Minetest logo

IRC log for #minetest-dev, 2016-01-15

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

All times shown according to UTC.

Time Nick Message
00:02 Skyduskguy anyone care to brainstorm how i should generate these nodes with smoothhills? i know i should have abms spammed everywhere
00:07 sofar don't use ABM's
00:07 sofar https://github.com/minetest/minetes​t/blob/master/doc/lua_api.txt#L1831
00:07 sofar do it at mapgen time instead
00:07 kahrl RealBadAngel: I don't see what's wrong with the function that you linked
00:08 RealBadAngel kahrl, do you remember those wrong vertices i posted?
00:08 kahrl RealBadAngel: yeah
00:08 RealBadAngel thats the reason
00:08 kahrl RealBadAngel: the one you mentioned earlier, cloneMesh, seems wrong to me, but not scaleMesh
00:09 RealBadAngel that trick with stride deforms  all the vertices
00:09 RealBadAngel each one that touches vertices is affected
00:09 RealBadAngel not only cloneMesh
00:09 RealBadAngel translate, rotate etc too
00:10 kahrl what's the error in those?
00:10 RealBadAngel whenever you are doing calculations on Pos, using that stride thingy, they get deformed
00:11 kahrl no, the thing with stride is exactly the correct way to index vertices
00:11 RealBadAngel may look like so
00:12 RealBadAngel but in the end you get deformed coords, thats unaceptable
00:12 kahrl are you sure that it's not only cloneMesh?
00:12 RealBadAngel completely
00:12 kahrl any evidence?
00:12 RealBadAngel im using all those manipulators to get meshes for inventory
00:13 RealBadAngel i had to fix them all to get correct, non broken meshes
00:13 kahrl what's the fix?
00:13 RealBadAngel i can provide you dumps of the coords if you like
00:13 RealBadAngel hold on
00:13 kahrl sure
00:14 RealBadAngel http://pastie.org/10690722
00:14 RealBadAngel thats whole mesh.cpp
00:15 RealBadAngel lemme prepare dumps of the vertices using non-modified manipulators and the new ones
00:18 kahrl I'm pretty sure it's the methods that do *not* do the stride thing (cloneMesh, rotateMesh*) that are buggy
00:19 kahrl otherwise it makes no logical sense (the version with the stride is equivalent to your switch implementation)
00:19 kahrl can you try just replacing the code of cloneMesh and rotateMesh* with your new versions?
00:20 RealBadAngel http://pastie.org/10690729
00:20 RealBadAngel ive used same manipulators in both cases
00:21 RealBadAngel only difference is mesh.cpp
00:21 RealBadAngel as you can see in case of old ones, coords are completely wrong
00:21 kahrl which manipulators did you use?
00:21 RealBadAngel hold on
00:22 RealBadAngel http://pastie.org/10690731
00:22 RealBadAngel here you have all the code to get inventory mesh (WIP ofcourse)
00:23 kahrl right, it does a cloneMesh in pretty much every case :)
00:23 RealBadAngel but believe me, each one was deforming vertices
00:23 RealBadAngel even rotate or translate
00:24 kahrl yeah like I said, rotate is definitely wrong (it casts to S3DVertex* and doesn't care about stride), but translate should be correct
00:25 RealBadAngel it is not
00:25 RealBadAngel i wouldnt be coding it all if it was ok
00:27 RealBadAngel i can upload my working branch and you can see the dumps for yourself if you like
00:27 RealBadAngel easy to comment out all the manipulators used and see the difference
00:28 kahrl I'll try it out
00:28 RealBadAngel ok, gimme a few minutes then, im on it
00:32 RealBadAngel https://github.com/RealBadAngel​/minetest/tree/inventory_meshes
00:32 RealBadAngel i will compile it just to check if i moved there everything
00:34 RealBadAngel mesh.cpp here is untouched, all changed manipulators you have in mesh1.cpp
00:34 RealBadAngel so you can just swap files
00:35 kahrl cool, let me try it
00:35 RealBadAngel to get one mesh dumped, put regular node in 1st inventory slot
00:35 RealBadAngel and restart the game
00:35 RealBadAngel use dirt or something to get mesh passed from mapblock mesh
00:36 RealBadAngel ok, sources are complete
00:37 RealBadAngel heres the dump when i started minimal game:
00:37 RealBadAngel http://pastie.org/10690743
00:38 RealBadAngel there are two cubes in minimal at the start
00:40 RealBadAngel and after replacing mesh.cpp: http://pastie.org/10690744
00:41 RealBadAngel btw, offtopic, how do you like the inventory changes?
00:41 kahrl ok good so far, I get the same output
00:43 kahrl looks good, though some things I've noticed:
00:43 kahrl 1. the rotation angle should reset when you select/hover a different item
00:44 RealBadAngel agree
00:44 kahrl 2. inventorycube textures look weird when rotated (not sure what to do about this one)
00:44 RealBadAngel i was just stopped with the manipulators issue, it took me a few days to hunt down the reason
00:44 RealBadAngel ad 2. drop them
00:44 RealBadAngel theyre not usefull anymore
00:45 kahrl 3. in general item meshes should be drawn a little smaller (in the inventory) IMHO
00:45 RealBadAngel easy to fine tune it
00:45 RealBadAngel you can now try it with dreambuilder and UI
00:46 RealBadAngel you will see how it changes the speed of displaying pages
00:46 kahrl the code seems to be calling mesh manipulators every frame, is that intended?
00:46 RealBadAngel showing the mesh is a bit slower than the texture, but theres no call to create the texture at all
00:47 kahrl that seems like it could be a huge slowdown on machines where upload to the GPU is slow
00:47 RealBadAngel ive measured timings for single item hold on, i do have it somwhere
00:48 RealBadAngel http://pastie.org/10685891
00:48 RealBadAngel average is 50,55 for meshes, 28,57 for textures
00:48 kahrl RealBadAngel: what is "...code..." doing?
00:49 RealBadAngel showing the item
00:49 RealBadAngel ive measured drawtime
00:49 RealBadAngel hold on a sec, will point you to it
00:49 kahrl well like I said, it hugely depends on how fast upload to the GPU is
00:49 kahrl I'm pretty sure yours is very fast
00:50 RealBadAngel hud.cpp, line 488
00:50 RealBadAngel showing texture is disabled with if (0)
00:51 RealBadAngel you can replace it with if (texture) to display them again
00:51 RealBadAngel and disable meshes with if (mesh) -> if(0)
00:51 RealBadAngel timers are here
00:52 RealBadAngel idk if that fast, its 6yrs old nvidia
00:52 RealBadAngel gtx550ti
00:53 RealBadAngel it costs circa 50usd right now
00:55 RealBadAngel we can measure the whole thingy on other platforms too
00:55 RealBadAngel what timings do you actually get?
00:56 kahrl I'm doing the mesh.cpp stuff at the moment
00:59 RealBadAngel trying to get stride thingy to work correctly?
01:02 kahrl got it
01:03 kahrl I replaced cloneMesh with your version and changed the other methods that didn't use stride to use stride
01:03 kahrl now the coordinates in minimal are correct
01:04 kahrl https://gist.github.com/kahrl/bd5c86abb5c08d61984d
01:05 RealBadAngel can you paste whole modified mesh.cpp?
01:06 kahrl https://gist.github.com/kahrl/2b8daf93af30462a0de3
01:08 RealBadAngel everything seems to be working fine
01:11 RealBadAngel will you make a PR with it?
01:11 kahrl I'll do that
01:11 RealBadAngel ok
01:12 kahrl wasn't sure if you wanted that or not, and whether it conflicted with your changes :)
01:12 RealBadAngel my code will only depend on it (to work correctly)
01:12 RealBadAngel but shouldnt be a part of my PR imho
01:13 RealBadAngel also there will be no conflicts, the inventory code doesnt add anything to mesh.cpp at all
01:15 RealBadAngel so, can you check the timings?
01:16 RealBadAngel i need to cleanup the code so we can measure memory usage after it
01:17 RealBadAngel im hoping to get huge gain there in case of dreambuilder and HDX
01:17 RealBadAngel atm its 6-8 gb
01:21 kahrl PR made: #3567
01:22 kahrl https://github.com/minetest/minetest/pull/3567
01:22 * kahrl tickles ShadowBot
01:23 sofar would this fix waving mesh nodes?
01:23 sofar RealBadAngel: ?
01:23 RealBadAngel sofar, what issue you mean?
01:24 sofar there's a patch pending of mine to make mesh and nodebox nodes allow to be "waved" like plants/leaves
01:24 sofar but that doesn't work for many mesh nodes
01:25 RealBadAngel im not sure, but evertyhting that was using our mesh manipulators were affected
01:26 RealBadAngel kahrl, btw, now you can safely use cloneMesh instead of meshmanip and its copy one
01:26 RealBadAngel i mean in wieldmesh
01:26 sofar oh
01:26 sofar is this related to the wield glitch?
01:27 RealBadAngel some1 will have to test it
01:27 RealBadAngel i wont be suprised it that will fix it
01:28 RealBadAngel anyway good we have solved that issue, it took me a few days and delayed me a lot
01:29 kahrl yeah I can't even imagine how frustrating something like that is
01:31 RealBadAngel i hope to get the meshes patch finished today then, need to clean all the debugging mess
01:33 RealBadAngel btw, throwing 256x texture pack at it doesnt seems to have any impact on frame rate
01:39 paramat joined #minetest-dev
01:40 kahrl RealBadAngel: timings: I'm getting an average of 105.252 with meshes, 56.976 with textures
01:40 RealBadAngel whats your GPU?
01:41 kahrl radeon hd 6850
01:41 paramat kahrl does 3567 need extra review or shall i +1 and merge it?
01:42 RealBadAngel paramat, its rather an urgent bugfix and its tested
01:42 paramat ok +1
01:42 RealBadAngel kahrl, can you see any slowdowns using it?
01:43 kahrl paramat: maybe a code style review
01:44 paramat ok
01:44 kahrl I didn't change the code style of lines adjacent to what I've changed, not sure if that's needed
01:44 kahrl RealBadAngel: well with the dstream output enabled it's really sluggish ;)
01:45 RealBadAngel indeed, i will rather polish it first
01:45 kahrl it's because outputting to terminal is slow
01:45 RealBadAngel it dumps a shitload of things atm :)
01:46 RealBadAngel but im really happy i can move on with my code now :)
01:46 kahrl without the dstream output, I always get >55 FPS in minimal with some noticeable stutter
01:47 kahrl but I had that stutter before, so I can't say it's because of your code
01:47 RealBadAngel trying now to measure memory usage before/after
01:47 paramat kahrl what's your opinion on creating a split renderer as a solution to https://github.com/minetest/minetest/issues/3530 ?
01:48 paramat so far no-one has detailed why it's a bad idea, i'm neutral
01:48 kahrl paramat: I would agree to splitting it, but RBA said it's not possible. I didn't look into the specifics
01:49 kahrl paramat: oh, err
01:50 RealBadAngel it could be possible
01:50 paramat ah
01:50 kahrl paramat: I thought you meant making the renderer a template with the vertex type as template argument
01:50 paramat i see
01:50 RealBadAngel i will just need to code mesh stuff the same way as cloneMesh is now
01:51 RealBadAngel anyway its the way its done in irrlicht itself
01:51 kahrl yeah, that would be another way
01:51 kahrl I would be ok with either one
01:51 paramat hmmmm what's your objection to renderer split?
01:51 RealBadAngel its not renderer split at all, just coding 2 possible cases
01:52 RealBadAngel and that applies only to mesh collector
01:52 kahrl (with the template solution, just avoid including headers with templates as much as possible, to avoid excessive compile times)
01:52 RealBadAngel i tried to code that with templates before, gave up after a week or so ;)
01:53 kahrl yeah templates aren't fun in the slightest
01:53 RealBadAngel instead of adding some cases, whole mapblock mesh shall be rewritten
01:53 RealBadAngel thats not fun at all ;)
01:54 paramat it would be ideal to be able to use some of RBAs effects if we want, so if no-one has a good reason not to i feel we should do it
01:55 paramat i can't judge obviously
01:57 RealBadAngel paramat, "some of the effects" are basically all shaders things and soon the lighting too
01:58 paramat yes i realise it's fundamental
01:58 RealBadAngel the lighting will also require another kind of vertices too
01:59 RealBadAngel if i will put in the meshcollector case for it to be able to collect different kind of vertices you will propably get what you want
02:03 paramat there needs to be the option to use the previous simple verticies, with shaders disabled obviously
02:04 RealBadAngel when i code it with cases that will be possible
02:07 paramat cool. the extra data in each vertex will inevitably have an effect, so if you can do that i will probably support it
02:18 RealBadAngel but keep in mind that 25% FPS drop is not a general drop for everybody
02:18 RealBadAngel its just your case
02:19 RealBadAngel dont expect that everybody will get then fps gain
02:20 RealBadAngel on my box non-shaders version will become slower even
02:26 paramat my case and hmmmmm's. but yes, i hope others test the change in FPS
02:27 RealBadAngel i will code that in next few days once im done with inventory meshes
02:27 paramat great thanks
02:28 paramat maybe it's because i have integrated graphics, no card
02:28 RealBadAngel btw, what about https://github.com/minetest/minetest/pull/3547 ?
02:30 paramat reading..
02:34 paramat +1
02:38 sofar yeah that one is an easy one
02:38 sofar RealBadAngel: I will try and test your changes on a new intel HD graphics card if I can, soon
02:38 RealBadAngel kahrl, btw no wonder you had twice the time of rendering mesh item, in current code it was displaying 2 meshes at once, there was an attempt to display halo mesh
02:39 kahrl RealBadAngel: yeah, I saw something about that, wasn't sure what it was
02:39 kahrl (never saw a halo)
02:42 RealBadAngel tried to show selected the same way as highlighted node but that doesnt work any good
02:43 RealBadAngel btw, with the change preloading item visuals can be completely removed
02:44 RealBadAngel texture is demanded only as a fallback in case of getting mesh failed
02:45 paramat i noticed you changed the angle of the inventory cubes? please can they be isometric as before?
02:45 paramat then top is as clear and visible as the sides
02:45 RealBadAngel paramat, they do rotate when selected
02:46 RealBadAngel when still, they look the same way as before
02:46 paramat i must have seen an out of date screenshot of yours
02:46 RealBadAngel hold on a sec
02:47 paramat the screenshot i saw had them at a lower angle, top not so visible
02:48 paramat and the inventory cubes spin around y-axis?
02:50 RealBadAngel indeed the angle is a bit lower, will change it
02:50 RealBadAngel yes
02:54 paramat ok
03:03 paramat i don't mind the spinning of the selected item in the inventory, it looks just like the item on the ground, and with inventory open it's not affecting in-game performance. but i'm not keen on spinning the wielded item in the hotbar, because this might reduce in-game performance, it's also visually distracting
03:06 paramat in inventory spin can be useful for seeing all sides of the cubes, but not much needed for wielded item
03:07 RealBadAngel displaying a bunch of meshes more when you display whole the world around wont affect the performance rather
03:08 * sofar builds RBA's inventory meshes branch
03:08 RealBadAngel sofar, its highly bloated atm, lotsa debugging stuff etc
03:08 RealBadAngel but you can see for yourself how it works
03:09 sofar yeah, it works great
03:09 sofar no page flipping delay either :D
03:09 sofar at the cost of some compile warnings ;)
03:09 RealBadAngel hehe, im cleaning the code atm
03:10 RealBadAngel paramat, does items spinning on the ground distract you too? ;)
03:10 paramat heh no i'm okay with that
03:11 RealBadAngel paramat, just wait when you will be able to test polished code
03:12 sofar it's not distracting
03:12 sofar not even in the action bar below
03:12 sofar actually helps me to see what I'm wielding
03:12 paramat ok
03:13 paramat i'm just concerned about the possible extra processing being needed to spin it, hopefully it's insignificant
03:15 kahrl paramat, est31, thanks for the review on 3567, will merge it now
03:15 paramat ok
03:34 crazyR joined #minetest-dev
03:45 paramat i'll merge aspen trees and new creative inventory later
03:48 paramat i'll also update saveschems
03:48 Skyduskguy sofar but if i do it at mapgen then how will they generate after map gen?
03:52 paramat use 'on place' and 'on dig' functions for after mapgen
03:53 Skyduskguy that at min doubles all of my if's
03:53 paramat .. should this discussion be in the other channel?
03:53 Skyduskguy I dont know what should be where, as i see it talking about developing a mod
03:53 Skyduskguy other is minetest general isnt it?
04:42 paramat joined #minetest-dev
04:52 paramat mod discussion goes to #minetest channel
04:53 paramat now merging game 771 777
04:53 Skyduskguy so what is this?
04:53 Skyduskguy thats confusing
04:54 kahrl Skyduskguy: core development
04:54 kahrl core as in engine, not mods
04:55 kahrl (the exception that minetest_game development can be discussed proves the rule)
04:57 Skyduskguy welp, still think there should be a general and a modding channel. I'm going to leave now
04:57 Skyduskguy left #minetest-dev
05:04 sofar paramat: ahh lol the written book image finally merged, thanks :)
05:05 sofar paramat: did you have time yet to dig up those mtr saveschems?
05:05 paramat merged
05:05 paramat will update savaschems now
05:06 sofar thanks, I'll reconvert them to the new format in minetest-mods/saveschems then as well
05:06 paramat i'll also correct the "mts" thing
05:07 sofar yeah, I pulled that in as well
05:10 sofar damn! I never saw 777
05:10 sofar nice one
05:10 paramat lol
05:19 sofar 2016-01-14 21:13:53: WARNING[Main]: Irrlicht: Could not open file of texture: character.png
05:19 sofar still happening, no clue why
05:20 kaadmy because of the model
05:20 kaadmy put the character.png in the textures/ dir
05:21 sofar $ find . -name character.png
05:21 sofar ./mods/default/models/character.png
05:21 sofar huh
05:21 sofar why is that in the models folder?
05:21 sofar lemme fix that :)
05:23 sofar doesn't even fix the problem
05:24 kaadmy oh
05:24 sofar [pid 20167] 21:18:42.942261 open("/home/sofar/git/mine​test/build/character.png", O_RDONLY) = -1 ENOENT (No such file or directory)
05:24 sofar ah, it's because it's my git build
05:27 sofar sweet, valleys_c with aspen and logs look great
05:35 paramat my saveschems repo is updated
05:35 sofar thanks
05:35 sofar I'll double check to compare it actually creates the same binary files as in-game
05:38 paramat there's even a 1 node waterlily schem heh
05:39 paramat might return to lua table schems for the small ones
05:39 sofar yeah, that's probably better
05:41 paramat papyrus and waterlily, large cactus is big enough for an mts file
05:42 sofar logs are 3x3
05:42 sofar anything under that size back to lua tables
05:42 paramat yes
05:50 sofar 783c71fa8083f25752a3834c96f945c1  /home/sofar/git/minetest_game/mods/​default/schematics/acacia_tree.mts
05:50 sofar 8ccd2c7a512b34adb484a84ddc842363  schematics/acacia_tree.mts
05:51 sofar I probably made an error in that one
05:51 sofar the rest now matches
05:52 sofar yup
05:55 sofar paramat: you probably can't merge from minetest-mods/saveschems, but I suspect you may want to use the new table format to make new schematics easier to read and update
06:06 Player2 joined #minetest-dev
06:35 Hunterz joined #minetest-dev
06:40 blaze joined #minetest-dev
06:55 paramat hey hmmmm is there a reason why all members of the mgv5/v6/v7 classes are public? i've been asked to make as many as possible private
06:56 hmmmm it doesn't really matter
06:56 hmmmm they're public for convenience
06:57 hmmmm access modifiers in C++ are a bunch of crap without any real utility
06:58 paramat nerzhul told me it's important to have as many as possible private
06:58 hmmmm it's not really that important...
06:59 paramat ok
06:59 paramat thanks
07:09 hmmmm you know what I feel like coding right now
07:09 hmmmm what things need to be done in minetestland?
07:09 hmmmm that don't consist of code reviews*
07:11 sofar need an idea?
07:11 sofar I've got some that I was saving
07:12 sofar use the font engine to paint text on signs client-side rendering
07:12 sofar for SIGNLIKE nodes
07:12 sofar so we don't have lua shove pixel macaroni
07:12 hmmmm wasn't RBA doing something like that already?
07:12 sofar just fetch text from node meta infotext
07:13 sofar and pick a nice bitmappy but ye olde font style that the font engine can paint
07:13 sofar it'll make a TON of server lag go away
07:13 paramat erm, the water over ignore thing?
07:13 sofar ahahahahaha
07:14 sofar yeah that one too
07:14 sofar that one isn't fun though
07:14 sofar ;)
07:14 hmmmm ugh transformLiquidNodes
07:14 paramat yeah not much fun
07:14 sofar hmmmm: just admit, my idea is better
07:14 hmmmm sure, it shouldn't take more than 2 hours
07:15 paramat anyway gregorycu might attempt the water ignore thing sometime
07:16 kahrl hmmmm: change CNodeDefManager::set() so that it removes the node from groups (in m_group_to_items) it is no longer in, when a node is redefined
07:17 paramat mgv8 heh, aether and nether realms ..
07:19 hmmmm well
07:19 hmmmm valleys mapgen starts me underwater
07:19 hmmmm that's a big no-no
07:19 paramat ah someone reported that, we'll tackle it
07:20 paramat will post in the PR thread
07:21 paramat perhaps even hide it for a while
07:21 hmmmm don't... that's tantamount to commit spam
07:22 hmmmm ahh i like how you guys got rivers working without overflowing all the time
07:22 sofar the river water is getting good use now
07:22 hmmmm mapgen v7 was supposed to have rivers defined by log(noise^3) but they were too finnicky in overflowing
07:23 hmmmm i take it that river water just floods over a shorter distance or something?
07:23 hmmmm only looks like two nodes
07:23 paramat yeah range 2
07:23 hmmmm decent solution
07:24 paramat and non-renewable
07:25 hmmmm PTAL https://github.com/kwolekr/minetest/commit​/f8b3395227b2f15ea67844415c96e076f6c097ef
07:25 paramat did you spawn underwater at 0,0,0?
07:25 hmmmm i did not take note of the exact position
07:27 paramat no problem i'll test
07:27 hmmmm no code review?
07:27 hmmmm i already manually tested it and it works fine, it's just the code review would be nice
07:33 paramat i looked and code style seems ok, implementation is a little beyond me though, sorry
07:35 paramat i could +1 on trust
07:35 hmmmm you can't...
07:35 hmmmm the point of a code review is to review the code, not review your trust
07:37 paramat exactly
07:46 hmmmm do you know how much trouble it is to intentionally place content ignore
07:57 Alduin_ joined #minetest-dev
07:58 Obani joined #minetest-dev
08:05 Calinou joined #minetest-dev
08:17 paramat left #minetest-dev
08:34 sofar huh, wth
08:34 sofar left click trapdoor, param2 goes 0 1 2 3
08:34 sofar right click it
08:34 sofar screwdriver mod gets all the way to minetest.swap_node(pos, node) where node has param2 = 4
08:35 sofar never does it. after that param2 just stays 0
08:35 sofar oh, huh
08:36 sofar nvm something is interfering
08:44 nrzkt joined #minetest-dev
08:47 Calinou joined #minetest-dev
09:02 Calinou_ joined #minetest-dev
09:29 Megaf joined #minetest-dev
09:45 Hunterz joined #minetest-dev
09:58 sfan5 10/10 https://forum.minetest.net/v​iewtopic.php?f=3&t=13891
09:58 sfan5 ~title https://forum.minetest.net/v​iewtopic.php?f=3&t=13891
09:58 sfan5 damn you ShadowBot
09:58 ShadowBot sfan5: Error: The command "title" is available in the Format and Web plugins.  Please specify the plugin whose command you wish to call by using its name as a command before "title".
10:25 Soni joined #minetest-dev
10:28 Alduin_ sfan5, hum, is it as worrysome as it sounds? I was thinking about a new hobby project where I can contribute/spend my hours of subway time each week on...
10:34 Megaf joined #minetest-dev
10:52 everamzah https://github.com/minetes​t/minetest_game/issues/786
11:12 RealBadAngel sofar, here?
11:16 RealBadAngel sofar, until 3166 is fixed and merged you cant make signs that way
11:17 RealBadAngel each meta change (simple as moving item in the chest) is causing source mapblock to be sent -> then regeneration of meshes for that and all the adjacent blocks
11:18 RealBadAngel also, all metas for the source block are deleted and restored from incoming block
11:19 RealBadAngel i do have almost finished code for the signs, but its useless without 3166
11:22 PenguinDad joined #minetest-dev
11:32 Megaf joined #minetest-dev
11:38 Megaf joined #minetest-dev
11:38 everamzah joined #minetest-dev
11:40 sfan5 Alduin_: would you believe a conspiracy that says that the jews control everything in the world?
11:41 sfan5 if you answer no then don't take the forum topic seriously either
11:45 Calinou joined #minetest-dev
11:56 kahrl everamzah: commented
12:05 proller joined #minetest-dev
12:07 sdzen joined #minetest-dev
12:08 sfan5 kahrl: http://sprunge.us/VQjG?diff fixes the infinite placing of slabs
12:08 sfan5 but you can still place 1 full block + 1 slab from 2 slabs
12:08 sfan5 the problem was that it returned an itemstack of 0 items
12:08 sfan5 those are apparently not rendered in the inventory and thus invisible
12:09 sfan5 but you can still use them
12:09 sfan5 perhaps itemstack:set_count(0) should delete the itemstack
12:22 H-H-H joined #minetest-dev
12:25 Fixer joined #minetest-dev
12:33 sfan5 ok, found the bug
12:34 sfan5 https://github.com/minetest/minetest/blob/m​aster/src/script/cpp_api/s_item.cpp#L75-L81
12:34 sfan5 when on_place returns nil the itemstack is not modified
12:35 sfan5 maybe this is intended
12:35 sfan5 but i doesn't seem to be because a lot of on_place handlers explicitly return the itemstack
12:37 jin_xi joined #minetest-dev
12:45 sfan5 pushing http://sprunge.us/AiDi?diff in 10 minutes
12:45 sfan5 what i'm fixing is :set_count(0)
12:45 sfan5 (not sure whether the nil thing i mentioned above is the real problem)
12:46 sfan5 s/the real problem/a bug or a feature/ :D
13:16 behalebabo joined #minetest-dev
13:22 proller joined #minetest-dev
13:24 sdzen joined #minetest-dev
13:46 sdzen joined #minetest-dev
14:05 Fixer RealBadAngel, did you manage to reproduce weild mesh bug? You mentioned it earlier few days ago
14:17 Hunterz joined #minetest-dev
14:39 Amaz joined #minetest-dev
14:47 Taoki joined #minetest-dev
14:49 Taoki joined #minetest-dev
15:00 Calinou joined #minetest-dev
15:07 kaadmy joined #minetest-dev
15:20 proller joined #minetest-dev
15:37 Taoki joined #minetest-dev
15:37 H-H-H <<+RealBadAngel>> thePalindrome, theres a mod that adds animated chests for example
15:37 H-H-H <<+RealBadAngel>> or item pedestals, with animated, rotating item
15:37 H-H-H does anyone have a link to these?
15:38 RealBadAngel H-H-H, that chest is part of Jordach's game i believe
15:38 H-H-H i did find some in Majichet game
15:39 Taoki joined #minetest-dev
15:39 H-H-H and ty for your reply RealBadAngel
15:39 RealBadAngel https://forum.minetest.net/viewtopic.php?​f=9&amp;t=13036&amp;hilit=animated+chest
15:40 H-H-H awesome thanks :)
15:42 sfan5_ joined #minetest-dev
15:44 RealBadAngel H-H-H, and item pedestals : https://forum.minetest.net/viewtopic.p​hp?f=11&amp;t=5600&amp;hilit=pedestals
15:45 H-H-H thank you very much apreciated
15:46 CraigyDavi joined #minetest-dev
15:47 RealBadAngel np
15:57 Calinou_ joined #minetest-dev
16:02 Hunterz joined #minetest-dev
16:17 hmmmm joined #minetest-dev
16:21 proller joined #minetest-dev
16:26 Darcidride joined #minetest-dev
16:29 Player2 joined #minetest-dev
16:35 H-H-H now to figure out how to edit the builtin_items mod to work with your entity meta patch
16:36 RealBadAngel theres simply new field in items entitity properties, "infotext"
16:36 RealBadAngel mods can freely alter it
16:36 Elinvention joined #minetest-dev
16:37 H-H-H ty will look into it as i like the way itesm flow with water and get burned in lava in the builtin_items mod i use but it overides the definition
16:41 Fixer RealBadAngel, did you manage to reproduce weild mesh bug? You mentioned it earlier few days ago
16:43 RealBadAngel fixer, ive found out that it is always present when rendering to texture
16:43 RealBadAngel but you can also check the latest patch made by kahrl, it may be related to that glitch
16:44 Fixer RealBadAngel, https://forum.minetest.net/downlo​ad/file.php?mode=view&amp;id=1748 you seen it like that on a left?
16:45 Fixer RealBadAngel, it was pushed already, i will look at it when Krock/sfan5 compiles a new one eventually
16:48 Fixer RealBadAngel, if you seen that weild mesh bug than you can probably observe texture tear problem, to see it, go over snow biome with snow over trees, fly above them increasing heigth and sligthly changing mouse angle, snow on trees should start to look edgy (like crap)-depending on mouse angle
16:53 H-H-H ty RealBadAngel ive just fixed my builtin_items mod and it now shows the infotext wich imho is an extremely usefull feature
16:53 RealBadAngel i doubt those two are related in any way
16:54 Fixer i will report if kahrl thing fixes that weild mesh thing
16:55 RealBadAngel im not sure it will, just saying it was a nasty bug, hard to find
16:55 RealBadAngel and meshes related
17:08 Krock joined #minetest-dev
17:43 Calinou joined #minetest-dev
17:46 rubenwardy joined #minetest-dev
18:22 Megaf joined #minetest-dev
18:44 est31 joined #minetest-dev
18:54 Megaf joined #minetest-dev
18:58 luizrpgluiz joined #minetest-dev
18:58 sdzen joined #minetest-dev
19:00 DFeniks joined #minetest-dev
19:07 Amaz joined #minetest-dev
19:20 Gael-de-Sailly joined #minetest-dev
19:36 Terusthebird joined #minetest-dev
19:38 Fixer joined #minetest-dev
19:39 Elinvention joined #minetest-dev
20:05 Darcidride joined #minetest-dev
20:41 Elinvention joined #minetest-dev
21:35 Taoki joined #minetest-dev
22:26 Elinvention joined #minetest-dev
22:27 sdzen joined #minetest-dev
22:37 proller joined #minetest-dev
22:39 louiloui joined #minetest-dev
22:48 Fixer joined #minetest-dev
23:23 GhostDoge joined #minetest-dev

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