Minetest logo

IRC log for #minetest-dev, 2014-07-02

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

All times shown according to UTC.

Time Nick Message
00:00 sapier left #minetest-dev
00:06 ShadowNinja Comments?  Fixes the server list announcing ignoring --port: http://ix.io/de7/diff
00:06 ShadowNinja Perhaps it should be stored in the connection.
00:10 Megaf fix it already =/
00:44 smoke_fumus joined #minetest-dev
02:10 kaeza joined #minetest-dev
03:21 us`0gb joined #minetest-dev
03:24 Weedy joined #minetest-dev
03:24 Weedy joined #minetest-dev
04:31 kaeza joined #minetest-dev
04:32 kaeza1 joined #minetest-dev
04:50 Hunterz joined #minetest-dev
05:33 kahrl joined #minetest-dev
05:43 kaeza joined #minetest-dev
05:57 grrk-bzzt joined #minetest-dev
05:57 grrk-bzzt joined #minetest-dev
05:59 LemonLake joined #minetest-dev
06:16 Garmine joined #minetest-dev
06:34 darkrose joined #minetest-dev
06:34 darkrose joined #minetest-dev
07:03 nore joined #minetest-dev
07:14 Hunterz joined #minetest-dev
07:23 kaeza joined #minetest-dev
08:17 Calinou joined #minetest-dev
09:32 Taoki joined #minetest-dev
09:49 VanessaE joined #minetest-dev
09:53 PenguinDad joined #minetest-dev
09:55 ImQ009 joined #minetest-dev
10:28 us`0gb joined #minetest-dev
10:43 proller joined #minetest-dev
10:55 Krock joined #minetest-dev
11:32 Krock servers.minetest.net/announce returned status code 100
11:32 Krock 100 Continue
11:32 Krock !?
11:36 Krock gone.
11:36 Krock left #minetest-dev
11:39 Megaf joined #minetest-dev
12:02 PenguinDad joined #minetest-dev
12:08 Megaf__ joined #minetest-dev
12:17 Jordach joined #minetest-dev
12:53 proller joined #minetest-dev
13:11 proller joined #minetest-dev
13:13 hmmmm joined #minetest-dev
13:28 vifino joined #minetest-dev
13:30 Taoki[mobile] joined #minetest-dev
13:33 daswort make throws a compiling error: http://paste.ubuntu.com/7736795/
13:35 VanessaE make clean and try again?
13:36 nore anyone against merging #1437? (I'll merge it in 1 hour if no one complains)
13:41 PilzAdam joined #minetest-dev
14:08 proller joined #minetest-dev
14:24 LemonLake joined #minetest-dev
14:44 nore thoughts on https://github.com/minetest/minetest/pull/1438 ?
14:44 nore anyway, merging #1437 now since no one complained
15:08 Garmine joined #minetest-dev
15:20 deltib joined #minetest-dev
15:22 Jordach joined #minetest-dev
15:27 Calinou joined #minetest-dev
15:47 deltib joined #minetest-dev
15:51 zat joined #minetest-dev
16:08 Hunterz joined #minetest-dev
16:11 Calinou https://github.com/minetest/minetest/issues/200 → close maybe? maybe fixed when using shaders…
16:14 Calinou https://github.com/minetest/minetest/issues/233 → fixed maybe? it seems to update properly, except maybe in creative mode where water flowing is maybe disabled (at least I heard of)
16:14 Calinou https://github.com/minetest/minetest/issues/244 → close, invalid? I don't see why dead players shouldn't move and get_objects_inside_radius should return everything probably
16:24 Piggybear87 joined #minetest-dev
16:25 Piggybear87 left #minetest-dev
16:41 grrk-bzzt joined #minetest-dev
16:42 werwerwer_ joined #minetest-dev
16:44 proller joined #minetest-dev
17:32 asl joined #minetest-dev
17:40 cg72 joined #minetest-dev
17:44 tomreyn joined #minetest-dev
18:04 sapier joined #minetest-dev
18:12 RealBadAngel heh, hi sapier
18:12 RealBadAngel why so many questions?
18:12 sapier questions?
18:12 RealBadAngel close maybe? fixed maybe?
18:13 sapier sorry don't know :-)
18:13 RealBadAngel lost self confidence or what?
18:13 sapier nope just error reports without real information in it
18:14 sapier #1435 those traces look like multiple bugs
18:20 RealBadAngel that also applies to lighting/shading faces
18:20 RealBadAngel from what i can see, thats irrlicht bug
18:21 RealBadAngel normals seems to be sometimes incoreclty calculated
18:21 RealBadAngel but when, on what conditions, only santa clause knows
18:22 sapier well the tested version was prior my fix for ambient occlusion too so no idea if it's same or not
18:22 RealBadAngel its related
18:22 RealBadAngel and you cant do shit about it
18:22 RealBadAngel normals are fucked up on the engine side
18:22 sapier well it has to be as good as before that's our minimum requirement
18:23 RealBadAngel nore figured out way to go around
18:23 RealBadAngel pick different default value
18:24 RealBadAngel but thats temporary
18:24 RealBadAngel that math has to be checked
18:24 RealBadAngel normal vector is kinda basis for whatever i touch within shaders
18:25 sapier if it's as good as before it's fine, right now those crashes are what I consider to be a potential problem
18:25 RealBadAngel mine problem are not crashes, just incorect shading
18:26 RealBadAngel but still a fuckin bug :)
18:26 sapier well some incorrect shaders wont block a release but crashes would
18:27 werwerwer joined #minetest-dev
18:32 RealBadAngel we have found a way
18:32 RealBadAngel i also solved transparency materials issue
18:33 RealBadAngel but both solutions are temporary
18:34 RealBadAngel minetest is abusing irrlicht in worst possible way
18:34 Calinou every game abuses every engine
18:34 sapier 4 days prior release changing everything isn't an option ;-)
18:34 RealBadAngel but we made it bad to the bone
18:35 RealBadAngel im not talkin this release
18:35 sapier there are more ugly things in enginge don't try to change everything at once but step by step
18:35 RealBadAngel but about the future
18:35 RealBadAngel makin whole world a single scene node was worst decision ever
18:35 proller joined #minetest-dev
18:36 sapier no need to repeat commonly accepted facts ;-)
18:36 RealBadAngel until i revert all that shit
18:36 RealBadAngel we wont go any further
18:37 sapier you could be a little bit more precise about what you plan to do ;-)
18:37 RealBadAngel this is real blocker
18:37 sapier stop complaining start explaining rba
18:37 RealBadAngel we cant use the engine with it
18:37 RealBadAngel materials are fucked up
18:37 RealBadAngel no sorting
18:37 sapier << not continuing to read
18:38 RealBadAngel hehe
18:38 RealBadAngel we just have to go back
18:38 RealBadAngel node = node
18:38 sapier I don't wanna know what doesn't work everyone complains constantly
18:38 RealBadAngel then everything will start to work out of the box
18:38 RealBadAngel i will be able to use all the materials
18:39 sapier sorry but something mysticaly starting to work by a single chane never was true ;-)
18:39 RealBadAngel water surface will become aviable
18:39 RealBadAngel with reflections
18:39 RealBadAngel also we will be able to use irrlicht bullet
18:39 sapier make a branch show it to us, no one will stop you
18:39 Calinou I hate when people complain about software limitations they don't understand
18:40 RealBadAngel theres a branch already
18:40 Calinou if people actually understood this, software choices would be insanely more sensible
18:40 RealBadAngel which is light years ahead
18:40 RealBadAngel its called Build A World...
18:40 sapier rba url?
18:40 sapier pull request?
18:40 RealBadAngel its a project
18:41 RealBadAngel closed source, built on Irrlicht engine
18:41 sapier that's as helpfull as knowing there's water on moon
18:42 Calinou in before it changes so much of Irrlicht it's not even Irrlicht anymore (see STK)
18:42 sapier calinou sorry but what are you talking about? ;-)
18:42 Calinou <RealBadAngel> its called Build A World...
18:42 RealBadAngel https://www.youtube.com/watch?v=jWR-RKGY8Pw
18:42 RealBadAngel for example
18:43 sapier RealBadAngel: since you're here today you're just complaining what's up? ;-)
18:44 sapier noone ever told minetest is perfect ... most parts are way more near to crap then to perfect ;-)
18:44 sapier -most
18:45 sapier + some
18:45 RealBadAngel i think, why the fuck we cant go the right way
18:45 sapier did anyone say we can't?
18:45 RealBadAngel yes, in most certain cases
18:45 Calinou prefer evolution to revolution; if you want a task to do, fix code consistency and styling
18:45 RealBadAngel built-in logic is one of them
18:45 Calinou it's a long one, a boring one, but rewarding
18:45 sapier but you have to keep in mind client lighting is only one part
18:46 sapier you need server to know about light too
18:46 RealBadAngel ofc
18:46 RealBadAngel the light level
18:46 Calinou ^ please do remember this when you think about alternate lighting systems
18:46 RealBadAngel rest of the code can be safely trashed
18:47 RealBadAngel for hardware lights you dont even need shaders
18:47 RealBadAngel engine will do that
18:47 sapier I still don't understand what you wanna tell
18:47 RealBadAngel calinou already used that word for me
18:48 RealBadAngel continously fixin broken stuff is a road to nowhere
18:48 sapier revolution? you're free to try it but I fear you'll be even more annoyed if you're working for weeks and still lack of result
18:49 RealBadAngel we need a revolution, and do gfx again
18:49 sapier that's your believe
18:49 RealBadAngel cmon, seeing caves being rendered all the time
18:49 sapier you can do same as pilzadam
18:49 RealBadAngel is not even worth a laugh
18:49 sapier he wrote a irrlicht replacement
18:49 RealBadAngel its a bad joke
18:49 sapier yes but it's a lot of time spent for it
18:50 sapier only he can tell but I'd guess it's about 2 months of real time spent to it
18:50 RealBadAngel do you realize that every single mesh is being rendered?
18:50 RealBadAngel and thats why mt is so slow?
18:50 RealBadAngel all the visible and invisible stuff is RENDERED
18:50 sapier can you stop complaining and start explaining as I already said?
18:51 Calinou MT is far from slow.
18:51 sapier except pa here's noone capable of helping you, that's sad but reality
18:51 RealBadAngel im telling the facts
18:51 Calinou it's fast, even faster without shaders
18:51 Calinou very smooth (small amount of batches?)
18:52 RealBadAngel for newly created world
18:52 Calinou get some PC that can render 60 FPS constantly (not so hard); compare Minecraft and Minetest on these
18:52 RealBadAngel an empty one
18:52 Calinou no, one with decent amount of builds
18:52 Calinou of course it's also your problem if you spam allfaces nodes :P
18:52 RealBadAngel sure, login on Ve servers
18:52 Calinou (this is why jungle leaves are glasslike in Carbone, and use a dense texture to compensate)
18:52 Calinou spam of complex nodes is the issue
18:52 sapier you tell them back and forward about 2 months now RealBadAngel using that time you'd already have a signigicant part of what you wanna do
18:52 Calinou pipeworks are fairly insane when it comes to nodebox count
18:53 RealBadAngel all the recent mods does that
18:53 Calinou my mods have 5 nodeboxes at most for circular saw, an uncommon block
18:53 Calinou 3 at most for Stairs+ nodes
18:53 RealBadAngel have you saw proper mesecon wires?
18:53 Megaf__ hi
18:53 RealBadAngel that can climb the walls etc?
18:53 Calinou each of them is a few nodeboxes
18:53 RealBadAngel they do need over 1000 nodeboxes
18:53 Calinou if you want high performance with lots of them, make them raillike
18:54 RealBadAngel 1 fucking thousand
18:54 sapier I'll leave this duscussion now.
18:54 Calinou Jeija could have simplified wires by not putting a central dot
18:54 RealBadAngel instead of ONE
18:55 RealBadAngel and you guys still thinkin that the code that alllows that is not needed
18:55 sapier calinou what are you arguing about, if RealBadAngel wants to do his don't take his time let him do it
18:56 RealBadAngel sapier, its not that im angry or something
18:57 RealBadAngel but we cant all the time take care of older versions
18:57 RealBadAngel we have to move on
18:57 RealBadAngel if somethins is wrong, leave it and make it better
18:58 sapier RealBadAngel: noone stops you from rewriting the graphics
18:58 sapier there's not even a compatibility issue
18:58 RealBadAngel from my point of view it is
18:58 sapier graphics are absolutely independent from client server communication
18:58 sapier so what are you complaining about?
18:58 RealBadAngel drop the node definition
18:59 sapier no
18:59 sapier how is node related to graphics?
18:59 RealBadAngel that breaks all the gfx
19:00 sapier that's like telling definition of length breaks golden gate bridge
19:00 RealBadAngel :)
19:00 nore ShadowNinja, what do you mean in your comment of #1438?
19:00 RealBadAngel not like that
19:00 sapier a node is a abstract definition how is this supposed to break a graphical representation?
19:00 RealBadAngel we are collecting meshes into table
19:01 RealBadAngel and making a node out of them
19:01 RealBadAngel to show a world
19:01 RealBadAngel so, we cant have physics
19:01 RealBadAngel z-sorting
19:01 sapier so yowe can't have physics anyway
19:01 RealBadAngel irrbullet, just merge it :P
19:02 sapier ok maybe you should call a scenenode scenenode to not cause confusion to nodes
19:02 RealBadAngel node shall be a single mesh
19:02 sapier you're free to do that
19:02 RealBadAngel with own properties
19:03 sapier try it but I guess you're causing even big machines to run out of memory
19:03 sapier I'd be glad if you proove me wrong
19:03 RealBadAngel shall i again show you build a world?
19:03 RealBadAngel they did it proper way :P
19:03 sapier you already said it's closed source
19:03 sapier so you don't know what tricks they did
19:03 sapier and usually it IS tricks if someone claims "I solved it"
19:04 RealBadAngel yup, they use nasa mainframes to render the gameplay :P
19:04 sapier as I said you don't know how they did it ... you'd not even know if they did it similar to we do
19:04 RealBadAngel they made game like we do
19:04 RealBadAngel they do use irrlicht
19:05 RealBadAngel what else do you need to know?
19:05 sapier and that's all you know
19:05 sapier how they did it
19:05 RealBadAngel not
19:05 RealBadAngel they did it better
19:05 sapier sorry but I don't see any progress in this discussion
19:05 PenguinDad joined #minetest-dev
19:05 RealBadAngel while i can
19:06 sapier you wont get an agreement to break everything for a vague hope that this is better
19:06 RealBadAngel when something can be done better, its our turn to think on how
19:06 sapier and I don't even believe you HAVE to break everything to get it done
19:06 sapier imho you could do it without significant changes to non graphics code
19:06 RealBadAngel not really everything
19:07 RealBadAngel just the main display loop
19:07 RealBadAngel collect meshes
19:07 sapier and noone will stop you from changing that code ... except you push alpha code to master
19:07 RealBadAngel i am at the doorstep to do so, i mean changing the way
19:08 RealBadAngel i started with new drawtype
19:08 sapier and in case your code is really faster I don't think anyone will stop this from beeing merged
19:08 RealBadAngel mesh
19:08 sapier but if it's same speed or slower ... sorry but then your work may have been useless
19:09 sapier let's hope it's not gonna be that way
19:09 PenguinDad joined #minetest-dev
19:10 RealBadAngel idk if it will be 2.345 frames per second slower or faster
19:11 RealBadAngel but i am sure it will be done proper way
19:11 sapier android has 12 fps I do care about 20% fps decrease
19:11 sapier a proper solution beeing horribly slow isn't an option too ;-)
19:11 PenguinDad joined #minetest-dev
19:11 RealBadAngel trashing all the current meshes code will propably bring back more fps you can even imagine
19:12 sapier as I said if it's faster that'd be best to happen
19:12 RealBadAngel removing shaders compilation out of runtime brought me 50% gain
19:12 RealBadAngel and we are still full of such nonsenses
19:12 sapier well it's good to have a fast high quality option but don't forget about slow machines
19:13 RealBadAngel asking core 50 times a second, or whatever ur framerate is, about texture  names is wise?
19:14 sapier RealBadAngel: why do we discuss about maybe things, if you believe your way of doing is faster do it, if you don't let it be
19:15 RealBadAngel you know, when i was coding for demoscene i used to count cycles
19:15 RealBadAngel to get faster code
19:16 RealBadAngel now i can see fileacess in rendering code
19:16 RealBadAngel monthy python at its best
19:18 sapier I appreciate your experience RealBadAngel and wont defend crap like that but plz keep in mind you're not writing for a exactly specified hardware thus have to keep code generic ;-)
19:19 RealBadAngel ofc i want it to be generic
19:19 RealBadAngel irrlicht compatible in the first place
19:20 RealBadAngel so for example i could use engine bulit in stuff
19:22 sapier no ... don't use irrlicht internals but only the api
19:24 RealBadAngel you think we are using something outta here?
19:24 RealBadAngel we are using irrlicht but we managed to mess it up
19:25 RealBadAngel thats even impossible to use two kinds of transparent materials at the same time
19:25 RealBadAngel using one disables the other
19:26 RealBadAngel water flickers like hell, with z-fights
19:26 RealBadAngel its impossible to use tangent space for meshes
19:27 RealBadAngel its impossible to use water surface node
19:27 RealBadAngel we cant use reflections
19:27 RealBadAngel we cant use lighting
19:29 RealBadAngel it that enough?
19:31 RealBadAngel anyway, its a good start
19:31 RealBadAngel i will come back with rebuild of it or nothing
19:31 proller no, good start - rewriting anything from scratch
19:31 RealBadAngel thats a perfect start imho
19:32 RealBadAngel motivation to do it better
19:32 RealBadAngel what i could really appreciate would be a coder to help me
19:32 proller write all big important improvements for last year
19:32 RealBadAngel but if not, i cad do it on my own
19:33 RealBadAngel *can
19:33 RealBadAngel you know, with all the code i wrote
19:33 RealBadAngel i barely touched a surface
19:33 sapier (20:51:24) sapier: except pa here's noone capable of helping you, that's sad but reality
19:34 proller ;)
19:34 RealBadAngel sad or not
19:34 RealBadAngel be it a challenghe
19:35 RealBadAngel im takin it
19:35 sapier (20:52:38) sapier: you tell them back and forward about 2 months now RealBadAngel using that time you'd already have a signigicant part of what you wanna do
19:35 RealBadAngel i wont need 2 months :p
19:35 RealBadAngel revolutions were made way faster before
19:35 sapier (21:08:11) sapier: and in case your code is really faster I don't think anyone will stop this from beeing merged
19:36 RealBadAngel one head to fall, was sometimes enough ;)
19:36 sapier (21:11:54) sapier: a proper solution beeing horribly slow isn't an option too ;-)
19:36 RealBadAngel since we are advanced lets do that one node to fall ;)
19:37 sapier (21:07:23) sapier: and noone will stop you from changing that code ... except you push alpha code to master
19:37 RealBadAngel once done, it wont be alpha
19:37 RealBadAngel it will just work
19:38 sapier (20:58:53) sapier: so what are you complaining about?
19:38 sapier ;-)
19:38 RealBadAngel lol
19:39 RealBadAngel no more comments :)
19:39 RealBadAngel im just gonna do that
19:53 casimir joined #minetest-dev
19:55 kaeza joined #minetest-dev
19:59 sapier RealBadAngel: do ypi know why m_shader_callback isn't dropped in ShaderSource?
20:04 sapier hmm most likely not as it seems to be that way as of merging that code
21:04 EvergreenTree joined #minetest-dev
21:07 Megaf joined #minetest-dev
21:08 Megaf left #minetest-dev
21:18 kaeza forums are giving 502 Bad Gateway
21:18 kaeza celeron55, ^
21:24 celeron55 ack
21:26 sapier https://gist.github.com/sa​pier/1a661cf78f29ca9f03b2 is fixing memory leaks bugfixing or feature addon?
21:34 sapier ok no complains so I'm gonna push the bugfix
21:34 sapier commit 3700
21:44 paramat joined #minetest-dev
21:53 jp joined #minetest-dev
22:01 kahrl /give RealBadAngel http://irrlicht.sourceforge.ne​t/forum/viewtopic.php?t=39956
22:01 kahrl (the first answer especially)
22:15 asl left #minetest-dev
22:30 celeron55 kaeza: 01:28:00 <xyz> fixed it
22:30 kaeza alright thanks
22:59 sapier left #minetest-dev
23:16 proller joined #minetest-dev
23:17 ShadowNinja celeron55: Can you ask xyz to drop the serverlist subdomain?  He wants you to ask since you'll manage it.  (he's online as "xyz")
23:23 ShadowNinja kahrl: They mention that an octree is probably the best way to do it since it's fast but lets you do culling.
23:24 celeron55 sfan5: i pointed servers.minetest.ru to 144.76.208.120/2a01:04f8:01​92:22e5:0000:0000:cf1c:4c46 now
23:24 celeron55 maybe it updates after a while
23:25 celeron55 it's definitely broken right now though, and it's broken for the minetest client too
23:28 celeron55 it looks like it will work after DNS servers propagate the change
23:28 celeron55 (based on 'wget -O- --header="Host: servers.minetest.net" 144.76.208.120')

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