Minetest logo

IRC log for #minetest-dev, 2014-11-21

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

All times shown according to UTC.

Time Nick Message
00:00 Megaf joined #minetest-dev
00:01 Megaf joined #minetest-dev
00:08 paramat joined #minetest-dev
00:10 khonkhortisan joined #minetest-dev
00:39 shadowzone joined #minetest-dev
00:54 khonkhortisan joined #minetest-dev
00:57 ShadowLadyXD joined #minetest-dev
01:00 Zeno` joined #minetest-dev
01:14 shadowzone joined #minetest-dev
01:39 khonkhortisan joined #minetest-dev
01:58 paramat hey hmmmm please could you check code/merge https://github.com/minetest/minetest/pull/1834 ? it is rebased and squashed, details in last comment
01:59 hmmmm sure
02:00 hmmmm ahh
02:00 paramat thanks, also preparing the overgen commit
02:01 hmmmm you realize that f * noise_ground->np->octaves * noise_ground->np->scale isn't the min/max of that noise?
02:01 hmmmm I am going to be adding lacunarity soon as a noise param, and if that is greater than 1, it would blow your entire assumption away
02:02 paramat ah okay
02:03 hmmmm hmm
02:03 paramat i assumed f * octaves * scale because persist is usually set below 1
02:03 paramat so i think it's okay currently
02:03 hmmmm i should probably add a utility function to Noise to calculate the min/max possible values
02:04 paramat i do that often with a calculator
02:05 paramat okay so i'll need to rewrite in future, but this works for spawning in the meantime
02:24 paramat hmmmm, here's the new pull request for a simpler (temporary) mgv5 overgeneration https://github.com/minetest/minetest/pull/1855
02:36 paramat hmmmm, biomes certainly does need work, i have a suggestion for being able to set biome heat/humidity noise parameters in .conf, it's essential to tune the spreads depending on number of biomes, i'll probably attempt this myself
03:01 Megaf joined #minetest-dev
03:03 mrtux joined #minetest-dev
03:37 cg72 joined #minetest-dev
03:50 Megaf joined #minetest-dev
04:10 hmmmm just wondering how the overgeneration affects speed
04:10 hmmmm did you check that?
04:10 paramat no, but its from 80 to 82 node layers
04:10 hmmmm oh okay
04:11 paramat just 1 up 1 down
04:11 hmmmm i like the overgeneration idea, but the implementation could probably be improved, or generateBiomes could be improved, or both
04:11 hmmmm notice how at y=47 the biome 'top node' is discontinous?
04:11 hmmmm it forms a right angle
04:12 paramat yeah
04:12 hmmmm yea
04:12 paramat you mean in my commit?
04:13 hmmmm the screenshot on the pull request
04:13 paramat ah
04:14 paramat yes i think i see what you mean
04:15 hmmmm oh
04:15 paramat this is just a very minimal temporary overgen, because ive been working on deeper overgen and its complex, many weeks work perhaps
04:15 hmmmm when you write your commit messages, could you do it in the format of
04:15 hmmmm component: Add thing here
04:16 hmmmm for example
04:16 hmmmm Mapgen V5: Add overgeneration support
04:16 paramat okay
04:16 hmmmm thanks
04:16 paramat oh wow thanks for the merge
04:16 hmmmm what
04:17 paramat only just seen
04:27 Zeno` Ok I'm happy with #1852 now. Some good improvements in speed with few changes
04:27 ShadowBot https://github.com/minetest/minetest/issues/1852 -- Optimise getTileInfo(), CNodeDefManager::get() and VoxelManipulator::addArea() by Zeno-
04:28 hmmmm excellent
04:28 hmmmm addArea doesn't modify area any, right?
04:30 hmmmm you should probably add a const qualifier to those if so
04:30 Zeno` good point; I'll check
04:31 hmmmm yea, VoxelManipulator::addArea(const VoxelArea &area)
04:31 hmmmm it looks great!
04:31 Zeno` ok, will do
04:31 hmmmm i love when people make things faster
04:31 hmmmm minetest is too damn slow
04:31 Zeno` thanks. So you're ok with a merge once I fix that const qualifier?
04:31 hmmmm yea
04:32 Zeno` thanks
04:36 Zeno` area is modified
04:36 Zeno` hmm no it's not... will have to change some other functions
04:46 mrtux joined #minetest-dev
04:47 Megaf joined #minetest-dev
04:54 NakedFury joined #minetest-dev
04:57 Megaf joined #minetest-dev
05:00 selat joined #minetest-dev
05:31 sol_invictus joined #minetest-dev
05:37 VanessaE joined #minetest-dev
05:38 Megaf joined #minetest-dev
05:47 paramat left #minetest-dev
05:53 Megaf joined #minetest-dev
06:12 Megaf joined #minetest-dev
06:25 Weedy joined #minetest-dev
06:26 SmugLeaf joined #minetest-dev
06:36 VanessaE joined #minetest-dev
06:41 Hunterz joined #minetest-dev
06:46 Miner_48er joined #minetest-dev
06:47 Miner_48er joined #minetest-dev
06:57 Weedy joined #minetest-dev
06:57 Weedy joined #minetest-dev
07:06 Weedy joined #minetest-dev
07:11 Weedy joined #minetest-dev
07:14 jin_xi joined #minetest-dev
07:15 SmugLeaf joined #minetest-dev
07:20 Weedy joined #minetest-dev
07:20 Weedy joined #minetest-dev
07:21 Megaf_ joined #minetest-dev
07:32 VanessaE joined #minetest-dev
07:45 RealBadAngel joined #minetest-dev
07:46 RealBadAngel hi
07:46 RealBadAngel #1856
07:46 ShadowBot https://github.com/minetest/minetest/issues/1856 -- Port createForsythOptimizedMesh from Irrlicht 1.8 by RealBadAngel
07:47 RealBadAngel i need that function to speed-up mapblock mesh, unfortunatelly its aviable only in irrlicht >=1.8, so i merged it into our codebase
07:49 Zeno` more performance improvements? :D
07:55 VanessaE RealBadAngel: how much of an improvement will it lead to?
07:56 VanessaE corollary to that, is there a good reason to maintain compatibility with 1.7.x at this point?
07:57 RealBadAngel some distros still dont have 1.8
07:58 RealBadAngel and this is second case i ported back needed code, first time it was when i added l-system trees
07:58 RealBadAngel read: ages ago ;)
07:58 VanessaE well for those distros where 1.8 isn't yet available, it's not that hard to compile irrlicht
07:59 VanessaE just wondered is all
07:59 RealBadAngel https://github.com/minetest/minetest/blob/master/src/treegen.cpp#L496
08:00 RealBadAngel this code is aviable in 1.8 too
08:01 RealBadAngel anyway, either we switch to 1.8 now or i need patch above
08:02 Zeno` What's the performance gain?
08:02 RealBadAngel it depends on the models(meshes)
08:02 Zeno` lots or moderate? :)
08:02 RealBadAngel it simplifies the meshes, so they have less polygons
08:03 Zeno` Oh, there is a link to a paper. /me reads
08:03 RealBadAngel and its quite fast
08:03 RealBadAngel im already using that for all meshnodes and wielded
08:04 Zeno` In your own repo?
08:04 RealBadAngel in my worktree
08:04 Zeno` yep
08:04 RealBadAngel instead of cloneMesh im using this function
08:04 RealBadAngel i get a cloned mesh but optimized at the same time
08:05 Zeno` well if it's needed and we don't want to move to 1.8 then there cannot be a problem with this I'd expect :)
08:05 Zeno` And when we do move to 1.8 it would simply get removed (correct?)
08:05 RealBadAngel yup
08:06 RealBadAngel same as setRotationAxisRadians
08:07 Weedy joined #minetest-dev
08:07 Weedy joined #minetest-dev
08:11 kilbith joined #minetest-dev
08:29 VanessaE bbl
08:39 Zeno` I'm all for something that improves performance or leads to improved performance
08:41 kilbith fix the mem. leaks reported by VE, Sokomine, Megaf
08:42 Zeno` kilbith, I've started hunting
09:00 GrimKriegor joined #minetest-dev
09:02 FR^2 joined #minetest-dev
09:11 jin_xi Zeno: one def mem leak is the one with signs and such. celeron also mentioned a solution, which was to reference count texture ourselves. that would be a very worthwile addition imo
09:13 jin_xi http://irc.minetest.ru/minetest-dev/2014-10-11#i_3976793
09:16 Zeno` hmm
09:17 Zeno` I dunno if signs would cause the kind of leak I'm seeing though
09:17 Zeno` I can get the client to use 10GB of RAM in about 15 minutes
09:17 jin_xi ugh
09:17 Zeno` maybe they can
09:17 Zeno` I'm going to try something
09:17 Zeno` it's hard to track down because at end of execution everything is cleaned up correctly
09:18 Zeno` I'll deliberately crash it
09:25 OldCoder Zeno`, Hi
09:25 Zeno` Hi
09:26 OldCoder Most people asleep?
09:26 Zeno` They've had several days of sleep. Must have eaten an enchanted apple
09:26 OldCoder Ah
09:30 selat joined #minetest-dev
09:40 Megaf_ heh, yesterday I was asking about Irrlicht 1.8 features here, and today we seem to soon get some https://github.com/minetest/minetest/pull/1856
10:00 ImQ009 joined #minetest-dev
10:09 Megaf_ Comments? https://github.com/minetest/minetest/pull/1854#issue-49590015
10:31 Zeno` I have no issue with that Megaf_ (in theory) except for the fast_fly
10:31 Zeno` RealBadAngel, if your PR is needed to improve things then I say go for it
10:33 Megaf_ If the whole problem is with fly fast then I can just remove that...
10:34 Megaf_ That does seems like a personal opinion and seems like people like the way it is not because they find it better, but just because it is like that at the moment
10:40 selat joined #minetest-dev
11:05 Megaf_ Why this havent been marged yet? https://github.com/minetest/minetest_game/pull/336
11:05 Megaf_ #336
11:05 ShadowBot https://github.com/minetest/minetest/issues/336 -- Settings checkboxes not saved if you quit without first leaving the Settings tab
11:06 Megaf_ not that
11:06 Megaf_ Pine tree, pine needles, pine saplng and pine wood nodes, stairs and slabs
11:08 Zeno` Dunno. I gave it a thumbs up just now
11:09 Megaf_ https://github.com/minetest/minetest/issues/1643#issuecomment-63956484
11:09 Megaf_ #1643
11:09 ShadowBot https://github.com/minetest/minetest/issues/1643 -- Single player worlds should be visible regardless of selected game
11:10 Megaf_ !tell PilzAdam https://github.com/minetest/minetest/issues/1643#issuecomment-63956484
11:10 ShadowBot Megaf_: O.K.
11:27 zat joined #minetest-dev
11:28 proller joined #minetest-dev
11:32 kahrl joined #minetest-dev
11:43 kahrl so... wieldmesh
11:43 kahrl I had an idea for fixing #1844
11:43 ShadowBot https://github.com/minetest/minetest/issues/1844 -- odd wield item glitch 2
11:44 kahrl if the extrusion mesh is split into two meshbuffers, one for the Y slices and one for everything else, anisotropic filtering can be forced on for only the problematic Y slices
11:44 kahrl so I implemented that: https://github.com/kahrl/minetest/commit/5709750ff403dfb68d2d7185888ef7c85e249c28
11:45 kahrl seems to work fine for me, but of course I never had the blurring problem so that's kind of pointless
11:45 kahrl if someone who has that blurring problem could test ^, that would be great
11:48 Zeno` kahrl, refactor_main is updated again :p
11:48 Zeno` kahrl, I'm also thinking of taking limitFps() out of thegame.cpp and putting it into something like utils/gfx.cpp (so it can be shared with main.cpp mainly). Thoughts?
11:51 Zeno` I
11:51 Zeno` I'd test your changes but I've never experienced the issue(s) either
11:51 kahrl Zeno`: yeah it makes sense to share a function like limitFps
11:52 kahrl how does main.cpp do it right now?
11:52 kahrl ah, just sleep_ms(25)
11:54 Zeno` yeah
11:54 Zeno` which is kind of fine but not great
11:55 Zeno` I mean it doesn't hurt, but I'd just like it consistent if that makes sense
11:55 Zeno` I'll split limitFps() out into it's own file some time (in the next week, or three heh)
11:56 kahrl Zeno`: maybe it would fit into drawscene.h
11:56 kahrl dunno though
11:57 Zeno` maybe, I'll keep thinking about it. It does seem strange having a file for just one function but maybe others can be moved into it as well
11:57 Zeno` will have to look more closely at some point
11:58 Zeno` Then again... maybe the whole irrlicht thing can be abstracted
11:59 Zeno` if it doesn't affect performance. If that happened then maybe a port to, I dunno, Ogre3d (just an example!) for the client could become possible
12:00 Zeno` that's very long term though
12:00 kahrl well I doubt that's feasible
12:00 kahrl at the least such an abstraction would be very leaky
12:00 Zeno` I think it could be done
12:00 Zeno` I got told that refactoring the_game() wasn't really feasible as well :P
12:01 kahrl well I don't want to tell you not to try it ;)
12:01 Zeno` Anyway, this is not anywhere near the top of my TODO list hehe
12:01 Zeno` It's way down the bottom and my TODO list is about to overflow
12:01 kahrl things like the formspec menu would have to be essentially rewritten probably
12:02 Zeno` yeah exactly
12:02 Zeno` I didn't say it would be *easy*
12:06 proller joined #minetest-dev
12:12 Zeno` I think #1847 is ok to merge unless there are objections
12:12 ShadowBot https://github.com/minetest/minetest/issues/1847 -- Use native byte-swapping for de/serialization by rafael7
12:13 Zeno` two core devs agreed; I'll merge in 30 minutes or so (assuming there are no further comments)
12:16 guest365 joined #minetest-dev
12:27 Megaf_ Zeno`: go for it
12:37 ImQ009 joined #minetest-dev
13:21 kaeza joined #minetest-dev
13:23 Zeno` what are people's thoughts on a feature freeze until 0.4.11 is released?
13:24 Zeno` i.e. only bug fixes and possibly performance issues addressed
13:25 Zeno` RealBadAngel, #1705    <-- does adding that back to shaders until 0.4.11 is released work for you?
13:25 ShadowBot https://github.com/minetest/minetest/issues/1705 -- Regression in smooth day-night transition
13:28 Zeno` kahrl, https://github.com/kahrl/minetest/commit/d157219cb3ab3949147a9a87de6276341d93707a.patch adds one whitespace error :P
13:29 Zeno` other than that it seems ok to me... but I didn't experience the issue in the first place
13:31 Zeno` it does seem complicated though heh
13:32 Zeno` hmm, there may be a performance regression with it
13:32 Zeno` not sure
13:33 Zeno` fps difference with and without *seems* to be about 10fps on my computer
13:34 Zeno` hard to tell... I'll test with my ram-database later
13:41 Zeno` I guess feature freeze is boring, but surely it has to be done some time
13:42 shadowzone joined #minetest-dev
13:43 Megaf_ +1
13:43 Megaf_ :)
13:46 Megaf_ just merge my PR before freezing things
13:47 Megaf_ If only a core dev agreed with it
14:01 kilbith RBA is doing such things on signlike, so not now IMO
14:01 kilbith and plantlike
14:04 kilbith .11 for xmas, .12 for july, etc. 2 releases/year
14:04 MinetestForFun joined #minetest-dev
14:05 Zeno` possibly
14:05 Zeno` although signlike is a performance improvement AFAIK
14:06 Zeno` but where does it stop?
14:07 kilbith we can raise a major versionning in waiting more
14:07 Zeno` the next release is 0.4.11 though
14:07 Zeno` I'm not sure that should be changed
14:08 Zeno` The more new features get added the more there is to (potentially) debug
14:08 daswort joined #minetest-dev
14:10 Zeno` I guess a related question is "does the current state of HEAD warrant a release"?
14:11 Zeno` (assuming there is a feature freeze for a bit so bugs can be squashed)
14:12 Megaf_ every x.x.n release could be bug fix and every x.n.x release could be new features
14:12 Megaf_ and n.x.x could be major changes
14:13 Megaf_ like new mapgen, new textures, new rendering engine
14:13 kilbith MT 1.x in ~15 years lol
14:14 n4x how are new textures a major change?
14:15 Megaf_ alone it isnt, be is should be included on with another major change
14:15 Megaf_ for a player view can be a major change
14:15 Megaf_ in a player point of view*
14:16 Megaf_ Im not talking about little textures changes.. Im talking about a total change in all textures at once
14:16 daswort joined #minetest-dev
14:17 Zeno` well that then begs the question: "how do minetest_game changes affect the engine versioning?" :P
14:17 kilbith unrelated
14:18 prozacgod joined #minetest-dev
14:19 Megaf_ so, when we say. Minetest version 3.4
14:19 Megaf_ to what are we referring to?
14:19 Zeno` no idea
14:19 Zeno` minetest_game, however, is not the engine
14:20 Zeno` and not part of the engine
14:22 Megaf_ -- Looking for include file endian.h
14:22 Megaf_ -- Looking for include file endian.h - found
14:22 Megaf_ interesting
14:22 Megaf_ and I will be testing this https://github.com/minetest/minetest/pull/1856 in a couple of minutes
14:26 jin_xi Megaf_: this adds a feature, but its not used yet afaict
14:26 jin_xi so you wont see any change
14:27 Zeno` yeah, it's not used yet
14:27 Megaf_ jin_xi: Im happy enough if it doesnt crass the game
14:27 Megaf_ crash*
14:27 jin_xi well, its not used in the game. you can test if it builds.
14:30 Megaf_ I will test some PR by Zeno` too
14:30 Megaf_ mostly tests like if it builds properly and doesnt breaks things
14:30 Zeno` which PR?
14:34 Zeno` :3
14:35 * Zeno` panics
14:35 Hunterz joined #minetest-dev
14:36 VanessaE Megaf_: you don't need to test if it can build, Travis CI does that
14:37 VanessaE but testing if a commit breaks your normal workflow is always a good idea.
14:38 Megaf_ Yep, I do the second too
14:38 Megaf_ Zeno`: #1851
14:38 ShadowBot https://github.com/minetest/minetest/issues/1851 -- Cleanup updateCameraDirection and fix random input not working by Zeno-
14:38 Megaf_ and #1853
14:38 ShadowBot https://github.com/minetest/minetest/issues/1853 -- Add name of node pointed at to debug by Zeno-
14:41 Zeno` oh yeah, that needs to be merged (the first one)
14:42 Megaf_ let me test it first
14:42 Zeno` I can't merge without another core dev approving, but yes testing is most welcome
14:44 Megaf_ Ok, RBA thingy didnt cause any drawback
14:45 Megaf_ actualy it did
14:45 Megaf_ the field of view seems to be messed up now
14:45 Zeno` It cannot have done that... it's not used
14:46 Megaf_ hm
14:46 Megaf_ something looks weird
14:46 Zeno` so if FOV is messed up then it's something else :)
14:46 Megaf_ it noticed that after applying RBAs patch
14:47 * Zeno` looks again
14:47 Zeno` it's not used yet... must be a coincidence
14:48 Zeno` it's just "dead code" waiting to be put to a use
14:49 Megaf_ ok, so it must be something in my head...
14:51 ImQ009 joined #minetest-dev
14:51 Megaf_ Zeno`: Compiling #1851 agora
14:51 ShadowBot https://github.com/minetest/minetest/issues/1851 -- Cleanup updateCameraDirection and fix random input not working by Zeno-
14:52 Megaf_ s/agora/now
14:53 ImQ009 joined #minetest-dev
14:54 selat joined #minetest-dev
15:09 Megaf_ Zeno`: testing 1851 now
15:11 proller joined #minetest-dev
15:12 Megaf_ Zeno`: 1851 have passed Megaf's test
15:14 Zeno` thank thee
15:25 Megaf_ sfan5: did you do a rm CMakeCache.txt and make clean before compiling?
15:25 Megaf_ try a fresh git clone in another folder
15:25 VanessaE ...
15:25 Megaf_ #1858
15:25 ShadowBot https://github.com/minetest/minetest/issues/1858 -- std::bad_alloc
15:25 sfan5 I did a make clean
15:25 sfan5 didn't clean the cmake cache
15:25 Megaf_ I mean, I have no problem at all with that
15:26 Megaf_ sfan5: you have to remove CMakeCache too
15:26 Megaf_ and rerun cmake
15:26 sfan5 why
15:26 sfan5 I re-ran cmake multiple times
15:26 Zeno` meh, just do out of source builds... it's easier
15:27 Zeno` rebuilding with clang to test again
15:27 NakedFury joined #minetest-dev
15:29 Zeno` are you sure that commit is the issue sfan5?
15:29 sfan5 no
15:29 sfan5 thats just the commit im on
15:29 Zeno` I think that needs to be cleared up first ;)
15:30 sfan5 hm
15:30 Zeno` ok with clang and your settings I don't have the issue
15:30 sfan5 can't reproduce with another clone
15:30 sfan5 but there's still the "no worldname given or no game selected" bug when you start minetest for the first time
15:31 ImQ009 joined #minetest-dev
15:31 Zeno` next!
15:31 Zeno` lol ;)
15:32 sfan5 nah
15:32 sfan5 might be because of -march=native
15:33 Zeno` building again
15:34 ImQ009 joined #minetest-dev
15:35 sfan5 yep
15:35 sfan5 crashes
15:35 Megaf_ I have found another bug, related to saving world localy
15:38 Zeno` sfan5, I still can't reproduce
15:38 Zeno` -g -O1 -Wall -march=native
15:38 sfan5 Zeno`: what does "echo | clang -### -E - -march=native" output?
15:39 Zeno` http://pastebin.com/Cci0ZAvE
15:40 Zeno` excuse the world name hehe
15:40 sfan5 >"-target-cpu" "core-avx2"
15:41 sfan5 it's "-target-cpu" "corei7-avx" for me
15:41 Zeno` hmm
15:42 Zeno` Ok, can you check if the commit prior to f7d6509 works?
15:43 sfan5 sure
15:43 Megaf_ Im downloading clang here
15:43 Megaf_ and I can test on an i7 too
15:43 Zeno` If it does then I guess I should revert f7d6509
15:44 Zeno` although I have nfi why it's an issue
15:44 sfan5 Megaf_: I have an i5 here, but the corei7-avx is probably for the whole i[357] series
15:44 Megaf_ isnt it better to fix instead of reverting?
15:44 sfan5 ofc
15:44 Zeno` Megaf_, if the game is not playable it's better to revert until why it's not working is determined (IMO)
15:45 sfan5 but only if we can figure out how to fix it
15:45 Megaf_ sfan5: still, my laptop is an AMD-E1, I believe I will not have the same issue, but I will try on another laptop we have here
15:46 Zeno` I have i7
15:46 sfan5 Zeno`: happens with d406ac9 too
15:46 sfan5 Zeno`: why is your target_cpu core-avx2 then?
15:46 Zeno` ok, well it's not the latest commit
15:46 sfan5 oh wait
15:47 Zeno` I have no idea
15:47 sfan5 you have an i7, those have avx2
15:47 sfan5 Zeno`: try -DCMAKE_CXX_FLAGS=-target-cpu corei7-avx
15:47 sfan5 that should work for you
15:47 sfan5 avx is backwards-compatible IIRC
15:47 Zeno` should work or break?
15:47 sfan5 work as in not a SIGILL
15:48 Zeno` ok one sec
15:48 sfan5 and breaks as in std::bad_alloc
15:48 sfan5 -s
15:48 Megaf_ my CPU does have avx too
15:49 Zeno` umm no I can't even run cmake with that
15:50 sfan5 why not
15:50 Zeno` target does not exist
15:50 sfan5 clang version?
15:50 Megaf_ my target here is "-target-cpu" "btver2"
15:50 Zeno` 3.4
15:50 sfan5 i have 3.5
15:50 sfan5 upgrade
15:50 Megaf_ sfan5: maybe you can use other target that is not your cpu?
15:51 sfan5 you can use any target you want
15:51 sfan5 the question is whether it will run on your current cpu
15:54 Megaf_ how can you see the list of available targets?
15:55 Zeno` 3.4.2 is the latest version I can get
15:55 Zeno` maybe bisect the issue and see where it started?
15:56 Calinou joined #minetest-dev
15:56 Zeno` sfan5 ^^
15:57 sfan5 I'll do that
15:57 Zeno` I'm sorry I can't reproduce
15:57 Zeno` I'll install another distro tomorrow for testing and stuff
15:58 Megaf_ sfan5: can you give me your cmake line please?
15:58 sfan5 Megaf_: http://clang.llvm.org/doxygen/Targets_8cpp_source.html#l01969
15:58 sfan5 Megaf_: CC=clang CXX=clang++ cmake . -DBUILD_SERVER=0 -DENABLE_REDIS=0 -DENABLE_LEVELDB=0 -DENABLE_SOUND=0 -DENABLE_GETTEXT=0 "-DCMAKE_CXX_FLAGS=-D__extern_always_inline=inline -march=native"
16:00 Calinou -DENABLE_LEVELDB=0 and -DENABLE_REDIS=0 are redundant
16:00 hmmmm joined #minetest-dev
16:00 sfan5 wrong
16:00 sfan5 redis is enabled by default
16:00 sfan5 and I actually wanted to write -DENABLE_LEVELDB=1
16:00 Megaf_ Calinou: wrong
16:01 Megaf_ sfan5: maybe thats why it crashes?
16:01 Calinou someone broke smooth lighting in Git?
16:01 sfan5 Megaf_: because leveldb is disabled?
16:01 Calinou why aren't all backends enabled if found?
16:01 Megaf_ if you map is leveldb
16:01 Calinou stop optimizing getTileInfo without testing the game…
16:02 Calinou 2nd time this happens
16:02 Megaf_ sfan5: and to compile with clang I run make?
16:02 sfan5 Megaf_: my map is newly created and sqlite3
16:02 Zeno` I was using a newly created map as well
16:02 sfan5 and it won't std::bad_alloc because it can't find the backend
16:02 sfan5 Megaf_: run cmake the way I did
16:02 sfan5 then run make
16:02 Megaf_ ok
16:03 Megaf_ compiling
16:03 Megaf_ we forgot this  -DCMAKE_CXX_FLAGS=-target-cpu corei7-avx
16:03 Zeno` I created the world, changed map_meta.tx to sfan5's values, deleted the original *.sqlite files and re-launched
16:04 sfan5 Zeno`: the map meta is not relevant
16:04 sfan5 it happens with every v6 world
16:04 Zeno` well I wanted to make sure :)
16:06 Zeno` Megaf_, you were affected by the bug before #1851 ;)
16:06 ShadowBot https://github.com/minetest/minetest/issues/1851 -- Cleanup updateCameraDirection and fix random input not working by Zeno-
16:06 Zeno` add random_input = true to minetest.conf and try before and after the patch
16:07 Calinou https://github.com/minetest/minetest/issues/1859 → tag as bug
16:07 Megaf_ how do I add -target-cpu corei7-avx?
16:08 Megaf_ my cmake is like this at the moment cmake ../../ -DRUN_IN_PLACE=1 -DBUILD_SERVER=0 -DENABLE_LEVELDB=0 -DENABLE_REDIS=0 -DENABLE_FREETYPE=1 "-DCMAKE_CXX_FLAGS=-D__extern_always_inline=inline -march=native"
16:08 Zeno` Calinou, I see no problem. Can you be more descriptive?
16:08 Calinou GRR
16:09 VanessaE a screenshot would be good
16:09 Calinou you're a core developer and not seeing any difference between new and old?
16:10 Calinou OK, one sec
16:10 Zeno` Apparently so
16:10 Calinou https://cdn.mediacru.sh/gMxznO-H5I4i.png
16:10 Calinou broken
16:10 Calinou no ambient occlusion at node edges.
16:10 Calinou but light fades smoothly (see cave)
16:10 Calinou working: https://cdn.mediacru.sh/E_mXDbjD4PkF.png
16:12 CraigyDavi joined #minetest-dev
16:12 VanessaE ah, yeah I see the difference
16:12 Calinou https://github.com/minetest/minetest/issues/1860 → also tag as bug
16:12 Calinou (that's fine for torches, but not non-glowing stuff :P)
16:12 VanessaE Calinou: you probably ought to put those screenshots in the bug report
16:12 Calinou OK
16:12 VanessaE but crop to subject
16:13 VanessaE 1860 is already known
16:13 VanessaE RBA's fixing it
16:14 Megaf_ Calinou: I confirmed your issues, just tested here
16:14 Calinou thanks :)
16:14 Calinou too lazy to crop :( I hope it's not too hard to notice
16:14 Megaf_ Zeno`: sorry to interrupt you, but how do I add the target CPU thing?
16:15 Megaf_ Calinou: I created a Minimal world and even there you can confirm, using v6
16:15 Megaf_ I go under the shadow of a tree and my wield item still at full brightness
16:15 Megaf_ also, theres no smoothing on the "corners" of nodes
16:16 Calinou [by the way: smoothing of node corners was worsened by some commit, but I didn't bother reporting it; it looks okay anyway]
16:16 Calinou try placing a single node in a flat area
16:16 Calinou there is (nearly) no smoothing
16:18 Megaf_ https://cloud.githubusercontent.com/assets/6201512/5145348/e477d398-7199-11e4-8da6-e99d39564e06.png
16:18 VanessaE hm.  I'd say smooth, but more subtle than it used to be; it's hard to tell
16:18 VanessaE I'm not at current HEAD, rather a couple weeks behind
16:19 Zeno` Calinou, does this look better: https://github.com/Zeno-/minetest/tree/test
16:20 VanessaE Calinou: is this what you're seeing?  http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/Screenshot%20-%2011212014%20-%2011%3a21%3a29%20AM.png
16:20 VanessaE (that's a full block, the top just gets "lost" in the background because the shading is the same)
16:21 Calinou yes!
16:21 Calinou exactly
16:22 Calinou was broken by some April-June commit
16:22 Calinou (the one that was a speedup too)
16:23 VanessaE the same thing, with more nodes placed:  http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/Screenshot%20-%2011212014%20-%2011%3a24%3a10%20AM.png
16:23 VanessaE (for reference)
16:23 Calinou I'll post an issue
16:23 VanessaE maybe it's just that the shading is too subtle?
16:24 Calinou there is none at all
16:24 VanessaE no, look again
16:24 VanessaE it's there
16:24 Calinou the way the checks are made make it impossible to have any shading
16:24 VanessaE it's just very subtle
16:24 VanessaE at least in my screenshots
16:24 Calinou or yes, it is too subtle due to the lack of checks
16:25 VanessaE feel free to copy&paste my screens into your bug if you want
16:26 VanessaE RealBadAngel would probably say to forget all lighting bugs because hardware-assisted lighting is coming Real Soon Now :)
16:26 Calinou done
16:26 Calinou yeah… but when will it?
16:26 VanessaE idk
16:28 * Megaf_ still downloading Zeno` branch
16:29 Zeno` I'll revert it tomorrow
16:30 Zeno` 3AM here, goodnight all
16:31 ImQ009 joined #minetest-dev
16:31 Zeno` I won't revert the whole thing though
16:32 Zeno` just that one function
16:32 Megaf_ $ clang --version
16:32 Megaf_ Debian clang version 3.5.0-6 (tags/RELEASE_350/final) (based on LLVM 3.5.0)
16:32 Megaf_ I dont have that target either Zeno`
16:32 Megaf_ good night Zeno`
16:32 Zeno` Thanks for pointing it out Calinou
16:35 Megaf_ Calinou: I'm compiling Zeno` branch to see if he fixed or reverted or whatever that
16:36 Zeno` Megaf_, it didn't
16:37 Megaf_ no?
16:37 Zeno` no
16:37 Krock joined #minetest-dev
16:37 Zeno` I'll revert it and work out how to make it faster another time
16:38 Megaf_ Zeno`: I kinda like the way it is now
16:38 Megaf_ I dont why Calinou is so upset
16:38 Hunterz joined #minetest-dev
16:40 rubenwardy joined #minetest-dev
16:41 Calinou smooth lighting is completely broken
16:41 Calinou we do not break features like this
16:41 Calinou do you developers play the game you develop?
16:41 Megaf_ some dont
16:43 Megaf_ sfan5: I cant confirm the clang issue, version 3.5 here and I dont have that target either... Just use gcc...
16:43 Megaf_ or other target
16:43 sfan5 ¯\_(ツ)_/¯
16:43 Megaf_ maybe its a clang bug...
16:45 Zeno` yes I play the game Calinou
16:45 Zeno` geez
16:45 Megaf_ Whos messing with signs again?
16:45 Megaf_ Every time someone messes with sign I have to remove them all and replace so I can actually read the text on them
16:45 Zeno` I acknowledged your issue and said I'd fix it. What more would you like me to do?
16:47 Calinou talking to Megaf_, RBA, sapier and all that
16:47 VanessaE Megaf_: what sign?
16:47 Zeno` My motivation for that change was to increase performance and I think I've done a pretty good job at that so far
16:47 Zeno` anyway, I am asleep :)
16:48 Megaf_ VanessaE: the one we can write using colors
16:48 VanessaE that's signs_lib and is not related to this channel
16:48 Megaf_ good night Zeno`, sleep well
16:48 Megaf_ VanessaE: ok
16:49 Megaf_ Calinou: I dont develop minetest, but I do play it, lol
16:49 VanessaE besides which, if you can't see the text, either an engine issue is preventing the entities from being displayed, or your entities-per-mapblock limit is set too low
16:49 Calinou Zeno`, then you may also break all shaders
16:49 Calinou that is good for performance
16:50 Megaf_ VanessaE: I will take the second option
16:50 Megaf_ Calinou: As matter of fact it is...
16:50 Calinou when you optimize such things, always test. no excuses
16:50 Calinou it happened once already
16:50 Calinou exact same issue
16:52 Zeno` I tested
16:52 Zeno` Calinou, can you please test https://github.com/Zeno-/minetest/tree/test ?
16:52 Zeno` I've just updated it
16:53 Zeno` I test everything as much as anybody :/
16:53 Zeno` or more
16:53 Zeno` sheesh
16:53 Zeno` I'm not here to ruin the game
16:53 Calinou cloning
16:53 Zeno` and was that exact same issue my fault last time? :(
16:53 Calinou how did you break smooth lighting like that then?
16:53 Calinou no, was sapier's, who did similar stuff
16:53 Megaf_ Calinou: stop complaining to Zeno`, he may be the only one who test this and the one who is working more now
16:54 Zeno` I didn't notice
16:54 Zeno` I tested nearly everything. I did not notice the subtle difference in rendering
16:54 Zeno` I can see it now though
16:54 mrtux joined #minetest-dev
16:54 Megaf_ Zeno`: Im compiling your new commit on test branch
16:55 Calinou subtle? :/
16:55 Megaf_ Calinou: Yes, subtle
16:55 Calinou in theory it's subtle, in practice it's ultra obvious
16:55 Megaf_ VanessaE said that, Zeno` said that and now Im saying that too
16:55 Calinou except on a flat world
16:55 Megaf_ Zeno`: that fixes the smooth ligthing
16:56 Megaf_ Zeno`: wield item still at full brightness
16:56 VanessaE the wield item brighness issue is unrelated
16:56 Megaf_ maybe it doesnt fix, hard to tell
16:56 Calinou your fork does not work
16:57 Calinou in fact, smoothing works only if there is air below the smoothed nodes
16:57 Calinou _ X X _
16:57 Calinou X _ _ X
16:57 Calinou X = node, _ = air (side view)
16:57 Calinou then you will have smoothing…
16:57 VanessaE lemme try that
16:58 Megaf_ there is smoothing everywhere
16:58 Megaf_ but at the corners
16:58 VanessaE that screenshot is solid stacks of sandstone, I didn't think to make the stacks hollow
16:58 Zeno` Ok, well I'll put the slow code back
16:59 Zeno` problem solved
16:59 Zeno` wait air below it?
16:59 Zeno` I just tested with sand below the nodes
16:59 Calinou no one has managed to make it faster without breaking stuff
17:00 Calinou yes, I'd like improvements to be able to use view range 150-200 with good FPS, but not if it breaks smooth lighting
17:00 Zeno` I have
17:00 Calinou https://cdn.mediacru.sh/83Sz-kj9OCnl.png
17:00 Calinou how it looks with your build ^
17:00 Calinou clearly lacks smoothing
17:00 VanessaE wait, I see what calinou's saying, Zeno`
17:00 Zeno` ignoring that commit I have already increased FPS a lot
17:00 Zeno` VanessaE, yeah, I will revert it
17:01 VanessaE nono
17:01 Zeno` are you sure that's on the test branch?
17:01 VanessaE there's two issues in play here, I'm talking about the one above with the _ X X _ / X _ _ X  diagram
17:01 Zeno` because that is not what I am seeing
17:01 VanessaE that predates your commits, I think
17:02 VanessaE I can confirm the effect he's talking about, and I'm pretty sure you did NOT break it
17:02 Megaf_ why dont we all use the same mapgen and seed and go to the same place?
17:02 Calinou oh, branch
17:02 Calinou I forgot
17:02 Zeno` on my test branch I see shading in the 45 degree angles
17:02 Megaf_ lets say seed shader, mapgen v6, spawn position
17:02 Calinou I'll test performance too
17:03 VanessaE Zeno`: go in-game and build this 8-node structure:  http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/Screenshot%20-%2011212014%20-%2012%3a04%3a05%20PM.png
17:03 Calinou sorry for forgetting about branch
17:03 Calinou and in general for being angry
17:03 Calinou (but I'm extremely annoyed when people break things and don't notice it)
17:03 Zeno` so it's better now?
17:03 Calinou I'm compiling
17:03 Calinou wait
17:03 shadowzone What is that for?
17:03 VanessaE then place a block underneath and watch how the shadow changes on the tops of the two blocks at either end
17:03 Calinou your branch works
17:04 Calinou individual nodes are slightly smoothed
17:04 Zeno` ok, I will use the test branch tomorrow. Thanks for the feedback :) If someone (or you) finds something else wrong I'll just revert the changes completely. I really don't want to break stuff
17:05 Calinou now who volunteers to fix the crap light table? :P
17:05 CraigyDavi joined #minetest-dev
17:06 Calinou far from extraordinary: 90-110…
17:06 VanessaE Calinou: why can't we just use the one you proposed like 6 months ago? :P
17:06 Calinou are you sure you didn't worsen performance?
17:06 Calinou VanessaE, c55 doesn't want it due to reasons like gamma and stuff like that
17:06 Zeno` actually it's crap
17:06 Zeno` I'll revert it hehe
17:06 Megaf_ VanessaE: Calinou Zeno`
17:06 Megaf_ https://cdn.mediacru.sh/z7mTIBwWcCJO.png
17:07 Megaf_ Thats on the test branch
17:07 Calinou performance is more or less the same
17:07 Calinou 140 FPS on both, same place
17:08 Zeno` night looks horrible when placing a torch
17:08 Zeno` I think it's best that I just take those changes out completely
17:08 Calinou maybe I gained 3 FPS
17:08 Calinou how does night look different?
17:08 Zeno` oh there is an increase, but if it doesn't look right it's broken
17:08 Calinou this doesn't affect the light table
17:08 Megaf_ Zeno`: isnt it fixed now?
17:08 Zeno` Megaf_, not at night apparently
17:09 Zeno` the light torches emit look diamond shaped
17:09 Zeno` I'll work on this tomorrow
17:09 Calinou [carbone torches emit light level 14 instead of 13 in minetest_game] but notice the dots, they are due to the light table, not due to your changes
17:09 Calinou uploading screenshot
17:09 Calinou this isn't due to your changes
17:10 Calinou https://cdn.mediacru.sh/5Z0astRCCwHg.png
17:10 Megaf_ Zeno`: at night https://cdn.mediacru.sh/2RUD50QVWOdL.png
17:10 Calinou you just need to tweak the light table to be smoother
17:10 Zeno` hmm
17:11 Zeno` the lights look good at ground level but I'm not sure they look very good from afar
17:11 Megaf_ hold on
17:11 Calinou this is truly a light table problem
17:12 Calinou https://github.com/minetest/minetest/issues/1362
17:12 Zeno` I'll tweak the light table to see if I can make it nicer
17:12 Calinou my table is linear, so not necessarily a good idea
17:12 Calinou personally I use the debugging-for-dark table as my screen makes stuff too dark
17:12 Calinou works great
17:12 Calinou see light.cpp for examples
17:13 Megaf_ https://cdn.mediacru.sh/Bli3kLk4W8Br.png
17:13 Megaf_ from distance
17:13 Megaf_ looks like the screenshot from Calinou
17:13 Megaf_ thats not due to your changes Zeno`
17:13 Megaf_ thats just the it is
17:13 Calinou since 0.4_20120318
17:13 Calinou so, long time ago
17:13 Calinou 0.3 had a better table overall
17:13 Zeno` the transition is too sharp
17:14 Megaf_ Calinou: whats wrong with linear light table?
17:15 Calinou it looks diamond-y from far away
17:15 Calinou it doesn't make sense
17:15 Calinou light is not linear in real life
17:15 Zeno` I don't like that diamond look
17:16 Calinou so, keep your changes, they give me 3-4 FPS
17:16 Calinou I thought it was a large improvement
17:19 Zeno` :)
17:20 Calinou I'd like a big bump, like Minecraft recently saw :(
17:21 Calinou view range 256 (equivalent to Minecraft's Far) is still a dream
17:21 Calinou (especially if you want server to send that)
17:21 Zeno` ok this is test branch and a build from a few weeks ago: http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/Screenshot%20-%2011212014%20-%2012%3a19%3a44%20PM.png
17:21 VanessaE master upstream on left, zeno's test branch on the right ^^^^^^^^^
17:22 Zeno` so I think tomorrow I'll work on the night "diamond-y" thing tomorrow
17:22 Zeno` I think I can make that better
17:22 VanessaE as for diamond-y light, it's always looked diamond-y to me
17:22 VanessaE since, forever
17:22 Zeno` I'll still try :)
17:22 * Zeno` is paranoid now
17:23 Zeno` 3:30AM I am not going to be able to watch ALF :(
17:23 Krock Zeno`, I like the left one more
17:23 Calinou a light table change will probably need several core devs to agree
17:23 VanessaE http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/Screenshot%20-%2011212014%20-%2012%3a25%3a05%20PM.png
17:23 Calinou this is big stuff
17:23 VanessaE ^^^^ upstream
17:24 Calinou how much FPS in each?
17:24 Calinou /set -n fps_max 0
17:24 Calinou to temporarily disable FPS limit
17:25 Zeno` Oh, I can show you profiles to prove it's faster now. As long as I can get it to look acceptable it's certainly faster
17:25 Calinou profile doesn't mean it is necessarily faster in practice :P
17:25 Megaf_ [17:22] <VanessaE> as for diamond-y light, it's always looked diamond-y to me
17:25 Megaf_ I agree
17:26 Zeno` Calinou, yes, but I also look at sleep time :)
17:26 sapier joined #minetest-dev
17:26 VanessaE sapier!
17:26 Megaf_ but, I mean, everyone on these games needs to be squares
17:26 Calinou drawtime is another solution
17:26 sapier hello
17:26 Calinou hi sapier, you are welcome
17:26 Megaf_ so, whats wrong with that shape?
17:26 Zeno` and my test world is the same every time (RAM database)
17:26 Zeno` So I'm not comparing apples to oranges
17:27 VanessaE Megaf_: light propagates in straight lines, so absent an obstruction, it should be spherical, not diamond.
17:27 Megaf_ That in real life, but we are in a voxel world
17:27 Calinou it is impossible for it to be really spherical
17:27 Calinou hardware lighting will fix this
17:27 Calinou it can try looking spherical using a non-linear table
17:27 VanessaE Calinou: sure, but it's possible to calculate an approximation
17:28 Calinou eg. use 85 % of each light level
17:28 sapier "VanessaE: Megaf_: light propagates in straight lines," << only exactly true in absence of gravity
17:28 VanessaE Megaf_: yeah but we have many non-voxel things like fences, torches, etc.
17:28 Calinou each lower light level uses 85 % of the previous one
17:28 Calinou or 80 % (that's what Minecraft did for a while)
17:28 VanessaE sapier: well let's assume Earth-like gravity where light effectively travels in straight lines.  we're not in range of any supermassive objects here ;)
17:28 Calinou or 82 % (a middle ground)
17:29 VanessaE sapier: besides, gravity would count as an "obstruction" ;)
17:29 khonkhortisan joined #minetest-dev
17:30 Zeno` light at night is brighter now but I think that may be how it used to be
17:30 sapier VanessaE: are you sure? ;-) gravity doesn't modify light but the room light travels through
17:30 VanessaE sapier: ok now you're just splitting hairs ;)
17:30 Zeno` I remember having to change all my mods to increase node brightness about 4 months ago
17:30 sapier true ;)
17:30 Zeno` sapier!
17:30 * VanessaE dives out of zeno's way
17:31 Zeno` lol
17:31 khonkhortisan and slowly doth the two great lights become one
17:31 Krock e:\programme\mt_compiler\minetest\src\util/serialize.h(25): fatal error C1017:Unknown const of a integer
17:32 Krock Current GIT is broken for MSVC
17:32 sapier OldCoder did the patch I sent you quite some time do same thing as your variant of fixing same error?
17:32 Calinou I tweak light table if I need bright stuff :P
17:32 sapier hello zeno how are you?
17:33 Megaf_ Hi sapier, how are things? Have you been busy?
17:33 Krock okay. the error seems to be a cmake fault
17:34 Megaf_ Krock: sfan5 have a problem with that commit too
17:34 sapier yes extremely busy ... ppl always want things to be ready prior christmas ... all at once
17:34 Krock Megaf_, mhm okay
17:34 Zeno` Megaf_, I don't think sfan5's issue was related because the commit before that didn't work for them either
17:34 * Krock slaps Zeno`
17:35 Megaf_ Krock: https://github.com/minetest/minetest/issues/1858
17:35 Megaf_ Zeno`: there were several commits in a short period
17:35 Megaf_ perhaps we have to slow down a little
17:36 Zeno` and none of those commits would cause that issue
17:36 Zeno` still waiting on a bisect
17:37 VanessaE sapier: speaking of prior-to-christmas, we have an initial call for feature freeze :)
17:37 VanessaE but no one's agreed yet
17:37 Krock gonna run cmake config and compile everything from scratch
17:37 Calinou 0.4.11 was supposed to be released on November
17:37 Calinou so a feature freeze today or tomorrow makes sense
17:37 VanessaE and I guess xmas is now the target for 0.4.11 isn't it?
17:37 Calinou aim for release on late November, I say
17:37 Calinou don't wait too much
17:37 Zeno` I suggested a feature freeze a few hours ago
17:38 Zeno` anyway, sapier #1826
17:38 ShadowBot https://github.com/minetest/minetest/issues/1826 -- Fix for various Android build errors. Enable landscape rotation. by KodexKy
17:39 sapier zeno already looking at it ;-)
17:39 Zeno` hehe cool
17:39 Zeno` night all (again)
17:40 Zeno` geez, the sun is coming up
17:41 ShadowNinja I think 0.4.10 os too soon for #1409, move to 0.5?
17:41 ShadowBot https://github.com/minetest/minetest/issues/1409 -- MIDI music support
17:42 Calinou 0.5 will never be out at this pace…
17:42 Calinou this is for 0.5 anyway
17:42 Calinou and quite low-priority
17:43 Calinou low-quality OGG does wonders :P
17:43 Calinou (or support Opus? :D)
17:43 VanessaE ShadowNinja: feature freeze has been called for.
17:43 Calinou would be cool to have.
17:43 VanessaE (having MIDI or module music would be nice, but not for 0.4.11)
17:44 ShadowNinja Also: Sokomine's item move out of read-only bug, and signals API can be delayed.
17:44 ShadowNinja VanessaE: But not agreed to yet.  ;-)
17:44 VanessaE ShadowNinja: agreements/disagreements are requested :)
17:45 Calinou however… prediction for node inventories?
17:45 Calinou this is urgent
17:45 Calinou gameplay suffers a lot from this
17:45 ShadowNinja Hmmm...
17:45 VanessaE the move-out-of-read-oly bug needs fixed before 0.4.11
17:45 Calinou there was a patch, which had occasional false positives but otherwise worked
17:45 VanessaE that's persisted for way too long.
17:45 VanessaE only*
17:45 ShadowNinja Calinou: No, not urgent, just annoying.
17:46 VanessaE er... when it comes to gameplay, annoying == urgent
17:46 ShadowNinja Anyway, we don't want to do that now.
17:46 VanessaE *never* annoy your users.
17:46 VanessaE but whatever
17:46 ShadowNinja VanessaE: I classify urgent as game/build-breaking.
17:48 Krock Okay, could someone help me to solve this error: "src\util/serialize.h(25): fatal error C1017: Invalid expression
17:48 Krock for integer constant" quickly?
17:49 VanessaE Krock: git reset to commit d406ac994b8092c5bd2dc32eda1a2eafbf95a30c and see if that works
17:49 selat joined #minetest-dev
17:50 Krock compiling
17:51 Krock works
17:51 Krock client.cpp compiled finw
17:51 Krock ^ VanessaE
17:52 Zeno` ok it needs to be reverted
17:52 Krock but maybe it's also a fault of cmake
17:53 Megaf_ you still here?
17:53 Megaf_ Krock: whats your CPU?
17:53 Krock Megaf_, AMD Athlon 4000+
17:54 Megaf_ oh, ok
17:54 Zeno` what line does the error occur on?
17:54 Krock "CMAKE_HAVE_ENDIAN_H:BOOL=OFF" is in CMakeCache.txt
17:54 Krock rror: "src\util/serialize.h(25): fatal error C1017: Invalid expression
17:54 ShadowNinja Zeno`: The android patch seems fine BTW.
17:54 Megaf_ Krock: if change that to ON?
17:55 ShadowNinja It's not optimal, but it's better than before.
17:56 Krock same error, Megaf_
17:56 Zeno` change line 25 to #ifdef HAVE_ENDIAN_H and see what happens?
17:57 Krock :\programme\mt_compiler\minetest\src\util/serialize.h(26): fatal error C1083: Datei (Include) kann
17:57 Krock nicht geöffnet werden: "endian.h": No such file or directory [E:\Programme\MT_compiler\bin\src\min
17:57 Krock edian not found
17:57 Krock need to set to false now
17:58 Krock okay.
17:58 Krock that thing is defined but it keeps throwing errors
17:58 Krock (same as latest posted error)
17:59 cg72 joined #minetest-dev
18:00 Megaf_ Zeno`: I'm afraid that commit only works on Linux
18:00 Megaf_ does windows have this .h files?
18:00 Megaf_ these*
18:00 Krock removing those https://github.com/minetest/minetest/commit/f7d65091f83abe1cb1a70d6823291df8accbe6ab#diff-ebab3775208ceccb934d0d29f384a367R77 both lines helped
18:01 Krock I also added an "#ifndef HAVE_ENDIAN_H 0 #define HAVE_ENDIAN_H 0 #endif" on top
18:01 Zeno` did you remove line 39 as well?
18:02 Krock I added an #ifndef around it
18:02 ImQ009 joined #minetest-dev
18:02 Zeno` might be able to remove it. Check with ShadowNinja :)
18:02 Zeno` I'm asleep :)
18:02 Krock ok gn.
18:03 gravgun joined #minetest-dev
18:03 Megaf_ Zeno`: is it for real now?
18:04 Megaf_ :)
18:04 gravgun Too shy of answering :3
18:07 RealBadAngel i just got new keyboard, damn it looks nice: http://a4tech.com/images/products/Accessary/KD-800L_410X220.jpg
18:07 RealBadAngel irl it looks even better :)
18:17 PenguinDad joined #minetest-dev
18:25 FR^2 joined #minetest-dev
18:27 shadowzone joined #minetest-dev
18:43 shadowzone joined #minetest-dev
18:43 GrimKriegor joined #minetest-dev
18:48 Megaf_ Calinou: Here we can see your issue with smooth light very clearly https://cloud.githubusercontent.com/assets/6201512/5147471/bceea874-71ad-11e4-9c94-418cede76284.png
18:50 Megaf_ RealBadAngel: is the sun a light source?
18:50 Megaf_ I mean, does the light actually comes from the sun on minetest? I don't think it does
18:51 Calinou sun is a light source, emits a reserved light level (15)
18:51 Megaf_ it looks like there is a gigantig torch always at 0 degrees
18:51 Megaf_ shadows dont move
18:51 Calinou on the two screenshots, why don't the trees generate at the same place?
18:51 Calinou worlds should always generate the same way with the same seed…
18:51 Calinou all light is fake…
18:52 Megaf_ Calinou: no idea, I was asking myself the same thing
18:52 Megaf_ I compiled both exatectly the same way, same world name, same seed
18:53 Megaf_ they are 8 commits appart
18:54 RealBadAngel Megaf_, shaders are using sun for calculations
18:54 RealBadAngel atm it is located always 100 nodes above the player
18:55 RealBadAngel temporary solution, it just waits for real sun in game
18:56 RealBadAngel the moment we will switch to irrlicht lighting, shaders will use them with almost no changes at all
19:00 Megaf_ Im thinking about redoing this and sending a pull request, https://github.com/minetest/minetest/pull/844/files
19:00 Megaf_ anyone agrees with that?
19:01 RealBadAngel partially i do
19:01 shadowzone joined #minetest-dev
19:01 RealBadAngel the setting is needed but i think that we need new menu design
19:02 RealBadAngel simply theres no place for more settings
19:04 RealBadAngel and we shall redo them the way c55 proposed: simple menu with single/multi/settings/whatever and then open right formspecs, designed for tasks
19:04 Calinou I agree ^
19:04 Calinou also, move credits to a text file?
19:04 Calinou (to save up one tab)
19:05 RealBadAngel i see it as 4-5 lines navigated by up/down/enter (or mouse)
19:05 RealBadAngel the main menu ofc
19:05 RealBadAngel credits can be put there too
19:06 RealBadAngel with such layout you can make the memu nice, with big fonts
19:06 Megaf_ and some commit have actually messes with field of view
19:06 Calinou we could have a quit button too, for fullscreen players
19:06 Megaf_ I get distortions now
19:06 Calinou in the main-menu
19:06 proller joined #minetest-dev
19:07 RealBadAngel i asked google for "main menu layout"
19:07 RealBadAngel http://media.moddb.com/images/games/1/16/15864/UI_menu_mainMenu_guide.png
19:08 RealBadAngel i was thinking about something like that
19:08 Calinou that is mediocre
19:08 Calinou this isn't an action game by the way :P
19:08 shadowzone joined #minetest-dev
19:08 RealBadAngel ofc not, but main menu style
19:09 RealBadAngel or rather funcionality. simple. just select what you need\
19:10 RealBadAngel tabs are functional but looks like from 20 yrs ago
19:11 RealBadAngel and are overloaded with settings
19:11 RealBadAngel this doesnt look either nice nor friendly
19:12 RealBadAngel also the menu is unusable on android devices
19:13 RealBadAngel looks like shit and works the same way :(
19:13 Calinou too many outlines, borders is the problem
19:13 Calinou flat design is the new cool thing ;)
19:13 Calinou having a lot of settings is fine, if they are well-organized
19:14 Calinou (and don't add settings that don't change much/anything)
19:15 Megaf_ maybe a simple single window thing, in portrait would be ideal
19:15 Megaf_ a scrolling window
19:15 Megaf_ divided in areas
19:21 GrimKriegor joined #minetest-dev
19:21 n4x joined #minetest-dev
19:26 kilbith what about a HTML/CSS mainmenu ?
19:26 kilbith (idea from ShadowNinja)
19:26 jin_xi joined #minetest-dev
19:28 gravgun joined #minetest-dev
19:29 kaeza and JS for modding!
19:30 rubenwardy I would love HTML and CSS instead of formspecs
19:30 rubenwardy Not full HTML and CSS, but inspired by it
19:30 rubenwardy very similar elements but a few custom ones
19:30 VanessaE kaeza, rubenwardy:  channeling lamefun? ;)
19:33 Megaf_ where does form spec comes from? Is it something from Lua?
19:34 Calinou libRocket makes it possible to do that, kilbith
19:34 Calinou the CPU usage is high though
19:34 Calinou both in Warsow and Unvanquished…
19:34 Calinou which implement it (quite well)
19:34 Calinou the Unvanquished HUD is libRocket ;)
19:34 kilbith but that ought very flexible
19:34 ShadowNinja HTML is big, but it implements just about every feature you could want.
19:34 Calinou Megaf_, yes, and the C++ code
19:34 Calinou http://librocket.com/
19:34 rubenwardy HTML and Django style templates. Will probably be quite large.
19:35 PenguinDad libRocket causes really high cpu usage at high framerates for me
19:36 kaeza what I'd like instead of HTML is some kind of packing/layout manager
19:38 jin_xi kaeza: its ok for generated gui but hard to get to look right in a game i think
19:38 Calinou using something like Tesseract or Sauerbraten's UI?
19:38 Calinou http://tesseract.gg/ look at menu code on SVN web view
19:42 GrimKriegor joined #minetest-dev
19:49 asl joined #minetest-dev
19:52 proller joined #minetest-dev
19:56 lukevph joined #minetest-dev
19:56 lukevph left #minetest-dev
19:58 proller joined #minetest-dev
19:59 Krock I think gsmapper needs a rewrite which only updates the complete minimap every 2 seconds
20:00 PenguinDad and changes the license
20:00 Krock I don't get th logic of the current draw-style
20:02 casimir joined #minetest-dev
20:03 Calinou every 5 seconds is probably enough, or even 10
20:03 Calinou I'd like a round/rotatable map, so it'd have to be entirely redone
20:03 Calinou à la Rei's minimap
20:03 kaeza PenguinDad++
20:07 paramat joined #minetest-dev
20:10 paramat concerning feature-freeze, hmmmm usually organises that. personally i would like at least 2 weeks before freeze to finish mgv5 default biomes (with new default snow biomes and pines)
20:15 paramat sfan5 please could you consider https://github.com/minetest/minetest_game/pull/336 ? see last comment i would like to use the pines from mg, is that okay? but i need 336 merged very soon to give me the nodes to make the pine schematic
20:15 ImQ009 joined #minetest-dev
20:16 sfan5 paramat: why a schmatic?
20:16 sfan5 the pines in mg are generated using randomness
20:17 sfan5 s/randomness/lua/
20:17 paramat yep i know
20:17 sfan5 (note to self: think before you type
20:17 sfan5 )
20:17 paramat because mgv5 uses the biome API and schematics for trees
20:17 paramat although i would prefer to grow them by random code
20:20 paramat so yeah i need to make tree schems for appletree and jungletree too
20:21 paramat although the lua sapling-growing code in default will alow me to generate sapling pines exactly as in mg
20:22 Krock This solved the compiling bugs where HAVE_ENDIAN_H was not defined as an constant integer http://pastebin.com/5gewXYjK
20:22 Krock actually, it just turns off the feature in the source codes
20:23 paramat i feel a release for winter solstice (21st Dec) is good and realistic, people wil be busy the few days before xmas
20:23 paramat *will
20:24 PenguinDad Krock: you shouldn't solve bugs by simply disabling the feature
20:25 paramat a lot seems broken so an immediate feature freeze seems too soon
20:27 proller joined #minetest-dev
20:29 paramat ive waited 2 weeks for Blockmen's input on MTG336, perhaps he's busy or a taking a break from MT, so i need Nore and sfan5 to merge 336
20:39 Krock PenguinDad, I know
20:42 Megaf_ And I have just messed up my pull request
20:42 Megaf_ https://github.com/minetest/minetest/pull/1854
20:46 * Megaf_ gives up
20:47 ImQ009 joined #minetest-dev
20:47 Krock Megaf_, it's so easy with tortoise git :P
20:47 Megaf_ I just wanted to merge to minetest master...
20:48 Megaf_ then I squashed everything to a single commit
20:48 Megaf_ and fucked up things =/
20:48 Megaf_ and besides that, no one wants my pull request anyway...
20:49 Krock exactly
20:55 proller joined #minetest-dev
21:03 OldCoder sapier, your patch is incomplete
21:04 OldCoder sapier, please PM me your email address. The address I was given was apparently incorrect.
21:04 OldCoder sapier, I have a new combined patch that fixes the problem entirely. PM me email address and I will send it.
21:04 OldCoder I am away today; traveling to Silicon Valley o/
21:04 OldCoder
21:05 OldCoder If somebody else sees this, please PM me his email address, the patch is important
21:09 NakedFury joined #minetest-dev
21:15 zat joined #minetest-dev
21:16 asl left #minetest-dev
21:20 RealBadAngel paramat, if you want random trees, use l-system ones, not schematics
21:22 RealBadAngel http://en.wikipedia.org/wiki/L-system
21:23 RealBadAngel you got the most powerful tool to model the plants and you still want schematics. thats a bit weird, do you realize that?
21:24 kilbith take a look into MoreTrees, though
21:24 RealBadAngel those are just examples
21:24 RealBadAngel imagine what l-systems could do with micro blocks
21:26 paramat RBA try to convince hmmmm instead =) im just following orders
21:27 RealBadAngel best way to convince here is to show working code
21:27 RealBadAngel and a beautiful landscape
21:27 VanessaE "I was just following orders!" is never an excuse, paramat  ;)
21:28 paramat heh agreed VE
21:28 RealBadAngel lol indeed
21:29 RealBadAngel paramat, i wish i could see mountains high, big trees and world proportional
21:29 RealBadAngel and nice shaped
21:29 paramat i guess schems are faster and that's the most important thing
21:30 RealBadAngel schems are for villages and the like
21:30 paramat my mapgens can be scaled up to unlimited sizes
21:30 RealBadAngel not for the flora for christ sake
21:30 paramat trees need to be proportional to the terrain, i find moretrees too big for mgv6
21:30 RealBadAngel make proper landscape then
21:31 paramat indeed (watershed)
21:31 RealBadAngel atm we do have human beings nearly same sized as trees
21:31 RealBadAngel do you think thats realistic?
21:31 VanessaE well to be fair,
21:31 VanessaE the default trees are more like 2-3x the height of a human in the game
21:31 VanessaE and in the real world, there are many kinds of trees that are that height.
21:32 VanessaE *looks out the window*
21:32 paramat mgv5 is small scaled too, needs small trees, also to avoid getting chopped during mapgen trees need to be <16 nodes high
21:32 VanessaE I see a few of them right here in the yard around me
21:32 RealBadAngel oh cmon
21:33 VanessaE so in the case of moretrees, the terrain just needs to scale up to match
21:33 VanessaE the trees aren't too big - the terrain is too flat.
21:33 RealBadAngel http://www.visitsequoia.com/img/3_1_1_The_Giant_Forest.jpg
21:33 VanessaE RealBadAngel: exactly.
21:34 RealBadAngel http://www.cel-mai.ro/imagini/sequoia-m/sequoia.JPG
21:34 paramat well i hope to add a large scaled mapgen, so you can use that
21:34 RealBadAngel exactly, large scaled one
21:35 RealBadAngel with big lakes, trees, mountains and all that stuff
21:35 RealBadAngel not just funny puddles with some bonsai
21:35 VanessaE wouldn't it be simpler to offer an mgv6 option with large-scaled map_meta.txt defaults?
21:36 VanessaE (or mgv7, either way)
21:37 paramat i'll tr to fix the mgv6/mgv7 discontinuities at y = 47 etc so you can do that
21:37 paramat ^ now fixed in mgv5
21:39 RealBadAngel paramat, have you saw terasology landscape with those trees?
21:41 paramat yes and investigated their mapgen
21:41 VanessaE paramat: there may be another issue.  Isn't mgv6 supposed to be continuous all the way back to pretty much near the beginning of the 0.4.x series when it was introduced?
21:41 VanessaE I mean, how the default settings are interpreted that is
21:41 RealBadAngel http://blog.movingblocks.net/wp-content/uploads/Terasology-130707004254-3840x2160-1800827_960x332.jpg
21:41 paramat you mean not breaking old worlds?
21:42 RealBadAngel please keep in mind one thing: single l-system tree will be slower than a bonsai or schematic one
21:42 paramat i can make terrain like that, already have
21:42 VanessaE paramat: yeah, more or less - I mean like bring in an old world, load it, go generate some new terrain and it just blends right in with the old terrain, without mapblock edges
21:42 VanessaE assuming the same seed as before
21:42 RealBadAngel but you dont need so much of big trees
21:43 RealBadAngel a few around are better than 100 hundred of small ones
21:43 paramat yes i will have to make sure old world's don't break
21:43 RealBadAngel and then you can start to think if a few could be done faster than 100
21:44 VanessaE paramat: I ask because I'm talking with OldCoder elsewhere about it.  unless an error was made in importing the old world data or the settings thereof, it seems that this is no longer the case.
21:44 paramat l systems is something i need to play with and time etc
21:44 VanessaE paramat: you are aware of the efforts to bring Redcrab's server up-to-date (with the result being called Silvercrab), yes?
21:45 paramat Vanessa, old worlds being broken is bad news, please link me to the discussion
21:45 paramat ehm sort of
21:45 RealBadAngel paramat, in the future i do plan (after current work is done) to extend, or rather shrunk l-systems
21:46 RealBadAngel and let it generate grass, bushes etc
21:46 OldCoder VanessaE, I must leave but I will put up the world that resulted; this will explain what I mean
21:46 OldCoder o/
21:46 VanessaE paramat: can't link to it (#minetest-staff, so it's not publicly-logged) but the short version is that efforts to recover the map have been successful, but efforts to ensure that the map blends seamlessly into modern iterations of mgv6 result in "chaos" which I take it means blocks being....rather poorly matched.  OldCoder will have to fill you in on the rest.
21:47 OldCoder I mean, I will put the resulting world online soon for your review
21:47 OldCoder Probably tonight or tomorrow
21:47 OldCoder VanessaE, paramat ^
21:48 VanessaE OldCoder: one quick question:  are you dead certain that the spawn building is exactly in the same place in the map as it was in the original world, and that the seed I gave you was correctly transcribed?
21:52 VanessaE in fact I'll just check the former myself :)
21:53 VanessaE yep, it's correct.
21:54 paramat okay, how old is the redcrab world? perhaps it was started before hmmmmm decided to keep mgv6 consistent
21:55 shadowzone joined #minetest-dev
21:55 OldCoder paramat, VanessaE I will boot Chaos World tonight or tomorrow for review
21:56 paramat personally i prefer coded trees (as in mgv6) to schems or l-systems, because they can have any generation method, therefore more character
21:56 paramat oldcoder okay
21:57 VanessaE OldCoder:  --> msg
22:11 Sokomine RealBadAngel: hmmm seems to see the usage of schems for villages as almost abusive. afaik he intended it to be used for trees, special hills, stalagmites/stalagtites and the like
22:13 blaise joined #minetest-dev
22:13 Sokomine VanessaE: with large trees, it's important to be able to sustain a sufficiently high view range so that the trees can be seen
22:14 paramat to be clear schems do have randomness, but this is set per node
22:15 Sokomine RealBadAngel: small trees are better for gameplay. a 5-nodes-high tree trunk may in most cases be cut down by a lazy player. higher trees - they get cut up partly by players
22:17 paramat Sokomine i'm very happy you managed to get 'force place' added to schems, essential for my own (ab)use of schems for generative architecture etc
22:18 paramat to be artistic or creative abuse (non-standard use) is essential =)
22:18 Sokomine i'm not sure if anyone has ever actually used randomness for schematics so far?
22:19 Sokomine paramat: yes :-) that force parameter kind of saved the ethereal mod/mapgen and the xanandu server. else they would have had no real option to re-grow their trees (they're all schematics)
22:27 Fritigern Sokomine: Any chance that the markers mod will get updated any time soon?
22:28 RavonTUS joined #minetest-dev
22:32 RavonTUS_ joined #minetest-dev
22:34 zat2 joined #minetest-dev
22:37 proller joined #minetest-dev
22:47 proller joined #minetest-dev
23:03 blaise joined #minetest-dev
23:10 DFeniks joined #minetest-dev
23:17 khonkhortisan joined #minetest-dev
23:27 Sokomine it would be really great if commit 4896d4b where optional and could be disabled so that players can again place nodes where they stand. this is very important for any people who use the game for building (some do!). i regulary fix that in my local version
23:28 PenguinDad Sokomine: what's so bad about https://github.com/minetest/minetest/commit/4896d4b?
23:30 shadowzone joined #minetest-dev
23:32 VanessaE I think she means 1b4908b :)
23:32 VanessaE https://github.com/minetest/minetest/issues/85
23:35 paramat what was the argument against an option for this? can it be done in creative? i guess you want it in non-creative mode too
23:36 paramat placing signlike drawtypes where you stand seems allowable
23:37 PenguinDad Just removing the client.checkPrivilege("noclip") would be an acceptable solution imo
23:38 DuDraig joined #minetest-dev
23:38 paramat i consider it important to keep sokomine happy =D
23:40 proller joined #minetest-dev
23:41 paramat i'd code it myself if i knew how, hope it happens
23:44 DFeniks i probably shouldnt talk about this because i know nothing about it. but maybe player coordinates can be used to see if he is inside cube , not radius or something ?
23:45 DFeniks because cubes are not round
23:45 PenguinDad DFeniks: that wouldn't change anything
23:46 DFeniks well, you know better
23:47 shadowzone joined #minetest-dev

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