keeps the unnecessary information widely wisible and squeezes the titles to completely unreadable mess 8)
17:35 thexyz yeah, the code won't change itself
17:36 celeron55 i'm too tired to really think or code at the moment
17:36 celeron55 i'd guess it would work out better if it used simply percentages for column sizes
17:37 thexyz lol
17:37 thexyz so naive
17:37 thexyz mind me asking what colums are you talking about?
17:37 celeron55 the topics/forum/replies/last post ones
17:38 thexyz did you check the layout?
17:38 thexyz it's a mess of
s and
s
17:39 celeron55 i guess i'll have to...
17:40 celeron55 okay so while it's clearly table data, some clever person thought it shouldn't use tables
17:40 celeron55 well this is still probably doable *attempts something*
17:46 celeron55 okay so what the fuck
17:46 celeron55 left:952em
17:47 celeron55 i wonder what kind of drugs the someone creating this layout was using
17:48 celeron55 "text-indent:-999em;"
17:48 hmmmm speaking of interface layouts
17:48 celeron55 i'm not touching this crap; let's forget trying to fix it
17:49 hmmmm do you think it's a bad idea to have each tabbuilder function report the formspec size
17:50 hmmmm what i really want is to be able to modify it to be whatever for dialogs, but be the same size for all tabs, otherwise that'd go against the idea of a tab
17:51 ShadowNinja thexyz: What forum software would you prefer?
17:55 hmmmm nevermind, i decided on an elegant way to do what i'd like
18:00 thexyz ShadowNinja: something popular with nice code and clean interface (hint: it doesn't exist)
18:01 hmmmm man, where is sapier
18:06 ShadowNinja thexyz: vBulletin. :-D phpBB is popular and themable, but I don't know about the code, other than that it is more bloated than PunBB.
18:06 Calinou vbulletin sucks
18:06 Calinou the end
18:06 celeron55 we're kind of waiting for at least the one django-based thing to mature
18:07 celeron55 i don't remember the name though
18:10 thexyz ShadowNinja: vb 3 was good
18:11 thexyz then it went shit and xenforo was created
18:12 ShadowNinja vBulletin == Proprietary == I don't care how good it is. Was the Django forum DjangoBB?
18:13 thexyz lol
18:13 thexyz no
18:13 thexyz there are no good enough forums written in python, period
18:14 celeron55 it was some fancily named and nothing you can find by googling "django" and "forum"
18:14 celeron55 if it even was django
18:14 celeron55 also it could turn out shit as well, can't know
18:15 thexyz c55 has got a bad memory but he's talking about http://misago-project.org/
18:16 proller hmmmm, one timer weather_update_time but 2 separate calls heat and humidity - sometimes calculations of one can lost, but it not big problem
18:17 proller best way to update heat and humidity together
18:17 thexyz it looks modern, it feels modern, unlike whateverBB's
18:17 ShadowNinja That one looks nice.
18:22 celeron55 maybe someone of us should give them a heads-up so they know there will be use for it if they get it right
18:23 celeron55 altough they could like being unnoticed
18:23 celeron55 it really depends on what kind of person the main dev is...
18:24 thexyz I don't think the guy who's making it will care
18:39 hmmmm Oh
18:39 hmmmm you're right, I didn't consider the case of the individual lua api calls
18:40 hmmmm how do you want to proceed, would you rather call both updateBlockHeat and updateBlockHumidity, and throw away the result of whichever one isn't needed?
18:42 proller its only way if one timer used
18:43 proller liquid63 and weather updated
19:39 proller #0 0x0000000000642e50 in ServerMap::updateBlockHeat ()
19:39 proller #1 0x00000000004500af in ModApiEnvMod::l_get_heat ()
19:41 * VanessaE peeks in
19:41 proller block = getBlockNoCreateNoEx(getNodeBlockPos(p));
19:41 proller block->heat = heat;
19:41 proller BOOM
19:41 proller in my version was if(block != NULL) { block->heat = heat;
19:41 VanessaE celeron55: it's worse than that - it cuts off topic subject if the zoom level is high enough
19:48 hmmmm oooooooooh
19:48 hmmmm here I was, under the presumption that it would never be null
19:49 hmmmm right, need to take into account misbehaving mods
19:49 hmmmm I'll fix that back
20:02 ShadowNinja Back to 73 pulls. I have a sumple new one: https://github.com/minetest/minetest/pull/929 sfan5 already agreed to it but never actually commited it.
20:02 ShadowNinja simple*
20:03 celeron55 that has already been discussed many times; also we don't use 2^x values without valid reason to use them
20:04 thexyz ShadowNinja: what's the point?
20:04 proller pi^x more reasonable
20:04 lanxu so, what is the maximum value of the variable which stores 'max_users'? ;)
20:04 ShadowNinja thexyz: Minetest is very vulnerable to DoS attacks.
20:04 celeron55 the result of those discussions is that the right way to fix stuff is to make the server able to handle a lot of players... but i guess 100 is still too much; it would be really useful to get some feedback from server admins about what is a usable number with the modifications in recent versions
20:05 thexyz I doubt it can handle 32 tbh
20:05 ShadowNinja >8 == lots of lag, at least on my server.
20:05 thexyz yeah
20:07 celeron55 well, how about defaulting to a limit of 10?
20:07 ShadowNinja IMO it should be able to handle at least 10 without much lag.
20:07 ShadowNinja Some servers go above that.
20:07 celeron55 that lag is probably caused by the crappy network stack (that handles congestion control and stuff)
20:08 celeron55 umm... or well, it's probably equally likely to be caused by CPU maxing out
20:08 thexyz yeah
20:08 thexyz it eats ~100% of cpu
20:08 ShadowNinja I use 16 on my server, I haven't reached it yet, but I have gotten 12 or so.
20:08 thexyz (well, it did last time we tested it)
20:09 ShadowNinja Does lowering it to 10-20 seem better? Perhaps 16 or 15?
20:09 celeron55 it should be lowered to a value that doesn't generally cause problems
20:10 celeron55 it sounds like 20 is bad
20:10 ShadowNinja Hmmm, by problems do you mean lots of lag? 8 will be exceded by a number of servers.
20:11 celeron55 this is the default, not something that anyone has to stick to
20:11 ShadowNinja And more powerfull servers like Vanessa's can probably deal with more players(If it weren't for all the mods)
20:11 celeron55 the purpose of default settings is to minimize the need for most users to change settings
20:12 celeron55 (and the "common server" is probably quite small)
20:13 thexyz which is impossible since we don't have "average" servers
20:13 ShadowNinja thexyz: How so?
20:14 thexyz yeah
20:14 VanessaE on my server, the most I've seen at one time was about 8 or 9 players. Given its specs and history, I'd say double that at the very most.
20:14 VanessaE when it gets to around that many players, people start getting *some* lag, but most of that comes from new terrain I think.
20:15 ShadowNinja Pull changed to 8.
20:15 VanessaE (I have my server capped to 20)
20:16 proller 8 too small
20:17 celeron55 10 or 15
20:17 proller 13
20:17 ShadowNinja Most servers have 10-32.
20:17 proller 8)
20:17 celeron55 12.5
20:18 ShadowNinja Changed to 15...
20:19 VanessaE in fact I'd say the worst lag other than mapgen actually happens when people are signing on/off
20:19 VanessaE (well, on)
20:19 celeron55 by the way, is there any real-life occasion where having a different default setting for this would have had any effect?
20:20 celeron55 people mention about DoS so i guess some server has been flooded with users
20:21 celeron55 it would be useful to have written-down documentation about such somewhere so that people know what to expect and what to possibly fix
20:21 VanessaE celeron55: several servers got DoS'd a few months ago
20:21 VanessaE that's the only thing I've seen personally
20:21 VanessaE (and it was hardly a DoS, more like a slow flood)
20:23 celeron55 so would this have had any effect on those? i guess they had set their user limits to lower already anyway?
20:23 VanessaE only to the extent of making the server flood out sooner I guess
20:24 VanessaE but if there are already legit users online, it would keep the server from just plain dying
20:24 celeron55 i think i did some DoS-preventing work after that, but didn't touch the user limit
20:25 celeron55 previously the server jammed and stuck completely, now it will recover at the exact moment the DoS stops
20:25 VanessaE yes
20:25 celeron55 so it's really less an issue
20:26 VanessaE s/dying/becoming lagged beyond usability due to routine CPU load/
20:26 VanessaE :)
20:27 celeron55 anyway that latest pull could be merged
21:23 hmmmm minetest's server should've been designed from the ground up to scale
21:24 hmmmm I feel as if things were slapped together in an ad-hoc way and now that it can't scale for shit is a result of the "oh well we'll optimize it later" sort of cavalier attitude
21:24 hmmmm it'd also help if there were more things were event driven rather all handled in one huge step
21:25 hmmmm s/there were //
21:26 hmmmm basically what i'm saying is that making the minetest server offer better scalability is an entirely new project and you'd have to start over from square 1
21:34 proller https://github.com/proller/minetest/commit/a92c880da913949f065a62f83b69db3bb950becb - fix segfault
21:34 proller hmmmm, you or me?
21:34 hmmmm you do it
21:34 PilzAdam "Fix segfault" is not a good commit message
21:35 proller Segfault fix - better?
21:35 hmmmm lol
21:35 VanessaE I think he means to explain why it fixes it :)
21:36 hmmmm how about, "Fix null dereference in weather update functions"
21:36 PilzAdam "Fix segfault in updating heat/humidity"
21:38 proller done
21:39 PilzAdam hmmmm, https://github.com/minetest/minetest/pull/927 MSVC needs this for M_PI
21:39 PilzAdam this is the 3rd or 4th time someone breaks it like this
21:40 hmmmm I don't get it
21:40 hmmmm why don't you just include that in util/numeric.h
21:40 hmmmm by far, that's the more appropriate place
21:41 hmmmm anywhere you use mathconstants.h, you used numeric.h as well, no doubt
21:41 hmmmm like only a handful of us use windows so we can't keep checking to see if it works or not
21:59 Sokomine why does the server have problems with many players? 10 players is not really much, even though many servers have less
22:00 Sokomine what happens there that eats up so much cpu time? abms? chunks sent out to players?
22:00 Sokomine i've saved a lot of redcrabs map and was amazed at the slow speed the map chunks arrived at my client
22:02 Sokomine please understand that i don't want to complain. just wonder what is going on and hoping that a solution can be found
22:02 Sokomine to me, all that graphics calculation of actually showing the world sounds way more complicated than players running around withhin the world on the server
22:02 ShadowNinja Sokomine: See logs within the few hours.
22:03 ShadowNinja Basically there are lots of reasons, and just bad design.
22:18 Sokomine thank you. i'll read it
22:21 johnDC hello everybody, I'm newbie here, I'm trying to compile minetest, but I've found web-documentation as not complete. would you please point me, where to find info about levelDB, cURL, GetText and freetype?
22:23 ShadowNinja johnDC: #minetest would be a better channel for that question.
22:23 johnDC thnx