Time  Nick         Message
00:25 VanessaE     can someone please do something about the mouse not "detatching" from the game world when the F10 screen is open?
00:25 VanessaE     (in linux)
01:23 ShadowNinja  celeron55: I think that it would look nicer if the globe icon next to every link was removed. Or is there a way to disable it for specific links?
12:48 sfan5        when did this happen? http://i.imgur.com/Ock09LO.png
13:37 PilzAdam     since the news section is created
13:37 PilzAdam     "Developers" can create topics there
14:37 RealBadAngel hi all
14:37 RealBadAngel hmmmm, how far are you with decorationsdefs?
14:38 hmmmm        still stuck on figuring out a good way to deal with cutoffs
14:38 hmmmm        what i originally had wouldn't work if you approach the ground being generated from above
14:39 RealBadAngel looks like most of developing time here is finding a way ;)
14:39 hmmmm        ?
14:39 hmmmm        oh
14:39 hmmmm        well i haven't been working on it much either, i've been busy with other RL things
14:39 hmmmm        not school though, job stuff
14:40 RealBadAngel me too
14:40 RealBadAngel thats why i can look like "slowed down" a bit
14:40 hmmmm        i think everybody slowed down
14:41 RealBadAngel but when you will need data (trees) for decorations just let me know
14:41 hmmmm        sure
14:42 RealBadAngel next few days, maybe week or two i would like to solve lights problem
14:42 RealBadAngel because all of my shaders work is hittin a wall
14:44 RealBadAngel i learned irrlicht and shaders enough to bring to mt all that fancy shaders stuff
14:44 RealBadAngel but cant without proper lighting
14:45 hmmmm        yeah, proper lighting is long overdue
14:46 RealBadAngel we have a few such steps we need to make
14:46 RealBadAngel lights, tcp/ip, more client side predictions
14:47 RealBadAngel the stuff can bring us to next level
14:47 RealBadAngel and cause mt lookin more as game than a test
14:49 RealBadAngel about shaders, funny fact
14:49 RealBadAngel there are great shaders out there for minecraft
14:49 RealBadAngel but very hard to install and use
14:50 RealBadAngel i had many problems trying them
14:50 RealBadAngel guys from mojang tried to buy the work, but author refused (not sure why, propably they offered too less)
14:51 hmmmm        hrmm
14:51 hmmmm        yeah, i recall there was that shaders mod
14:52 hmmmm        they probably had to use real lighting as well
14:52 RealBadAngel it is still
14:52 RealBadAngel yes
14:52 sfan5        RealBadAngel: are talking about the sonic shaders?
14:52 RealBadAngel yes
14:52 PilzAdam     <hmmmm> i think everybody slowed down <- nope
14:53 RealBadAngel im not sure how the situation looks atm, since i havent run mc for months
14:53 PilzAdam     RealBadAngel, reminder: you still need to fix the lighting bug with facedir nodes
14:53 RealBadAngel yes, i remember
14:53 RealBadAngel but i would like to do that when we fix the LIGHTS
14:54 PilzAdam     dont throw new code in if the old code doesnt work
14:54 RealBadAngel and then PROPABLY i wont have to do anything ;)
14:55 RealBadAngel im not throwing anythin new there
14:55 RealBadAngel my latest work are shaders
14:55 PilzAdam     dont avoid fixing bugs by saying: "we will use something different in the future"
14:56 RealBadAngel hrmmm
14:56 RealBadAngel if we could have real lighting this wont be a bug
14:56 RealBadAngel so: do you want me to adapt my code for old lighting system
14:56 RealBadAngel wait for proper lighting
14:57 RealBadAngel and reverse the changes?
14:57 RealBadAngel for sake of what?
14:57 RealBadAngel art of coding?
14:58 RealBadAngel sorry, i will rather spend a few extra days to make right lights
14:59 PilzAdam     the new lighting is something in the future and it takes work to code, while the little lighting bug you produced with your facedir could be fixed right now
14:59 PilzAdam     (I guess it can be fixed without that much work)
15:00 PilzAdam     we are talking about weeks/months vs. days
15:00 RealBadAngel ok, heres the promise. if i wont make new light system by sunday evening
15:00 RealBadAngel i will make patch for old code
15:00 RealBadAngel deal?
15:00 PilzAdam     I dont want to tell you what you have to do; just what I think how it should be
15:01 RealBadAngel we are talkin a few days now
15:01 RealBadAngel sunday evening cet
15:01 RealBadAngel then expect one or another
15:11 RealBadAngel do we have a deal?
15:12 PilzAdam     you can do what you want :-)
15:12 PilzAdam     but it sounds good
15:12 RealBadAngel we are a team
15:13 RealBadAngel and we shall agree (at least most of the time ;)
15:15 RealBadAngel sidenote: i like the thing i noticed last month or two
15:15 RealBadAngel we have taken our niches in the engine
15:17 RealBadAngel i realized that when some1 posted in the forums something about model textures
15:17 RealBadAngel and reading hmmms commits too
15:18 RealBadAngel i know everything bout irrlicht models but have no glue what and why hmmm is commiting
15:19 RealBadAngel even his namespace remains mystery to me
15:39 kahrl        anyone opposed to changing the delay at test.cpp:1455 to 100?
15:40 PilzAdam     what does this do?
15:40 kahrl        this causes the asserts below not to fail (on my computer at least)
15:41 kahrl        because somehow the connect message takes more than 50ms to arrive
15:41 PilzAdam     why not test it in steps, like 10, 20, 30.. 100ms and fail after the last test failed?
15:42 PilzAdam     or maybe 20ms steps
15:42 kahrl        I didn't want to overcomplicate the code
15:43 PilzAdam     run_tests() is already slow
15:43 PilzAdam     and there are more sleep_ms(50), do they need to be increased too?
15:43 kahrl        at least not here
15:43 kahrl        the asserts below 1455 are the only ones that fail
15:44 kahrl        not sure why that is
15:44 hmmmm        oh
15:44 hmmmm        i know why
15:44 hmmmm        the default server step was increased from 0.05 to 0.10
15:45 hmmmm        connections are only serviced every step, so it makes sense for the latency of connection to be equal to at least one server step
15:45 kahrl        this test doesn't use the server, only Connection
15:45 hmmmm        mm.
15:46 hmmmm        well, minetest got slower, i guess that's just how it is
15:46 PilzAdam     yea, the same asserts fail here too
15:47 PilzAdam     (they fail not always, though)
15:47 kahrl        PilzAdam: yep
15:48 PilzAdam     100ms seems to work fine
15:48 kahrl        as I said yesterday an ultimately better solution might be to abandon UDP for local connections and pass the messages directly in-process
15:48 kahrl        then all of the delays in this test function could be removed
15:49 kahrl        although it wouldn't test the same thing anymore
15:50 PilzAdam     how much work would that need?
15:50 kahrl        I don't think very much, but I don't want to do it before ipv6 is merged
15:53 PilzAdam     someone tested the IPv6 patch some weeks ago and it seems to work fine
15:53 PilzAdam     only the checkbox in the menu has to be removed
15:54 proller      and not working on my arm
15:56 kahrl        not working in what sense?
15:59 proller      server listen nothing
15:59 proller      but maybe it my bug
16:00 proller      lets merge and fix if it real problem
16:01 kahrl        has it been tested on windows?
16:01 proller      and after merge need to patch masterserver script
17:08 celeron55    kahrl: can you make a pull request of the updated scriptapi_separation branch? it'll just get forgotten otherwise
17:08 celeron55    (as far as i can see)
17:09 celeron55    (or could somebody please just merge it)
17:56 kahrl        celeron55: last I heard sapier disagreed with the use of metatables
17:56 kahrl        because it makes it harder to sandbox lua
17:57 kahrl        brb
17:59 celeron55    umm
17:59 celeron55    how is that a problem in builtin?
18:00 celeron55    it's not like you can pull setmetatable out your ass if you hide it after builtin has run
18:01 celeron55    sapier's security thinking seems to sometimes consist more of fear than reason 8)
18:07 kahrl        I said the same thing :) http://irc.minetest.ru/minetest-dev/2013-05-12
18:08 kahrl        well I didn't really say it as much as think it
18:37 PilzAdam     this seems to be fine: https://github.com/minetest/minetest/pull/734
23:44 PilzAdam     https://github.com/PilzAdam/minetest/commits/predict
23:44 PilzAdam     comments?
23:46 * Exio       likes it ;P
23:54 kahrl        PilzAdam: great
23:54 PilzAdam     ok, I merge it then
23:55 kahrl        hmm I could see some ways to make the code shorter
23:55 PilzAdam     how?
23:55 kahrl        instead of the switch I'd use a static v3s16[]
23:56 kahrl        and the if in line 873 can be removed if you create the MapNode before it and use MapNode::getWallMounted()
23:57 kahrl        both things are not a big deal though, feel free to merge as is if you don't want to change them