Minetest logo

IRC log for #minetest-dev, 2015-06-16

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

All times shown according to UTC.

Time Nick Message
00:45 paramat latest dev sems slow to render terrain in all mapgens, i can't remember being able to out-fly the visible mapgen and leave big gaps in the terrain behind me. did anything during the last 9 commits slow down mesh rendering?
00:47 est31 I think there is one commit which did it
00:47 est31 but only on very special HW
00:47 est31 VanessaE,
00:47 est31 ^
00:47 paramat have all verticies been changed to tangent/binormal now, perhaps RBA's commit?
00:47 est31 do you have shaders enabled
00:47 est31 yes
00:47 paramat no shaders, minimal settings
00:48 paramat start a world and it takes 2-3 seconds to appear
00:49 paramat i can't see my last commit slowing things down as it only adds biomegen to very deep mapchunks
00:50 est31 how can a mapgen commit slow down rendering?
00:50 paramat RealBadAngel ^ perhaps the vertex change had a bad effect after all?
00:50 paramat well i wasn't sure if it was just slow mapgen or slow mesh render
00:51 hmmmm the mapgen is almost always faster than mesh generation
00:52 hmmmm mapgen generates in 5x5x5 chunks, meshgen generates one block at a time
00:53 paramat i can now fly fast and outrun rendered terrain to the point of being in empty space most of the time
00:53 paramat leaving big empty spaces behind me
00:53 est31 press h problem solved
00:54 est31 :P
00:54 est31 you cant outrun rendered terrain then ):
00:54 est31 :)*
00:55 paramat i know the vertex change was tested to not affect fps much, but maybe it caused this
00:56 paramat at first i thought i had broken something =/
00:56 hmmmm hmm
00:56 hmmmm what might be interesting is a render mode where somehow mapblocks that exist on the client side are visually displayed regardless if their mesh exists yet or not
00:58 est31 perhaps if you press f5
00:58 hmmmm terasology has something like this, if you've ever tried it out
00:59 hmmmm it's a really fantastic way to debug things
01:00 paramat another strange thing, i start a world and the f5 info is showing already
01:00 hmmmm that's the default view mode if you compiled as debug
01:00 paramat aha i'll check my build
01:01 paramat and debug is slower too
01:01 hmmmm i run as debug with -O0 all the time
01:02 hmmmm but then again, i can whereas most people aren't able to because i have a cpu that's still top of the line 3 years later
01:02 hmmmm :p
01:03 paramat last night i compiled a debug build, then recompiled back to a normal build, but not a clean compile
01:03 paramat it may be stuck in debug somehow
01:03 hmmmm make clean first
01:06 paramat i've only ever recompiled letting it automatically detect what files have changed, how do i force a clean build?
01:07 hmmmm make clean && rm -rf CMakeFiles && cmake -DRUN_IN_PLACE=1 -DCMAKE_BUILD_TYPE=Release . && make -j16
01:07 hmmmm oh
01:08 hmmmm rm CMakeCache.txt too
01:08 paramat aha thanks
01:15 est31 I think other things have to be removed too
01:15 est31 ah no you remove CMakeFiles already
01:20 paramat okay first i'll 'make clean' then manually delete those files, then 'run in place 1, build type release' then 'make -j2'
01:35 paramat okay thanks that fixes it, sorry for false alarm, i was stuck in a debug build
01:38 hmmmm yeah good
01:38 hmmmm you did have me scared for a bit.
01:39 hmmmm there really were changes to mesh making without shaders enabled, but when reviewing the code I reckoned it wouldn't make a difference in terms of performance
01:42 paramat im happy it didn't, as i would love to see realtime shadows added
01:43 hmmmm it's pretty graphics intensive
01:44 paramat pity. most shaders don't interest me. perhaps it could be done with a voxel resolution, really low-res..
01:46 hmmmm heh
01:46 hmmmm that's actually how most games do it
01:46 hmmmm you might notice that real time graphics seem pixelated in those crazy cool FPSes with nice graphics
01:46 hmmmm real time shadows* i meant to say
01:47 paramat the shadows are yes
01:47 hmmmm so minetest is actually at quite an advantage, if we make sure the only things that can cast shadows are non-meshnode-nodes
01:50 paramat our current lighting is voxel resolution so low-res shadows would suit our style too
01:51 NakedFury joined #minetest-dev
01:51 nolsen Maybe I should attempt to work at #2788, but I'll probably give up instantianously.
01:51 ShadowBot https://github.com/minetest/minetest/issues/2788 -- Tab complete feature.
01:51 nolsen But where would I start?
01:53 est31 nolsen, edit src/guiFormspecMenu.cpp
01:53 est31 add a new formspec element
01:53 nolsen I'm not sure how much C++ knowledge would I require.
01:53 est31 called listselect
01:54 nolsen Guess I need to start PITA erm I meant Linux VM
01:55 est31 ah dammit thats the tab complete feature
01:55 nolsen It does has name tab support, but it's flawed a bit.
01:55 est31 flawed in which way?
01:56 nolsen erm
01:56 * nolsen googles flawed to make sure he used the right word
01:56 nolsen It has a few problems like unneccesary space.
01:56 nolsen Though that's just me.
01:56 nolsen I can fix that
01:57 nolsen If Virtualbox will ever start...
02:01 paramat hmmmm, i rewrote noise defined biome blend, now it compiles and runs but no terrain appears, player falls like in singlenode, i suspect my cargo-culting of getting blend noiseparams and seed into NoisePerlin2D() https://github.com/paramat/minetest/tree/noiseblend
02:02 nolsen est31: Though it doesn't ignore the space like /teleport Foo*space*, and thinks I'm talking about someone named as Foo*space*, instead of Foo.
02:04 nolsen Trying to see where's the current username tab completion is located.
02:05 est31 nolsen, what do you mean? entering /teleport foo*space* and then pressing tab yields nothing happening
02:06 nolsen If I do /teleport f, then tab, It becomes foo*space*
02:06 est31 ah
02:06 nolsen If I just press enter, it doesn't see the player.
02:06 hmmmm why are you adding MapgenParams *params to all the different mapgen methods
02:06 est31 "see"?
02:06 nolsen The current username tab is located at https://github.com/UltimateNate/minetest/blob/master/src/guiFormSpecMenu.cpp right?
02:06 nolsen erm
02:06 nolsen src/guiFormSpecMenu.cpp
02:07 est31 no
02:07 hmmmm and you never even modify makeChunk calls to pass along params...
02:07 hmmmm how do you expect this to possibly work?
02:07 nolsen oh, where is it then?
02:07 est31 nolsen, src/guiChatConsole.cpp
02:07 hmmmm come on paramat i expected you to try a teeny bit harder :/
02:08 hmmmm also, isn't the code style supposed to be one continuation indent, not two?
02:09 nolsen est31: Found it, I have no idea will my method work but what the hell, I'm bored.
02:12 paramat i had no idea what i was doing =/
02:12 paramat i'll check indents, i thought it was 1 for if and for, 2 elsewhere
02:13 nolsen Ugh, my id_rsa broke
02:14 hmmmm anyway
02:14 hmmmm all you need to do is add MapgenParams np_biome_humidity_blend and the other one to the class definition
02:14 hmmmm set them in the constructor
02:14 hmmmm then use those members instead of passing arguments
02:16 paramat okay. and yes indents are 1 unless that causes confusion with conditionals, will fix
02:16 jordan4ibanez joined #minetest-dev
02:19 nolsen est31: https://github.com/UltimateNate/minetest/commit/40830e3482737a7b8cd672268e4101fa78a439db Would that work?
02:20 nolsen Going to take a while to compile
02:21 nolsen Wait I think I see it
02:25 est31 nolsen, thats all non-semantic changes
02:57 nolsen Ugh didn't work, Not sure what am I doing wrong.
03:08 est31 nolsen, the changes you made in github dont change any functional code
03:19 nolsen I give up.
03:24 nolsen What's wchar_t wc = L'_';
03:26 nolsen Maybe I'm looking at the wrong file, This seems to not put it together.
03:35 paramat wow finally i got it working :P thanks for the help
03:36 nolsen I finally got make to successfully work
03:36 nolsen umm
03:37 nolsen 2015-06-15 22:36:38: ERROR[main]: A serialization error occurred:
03:37 nolsen 2015-06-15 22:36:38: ERROR[main]: ServerEnvironment::loadMeta(): EnvArgsEnd not found!
03:38 nolsen 2015-06-15 22:36:38: ERROR[main]: The server is probably  running a different version of Minetest.
03:40 nolsen All I edited is chat.h
03:40 est31 thats an error with the envargs file
03:40 est31 or however its called
03:40 est31 remove it
03:40 est31 or make a new world and try there
03:43 nolsen TIL that Minetest has an full screen option
03:43 nolsen and my skin is crappy
03:46 nolsen Time till give up: Now
03:46 nolsen Maybe I could try getting rid of bool backwards.
03:55 nolsen YES!!!!!
03:55 nolsen I DID IT
03:55 nolsen I FUCKING DID IT.
03:55 nolsen It wasn't the bool backwords thing.
03:57 nolsen Commiting and pull requesting...once firefox responds
04:03 nolsen Now trying to figure how to detect : isin replacement_index
04:05 nolsen Could anyone write me a quick line for that? :P
04:06 est31 what do you want to do?
04:06 nolsen est31: Well, I fixed (sorta) the picky command errors like I been talking about earlier, now I need to have a space for none commands, like foo:
04:07 nolsen s/none/non
04:07 nolsen So I need to detect if : isin the message, then add a space.
04:08 est31 look at util/string.h
04:08 est31 it has many useful functions
04:08 est31 usually is in string is done by using indexof string other string
04:09 nolsen Uhh, I don't see string.h
04:09 hmmmm :\
04:10 hmmmm nolsen, why don't you take some time to acquaint yourself with minetest's codebase and perhaps the language
04:10 hmmmm it's okay if you don't make any meaningful contributions right away
04:10 hmmmm there's no pressure
04:10 nolsen I'm doing this because it's the easiest thing that I could do.
04:11 nolsen Oh, src/util/string.h
04:13 nolsen Wait, I think I fixed it.
04:14 est31 I need a good name for a formspec element
04:14 est31 basically you should do ELEM[<inventory location>;<list name>]
04:14 est31 and that multiple times
04:14 nolsen est31, hmmmm: It works now! I'm commiting and sending pull request.
04:15 est31 then when you hold shift and click in some inventory at some stack thats covered by ELEM, stuff is sent to the next inventory specified by ELEM
04:16 est31 most often to be used for chest-to-inventory movements
04:16 est31 perhaps for other things too
04:16 est31 also I wonder whether to allow only two ELEM specifiers.
04:17 est31 so does anybody have a good name for ELEM?
04:22 nolsen Creating pull request.
04:22 nolsen But paranoia says not to.
04:23 est31 ?
04:23 est31 we will check your PR
04:23 est31 and if it looks good merge it
04:24 est31 if not, we will kill you with fire
04:24 est31 and roast you
04:24 est31 then we will eat you
04:24 nolsen Make sure you use ketchup
04:24 nolsen jk
04:25 * est31 writes down notes
04:25 nolsen est31: This PR will probably be a temporary fix.
04:25 est31 ?
04:25 nolsen #2793
04:26 ShadowBot https://github.com/minetest/minetest/issues/2793 -- Fixed space issues. by UltimateNate
04:26 nolsen "The Travis CI build is in progress." Probably will fail because Travis CI hates me.
04:26 nolsen I never got that damn thing to work on anything.
04:28 est31 ok some things:
04:29 est31 1. pull request titles and commits are present tense
04:29 est31 (not a biggie)
04:29 nolsen Meh, it was my first pull request I ever done I think.
04:29 nolsen I don't think I really contributed to a project.
04:29 est31 2. a pull request should only consist of one commit if it changes one thing
04:29 est31 its quite good for a first PR :)
04:30 est31 regarding 2: there are projects which don't like you to git commit --amend && git push -f
04:30 est31 we want you to do it
04:30 nolsen doesn't git push -f fucks things up?
04:30 est31 so if sb points something out in a PR, then dont add commits but update the commit itself
04:31 nolsen forget calling -f force, it should be -f, fuck.
04:31 est31 hehe
04:31 hmmmm profanities aside, git rebase -i HEAD~4
04:31 hmmmm pick, squash, squash, squash.
04:31 nolsen git was difficult at the first times to use it.
04:32 nolsen Only thing I use git is to push, pull, clone, init, whatever it says in a README or whatever.
04:32 nolsen commit.
04:35 nolsen est31: So will my PR be accepted?
04:39 nolsen Fixed what needed to be fixed.
04:41 est31 nolsen, we will decide this in 4 months after the 0.4.13 release
04:41 est31 the PR has to rot a bit
04:41 nolsen lol
04:42 nolsen and get outdated
04:42 * nolsen nodes
04:42 nolsen er
04:42 * nolsen nods
04:42 nolsen est31: Will it seriously take 4 months?
04:43 est31 no
04:43 est31 with current PR merging speed, one day, perhaps I look at it even now
04:43 est31 its a very small commit
04:44 est31 for those you dont need 2 coredevs to agree
04:45 nolsen Soon I might add command completion.
04:45 nolsen But that'll be soon
04:55 nolsen I'm curious will Minetest crash if I connect with ÀÀÀÀÀ ®®®®©¶
04:55 nolsen Does anyone see a space in ÀÀÀÀÀ ®®®®©¶
04:58 nolsen Nope, It doesn't.
05:03 NakedFury joined #minetest-dev
05:05 nolsen Now I'm counted at a contributor :D
05:09 hmmmm haha... not so fast
05:09 hmmmm it takes more
05:11 est31 I guess I'll merge 2793, squashed and the artifact removed
05:13 RealBadAngel joined #minetest-dev
05:17 hmmmm graet
05:17 hmmmm great*
05:18 est31 ahhh dammit
05:18 est31 wuzzies commit is in there too
05:19 est31 meh, merge it too
05:19 est31 will write "merged by accident"
05:22 hmmmm oh well
05:23 hmmmm it's not worth reverting
05:40 Hunterz joined #minetest-dev
05:49 Player_2 joined #minetest-dev
05:54 paramat left #minetest-dev
06:02 neoascetic joined #minetest-dev
06:17 est31 wow this irrlicht is cool
06:17 est31 it knows what I want before I want it
06:17 est31 in fact, a highly unlikely feature is added to an API i have access to
06:17 est31 meaning I dont have to use other methods
06:17 est31 really nice
06:17 est31 quite handy
06:44 kilbith joined #minetest-dev
06:52 NakedFury joined #minetest-dev
06:54 Darcidride joined #minetest-dev
07:09 Amaz joined #minetest-dev
07:11 paramat joined #minetest-dev
07:27 paramat sfan5, please could you review this sometime? https://github.com/minetest/minetest_game/pull/539 -- Fix is_ground_content settings for nodes
07:33 chchjesus_ joined #minetest-dev
07:33 Calinou joined #minetest-dev
07:40 err404 joined #minetest-dev
07:48 paramat left #minetest-dev
08:05 Yepoleb joined #minetest-dev
08:18 cib0 joined #minetest-dev
08:59 est31 #2795
08:59 ShadowBot https://github.com/minetest/minetest/issues/2795 -- Add list-rings by est31
09:00 est31 a real feature this time
09:00 est31 and useful
09:10 Dartmouth joined #minetest-dev
09:17 RealBadAngel joined #minetest-dev
09:55 proller joined #minetest-dev
10:51 kilbith celeron55, http://www.minetest.net/  <- all pages are empty
10:54 kilbith s/all/most of
11:01 celeron55 lol wut
11:02 celeron55 oh oops
11:03 Wayward_Tab joined #minetest-dev
11:06 celeron55 it's fine now
11:06 celeron55 the server ran out of disk space due to a silly mistake in some custom scheduled tasks
11:08 celeron55 (and dokuwiki likes to clear all of its pages when it runs out of cache space)
11:19 cib0 joined #minetest-dev
11:19 Darcidride joined #minetest-dev
12:05 est31 joined #minetest-dev
12:46 est31 gonna push #2795
12:46 ShadowBot https://github.com/minetest/minetest/issues/2795 -- Add list-rings by est31
13:02 LittleJoe joined #minetest-dev
13:29 proller joined #minetest-dev
13:57 Darcidride joined #minetest-dev
13:57 hmmmm joined #minetest-dev
13:58 Darcidride joined #minetest-dev
14:00 Darcidride joined #minetest-dev
14:11 CraigyDavi joined #minetest-dev
14:49 nolsen Aww it was closed.
14:50 nolsen est31: What about 3ae8b92?
14:51 est31 thats the commit your pr was merged :)
14:52 est31 RealBadAngel, what do you think, do we need a cablelike drawtype?
14:54 nolsen oh, yay.
14:54 nolsen Now I wonder what issue I should do next.
14:58 nolsen Now I need to delete my fork and create another one since it's outdated.
14:59 est31 you dont have to
14:59 nolsen ?
14:59 est31 what I do is add the upstream repo as remote
14:59 est31 then I do git remote update
15:00 est31 then I do git reset --hard upstream/master
15:05 nolsen Going to work on command completion by tab, I should be able to just copy and replace some things.
15:05 est31 how do you plan to send the command list to the client?
15:07 nolsen well, is there already a command list?
15:07 est31 no
15:08 est31 only on the server
15:08 nolsen well I need to use the server's commandlist then.
15:08 est31 yup
15:08 est31 so best thing you download the command list, and do the completion on the client
15:09 est31 I guess it would be a great start if you manually specified a test command list showcasing some features
15:09 nolsen Yep, though it'd be a bit harder, but it shouldn't be too hard.
15:09 est31 harder than?
15:10 nolsen harder than what I done last night.
15:10 nolsen on my PR
15:10 est31 yea
15:10 est31 you will need some "language" I guess
15:10 est31 we already have nick completion
15:11 nolsen I'm a pythoneer, Barely know C++
15:11 est31 If you have questions, ask
15:11 est31 I can write the sending of the list, you only have to do the completion part
15:12 nolsen I guess I don't need to update my fork since none of the files I'm going to use are probably not outdated.
15:12 nolsen est31: sure.
15:13 cib0 joined #minetest-dev
15:14 nolsen est31: I wonder, what editor you use?
15:14 nolsen I'm using gedit
15:15 est31 I'm using kate
15:17 Calinou Atom :-D
15:18 nolsen I need to replace mine because gedit is being glitchy.
15:18 nolsen I'm using lubuntu on Virtualbox to edit and run Minetest.
15:19 Calinou then use Geany
15:20 nolsen est31: You could write on the sending of the list, because I have no idea how to do that.
15:27 est31 nolsen, when sending something over the network, or generally when changing network protocol, you have to make sure old servers work with old clients and vice versa
15:27 est31 what I suggest doing is the following
15:27 est31 you create a new opcode
15:28 est31 TOCLIENT_CHAT_COMMANDS
15:28 est31 then the fun begins
15:31 est31 you will need a c++ method both server side and client side
15:32 est31 look at src/network/clientpackethandler.cpp
15:32 est31 and serverpackethandler
15:32 est31 and .h files too
15:32 est31 err
15:32 est31 no .h files
15:32 est31 they are managed in src/server.h and src/client.h
15:34 est31 then you should edit src/networkprotocol.h
15:34 est31 and add your opcode
15:34 Calinou|mobile joined #minetest-dev
15:34 est31 in the toserver list
15:36 est31 then you have to edit serveropcodes.cpp and clientopcodes.cpp inside src/network
15:36 est31 to add your handler methods
15:37 Amaz game#542
15:39 Amaz Does ShadowBot not have that functionality any more? :( https://github.com/minetest/minetest_game/pull/542
15:48 AnotherBrick joined #minetest-dev
15:52 H-H-H joined #minetest-dev
15:52 rubenwardy joined #minetest-dev
16:20 proller joined #minetest-dev
16:21 ElectronLibre joined #minetest-dev
16:22 selat joined #minetest-dev
16:32 Robert_Zenz joined #minetest-dev
16:48 cib0 joined #minetest-dev
16:51 james22 joined #minetest-dev
17:17 nolsen Maybe Minetest should have a minimum internet speed limit on servers and clients, to reduce ping timeouts.
17:18 Calinou that makes no sense
17:18 Calinou network throughput != latency
17:20 nolsen Probably lost any messages, my internet died.
17:20 Calinou <Calinou> that makes no sense
17:20 Calinou <Calinou> network throughput != latency
17:23 nolsen Then have a latency check?
17:24 nolsen Then have the server ease the walking too fast errors because of high latency, and ping timeouts, for the user.
17:24 Krock joined #minetest-dev
17:26 nolsen Calinou: Though I thought network speed also matters.
17:26 nolsen Because it needs to speed to send it
17:30 jordan4ibanez joined #minetest-dev
17:30 Calinou you need a bit of speed, but even 1 Mb/s download is enough
17:30 Calinou nolsen, we already use a lag pool system for speed checks
17:30 Calinou personally I'd remove speed checks entirely...
17:30 est31 I've successfully played minetest on a 64 kbit connection
17:31 Calinou disable_anticheat = true disables all anticheats, not just the speed one
17:31 est31 it got worse the more players joined
17:31 est31 but worked more or less
17:31 est31 heck, I've managed to clean up spawn grief with that bad connection
17:36 Hunterz joined #minetest-dev
18:13 nolsen Well it's annoying when people keep timeing out my server
18:13 nolsen I wonder if you get ping timeout on this server 188.165.3.51
18:13 nolsen you two
18:32 cib0 joined #minetest-dev
18:34 * est31 pushes in 10 minutes: https://github.com/est31/minetest/commit/7a90b31b30897c99adcbef0a6ea0c69f1c99c2b1
18:38 Robert_Zenz joined #minetest-dev
18:50 nolsen est31: I think I found a bug with the bucket
18:50 nolsen wait, is the bucket part of minetest, or is that a mod?
18:50 est31 mod
18:51 nolsen oh ok
18:58 Jbb joined #minetest-dev
18:59 jordan4ibanez_ joined #minetest-dev
19:02 darkrose joined #minetest-dev
19:03 darkrose left #minetest-dev
19:03 err404 joined #minetest-dev
19:14 MinetestForFun joined #minetest-dev
19:41 err404 joined #minetest-dev
19:46 VanessaE joined #minetest-dev
20:02 VanessaE_ joined #minetest-dev
20:17 cvtsx joined #minetest-dev
20:33 cib0 joined #minetest-dev
20:50 ElectronLibre left #minetest-dev
21:02 Builderman joined #minetest-dev
21:03 werwerwer joined #minetest-dev
21:19 proller joined #minetest-dev
21:32 soup_ joined #minetest-dev
21:33 soupertrooper joined #minetest-dev
21:34 soupertrooper hello, just looking at the source and two questions where would I go to look at the input and how would I go about starting gamepad input on android?
21:35 soupertrooper input handler that is
21:39 soupertrooper left #minetest-dev
21:43 proller joined #minetest-dev
22:17 Wayward_Tab joined #minetest-dev
22:28 jordan4ibanez joined #minetest-dev
22:32 paramat joined #minetest-dev
22:34 cib0 joined #minetest-dev
22:38 Niebieski joined #minetest-dev
22:39 crazyR_ joined #minetest-dev
22:39 crazyR_ joined #minetest-dev
22:55 johnnyjoy joined #minetest-dev
22:56 Aaron1011 joined #minetest-dev
23:00 johnnyjoy joined #minetest-dev
23:05 Aaron1011 joined #minetest-dev
23:26 Aaron1011 joined #minetest-dev
23:26 Aaron1011 joined #minetest-dev
23:42 paramat left #minetest-dev
23:55 proller joined #minetest-dev

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