Minetest logo

IRC log for #minetest-dev, 2014-12-08

| Channels | #minetest-dev index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
00:01 Taoki Any reason why #1266 wasn't merged yet? It's been there for a while... it's a needed correction to make tone maps work properly.
00:01 ShadowBot https://github.com/minetest/minetest/issues/1266 -- Make directional fog colors respect tonemap by MirceaKitsune
00:01 Taoki RealBadAngel: Perhaps you can merge that now. Especially since I created the directional fog, and this is additionally a patch to my code. And it shouldn't give anyone reasons to disagree.
00:11 RealBadAngel im ok with it.  kahrl, sfan5 ?
00:19 RealBadAngel Tesseract, and you?
00:20 Tesseract Um, I don't know exactly what that does.
00:21 RealBadAngel when theres sun or moon tone map present it lets fog use it
00:21 RealBadAngel which is logical
00:21 Tesseract Seems O.K.
00:22 RealBadAngel Tesseract,  and what about #1925 ?
00:22 ShadowBot https://github.com/minetest/minetest/issues/1925 -- Disable loading .mtl files. Theyre not used anyway. by RealBadAngel
00:24 Tesseract RealBadAngel: Fixes the "couldn't load foo.png for foo.x" errors?
00:24 RealBadAngel fixes could not load xxx.mtl
00:24 Tesseract Never seen that.
00:25 RealBadAngel try anything that loads a mesh
00:25 RealBadAngel i mean .obj files
00:26 RealBadAngel join VE's creative for example to see it for window shutters
01:21 shadowzon joined #minetest-dev
01:36 RealBadAngel https://github.com/minetest/minetest/pull/1927
01:36 RealBadAngel #1927
01:36 ShadowBot https://github.com/minetest/minetest/issues/1927 -- Fix lighting artifacts (Issue #1887). by RealBadAngel
01:37 RealBadAngel before we get hardware lighting we have to update all the blocks around (waste of resources :( )
01:52 hmmmm just realized a problem
01:53 hmmmm i test config file writing now for Settings, but how can i guarantee that writing a floating point value will match the expected output string?
01:54 hmmmm i mean how can I know that ftos(stof("2.0")) == "2" ?
01:55 hmmmm and then what about locale settings... i need somebody who knows more about this to verify TestSettings' correctness
02:10 RealBadAngel a = stof (x); y = stof(a); if (x==y) it is correct
02:10 RealBadAngel a = ftos (x); y = stof(a); if (x==y) it is correct
02:10 RealBadAngel oops ;)
02:14 RealBadAngel hmmmm, why do you need to know the output format anyway?
03:00 hmmmm because i'm comparing strings
03:00 hmmmm paramat:  https://github.com/kwolekr/minetest/commit​/2fd3d5202051e03303ac2b8e76976a7c4c8477f3
03:00 hmmmm erm he's not here
03:00 hmmmm everyone:  https://github.com/kwolekr/minetest/commit​/2fd3d5202051e03303ac2b8e76976a7c4c8477f3
03:13 paramat joined #minetest-dev
03:13 paramat thanks hmmmmm
03:14 paramat wow you've been working hard
03:15 paramat so lacunarity for no change to before is 2
03:17 paramat ah i just researched, lacunarity is exactly what i was thinking of adjusting in the noise code
03:17 paramat this will allow a wide range of detail with fewer 'octaves'
03:18 paramat this ties into my studies of microtonal music theory
03:19 paramat some people are making music where the 'octave' is not a doubling of frequency
03:20 paramat such as this https://www.youtube.com/watch?v=wBOxnO10KjE
03:28 hmmmm lacunarity is best between 1.5 and 3.5
03:30 paramat for the last week or so i get this error message when starting a new world in mgv5/v6/v7: ERROR[main]: ServerMap::loadMapMeta(): could not open/home/../minetest/bin/../worlds/x/map_meta.txt
03:32 hmmmm hrmm
03:32 hmmmm i personally haven't
03:33 hmmmm i've heard of that problem existing months ago when somebody made some other change
03:34 paramat will report as an issue
03:38 hmmmm i'm going to add a 'point noise value buffer' mode to NoiseParams
03:39 hmmmm in this mode, the noise calculation functions will bypass the intermediate buffers and write directly to the result buffer using the point-polled value functions.  the purpose of this is to accomodate "crazy" noiseparams settings such as a very low spread factor, without creating crazy requirements
03:39 hmmmm and higher octaves too
03:40 hmmmm i'm wondering if, on InvalidNoiseParamsException, if I should write an error message about invalid noise params but switch to this point-buffer mode
03:40 hmmmm and have execution continue on
03:41 hmmmm basically it'd make the Noise class equivalent to celeron's 0.3.x-era NoiseBuffer class
03:41 paramat interesting
03:52 paramat left #minetest-dev
03:57 hmmmm i actually kind of like that fractal song you pasted the link to
03:58 hmmmm it sounds creepy and off-tune
04:34 paramat joined #minetest-dev
04:36 paramat that music video and artwork is actually an old experiment of mine =} (i confess!)
04:39 paramat hmmmm im not sure crazy noise params should be accomodated, if the smallest spread is less than 1 it can't be expressed in nodes, if octaves are high that's also a waste of processing
04:40 paramat better to throw an error so people learn
04:41 paramat is 'point noise value buffer' useful for intelligent use?
04:42 CraigyDavi joined #minetest-dev
04:51 hmmmm it can allow minetest to use marginally less memory...
04:51 hmmmm i guess
04:52 hmmmm at the expense of vastly slower noise computation :-)
04:55 paramat i suggest not bothering then
05:14 hmmmm WOW
05:14 hmmmm no freaking way gcc is this stupid
05:15 hmmmm for (size_t i = 0; i != sx * sy * sz; i++) result[i] = result[i] * np->scale + np->offset;     is 4us slower than   for (size_t i = 0; i != sx * sy * sz; i++) result[i] = result[i] * np->scale + np->offset;
05:15 hmmmm with -O2
05:18 hmmmm erm not stupid... just very careful
05:18 hmmmm if I make sx, sy, and sz local variables, there is no difference in performance
05:19 hmmmm protip:  if you're checking a loop condition against a class member, cache the result first.  it might result in less pretty code but it's faster
05:21 hmmmm sx * sy * sz in this case was pretty blatant, but something less obvious might be like a size member.   for (size_t i = 0; i != m_size; i++) { do stuff here on data }   - bad
05:21 hmmmm size_t localsize = m_size;  for (size_t i = 0; i != localsize; i++) { ...   - good
05:44 sol_invictus joined #minetest-dev
05:57 kahrl nice
05:57 kahrl go to the multiplayer tab, type a nonexisting address and name ;_; and connect
05:58 kahrl then abort. voila: the name field is gone
05:59 kahrl lesson: _never_ dump user input into a formspec without escaping it
06:10 Zeno` joined #minetest-dev
06:14 selat joined #minetest-dev
06:16 kahrl will push this in 30 minutes: https://gist.github.com/kahrl/591a33feeccf4f901c58
06:16 kahrl fixed the above
06:28 Zeno` I had to look up "lacunarity" on wikipedia
06:28 Zeno` And I still have no idea what it means
06:39 NakedFury joined #minetest-dev
06:44 kahrl as far as I can tell, a fractal is lacunar if the sizes of the holes vary a lot if you look at different parts of the fractal
06:44 kahrl I had to read the wikipedia definition several times though ;)
06:45 RealBadAngel #1927
06:45 ShadowBot https://github.com/minetest/minetest/issues/1927 -- Fix lighting artifacts (Issue #1887). by RealBadAngel
06:46 paramat before, the ratio of the spreads of adjacent octaves was always 2 (hence octave, a doubling of frequency), now that ratio (lacunarity) can be set
06:48 Zeno` ok, I think I get it now
06:49 Zeno` RBA, I'll test now
06:49 paramat doubling spatial frequency = halving spatial wavelength = halving 'spread'
06:51 Zeno` hmm ok
06:51 Zeno` I guess the best way for me to understand properly is to experiment
06:54 Zeno` RBA, I can no longer reproduce the issues I was experiencing
06:54 Zeno` So I'm guessing it's truly fixed :)
06:55 Zeno` yep +1 !
06:56 Zeno` All the test locations I had "bookmarked" where I was experiencing issues no longer show any anomalies
06:56 kahrl RealBadAngel: I don't understand the addUpdateMeshTaskForNode changes
06:57 kahrl why can you skip those mesh updates?
06:57 Hunterz joined #minetest-dev
06:57 RealBadAngel simply because addUpdateMeshTaskWithEdge already does them
06:58 RealBadAngel no need to update mesh twice
06:58 kahrl but you call addUpdateMeshTask and not addUpdateMeshTaskWithEdge
06:58 RealBadAngel i call the last one
06:59 kahrl not in line 2647
07:00 RealBadAngel https://github.com/minetest/minetest/pull/1927/f​iles#diff-34f48ad91ac6c202ac60b0348ae90e30R2351
07:00 RealBadAngel https://github.com/minetest/minetest/pull/1927/f​iles#diff-34f48ad91ac6c202ac60b0348ae90e30R2373
07:00 kahrl I saw that but what does that have to do with addUpdateMeshTaskForNode?
07:01 RealBadAngel one calls another
07:01 RealBadAngel result was a few unnecesary mesh updates
07:02 RealBadAngel we need to update block where node is and then addUpdateMeshTaskWithEdge which will update surroundings
07:02 kahrl it's broken
07:02 kahrl if you dig a node at a block boundary the crack doesn't appear
07:03 RealBadAngel lemme check
07:03 * Zeno` finds block boundary
07:04 RealBadAngel checked, no problems with cracks
07:04 kahrl try it from the other side
07:05 kahrl only one direction is affected
07:05 kahrl well, 3 of the 6 directions to be exact
07:06 RealBadAngel huh, indeed theres problem with it
07:09 RealBadAngel but i can see crack missing only at bottom of the node
07:09 kahrl it should happen at the x-, y- and z- faces
07:14 RealBadAngel youre right, ive removed too much there, will fix that
07:14 FR^2 joined #minetest-dev
07:18 paramat left #minetest-dev
07:19 RealBadAngel kahrl, check now
07:19 kahrl looks good now, will test
07:22 kahrl all the artifacts are gone now
07:22 kahrl good job :)
07:23 kahrl if it works for Zeno` too it can be merged
07:24 Zeno` give me 5 minutes?
07:24 Zeno` Just on the phone
07:25 kahrl sure
07:29 RealBadAngel when we will switch to real lighting we will have to re-revert those changes
07:29 Zeno` ok, my test locations are now fine (lighting wise)
07:29 RealBadAngel only lighting forces us to make so many mesh updates
07:30 Zeno` so yes, it resolves the issue :)
07:31 Zeno` re-re-re-revert?
07:31 Zeno` heh
07:31 RealBadAngel updating surrounding blocks meshes i mean
07:31 kahrl revert the revert that reverted the reverting revert
07:31 RealBadAngel hehe
07:32 RealBadAngel only those in update node makes sense, because removing a node can make face in adjacent block visible again
07:33 RealBadAngel any other updates in adjacent blocks are forced by lighting
07:36 Zeno` Anyone looked at #1873 ?
07:36 ShadowBot https://github.com/minetest/minetest/issues/1873 -- Temporary fix for irrlicht bug that cause MT not going to be compiled on BSD-like by neoascetic
07:37 Zeno` If I was going to address that bug I probably would have first done it the way neoascetic has done. There is already an #ifdef for POSIX anyway. I'd like to get away from irrlicht types altogether but that is a different story.
07:38 kahrl Zeno`: why is there an #ifndef _MSC_VER?
07:38 Zeno` kahrl, simply to include stdint
07:38 Zeno` same as the #else below already did
07:39 kahrl oh I see
07:39 Zeno` http://sourceforge.net/p/irrlicht/bugs/433/ explains it nicely and I really don't have an issue with this patch
07:41 kahrl I think it would be better if stdint.h was always included
07:41 Zeno` some compilers don't have it
07:41 kahrl just to avoid the case where a dev writes code that relies on it that then doesn't work on windows
07:42 kahrl huh
07:42 Zeno` I don't think older versions of MSCV do. Do you know which versions do and which don't?
07:42 kahrl those aren't C++ compilers then
07:42 kahrl or do they only have cstdint
07:42 Zeno` *shrug* I read it somewhere or maybe Krock told me
07:43 kahrl MSVC 2010 definitely has it
07:43 kahrl do we support older ones?
07:43 Zeno` I guess that's a question to answer
07:43 Zeno` I don't know
07:44 Zeno` If we don't then it's a non-issue and we include stdint all the time
07:44 Zeno` Actually that might make sense
07:45 Zeno` Supporting a 2008 version of Visual Studio seems a bit much...
07:46 kahrl it would be good to have a wiki page that explicitly spells out the compilers we currently support
07:46 Zeno` A bigger question (related) is do we support non-conforming compilers? :)
07:46 kahrl well we support at least one version of MSVC so yes ;)
07:47 Zeno` :)
07:48 kahrl and I guess since we don't set -pedantic, g++ isn't conforming either
07:51 kahrl that reminds me, I need to fix some struct/class differences
07:51 Zeno` Just setting -Wextra is bad enough
07:51 Zeno` I won't even try with -pedantic
07:52 Zeno` Oh also... do you know why -Wall is not included as a flag for release builds?
07:52 kahrl shall I push RBA's lighting fix?
07:52 Zeno` yep
07:52 kahrl no idea
07:52 kahrl I don't do releases ;)
07:54 Zeno` https://github.com/minetest/minetest​/blob/master/src/CMakeLists.txt#L661
07:54 Zeno` to me it seems strange to omit just because you're building a release version, but maybe there is a reason
07:55 Zeno` Hmm, I wonder why it's not picked up from line 657
07:56 Zeno` nvm, maybe it was fixed already
08:02 kahrl https://gist.github.com/kahrl/08a2403e35b85b1ff12c
08:02 kahrl I'll fix those forward declarations
08:20 Zeno` who said you were allowed to use a script?
08:20 Zeno` seems like cheating!
08:21 kahrl heh
08:26 hmmmm https://github.com/minetest/minetest/commi​t/2a7c6d27b32711e2439bf56715d813ef96b8f904
08:41 jin_xi joined #minetest-dev
08:42 NakedFury joined #minetest-dev
09:03 Zeno` he missed an optimisation :(
09:04 Zeno` Noise::transformNoiseMap() is apparently called a lot with np->scale and np->offset ~= 1.0
09:04 Zeno` np->scale ~= 1.0 and np->offset ~= 0 sorry
09:52 kaeza joined #minetest-dev
10:32 ImQ009 joined #minetest-dev
10:57 lag01 joined #minetest-dev
11:03 Amaz joined #minetest-dev
11:13 alexxs joined #minetest-dev
11:34 proller joined #minetest-dev
11:47 kilbith joined #minetest-dev
11:55 MichaelRpdx joined #minetest-dev
12:01 PilzAdam joined #minetest-dev
12:09 kilbith joined #minetest-dev
12:14 selat joined #minetest-dev
12:38 Megaf joined #minetest-dev
12:52 Zeno` will merge #1930 when travis finishes
12:52 ShadowBot https://github.com/minetest/minetest/issues/1930 -- Fix compiler warning (signed vs. unsigned) by Zeno-
13:01 Zeno` done
13:01 ImQ009 joined #minetest-dev
13:04 Zeno` Can someone else look at #1873 please?
13:04 ShadowBot https://github.com/minetest/minetest/issues/1873 -- Temporary fix for irrlicht bug that cause MT not going to be compiled on BSD-like by neoascetic
13:04 Zeno` kahrl and I discussed this earlier (check logs) and it seems to be a perfectly reasonable solution to an irrlicht problem to me
13:05 Zeno` Perhaps the word "temporary" can be removed from the commit message
13:06 Zeno` All it's doing is making sure that stdint is included before irrtypes.h :/
13:48 shadowzon joined #minetest-dev
13:52 kilbith joined #minetest-dev
14:01 proller joined #minetest-dev
14:21 twoelk joined #minetest-dev
14:38 kilbith joined #minetest-dev
14:48 ImQ009 joined #minetest-dev
15:01 proller joined #minetest-dev
15:04 DFeniks joined #minetest-dev
15:04 ImQ009 joined #minetest-dev
15:06 electrodude512 joined #minetest-dev
15:30 hmmmm joined #minetest-dev
15:32 Hunterz joined #minetest-dev
15:39 Calinou joined #minetest-dev
15:40 ImQ009 joined #minetest-dev
15:40 hmmmm Zeno`:  yeah, most of the time offset == 0 and scale == 1, which is why it's a separate function you could optionally call
15:40 hmmmm you save a good 6 microseconds that way
15:51 hmmmm 6 * 80 = 480 microseconds if using 3d noise
15:54 Zeno` ahh, ok, it's already checked then? (my comment was half tongue-in-cheek btw)
15:56 electrodude512 joined #minetest-dev
16:03 PenguinDad joined #minetest-dev
16:04 Hunterz joined #minetest-dev
16:28 MinetestForFun joined #minetest-dev
17:03 Hunterz joined #minetest-dev
17:19 electrodude512 joined #minetest-dev
17:36 Krock joined #minetest-dev
17:50 lag01 joined #minetest-dev
17:50 RealBadAngel hi guys
17:51 VanessaE greetz
17:51 Calinou hi RealBadAngel
17:51 RealBadAngel and ladies :)
17:51 RealBadAngel shit, wrong order ;)
17:51 Calinou there is no wrong order here
17:52 RealBadAngel Calinou, ladies first, i should have check the list first ;)
17:52 RealBadAngel anyway, theres a call to fix visual_scale parameter
17:53 VanessaE but is there anything to fix?
17:53 RealBadAngel atm, at lua side its single float, copied onto the vector in the engine
17:54 RealBadAngel so you cant control desired dimension, just all of them
17:54 RealBadAngel problem is, this way is used everywhere
17:55 RealBadAngel we need (like really) to switch from single float to a vector of 3 floats
17:55 Calinou like wield_scale?
17:56 RealBadAngel like visual_scale too
17:56 VanessaE oh right, forgot about that idea.
17:56 RealBadAngel with Z=0 in scale we could have old flat plants for example
17:58 RealBadAngel and control thickness of all extruded things
17:59 RealBadAngel not to mention separate factor for x and y could be also useful
18:06 Problemascraft joined #minetest-dev
18:23 electrodude512 joined #minetest-dev
18:39 PilzAdam joined #minetest-dev
19:00 celeron55 joined #minetest-dev
19:02 chchjesus joined #minetest-dev
19:16 Taoki Since I'm seeing a lot of misc stuff is being merged, can someone also look at #1926 ? It's a small fix, which would allow me to add swimming sounds to minetest_game and already enables ladder climbing sounds out of the box.
19:16 ShadowBot https://github.com/minetest/minetest/issues/1926 -- Liquid & climbing footsteps by MirceaKitsune
19:18 Krock Btw, why was the feature with >right-clicking once to put a single item into the inventory< changed to 2 clicks?
19:23 Taoki Also, I tested how view bobbing while swimming and climbing looks, and IMO that is also a large improvement because it looks a whole lot better and more realistic. So this kinda improves two things simultaneously.
19:23 Calinou yes, I like view bobbing when swimming/climbing, good idea
19:23 Calinou (I have  view bobbing off but wield item bobbing on)
19:24 Calinou Taoki, also can you increase max bobbing speed?
19:24 Calinou m_view_bobbing_speed = MYMIN(speed.getLength(), 40);
19:24 Calinou change the 40 for a bigger value, I suggest 80
19:24 Calinou or 75
19:24 Taoki I think someone else made a separate pull request for that, which changes that max to 100
19:24 Taoki Might have commented on it just yesterday
19:24 Calinou someone_else = "Calinou",
19:25 Krock #1921 comments please
19:25 ShadowBot https://github.com/minetest/minetest/issues/1921 -- Ignore .name directories and files by SmallJoker
19:25 Calinou if only I was a core dev, I could merge simple stuff like that :(
19:26 Krock Calinou, and that's bad to merge on your own
19:26 nore joined #minetest-dev
19:26 Calinou I mean, other people's PRs
19:26 Taoki Another nice thing about view bobbing working while swimming, is that the bob cycle speed depends on movement speed. So the view and wield item bob more slowly when you swim or climb then when you walk, wich looks and feels very correct :)
19:31 VanessaE joined #minetest-dev
19:32 VanessaE_ joined #minetest-dev
19:38 VanessaE_ joined #minetest-dev
19:41 paramat joined #minetest-dev
19:43 paramat hi nore sfan5 please could you merge https://github.com/minetest/minetest_game/pull/366 ? roots are removed as requested, rebased
19:46 Taoki Looks nice! Would have been even nicer with non-mgv7 snow biomes though.
19:50 Guest82129 joined #minetest-dev
19:53 FR^2 joined #minetest-dev
19:54 Calinou https://github.com/minetest/minetest_game/pull/354
19:54 Calinou can this be merged? core dev has approved it
19:58 * shadowzon wonders how she can vote it up
19:59 shadowzon or do I have to be added?
19:59 kilbith need a Zeno clone for _game
20:02 Calinou feel free to vote it up, but it won't do much
20:03 shadowzon I'm new to github so I don't know how to vote up
20:06 paramat 4000 commits!
20:08 PenguinDad paramat: come back when we reached 9001 :P
20:16 nore joined #minetest-dev
20:36 electrodude512 joined #minetest-dev
20:39 twoelk paramat: thinking about them trees with roots of yours, how about being able to plant saplings on trunks similar to real life grafting https://en.wikipedia.org/wiki/Grafting
20:44 paramat idea is okay but personally i can't get enthusiastic about adding it =)
20:46 twoelk might be usefull to have cultivated varieties such as discussed here http://irc.minetest.ru/min​etest/2014-08-24#i_3884595
20:47 twoelk or maybe a way to make apple orchards and not wait for the random apple tree
21:06 Garmine joined #minetest-dev
21:15 Miner_48er joined #minetest-dev
21:29 proller joined #minetest-dev
21:29 Garmine42 joined #minetest-dev
21:44 shadowzon joined #minetest-dev
21:48 shadowzone joined #minetest-dev
22:01 kaeza joined #minetest-dev
22:12 VanessaE joined #minetest-dev
22:15 VanessaE joined #minetest-dev
22:18 paramat left #minetest-dev
22:20 Amaz joined #minetest-dev
22:21 VanessaE_ joined #minetest-dev
22:23 electrodude512 joined #minetest-dev
22:36 Amaz left #minetest-dev
22:44 exio41 joined #minetest-dev
22:53 Sokomine joined #minetest-dev
22:53 Wayward_One joined #minetest-dev
23:10 proller joined #minetest-dev
23:15 electrodude512 joined #minetest-dev
23:17 Guest66859 joined #minetest-dev
23:34 electrodude512 joined #minetest-dev
23:49 hoodedice joined #minetest-dev

| Channels | #minetest-dev index | Today | | Google Search | Plaintext