Time Nick Message 00:05 VanessaE greetz, RBA 04:18 cg72 ok if im checking a node to see whats under it and the chunk under its not loaded what will the check return???? 04:21 VanessaE "ignore" 04:21 VanessaE that is, the node name will be "ignore" if you do e.g. minetest.get_node() on an unloaded block 04:24 cg72 ok cuz my treedecay is ok till they get taller lol 15:07 celeron55 some recent optimization stuff in MCPE, if someone is interested: https://tomcc.github.io/2014/08/31/visibility-1.html https://tomcc.github.io/2014/08/31/visibility-2.html 15:08 * sfan5 saw that blog post but didn't read it yet 15:08 VanessaE THOSE are minecraft's caves? 15:21 VanessaE hm, breadth-first search eh 15:21 VanessaE didn't I suggest something like this a while back? :) 15:21 sfan5 sounds interesting 15:21 sfan5 but I dunno about 3D graphics stuff 15:22 VanessaE specifically to address the stuttering problem with the view range tuner, that is 15:22 sfan5 don't we have a wireframe mode in mt? 15:22 VanessaE start at the player, start rendering stuff close by, keep adding stuff that's further out until you run out of time at the current fps 15:22 sfan5 hmm 15:22 sfan5 no 15:22 VanessaE sfan5: I think so but how to turn it on, no idea. 15:23 sfan5 I made minetest use wireframes 15:23 sfan5 but as always I stored the code in /tmp 15:23 sfan5 because it was.. well.. temporary 15:23 VanessaE couldn't hurt to add that code back into mainline, might be useful for testing in the future. 15:24 VanessaE (if you can recreate it I mean) 15:24 sfan5 I probably can 15:24 sfan5 >setMaterialFlag(video::EMF_WIREFRAME,t 15:24 sfan5 there it is 15:25 VanessaE attach it to F9 or some such 15:28 kahrl I don't think we should add an X-ray to the standard client 15:28 twoelk could that wireframe view be tweaked to also study cave-gereating in air or maybe disblay chunk borders to detect border issues? 15:29 twoelk maybe add to minimal? 15:29 VanessaE kahrl: wireframe just showing what the user normally sees? 15:29 VanessaE (+ caves, e.g. similar to that screenshot) 15:30 VanessaE it's not like the user's gonna be able to see where all the mese is. 15:31 kahrl VanessaE: yeah it's not useful for finding ores in stone, but as you can see on the screenshot in that article you can see abandoned mineshafts and big lava-level caves 15:31 kahrl which are often very profitable 15:31 VanessaE hm. good point 15:31 VanessaE I guess that's what all the orange is? 15:31 kahrl yep 15:31 VanessaE didn't think about that - I've never played Minecraft. 15:32 sfan5 gur.com/q1ZUsLs.png 15:32 sfan5 fail 15:32 sfan5 http://i.imgur.com/q1ZUsLs.png 15:32 VanessaE \o/ 15:32 sfan5 upon entering the ground all the caves get rendered 15:32 VanessaE that was easy 15:32 sfan5 that could be improved 15:34 twoelk only node sides as viewed with noclip in that pic? 15:35 VanessaE shows how much gets drawn that shouldn't be huh. 15:35 VanessaE sfan5: what's that little bit of something just above and left of the target? 15:35 VanessaE (four very short lines) 15:36 sfan5 VanessaE: ¯\_(ツ)_/¯ 15:36 sfan5 probably desert grass that has changed it's rotation because wave shaders 15:37 VanessaE ah 16:48 hmmmm hrmmm 16:48 hmmmm i wonder if there will ever be a situation where we'll use VoxelManipulators on the client side 16:49 VanessaE finding surfaces for mobs to spawn on maybe? 16:49 hmmmm i don't want mobs to ever spawn on the client 16:49 VanessaE that and realtime sound effects are the only two things I can think of that client-side modding would really be appropriate for at this point 16:50 VanessaE well to *walk* on then. 16:50 VanessaE otherwise, no 16:50 VanessaE I don't see a reason for it 16:50 hmmmm hmmm 16:50 hmmmm no we don't need that 16:50 hmmmm we have lua pathfinding 16:51 VanessaE right 16:51 VanessaE so yeah, I would just not bother considering that possibility for now. there's no logical reason to use an lvm on the client 16:51 hmmmm good 16:52 VanessaE I could see an argument for distributed computing using clients - that's how MC is able to handle so many at once, I'm told - but that's WAY out of scope here. 17:10 VanessaE kahrl: *poke* 17:18 hmmmm I feel as if removeNode* functions are just a gigantic duplication of effort 17:18 hmmmm you could easily trim those to remove a couple hundred LoC easily 17:43 kahrl VanessaE: peek? 17:44 ShadowNinja sfan5: The wireframe view definitely seems usefull. Maybe bind it to a key and surround it with #ifndef NDEBUG? 17:46 VanessaE kahrl: just reminding you about that no-extrude patch. it still works. Would be nice to see it merged as an option in the settings menu. 17:46 Krock Bug: Direct3d8 isn't showing up in the settings tab - do I need to create an issue or does someone have free time to fix it? 17:47 sfan5 ShadowNinja: maybe 17:49 ShadowNinja Krock: Is your build compiled with Direct3D8 support? 17:49 kahrl VanessaE: I'd prefer to fix the extrusion routine, but I don't really have good ideas for that, so I guess the setting is fine 17:50 Krock ShadowNinja, yup. I replaced Irrlicht.dll with my new build 17:50 Krock ShadowNinja, and minetest works with it 17:51 ShadowNinja Krock: driver=direct3d8 works, but it isn't in the settings menu? 17:51 sfan5 ShadowNinja: why would you want to use d3d8? 17:51 ShadowNinja sfan5: Ask Krock. 17:52 Krock ShadowNinja, exactly. I possibly found the bug. In the title, it says "Direct3D 8.1" 17:52 Krock but the driver is set with direct3d8 17:52 ShadowNinja Krock: Odd, you should ask sapier about that. 17:53 Krock ShadowNinja, or I'll look into the codes first. I'll tall him about this if I can't find it 17:53 sfan5 Krock: are you complaining about the ".1"? 17:53 Krock sfan5, I don't know if that's the reason for not having it in the settings menu 17:53 ShadowNinja Krock: If you're using a different build of Irlicht than you compiled with that might be the issue though. 17:53 sfan5 Krock: thats not the reason 17:54 Krock ShadowNinja, I compiled the non-direct3d8 and the one with on the same PC 17:57 hmmmm has this thing ever been seen with splizzard's mod? http://pastebin.com/f0ZFd00p 17:57 hmmmm or did I make it do that :< 17:57 VanessaE not sure, hmmmm. 17:59 ShadowNinja Krock: Doesn't matter. Use the same library for building and using and see if it works. 17:59 Krock ShadowNinja, I found the bug . 17:59 Krock https://github.com/minetest/minetest/blob/master/src/script/lua_api/l_mainmenu.cpp#L1008 17:59 Krock possibly. 18:00 Krock lua requests core.get_video_drivers() and doesn't get direct3d8 18:00 hmmmm vanessae, shouldn't Pine Needles be added to the leaves group? they aren't decaying 18:00 hmmmm nevermind about that 18:00 VanessaE heh 18:01 VanessaE moretrees manages leaf decay on its own, yeah 18:02 kahrl Krock: perhaps you've compiled minetest with the non-direct3d8 Irrlicht's headers 18:02 Krock kahrl, I can play with direct3d8... 18:02 ShadowNinja ^ 18:02 Krock oh wait 18:03 Krock but yes, nvm, it doesn't matter because I can play it... 18:03 kahrl Krock: yeah, but isDriverSupported is in the header, not the DLL 18:03 ShadowNinja Krock: Minetest might not think it supports D3D8 because of the headers it was compiled with, but the function calls still work bacause the DLL that it uses actually does have support for it. 18:04 Krock kahrl, ShadowNinja: but well, there is still a way to detect if the driver is supported or not, without looking at headers 18:04 Krock because it throws an error when I use my older Irrlicht build 18:06 ShadowNinja Krock: But that's not the right way to do it. That's trying it and seeing if it breaks instead of just checking if it will work beforehand. 18:07 kahrl if fact, doing that would be stupid because some drivers might crash minetest when they are created (see sapier and video_driver = software) 18:07 kahrl in fact* 18:09 * Krock points at https://github.com/minetest/minetest/blob/master/src/main.cpp#L1385 18:09 * Krock unpoints cuz he noticed how he failed 18:24 hmmmm wtf 18:24 hmmmm land of the cut-off trees 18:24 hmmmm did somebody do something lately concerning mapgen because the last time I worked on anything this did not break 18:29 hmmmm jesus christ the regressions 18:31 Sokomine hmmm: do you have a screenshot at hand? 18:31 hmmmm no. I can see it perfectly fine 18:32 hmmmm happened sometime in between e9e9e42573a50dbf94370996d65b03dca1d5c284 and e14c4cdd4c3c9b554dc9cb91f8f29078ad337ded 18:33 Sokomine hmmmm: posting a screenshot might help to find out if the phenomenon was observed by others before 18:33 hmmmm perhaps. i don't really need to find that out though 18:35 hmmmm you know what, I bet this was my own stupid doing 18:36 iqualfragile hmmmm: lava flows into ungenerated mapblocks 18:41 hmmmm whoops, sorry guys, i broke the trees 18:42 hmmmm well derp 18:42 * VanessaE mumbles incoherently 18:42 hmmmm >tfw you accidentally griefed hundreds of peoples' servers 18:44 VanessaE hmmmm: shit happens. 18:45 hmmmm well i think it might be possible, pending the response from this one guy who made an issue like 3 weeks ago. 18:45 VanessaE is this perhaps the cause of the ordeal we've been arguing about for the past few days? 18:45 hmmmm no 18:45 VanessaE ok 18:45 hmmmm i have that fixed 18:45 VanessaE can't blame me for hoping ;) 18:45 hmmmm https://github.com/kwolekr/minetest/commit/f46695d02b17466281b4d957c12f1a55633029d2 18:49 hmmmm https://github.com/minetest/minetest/commit/a2e1b0fc7fa9a470f645e130115ce14a89aa2e02 this is a stupid commit 18:49 hmmmm there are about 500000 different cases where a generated block would still need to be saved 18:49 hmmmm (from the mapgen) 18:50 VanessaE oh no... 18:50 VanessaE that's not what I think it is...is it? 18:50 hmmmm I don't know what you think it is 18:50 VanessaE the glitch some folks had with the mapgen overwriting existing/user-built structures 18:51 hmmmm i'm not really sure how it'd play into that 18:51 VanessaE well just looking at the patch alone it looks like it 18:51 hmmmm the only time blocks wouldn't be saved back if they've been generated is on mapgen time 18:51 hmmmm you didn't look closely enough 18:51 VanessaE you've got the whole code, I've just seen the patch so far 18:51 VanessaE anyway, no matter 18:57 VanessaE I'm just glad stuff is getting fixed. 19:18 hmmmm when I first saw the third person view i got excited 19:18 hmmmm then i realized it's crap.. the walking animation doesn't work properly 19:19 hmmmm it's very desynchronized from the player's actual movement 19:19 hmmmm and then you're not able to see what item/node is being wielded 19:19 VanessaE there's a mod for the latter, 19:19 VanessaE but how do you mean desynched? it looks fine to me 19:20 hmmmm when i start walking the animation for walking does not start 19:21 VanessaE strange.. it works here 19:21 hmmmm looks like sam is hovering over the land 19:21 hmmmm and then he'll start erratically walking after i had already stopped 19:21 VanessaE and this is in singleplayer? 19:22 hmmmm i do all of my testing in singleplayer. 19:22 VanessaE strange. even the tiniest steps are instantly reflected in the player animation here. 19:22 VanessaE is your minetest_game up-to-date? 19:22 VanessaE I know it's a derp question but can't hurt to ask 19:22 hmmmm it isn't 19:23 hmmmm it's quite old. i don't update the game very often 19:23 * VanessaE checks dreambuilder also 19:23 VanessaE my mt_game is pretty close to HEAD, dreambuilder is much older as far as the player control code is concerned. 19:24 VanessaE however both behave as they should 19:24 hmmmm i'll check it out i guess.. 19:24 hmmmm well 19:24 VanessaE and db has the split-limb model, too, so that's not it 19:24 hmmmm i don't play minetest so it doesn't concern me much 19:24 hmmmm i'm just pointing out that it doesn't work well 19:24 PilzAdam hmmmm, you need to update minetest_game in order for the animations to work properly 19:25 PilzAdam a new function was added so it can tell the client which animation to play 19:25 hmmmm ahhh ok 19:25 hmmmm thanks 19:32 VanessaE ah yes, I forgot about that function. 19:44 paramat hmmmm, that error you saw in snow mod http://pastebin.com/f0ZFd00p i got that too, lots of bugs in that mod, my fixed and faster version is stored in here https://github.com/paramat/wieldhandsam2 but uses non-mapgen object vm 19:45 VanessaE "paramat/wieldhandsam2" ? 19:45 VanessaE strange name for a snow mod :P 19:46 VanessaE oh, multiple mods there , derp 19:46 paramat :) 19:49 VanessaE hmmmm: getting back to the wield object for a moment.. there HAS been some call for a proper in-client solution to showing the wielded tool/object for a while now. there's a mod by stu that does this but it's buggy under some circumstances (last time I used it, it left indestructible item entities all over the place). 19:55 paramat the code in snow mod was a perfectionist's OCD nightmare 8/ 19:55 VanessaE haha 19:57 paramat back in 2h o/ 20:45 hmmmm paramat: I fixed that thing 20:46 hmmmm https://github.com/minetest/minetest/commit/9e4e7072da8f565eef37da7558053a436b9cbba3 20:46 hmmmm try using a mapgen object vm again 21:10 iqualfragile in the player files is the position x,y,z? 21:10 iqualfragile position = (3078,135,1101.83) 21:10 iqualfragile height x, y? 21:10 VanessaE it is. 21:10 VanessaE in tenths. 21:10 VanessaE x,y,z that is. 21:10 VanessaE so that position is 307.8, 13.5, 110.183 21:11 iqualfragile why the fuck in tenths? 21:11 VanessaE beats me 21:11 hmmmm why not the fuck in tenths? 21:11 hmmmm what would you use? 21:11 iqualfragile because it will be a horrible nightmare to process in bash -.- 21:11 VanessaE hmmmm: I think he means why is it multiplied by 10. 21:12 hmmmm that's an awfully specific use case. i'm sorry celeron didn't predict your future plight when he designed that 5 years ago or whenever 21:12 iqualfragile why not in ninths? 21:12 iqualfragile sounds better to me 21:13 iqualfragile i like the number 21:13 VanessaE which DOES beg the question, why IS it multiplied by 10? wouldn't integers have been perfectly good for this? 21:13 hmmmm the reason why it's multiplied by 10 is so that it COULD be stored as an integer 21:13 celeron55 it's the glorious BullShit constant 21:13 VanessaE that's what I figured. 21:13 celeron55 the best thing ever 21:13 celeron55 (to be explained to anyone) 21:14 hmmmm VanessaE, the position needs to be stored in tenths simply because the player moves in increments of 1/10th 21:14 celeron55 there's no reason whatsoever for it to be applied to that saved value but it just happened to be and that's how it is then 21:14 hmmmm if the player really moved node by node, don't you think that would ruin the fluidity of motion..? 21:14 VanessaE seems fair. 21:15 VanessaE hmmmm: sure...if that file were constantly being re-read after every player movement. 21:15 hmmmm I think it would be annoying to start off in a slightly different spot on every player join 21:15 celeron55 moving in tenths of nodes wouldn't exactly be "fluid" motion 8D 21:15 hmmmm well, yeah, but being able to round to 1/10th makes it a lot more fluid than otherwise 21:15 hmmmm makes things much less perceptible i should say 21:16 hmmmm differences 21:16 VanessaE I doubt a player would notice if they're off by half a node when they re-join, were it storing a rounded integer position. 21:16 celeron55 there's no other reason for BS than it forcing conversion between integer and floating-point positions to be explicit 21:16 VanessaE doesn't matter anyway, it's just an academic curiosity 21:16 celeron55 _no_other_reason_ 21:16 hmmmm huh? 21:16 celeron55 when it happens to be in saved values, it's just laziness 21:17 iqualfragile so why isn't it stored in ints then? 21:17 hmmmm I thought the point of BS was to convert to irrlicht coordinates 21:17 hmmmm because a block was visually 10 whatevers 21:17 celeron55 and just the fact that nobody would have guessed that anything other than minetest would have cared about what minetest saves 21:17 hmmmm so it really is bullshit 21:18 celeron55 well it does that too, that has no meaning whatsoever; everything would work equally with BS=1, except for the reason i just said 21:19 iqualfragile ok, so im gona post the command i will use in the end for extracting the position and you will be forced to look at it 21:20 celeron55 i posted some bash stuff on my solo game development channel today 21:20 celeron55 $ grep '>' -Isn 201*/*8dromeda* | grep -v 'celeron55.*>' | cut -d/ -f1,2 | cut -d: -f1,3 | cut -c1-7,17,32-33 | xargs -n1 date -d | cut -d' ' -f1 | sort | uniq -c | sort -n 21:20 celeron55 i'm mentally prepared for your version of bash shit 8D 21:21 * celeron55 is the true evil 21:21 iqualfragile you could just sort -u 21:21 celeron55 maybe; i couldn't have cared less 21:22 hmmmm you could just write a Python script in 1/10th the time it takes to figure out the necessary command switches 21:22 iqualfragile hmmmm: nah, thats just basic stuff 21:22 celeron55 ...or just don't figure out the necessary ones and use three times the ones you remember 21:22 iqualfragile you usually know these ones by heart 21:22 hmmmm right, it all comes down to familiarity 21:22 hmmmm whatever you know best is going to be the fastest 21:23 hmmmm ...i used to write C utilities to do text processing occasionally 21:24 iqualfragile echo $(($(cat worlds/newworld/players/Iqualfragile | grep -F position | cut -d, -f2 | cut -d. -f1) /10)),$(($(cat worlds/newworld/players/Iqualfragile | grep -F position | cut -d, -f3 | cut -d. -f1) / 10)) 21:24 celeron55 absolutely lovely 21:25 iqualfragile not even all that ugly 21:25 iqualfragile imagined it would be worse 21:25 iqualfragile but it does open the file twice, which is ugly 21:26 celeron55 hmmmm: using C's scanf and printf is actually probably a relatively fast way of making these 21:26 iqualfragile sed all the way 21:34 hmmmm https://github.com/kwolekr/minetest/commit/f3eefeb7948b8b8d1a98f2f89baa39abc807f72d 21:34 hmmmm tell me if anybody would find this useful or not 21:34 hmmmm basically you can set a specific MapNode table at a specific position in a voxelmanip, for special cases, rather than having to set everything in bulk and then set with set_node after 21:35 hmmmm so if you have one goofy node in the middle of the map that you'd like to set, you can do vm:set_node_at(pos, {name="koolmod123:thingie", param1=whatever, param2=whatever})) 21:53 VanessaE might be useful 21:53 VanessaE sokomine had an idea or two for her villages mod that I think needed something similar to this actually 22:09 hmmmm i thought sokomine was a guy 22:14 Sokomine no, i'm female 22:15 Sokomine i'm sure those new functions will come in handy in many situations - especially as they allow to set a few nodes without bothering too much about vm 22:16 Sokomine for the villages, the only working solution i found was to write to the entire vm - including shell. that seems to work fine so far - no more caves inside houses, and no more mudflow problems 22:17 Sokomine it seems to be cruical to be able to generate such a mapchunk procedurally from the coordinates/seed entirely (minus perhaps some decorative plants) 22:20 Sokomine after the vm is written via vm:write_to_map(data), there are a few calls using set_node - for chests which are to be filled (though that could be done - apart from the metadata - with vm as well), and some ground grass node type changes for those buildings that are schematics 22:20 hmmmm oh yeah - these new functions clearly aren't going to work for metadata 22:21 hmmmm metadata are far beyond the scope of what luavoxelmanipulator is intended to do and requires an unreasonable amount of effort to handle 22:21 hmmmm s/are/is/ 22:21 Sokomine that's ok. those chests are an exception and can be handled that way 22:23 Sokomine what would be great, though, would be an option to insert a schematic into a voxelmanip. would seem cleaner than reading a worldedit file, doing some string comparison/deserialization, rotation etc. there 22:24 hmmmm the best would be if schematics were removed from the core and instead implemented in Lua using voxelmanips 22:24 Sokomine perhaps that as well. if it wouldn't make them slower? 22:25 Sokomine it's an incredibly useful, well-working format, that .mts format. only disadvantage is that it can't be read from lua as we don't have a lib to uncompress the data at hand 22:36 paramat hmmmm, i'll compile latest and test your work https://github.com/minetest/minetest/commit/9e4e7072da8f565eef37da7558053a436b9cbba3 soon, with plantlife and snow combination, and with 2 of my own mods 22:36 hmmmm well I already tested it but okay 22:38 paramat also 'Add LuaVoxelManip methods: get_node_at() and set_node_at()' seems very useful 22:58 iqualfragile how often are player files saved? 23:14 PilzAdam iqualfragile, every server_map_save_interval seconds, but only if they are modified 23:15 iqualfragile is that in secs? 23:16 PilzAdam yes, seconds is in seconds 23:25 hmmmm alright so how is this: local vm = VoxelManip(pmin, pmax); creates a LuaVoxelManip with a buffer allocated, filled with CONTENT_IGNORE 23:26 hmmmm if you allocate a VoxelManip in that manner, you don't need to pass the pmin and pmax parameters to VoxelManip:read_from_map() 23:26 hmmmm if you specify a pmin/pmax in the ctor and also in read_from_map(), the area will be reallocated for the area specified in the latter 23:27 hmmmm and of course, reverse compatibility is maintained by not specifying a pmin/max in the ctor but in the read_from_map: i.e. local vm = VoxelManip(); vm:read_from_map(pmin, pmax) 23:27 hmmmm any comments? 23:33 iqualfragile PilzAdam: heh, i should sleep 23:34 paramat so i guess this allows writing a vm to map without having to read from map first, so faster for certain uses 23:38 hmmmm actually i'd argue it's faster for most uses 23:39 hmmmm whenever you want to place some thing on the map 23:39 hmmmm if you know the coordinates 23:40 hisforeverkid hi I;d like to know how to plant a pinecone? 23:40 paramat oh yeah for mapgen in singlenode, you don't need to read the map at all 23:40 hmmmm like if you wanted to make sphere of stone, or something, you can just draw the sphere in the voxelmanip without loading map first, and then when you go to write it to map, it'll only write the pieces of the voxelmanip that aren't content_ignore 23:40 hmmmm the idea is that content_ignore is sort of like a transparent pixel in a drawing 23:40 paramat nice 23:42 paramat this is for both mapgen object vm and non mapgen object vm? 23:43 VanessaE now THAT should be useful for sure 23:43 hisforeverkid ok sorry 23:43 hmmmm well both, but the map has already been read for the mapgen object vm 23:43 hmmmm so what's the point 23:43 paramat yeah