17:16 < shoghicp> (those don't happen in MCPE :D)
17:26 < nickelpro> SinZ: One word: latency
17:26 < SinZ> isn't mobile just lan multiplayer
17:35 < shoghicp> SinZ: nope
17:35 < shoghicp> Go to Play -> Edit -> External
17:35 < shoghicp> they put it a bit hidden, but well, that works for external servers :)
17:36 < shoghicp> PocketMine fault :P
21:07 < Bibl> http://λ.pw/hef lol
21:10 < kamcio96> http://scr.hu/1s1h/xvq3a ;)
21:11 < SinZ> Bibl: your domain is annoying
21:11 < Bibl> why 
21:11 < Bibl> not mine
21:11 < SinZ> breaking my irc clients link detector
21:11 < Bibl> same
21:11 < Bibl> hexchat sux 
21:11 < Bibl> is the conclusion you should have came to 
21:11 < SinZ> nah, tot's the domains fault
--- Day changed sam. janv. 03 2015
20:50 < SopaXorzTaker> Tried devekoping for forge 1710
20:50 < SopaXorzTaker> co
20:50 < SopaXorzTaker> ol
20:55 < Bibl_> is there a way to scrape 1.8 block and item data?
20:56 < Bibl_> i'm making a gamepedia parser but it means it has to connect to like 300 links
21:02 < nickelpro> Bibl_: It's not everything but a hell of a lot of block data is here https://raw.githubusercontent.com/nickelpro/spock/master/spock/mcmap/mapdata.py Anything that isn't there is probably bouncing around in Mineflayer somewhere, but it probably needs to be updated
21:20 < Grum> Bibl_: what do you need to scrap?
21:21 < Grum> +e; probably easier to so a small mod that prints the data from the game?
21:32 < Bibl> sorry my computer crashed
21:33 < Bibl> grum: basically need block and item data, so ids, metadata, hardness, tools that can break the blocks, block sizes etc
21:34 < Grum> haha i wouldnt even know where to find that in the code >.>
21:38 < Xor_Boole> dammit Grum what did you break
21:46 < nickelpro> Bibl: All of those things are in here sorta, easier than scraping anyway https://raw.githubusercontent.com/nickelpro/spock/master/spock/mcmap/mapdata.py
21:47 < Bibl> nickel how did you get those?
21:48 < Bibl> nickelpro:
21:49 < nickelpro> Stole em from Mineflayer mostly
21:49 < Bibl> i thought it didnt have new block stuff?
21:50 < nickelpro> Doesn't, but its easier to update that than figure out scraping 'pedias
21:52 < Bibl> need all the data tho
21:52 < Bibl> i'll just do the scraper
21:52 < Bibl> like all the data
22:11 < Bibl> grum: was 1.8.1 a big update from 1.8?
22:11 < Bibl> like were items and blocks added?
22:32 < Dinnerbone> Nothing was added
22:33 < Dinnerbone> If two versions are protocol compatible (you can see on wiki.vg, the protocol didn't change at all and the numbers are the same) then nothing was added
22:34 < Dinnerbone> (You could also check on the minecraft wiki or a quick google and you'd find out in seconds :p)
22:34 < Grum> well we could add blocks that do nothing but being blocks :p
23:24 < Xor_Boole> <Grum> well we could add blocks that do nothing but being blocks :p
23:24 < Xor_Boole> you could please the people that are screaming blood for multipart-type stuff
23:25 < Grum> nah multipart is evil
23:25 < Grum> fun but very very evil
23:25 < Xor_Boole> agreed
23:25 < Xor_Boole> I'm sad the chickenbones' multipart is an ugly scala dialect
23:25 < Xor_Boole> hardly functional
23:26 < Xor_Boole> tbh, a good system for modularizing stairs and slabs would be good though
23:27 < Grum> nah those things become tile-entities
23:27 < Xor_Boole> tile entities is the wrong way to do it
23:27 < Xor_Boole> you should be making stairs a single material and putting it all in blockstates
23:27 < Grum> they have to be something 'not normal'
23:27 < Grum> yes that is the plan
23:28 < Grum> a single stair-block just used multiple times with another texture
23:28 < Xor_Boole> also you guys should get on vertical slabs, I get to hear people scream murder about it every day
23:28 < Grum> nope
23:28 < Xor_Boole> lel
23:28 < Xor_Boole> typical
23:28 < Grum> never vertical slabs
23:28 < Xor_Boole> idc myself
23:28 < Grum> there is no way that feels good to do them
23:29 < Grum> because once you start with vertical slabs, you need to add rotated stairs (on their side so 3/4 columns), and single-column ones
23:29 < Xor_Boole> is it because of the whole "notch doesn't like slabs" thing or just because there's no good way to implement them that isn't a multipart clusterfuck?
23:29 < Xor_Boole> yep, exactly
23:29 < Grum> and then we have to go about the user experience
23:29 < Xor_Boole> I made a post exactly like this on reddit about a week ago, people asking for vertical slabs
23:30 < Xor_Boole> if you guys cave to the vertical-slab people, then they'll just ask for more
23:30 < Sanqui> just upsize the whole world
23:30 < Sanqui> ×2
23:30 < Grum> just pretend we did
23:31 < Xor_Boole> I wonder, has stuff changed enough that hitboxes are not the clusterfuck they used to be?
23:31 < Grum> not yet, they will be though
23:31 < Xor_Boole> I remember a friend suggesting to you arbitrary mob resizes, you said something about hitboxes being a mess
23:32 < Xor_Boole> for fuck's sake nintendo, how am I not able to get a trade partner?
23:33 < Xor_Boole> oh, and has ryan started poking entity models with a ten-foot pole yet?
23:34 < Grum> yes
23:34 < Grum> slow going though
23:34 < xehbit> MC is totally in control of microsoft now right ?
23:34 < Xor_Boole> is he still sane?
23:35 < Xor_Boole> xehbit who cares?
23:35 < Grum> xehbit: yes the next windows will be windows11
23:35 < Grum> sorry, winbanana11
23:35 < Grum> i just changed my mind
23:35 < Xor_Boole> unless they set us modding people on fire I give no shits
23:35 < xehbit> winbanana haha :P
23:35 < Grum> and it will be free to click
23:35 < Grum> but you pay per mousemeter
23:35 < Xor_Boole> oh yea how's 1.banana coming along?
23:36 < Grum> we put our best coder on it, Marc!
23:36 < Xor_Boole> oh god
23:37 < xehbit> indeed, still hope MS allows us to mod the game
23:37 < Xor_Boole> well, if they block off modding, idc. grum is still fun to shout at
23:37 < Grum> xehbit: why would MS want to change that?
23:38 < xehbit> I dont know, i see MS as a stupid company, but that is more a personal thing :)
23:38 < xehbit> Just scared that they will fuckup the game
23:38 < Xor_Boole> Grum probably microsoft's history of assertiveness in copyright makes everyone scared
23:38 < xehbit> Xor_Boole: that indeed
23:38 < Xor_Boole> xehbit to "fuck up" the game they need to fuck up realms
23:38 < Xor_Boole> and they're not doing to do something stupid like that
23:38 < Grum> what has realms to do with this? :/
23:39 < Xor_Boole> perspective
23:39 < Xor_Boole> *shrug*
23:39 < Xor_Boole> realms is an income vector, modding is not
23:39 < Grum> also xehbit: you are about 2 months late with the doomsday scenarios
23:39 < Grum> no, modding is an income vector bigger than realms
23:39 < Xor_Boole> I don't think microsoft is stupid enough to kill a community
23:40 < Xor_Boole> Grum oh, really?
23:40 < Grum> because modding == community == sales
23:40 < xehbit> Grum: hehheh, had a busy time last 6 months, so not realy much time for modding and stuff :P
23:40 < Xor_Boole> I don't have any stats you might have, so I can only speculate
23:40 < Grum> i speculate too
23:40 < xehbit> The thing i see MS do is, ingame items for real cash
23:40 < Xor_Boole> I really don't think ms is that stupid
23:41 < Grum> how would you even do that xehbit?
23:41 < Xor_Boole> they maybe the Evil Empire but they're very shrewd. they paid good money for mojang and p2w is the easiest way to scare people off
23:41 < Grum> buy pickaxes you can just /give ? :/
23:41 < Xor_Boole> seriously nintendo
23:42 < xehbit> Hehheh idk, mayby they are going to kill the dedicated server stuff
23:42 < xehbit> and make it a party system
23:42 < Xor_Boole> Grum tell whomever's in charge to sell mollstam to nintendo, they need better server infrastructure
23:42 < Grum> xehbit: how would they do that when we keep providing the server software?
23:42 < Grum> mollstam doesn't work at Mojang anymore
23:42 < Xor_Boole> oh? when'd he leave?
23:42 < Grum> after the merge
23:42 < xehbit> Oh, so the old mojang guys are still on MC but for MS now ?
23:42 < Grum> takeover meh
23:43 < Xor_Boole> really? huh, I didn't notice. that makes me sad
23:43 < Grum> xehbit: no? we work for Mojang
23:43 < xehbit> Mojang is a sub-company of MS ?
23:43 < Xor_Boole> Mojang is Microsoft's <legalese>
23:44 < Grum> we're just a studio owned by Microsoft
23:45 < xehbit> Ah, i see!
23:45 < Xor_Boole> can we get back to ranting about how bad the codebase is? this microsoft conspiracy is boring
23:45 < xehbit> Xor_Boole: i agree :)
23:45 < xehbit> Got it now, thanks for the intel ^_^
23:51 < Xor_Boole> why does maxie look like such an idiot in this game
--- Day changed dim. janv. 04 2015
16:51 < pcfreak30> Is there a packet that needs to be sent to get a Join Game packet? Im using a networking library in node, and logging all incoming packets and no Join Game packet comes in.
16:53 < pcfreak30> after a login success, i only http://pastie.org/private/rdmxkjy83x7qxoahg9w7q with the states logged but theres nothing related to joining a game.. 
17:12 < gamingrobot> pcfreak30: https://gist.github.com/gamingrobot/e371c0be0f630fcd7018
17:35 < pcfreak30> gamingrobot: thats, turns out the events werent being related from client object to bot object.
17:35 < pcfreak30> relayed
17:41 < gamingrobot> ah
21:49 < Paprikachu> how does the server get the players name?
21:51 < Paprikachu> nvm
--- Day changed lun. janv. 05 2015
22:19 < Not-f18a> [Glowstone] SpaceManiac pushed 4 commits [+2/-0/±23] http://git.io/jVxSDQ
22:19 < Not-f18a> [Glowstone] SpaceManiac 5ebf0fe - Basic item drop entity support (closes #4)!
22:19 < Not-f18a> [Glowstone] SpaceManiac 957e6e3 - Updated shadow plugin to 1.2.0 for Gradle 2 and up.
22:20 < Not-f18a> [Glowstone] SpaceManiac deleted branch items
23:42 < c1yd3i> anyone fimilar with the mc protocol?
23:43 < c1yd3i> trying to figure out what packet might be sent with a player dies
23:43 < c1yd3i> er, the client
23:43 < c1yd3i> 0x06 (update health) didn't change when the client died for me
--- Day changed mar. janv. 06 2015
00:00 <+SpaceManiac> c1yd3i: I don't think it's anything else than that
00:00 <+SpaceManiac> you might find a death animation packet, but that doesn't cause the "you died!" screen
00:01 < c1yd3i> alright, it must be a different problem
00:01 < c1yd3i> thanks
00:05 < c1yd3i> hmm this is too bizzare, i'm really not sure what's causing it
00:06 < c1yd3i> im listening for the health to change, i kill myself but nothing is being sent to the client
00:06 < c1yd3i> must be the server's fault, i'm not sure
00:06 <+SpaceManiac> are you getting just nothing?
00:07 < c1yd3i> correct
00:07 < c1yd3i> im not getting anything at all
00:07 < c1yd3i> no feedback
00:14 < c1yd3i> is there another packet that could be potentially sent
00:14 < c1yd3i> if the client dies?
00:16 <+SpaceManiac> *possibly* EntityMetadata with health=0? I don't think so, though
00:19 < c1yd3i> my guess is update_health just isn't called on a suicide so im just working around that
00:19 < c1yd3i> because it doesnt seem to be called on respawn either
00:19 < c1yd3i> but im pretty sure on an individual injury it is called
00:19 < nickelpro> You're not updating your position
00:19 < c1yd3i> I need to update my position?
00:19 < nickelpro> you need to be sending position updates 20 tps in order for the server to dean you worthy of health updates
00:20 < c1yd3i> Can I sent a position update to my current position?
00:21 < c1yd3i> Er let me rephrase that
00:21 < c1yd3i> Er nevermind I was just thinking out loud
00:23 < nickelpro> Spock does it here: https://github.com/nickelpro/spock/blob/master/spock/plugins/helpers/move.py and here https://github.com/nickelpro/spock/blob/master/spock/plugins/helpers/respawn.py for respawn, uses the actual packet names so should be pretty self explanatory
00:35 < c1yd3i> so i'm updating my position but
00:35 < c1yd3i> update_health is only sent after respawn
00:36 < c1yd3i> when i kill myself it doesn't say my health is 0
00:36 < c1yd3i> i dont know if it's good practice or if it could have any affect but the position updates are at 50ms intervals and im sending back the same position the server sent me earlier
00:54 < btilm305> anyone have a link to an implementation of gravity and block heights?
05:29 < btilm305> anyone have a java example of getting blocks from raw chunk data?
05:50 <+SpaceManiac> btilm305: gravity is just acceleration; you can't really treat blocks as just heights because they have lots of bounding box shapes
05:51 < btilm305> yeah ignore that, I've got that
05:51 < btilm305> Just having trouble with the offset in packetmapchunkbulk
05:51 < btilm305> never can read the 2nd chunk correctly -.-
05:54 <+SpaceManiac> Here's how Glowstone writes the chunks, maybe you can reverse it to read them: https://github.com/GlowstoneMC/Glowstone/blob/c41bf0e34473d25d1c70440f8debd5bf98c7f973/src/main/java/net/glowstone/GlowChunk.java#L683
06:12 < Xor_Boole> nightmare material: notch originally writting minecraft using very poor scala macro skills
06:12  * Xor_Boole shudders
15:24 < Wuppie> Any .net developer here? (C#)
15:24 < Wuppie> Willing to join a new project?
15:54 < rom1504> Wuppie: probably not, people don't join project with no information
16:08 < Wuppie> rom1504 That is true. I'll provide some information later on , if i actually decide to do it. as it a lot of work
19:35 < xehbit> tell us what the project is about :p
20:27 < Wuppie> xehbit: Drew (SirCmpwn) is already doing basically the same thing. So it is not really important.
20:27 < Wuppie> i also have a question. How does kicking a client with a reason work?
20:27 < Wuppie> i don't see it any where in the protocol specification ATM, or i looked over it
20:47 < gamingrobot> Wuppie: why dont you just maintain his project?
20:48 < gamingrobot> or contribute depending on the status of said project
20:48 < Wuppie> gamingrobot: i was gonna, but he doesn't want me to as it is a clean room project. And i have looked at minecraft source code once
20:49 < gamingrobot> is it his craft.net project?
20:49 < Wuppie> nope, http://truecraft.io/
20:49 < gamingrobot> ah
20:49 < gamingrobot> oh hey he lives in the same city as me
20:49 < Wuppie> haha lol
20:50 < Wuppie> i loved the project, so i pushed a new command handler, he liked it. but didn't accept it as i looked at MC source code once
20:50 < Wuppie> but oh well
21:45 < nickelpro> Wuppie: http://wiki.vg/Protocol#Disconnect
21:45 < nickelpro> That'show you disconnect with a reason
21:45 < Wuppie> A thnx man! Didn't see that!
21:46 < nickelpro> I know the protocol to well, feel free to shoot me question about packets/file formats/client behavior
21:46 < nickelpro> I wish I had spent this time on something productive, but well, here we are
--- Day changed mer. janv. 07 2015
01:46 < Nasuno> Can I get a dialog with someone capable of WatchService with Thread Pools?
02:04 < Nasuno> Or rather interested in answering some questions on the subject would be less rude. My people skills are on par with my java skills
07:22 < btilm305> anyone know the velocity when you jump?
--- Log closed mer. janv. 07 07:49:21 2015
--- Log opened mer. janv. 07 07:55:09 2015
07:55 -!- Irssi: #mcdevs: Total of 159 nicks [1 ops, 0 halfops, 10 voices, 148 normal]
07:58 -!- Irssi: Join to #mcdevs was synced in 216 secs
17:13 < nickelpro> Can anyone explain how drag works in Minecraft? The wiki just says "there is a drag force proportional to velocity" and gives constants with the unit 1/tick that I'm not sure how to interpret
17:15 < nickelpro> Oh I'm retarded, nvm
17:34 < nickelpro> Anyone know the max ground speed of a walking client? A sprinting one? And if strafing is different that walking/sprinting?
18:06 < gamingrobot> nickelpro: doesnt the client send that information?
18:07 < gamingrobot> nickelpro: http://wiki.vg/Protocol#Player_Abilities ?
18:08 < nickelpro> I don't know if those values are blocks/tick which is what I really need
18:17 < Xor_Boole> the client orders names in the tab list case-sentive-lexicographically
18:18 < Xor_Boole> this bothers me
18:18 < Xor_Boole> is there a reason the oreder isn't case insensitive =<
18:32 < nickelpro> Ah, so the wiki has speeds in m/s, should be able to use that. Anyone know where the minecraft wiki speed measurements come from? Code or just measurement?
18:33 < Bibl> code probably
18:33 < Bibl> Element
18:34 < Bibl> http://minecraft.gamepedia.com/Transportation *
18:35 < nickelpro> I'm on that page right now but a lot of the value later on say "From experiment" which confuses me. Especially since the entity page gives its value in blocks/tick http://minecraft.gamepedia.com/Entity#Motion_of_entities
--- Day changed jeu. janv. 08 2015
05:17 < Fenhl> Minecraft Wiki's values are usually from experiment
07:45 < Not-f18a> [Glowstone] SpaceManiac pushed 1 commit [+0/-0/±3] http://git.io/4l5Y6g
07:46 < Not-f18a> [Glowstone] BigxMac aaa3e8a - Added damage due to cactus blocks.
--- Day changed ven. janv. 09 2015
05:23 < Not-f18a> [Glowstone] SpaceManiac pushed 1 commit [+1/-0/±5] http://git.io/zXsm6Q
05:23 < Not-f18a> [Glowstone] turt2live 39ca0ab - Added support for melee attacks by players.
20:20 < Xor_Boole> someone apparently asked for a snooper data API and got downvoted into oblivion. this makes me sad.
20:25 < Xor_Boole> I wonder what havoc I could wreck with suck an API...
20:30 < Dinnerbone> Not as much as you think, considering what little information is in snooper
20:30 < Dinnerbone> No personal info so you can't track users between sessions
20:30 < Dinnerbone> You can see average hardware specs, that's about it
20:30 < Dinnerbone> And I've tweeted some of those
20:31 < Dinnerbone> The interesting ones at least
20:34 <+Amaranth> Dinnerbone: Now that you have currently playing stats you should put them on minecraft.net/stats :)
20:34 < Dinnerbone> That decision would have to bubble up
20:36 < Xor_Boole> Dinnerbone tell your web people to make some nice visualization at least =p
20:36 < Dinnerbone> We can absolutely do it I just mean that's not a decision I have the authority to make
20:36 < Xor_Boole> like that java version pie chart you had a while ago
20:37 < Xor_Boole> naturally
20:37 < Dinnerbone> Grafana has nice visualisations, have you seen the other snooper data I've tweeted?
20:37 < Xor_Boole> some of it
20:37 < Xor_Boole> I'd just be nice if you could take a sharp marlinspike and aim it at the correct peoples' sides for encouragement
20:37 < Xor_Boole> I mean, it works with Thinkofdeath 
20:37 < Dinnerbone> http://media.dinnerbone.com/uploads/2014-12/screenshots/05_17-11-31_qhmCvhEgDS.png
20:38 < Xor_Boole> oh I think I might have seen that one
20:39 < Dinnerbone> http://media.dinnerbone.com/uploads/2014-12/screenshots/04_18-02-19_eInpCvR0YG.png is the pie chart one you're talking about, that was just done in google spreadsheets :p
20:39 < Dinnerbone> (Also the percentage of java 8 went up *a lot*)
20:39 < Dinnerbone> (Our launcher helped with that :D)
20:39 < Xor_Boole> because of the new launcher right?
20:39 <+Amaranth> lol OpenPandora
20:39 < Xor_Boole> still the most horrible and great idea I have heard in a long time
20:39 < Xor_Boole> > bundle your own java
20:39 < Dinnerbone> Horrible?
20:40 < Dinnerbone> What's horrible about that?
20:40 < Dinnerbone> Almost every serious application does it, I actually strugle to think of one that doesn't
20:40 < Xor_Boole> meh, extra copies floating around
20:40 < Dinnerbone> (Which uses java)
20:40 < Xor_Boole> *shrug*
20:40 < Xor_Boole> it's not like I would know
20:41 < Xor_Boole> blame other people puting ideas in my head
20:41 < Xor_Boole> a better question is when the launcher will equal the scrolls launcher in sexiness
20:41 < Dinnerbone> Do you have a java IDE?
20:42 < Xor_Boole> Dinnerbone intellij, yes. I know it comes with java
20:42 < Wuppie> Dinnerbone: what java IDE do you use if i may ask?
20:42 < Dinnerbone> :p
20:42 < Xor_Boole> intellij iirc ^
20:42 < Xor_Boole> only real ide
20:42 < Dinnerbone> Yeah intellij
20:42 < Xor_Boole> except he uses this ugly white theme which means he's not a Real Hacker
20:42  * Xor_Boole runs
20:43 < Wuppie> haha
20:43 < Dinnerbone> I tried darkula, it makes me feel ill
20:43 < Xor_Boole> it's grown on me
20:43 < Dinnerbone> (I'm not saying "it's so ugly omg it makes me sick thinking about it", it actually does make me feel ill. Looks darn cool though.)
20:43 < Dinnerbone> I like it, I just can't use it.
20:44 < Xor_Boole> oh, fucks with your eyes?
20:44 < Dinnerbone> Yeah
20:44 < Xor_Boole> that's too bad, it is a sexy theme =<
20:44 < Dinnerbone> Besides leet haxors have more green.
20:44 < Xor_Boole> that said intellij hates me now because I do a lot of scala on it
20:45 < Xor_Boole> Dinnerbone am I leet then? http://puu.sh/ei9U1/a4a80f1fa9.png
20:45 < Xor_Boole> my intellij colors are also very green =p
20:45 < Dinnerbone> A++
20:45  * Dinnerbone waddles off to do laundry
20:45 < Xor_Boole> you off to- oh wait fucking timezones
20:46  * Xor_Boole shakes fist at the sun
21:22 < ScruffyRules> Xor_Boole, Shouldn't you be shaking your fist at the Earth since it is the one going around the Sun? :P
21:22  * Xor_Boole sheds on ScruffyRules's lap
21:22 < Xor_Boole> go away
21:22 < ScruffyRules> :P
22:16 < Darker> Hello, just a quick question
22:16 < Wuppie> Sure! just ask :P
22:16 < Darker> Last time I was asking here, somebody told me that there's a working API for minecraft protocol in C++
22:17 < Darker> I disregarded that and implemented it by myself
22:17 < Darker> It was lot of pain
22:17 < Darker> A friend of mine is trying to do similar thing
22:18 < Darker> So is there updated C++/java implementation of protocol?
22:18 < Darker> Ideally providing API for a simple AI that already queries for pathfinding etc?
22:18 < Darker> He's surelly make use of it
22:24 < Darker> I take that for "no" :(
22:28 <+SpaceManiac> Glowstone has a server-oriented protocol implementation in Java, but no pathfinding or anything
22:28 <+SpaceManiac> Maybe you want to look at node-minecraft-protocol (JS)? I think there's bots for that
22:28 < MrARM> Implementing path finding is quite simple though 
22:29 < MrARM> I did that once for a test client but failed to code sending the movement correctly. 
22:44 < nickelpro> MrARM: Client-correct MC physics is the hard part IMHO
22:46 < MrARM> Yeah, that's not too simple 
22:47 < nickelpro> I've got jumping working :D https://github.com/nickelpro/spock/blob/master/spock/plugins/helpers/physics.py
22:48 < MrARM> :o
22:48 < MrARM> nice! 
22:48 < Darker> I think the box colission calculation for pathfinder won't be that easy
22:49 < nickelpro> I'm ignoring the collision problem for now
22:49 < MrARM> I meant a simpler one, with all blocks being assumed as full ones 
22:49 < nickelpro> ^
22:50 < MrARM> that's quite easy and enough for most of the stuff 
22:50 < Darker> That's OK until you encounter slabs or stairs...
22:50 < Darker> Btw does bukkit server check for collision with fences? I mean the extra 1/2 block of them
22:50 < nickelpro> Easy enough to code a special case in for when the algorithm encounters those blocks
22:51 < MrARM> Even then, pathfinder itself shouldn't really care about it, rather the moving part of ot
22:51 < MrARM> *it
22:53 < Darker> Well it should. If the server checks for fence collision, pathfinder mustn't plan the rote ouver fence (jumping on it)
22:53 < Darker> While it can plan path over other types of blocks.
22:54 < MrARM> Ah, yeah, that's still quite easy to make an special case for 
22:54 < nickelpro> Vanilla server doesn't check jump height, just hop over it :-P
22:54 < Darker> Funny thing I encountered when testing my C++ bot is, that walking over water is not considered flying, as water is also block
22:54 < Darker> I wonder if this was fixed or not
22:55 < nickelpro> When I was trying out jump accelerations, my bot did superman-style leaps into the stratosphere without being kicked
22:55 < nickelpro> Also leaving on_ground=False all the time negates fall damage
22:55 < nickelpro> And yes, walking on any block other than air is always valid
22:55 < nickelpro> So lava/water/plants all good
22:55 < MrARM> To be kicked you need to be in air by minimum 5 seconds I guess 
22:56 <+SpaceManiac> did you have allow-flight enabled in the server options?
22:56 < MrARM> Vanilla anti-hacks aren't too good 
22:56 < nickelpro> Nope
22:56 < Darker> I was talking about bukkit
22:56 < shoghicp> craftbukkit without plugins ~ vanilla
22:56 < nickelpro> Vanilla only checks in every 5-10 seconds and asks if you've obeyed the laws of physics lately
23:03 < Darker> Well, thanks for chat, bye
23:23 < powerstrip> hm
23:24 < powerstrip> does anyone have a simple minecraft query tool?
23:29 <+SpaceManiac> will this suffice: https://dinnerbone.com/minecraft/tools/status/ or do you mean the query protocol
23:29 < gamingrobot> for pathfinding cant you just assume fences are two blocks tall?
23:30 < Dinnerbone> That does the query protocol too
23:30 <+md_5> > http://media.dinnerbone.com/uploads/2014-12/screenshots/04_18-02-19_eInpCvR0YG.png
23:30 <+md_5> > OpenBSD
23:30 <+md_5> Can we spoof versions on that?
23:30 < Dinnerbone> Yes, you are asked not to. :p
23:31  * md_5 installs on dragonflybsd instead
23:31 < shoghicp> Is there a way to set the port on your tool?
23:31 <+md_5> host:port ?
23:31 < gamingrobot> wait what windows 95
23:31 < Dinnerbone> Type it in as you would in the game
23:31 < shoghicp> MCPE servers default to port 19132 and having query on
23:31 < shoghicp> did not work D:
23:31 <+md_5> is the query the same?
23:32 < shoghicp> it is
23:32 < shoghicp> works with other tools
23:32 < Dinnerbone> (Hi shoghicp!)
23:32 < shoghicp> o/
23:32 < Dinnerbone> MCPE servers do not have query protocol to my knowledge?
23:32 < shoghicp> PocketMine has it :P
23:32 < Dinnerbone> You'd know more than me in that area though
23:32 < shoghicp> anyway
23:32 < shoghicp> sg.lbsg.net:19132
23:32 < Thinkofdeath> I think he is talking about the udp query?
23:32 < shoghicp> works in other tools
23:32 < shoghicp> yep
23:33 < Dinnerbone> Thinkofdeath, my tool supports that.
23:33 < Thinkofdeath> oh
23:33 < Dinnerbone> shoghicp, fork the code and try to run it - see if you get any errors. :D
23:33 < shoghicp> alright, got a free night :D
23:33 <+md_5> we could install Minecraft on Solaris
23:34 < Dinnerbone> It can be either that I ignore port for udp query because barely anyone used it, my library may have a bug that wasn't caught until now, or your implementation may not match the official one close enough
23:35  * Dinnerbone installs Minecraft on md_5's toaster
23:35 < shoghicp> it did work before
23:36 < shoghicp> and yeah, it uses UDP Query
23:36 <+md_5> I swear one of these days Im gonna make a raspberrypi toaster
23:36 <+md_5> http://www.embeddedarm.com/software/arm-netbsd-toaster.php
23:37 < jython234> xD
23:37 < Dinnerbone> My website is just a frontend to https://github.com/Dinnerbone/mcstatus - mcstatus supports query (udp) or ping (tcp), do MinecraftServer.lookup("example.org:1234").query()
23:38 < shoghicp> :o, timeout
23:40 < Dinnerbone> It's entirely possible that I relied on too much "untested behaviour" from vanillas Query that it may not work in a decent environment
23:41 < Dinnerbone> (Our query sucks and I hate it so much. In my defense, it far predates me and I haven't had a reason to touch that code yet. Mostly out of fear.)
23:41 < shoghicp> alright, I'll do a PR once I find a fix
23:42 < Thinkofdeath> Dinnerbone: just quietly remove it one update and see if anyone notices
23:42 < Dinnerbone> It has been on the agenda to make it obsolete with something better for a while, just no time
23:43 < shoghicp> heh
23:43 < shoghicp> now it works
23:43 < Dinnerbone> Huh
23:43 < dx> something better implementing the same protocol, or something completely different?
23:43 < Dinnerbone> Curious
23:43 < Dinnerbone> Diff protocol
23:43 < Dinnerbone> A sane one. :p
23:43 < dx> heh
23:47 < shoghicp> https://github.com/shoghicp/mcstatus/commit/4c4239e6f88c1dd447de39c198b5133576ed885e
23:47 < shoghicp> that broke it
23:47 < Dinnerbone> So you take more than 3 seconds to start responding
23:47 < shoghicp> seems like some servers struggle to reply under 3s
23:47 < shoghicp> it's instant on my server
23:47 < Dinnerbone> That's a really long time. :D
23:47 < shoghicp> but hta guy has a setup that contacts a global server
23:47 < shoghicp> that*
23:47 < shoghicp> let me switch server :s
23:50 < shoghicp> it replies me in half a second, takes more with the lib
23:51 < Dinnerbone> Huh
23:51 < shoghicp> it might be the server location
23:52 < shoghicp> 127.0.0.1 at least works fine :)
23:53 < shoghicp> oh, this might be it
23:53 < shoghicp> DNS resolution times
23:55 < shoghicp> yep, resolves DNS each time a packet is sent
23:55 < shoghicp> and that server uses DNS as a way to load-balance
23:59 < Dinnerbone> I see
--- Day changed sam. janv. 10 2015
00:04 < shoghicp> https://github.com/Dinnerbone/mcstatus/pull/26
00:06 < shoghicp> hmm
00:06 < shoghicp> tests
00:28 < shoghicp> alright, tests passing and ready to be merged if you like it
00:30 < Dinnerbone> <3 will take a look when I get some time, probably this weekend
00:32 < shoghicp> nice!
00:33 < shoghicp> lots of people ask for a way to check if they port forwarded correctly, I used to give them that tool
02:09 < Xor_Boole> I think the microsoft paranoia has reached a point where it is just an amusing distraction, similar to the eula drama. makes me giggle.
02:09 < Xor_Boole> I am a twisted man
06:30 < nickelpro> My bot can now walk and sprint reasonably correctly
06:30 < nickelpro> To celebrate, he took a quick jog across the ocean
06:31 < nickelpro> Next, we teach him about liquids
08:02 < Not-f18a> [Glowstone] SpaceManiac pushed 1 commit [+0/-0/±1] http://git.io/6q9o3Q
08:02 < Not-f18a> [Glowstone] SpaceManiac 80ee09c - Added void damage, dealt to entities with y < -64 (thanks @BigxMac).
08:26 < gamingrobot> nickelpro: or maybe gravity
08:27 < nickelpro> gamingrobot: He knows about gravity naively
08:27 < gamingrobot> he just assumes water is solid
08:27 < nickelpro> Someone just needs to fix his "floor detection" which only searches for "not air" atm
08:28 < nickelpro> https://github.com/nickelpro/spock/blob/master/spock/plugins/helpers/world.py#L25-L35
08:28 < nickelpro> "Someone" will probably be me
08:28 < gamingrobot> "Someone" will probably be me
08:28 <+Amaranth> Unfortunately such data is not in the model files so you'll have to hard code it same as MC does
08:28 < nickelpro> I'd appreciate that :D
08:29 < nickelpro> Amaranth: I know, I actually have most of the data already. Just not using it
08:29 <+Amaranth> It'd be really nice if it was in the model files
08:30 < gamingrobot> new mc model format
08:30 < gamingrobot> its the mc models with the bounding box appended to the end
08:30 < gamingrobot> err collision box
08:31 < nickelpro> Anyway physics is basically done, we walk, sprint, jump, and fall reasonably correctly. Enough to implement a rudimentary pathfinding algorithm at least
08:32 < nickelpro> It needs a huge amount of polish, and the Nice-to-Have list is a mile long
08:32 < nickelpro> but hey it works
08:34 < gamingrobot> have we updated the Todo file?
08:34 < gamingrobot> or should github issues really be what the todo file is
08:35 < nickelpro> My ToDo file is just "big concepts so I can't claim I have nothing to work on" its motivational more than anything else
08:35 < gamingrobot> gotcha
08:35 < nickelpro> Github issues is where it should really live, go ahead and create issues and assign them as you see fit if you're looking for work that isn't coding
08:36 < nickelpro> The nice thing about issues is they let you cross-reference and micro manage feature development and polish. But that is way too much professionalism for me to implement personally, much like documentation and a stable API
08:38 < gamingrobot> A stable API would be nice
08:38 < gamingrobot> I am off to bed now tho
08:38 < nickelpro> cya, I should go too
14:12 < barneygale_> are mojang not releasing weeklies anymore?
14:13 < shoghicp> they will start soon, jeb said around January
14:13 < shoghicp> (saw that in reddit yesterday)
15:47 < Fenhl> weekly snapshots tend to stop for a while after a release
16:29 < Paprikachu> im a little bit confused about how minecraft stores/transmits block data
16:29 < Paprikachu> has metadata been removed yet?
16:30 < Paprikachu> i remember an announcement that jeb wanted to make block ids 2 bytes and remove the concept of metadata
16:54 < Bibl> blocks still have metadata atm
16:57 < Bibl> Paprikachu this is how i receive and parse it
16:57 < Bibl> 		id = in.readVarInt() >> 4;
16:57 < Bibl> 		metadata = id & ((1 << 4) - 1);
16:58 < Paprikachu> ah, i see
16:58 < Paprikachu> thanks
17:12 < nickelpro> Bibl: That's a completely incorrect metadata decoder
17:12 < Bibl> lol
17:12 < Bibl> are you sure?
17:12 < nickelpro> data = in.readVarInt()
17:12 < nickelpro> id = data >> 4
17:12 < nickelpro> metadata = data&0x0F
17:12 < Bibl> i tested this a while ago and the last 4 bits is the metadata
17:13 < nickelpro> Bibl: yes, but you're deriving the metadata from id, which you already shifted the last 4 bits out of
17:13 < nickelpro> metadata = id & ((1 << 4) - 1);
17:13 < Bibl> o crap
18:18 < SopaXorzTaker> for(a;;){while();}
18:38 <+SpaceManiac> The wiki lists 'F' and 'O' items in the debug menu as information regarding chunks, but I don't see them. Am I missing something or is the wiki just outdated?
18:38 < nickelpro> SpaceManiac:  link?
18:38 <+SpaceManiac> http://minecraft.gamepedia.com/Debug#Legend
18:40 < nickelpro> Probably outdated wiki
18:54 <+SpaceManiac> mm, that's unfortunate
18:54 <+SpaceManiac> I keep running into issues where the client doesn't render enough chunks under various circumstances and can't pinpoint the cause
19:10 < MrARM> So I've tried to make an MC client in c++ today 
19:11 < MrARM> Got server ping to work (my current sample code for that: " http://pastebin.com/MjwEKSZY") 
19:12 < nickelpro> I did something with a similar structure to yours, looks good so far
19:15 <+SpaceManiac> random C++ tip: avoid manual pointer things where possible
19:16 <+SpaceManiac> you could avoid the manual new/delete by putting 'handshake' and 'statusReq' on the stack and passing '&handshake' to writePacket
19:28 < MrARM> hmm
19:28 < MrARM> k, thanks 
22:04 <+SpaceManiac> okay, I can eliminate one class of rendering problems as a client bug because it happens in singleplayer
23:50 < gurun> Is it typical or not for a MCPC server to do culling of player positions based on LoS?
--- Day changed dim. janv. 11 2015
03:31 < gamingrobot> https://db.tt/Sv5kGXKo hehehe
03:58 < nickelpro> move entity function deals fire damage and plays sound
03:58 < nickelpro> https://gist.github.com/gamingrobot/c9bc30a7f39e2a21de57
03:58 < nickelpro> I'm done, Minecraft is dumb
08:44 < Not-f18a> [Glowstone] SpaceManiac pushed 1 commit [+0/-0/±5] http://git.io/tGUAIg
08:44 < Not-f18a> [Glowstone] SpaceManiac a1a7fc7 - Added actual height map storage to improve performance.
10:58 < MrARM> hmm, almost got logging in working
10:58 < MrARM> however for unknown reason I get this as the return data: [username: Test\377, uuid: 444cf323-978c-3e83-9288-612345bfec67\377]
10:58 < MrARM> Any ideas why?
10:59 < MrARM> (I meant that \377 part of that, I don’t know why it is there)
11:00 < MrARM> nvm
13:38 < MrARM> what compression level does MCPC use?
16:22 < MrARM> lol, you can sneak and sprint at same time
16:24 < jython234[ping]> xD
17:23 <+Amaranth> MrARM: You can also move at any speed you want while sneaking, walking, sprinting or some combination
17:23 <+Amaranth> At 0,0 and want to get to 1000000,1000000 to get away from everyone else? Just send a packet saying that's where you are
17:24 <+Amaranth> Combine with /msg letting you triangulate where people are and you have full client side /tp
17:24 < shoghicp> msg to triangulate?
17:25 <+Amaranth> shoghicp: Command block syntax
17:25 < shoghicp> and I think there is a limit of 100 blocks per tick hardcoded in the check
17:25 < shoghicp> ah, right
17:25 <+Amaranth> That check was broken in 1.5
17:25 <+Amaranth> It only triggers if the server tries to move you more than 10 blocks a tick which is exactly when it shouldn't
17:26 <+Amaranth> Well, assuming the server tried to move you that much and you agreed
17:26 < shoghicp> at least I implemented it right :D
17:26 <+Amaranth> It does Math.max(serverVelocity, clientVelocity) and serverVelocity is what the server wants you to move not what it thinks you're moving based on packets
17:26 <+Amaranth> Err, Math.min rather
17:27 <+Amaranth> For a quick fix I changed it to Math.max in CraftBukkit which breaks huge TNT explosions and ender dragon knockback
17:27 <+Amaranth> And huge Knockback enchant levels but you can't get those normally
17:28 < shoghicp> or disregard the client velocity and just force the server velocity
17:28 <+Amaranth> That really doesn't even help
17:28 <+Amaranth> You can just spam movement packets that are under the distance check anyway
17:28 <+Amaranth> I was just trying to make it a little harder
17:29 <+Amaranth> Server won't let you move more than 10 blocks at once? Just tell it you moved 9 blocks 10000 times in a single tick
17:29 < shoghicp> haha
17:29 < shoghicp> hmm, my server checks for the 100 block distance once per tick
17:30 < shoghicp> reverts it if more
17:30 < shoghicp> so I guess that works fine
17:30 <+Amaranth> That would take more diff than I was willing to commit to
17:30 <+Amaranth> Also it's not supposed to be 100
17:30 < shoghicp> 10 squared?
17:30 <+Amaranth> Vanilla is comparing distance squared to 100 to avoid a sqrt
17:31 < shoghicp> alright, patching :D
17:31 < shoghicp> I'm checking against 100 squared, 10000 D:
17:31 <+Amaranth> Always better to square the other side than have a sqrt, common optimization
17:32 < shoghicp> yep
17:33 < shoghicp> I check against squared distance on most of the cases
17:41 < shoghicp> 10 blocks per tick is still 200 blocks per second D:
17:48 <+Amaranth> Yeah but an explosion (or ender dragon knockback) could do it to you
17:48 <+Amaranth> Do you allow them to go over the limit if the server thinks they should be?
17:48 < shoghicp> yep
18:06 < MrARM> Amaranth: by sprint+sneak combo i meant the sprinting particles while the player is sneaking rather than the speed itself
18:32 < MrARM> I have a problem with chat reading :/
18:33 < MrARM> I can read strings I think (as I can read the LoginSuccess username/uuid fine)
18:34 < MrARM> but chat gets cut/I have empty string
18:34 < MrARM> {"translate":"chat.type.text","with":
18:34 < MrARM> ^ is all i get when I join
18:36 < MrARM> my string reading code: http://pastebin.com/9D28e0kk
18:37 < MrARM> I checked and the length is fine
18:41 < MrARM> hmmm
18:41 < MrARM> I know what’s wrong now
19:03 < MrARM> or maybe I do not
19:09 < MrARM> so it's an buffer overflow, but idk why 
19:11 < shoghicp> MrARM: strings are UTF-8
19:11 < shoghicp> they can contain null bytes
19:12 < MrARM> I fixed that 
19:12 < MrARM> but 
19:12 < MrARM> it's an buffer overflow 
19:12 < MrARM> as
19:12 < MrARM> I typed same message in chat a few times 
19:12 <+SpaceManiac> what does readCharArray look like?
19:12 < MrARM> the output bytes were changing 
19:13 < MrARM> readCharArray here memcopy's the a part of an byte array 
19:14 <+SpaceManiac> sure but what's the implementation? (also the updated readString)
19:14 < MrARM> k
19:16 < MrARM> http://pastebin.com/B2fM1tjs
19:16 < MrARM> ^ all stuff that may be revelant
19:18 <+SpaceManiac> hmm, nothing obviously amiss
19:18 <+SpaceManiac> you're sure 'len' is being read correctly?
19:19 < shoghicp> MrARM: what is your readVarInt implementation?
19:21 < MrARM> shoghicp: that's fine, I did a std::cout << len; and it gives correct values 
19:21 < shoghicp> do you have the raw packet?
19:21 < shoghicp> that would be useful to have, as maybe it's an off-by-one bug?
19:22 < shoghicp> you pasted here some username names before, they had extra data at the end
19:23 < MrARM> shoghicp: I fixed it, it works fine 
19:23 < MrARM> I can read username and uuid just fine right now 
19:25 < MrARM> and what's weird it displays first part of the join message 
19:25 < MrARM> but when I chat, it displays nothing 
19:25 < MrARM> (not nothing, I meant it displays junk) 
19:27 < MrARM> weird, it may be something with encoding, idk
19:27 < shoghicp> utf-8?
19:27 < shoghicp> that's why I'm asking for the raw packet :)
19:28 < MrARM> will run tcpdump as I get back to PC
19:28 < shoghicp> are you running on an encrypted stream?
19:28 < shoghicp> or online mode is of?
19:28 < shoghicp> off*
19:29 < MrARM> online mode is off
19:30 < shoghicp> alright then, that works :)
19:30 < shoghicp> otherwise tcpdump would only get encrypted data
19:49 < MrARM> aww
19:49 < MrARM> what params to use?
19:50 < MrARM> it captures everything else except the things on the port I specified
19:50 < MrARM> I tried:
19:50 < MrARM> sudo tcpdump -X -F /Users/Mac/Documents/tcpdump.txt 'port 25565'
19:50 < shoghicp> use wireshark
19:51 < MrARM> can’t I do that with just tcpdump?
19:52 < MrARM> but k, dl’ing it
19:53 < shoghicp> helps a lot in MCPE, we made a dissector there :D
19:54 < MrARM> anyway what did I do wrong with the command?
19:55 < shoghicp> I used tcpdump long ago, try the man pages
20:03 < MrARM> idk what I do wrong with tcpdump
20:03 < MrARM> anyway
20:04 < MrARM> it’s not a problem with encoding
20:04 < MrARM> if I’m having random bytes as output
20:04 < MrARM> (when message is same)
20:09 < shoghicp> maybe it is and you just stop parsing there :D
20:09 < shoghicp> it could be unitialized memory
20:10 < shoghicp> or you reading past the value due to not saving everything
20:11 < nickelpro> So GamingRobot looked through the MCP/decompiled client code and pulled out everything he thought was relevant to client-side physics. Now I'm more confused than ever. Client seems to have random values all over the place related to movement
20:12 < nickelpro> A lot in MoveEntity, some in LivingEntityBase, and handful in the block/material data
20:12 < shoghicp> that would be server-side physics :s
20:12 < nickelpro> There are no server side physics
20:12 < shoghicp> for example, dealing fire damage in movement code
20:12 < nickelpro> Ya, that through me for a loop
20:12 < MrARM> uh
20:12 < nickelpro> threw*
20:12 < shoghicp> no server-side physics?
20:13 < MrARM> just ran my old java-based test client
20:13 < shoghicp> server runs physics too
20:13 < nickelpro> For mobs, not clients
20:13 < MrARM> it crashed with OutOfMemoryError
20:13 < shoghicp> ah, right. The moveEntity() part is also used by players in the server
20:14 < shoghicp> to go from a position to another
20:14 < shoghicp> or at least, I made it use it
20:14 < shoghicp> and works fine :D
20:15 < nickelpro> Anyway, I'm close to giving up on perfectly mirroring Minecraft physics. A close facsimile seems to keep anti-cheat plugins happy enough
20:21 <+Amaranth> Most of the movement and physics code in the server is short circuited by the player subclass and then various bits of it are called in response to a movement packet
20:22 <+Amaranth> Way too many things only happen if the player moves
20:22 < shoghicp> I saw that, and I decided to tick players as normal entities
20:22 <+Amaranth> Normally that's not a problem because the client runs the same code so sees the condition and moves but if you take that out or have your own client... oops
20:22 < shoghicp> now movement only updates movement
20:23 <+Amaranth> Things like potion ticking, air when underwater, etc
20:23 <+Amaranth> Used to be a lot more
20:25 < MrARM> so my java client doesn't even get chat packets... wtf
--- Day changed lun. janv. 12 2015
07:58 < gamingrobot> wow chat is not easy to parse
14:11 < techkid6> hey, wiki.vg is down (assuming you didn't know)
14:15 < shoghicp> TkTech: Unknown MySQL server host 'db1.int.tkte.ch'
14:16 < shoghicp> tkte.ch gives a DNS error
14:17 < techkid6> shoghicp: right-o
14:18 < techkid6> on mob so can't really type much, also nust woke up
14:18 < techkid6> also, achievements, is that a packet sent to the client or?
15:48 < TkTech> shoghicp: Idiots at SWITCH allowed some fuckers to transfer my domain. It's back under my control and is currently locked.
15:48 < TkTech> Everything will be back up within 24-48 hours. All servers are still running.
16:05 < techkid6> .transfer... your... domain, despite you owning it?
16:38 < MrARM> yeah, how could it happen? 
16:44 < MrARM> techkid6: you can use the cached version from Google if you need 
16:51 < TkTech> By convicing the registrar support they were me and getting a transfer code.
16:56 < Weloxux> What registrar? Doesn´t sound like a very good one, if anyone can just pretend to be you. They are supposed to prevent this kinda stuff.
16:59 < TkTech> SWITCH. And no kidding.
17:07 < techkid6> Well, we've seen how registrars have fucked up in the past (ie @N) so this surprises me none
17:08 < MrARM> TkTech: can’t you make it use direct IPs instead?
17:08 < jast> sounds like a good argument for a... SWITCH of registrars
17:15 < Weloxux> ayyyyyyyyyyy, I see what you did there
17:15 < techkid6> jast: no, bad.
17:15 < techkid6> but how the hell are achievements even happening!
17:16 < techkid6> name changes
17:16 < techkid6> Feb 4
17:24 < TkTech> Just to be clear (I was just PM'd) - *ALL* servers and services are up, but my personal DNS is down.
17:24 < TkTech> Wiki.vg will be fixed up in a moment.
17:25 < techkid6> <3
17:27 < TkTech> Annnnnd she's up.
17:27 < TkTech> Sorry for the downtime.
17:28 < techkid6> awesome, thanks so much :P working on a suggestion and i need to know current status quo for it xD
17:38 < MrARM> k, so about that string reading thing
17:38 < MrARM> i had yesterday
17:38 < MrARM> it reads some the text actually, it just has junk between it, idk what it comes from though
17:39 < MrARM> *some of
17:44 < shoghicp> TkTech: great!
17:44 < MrARM> debugging it further
17:45 < MrARM> made it put all the output into a file
17:45 < MrARM> I can see the string fine there
17:45 < MrARM> so problem is somewhere else
17:46 < MrARM> either my byte array input stream is wrong
17:46 < techkid6> how are achievements even sent to the client, huh
17:46 < MrARM> or my char array reader
17:46 < techkid6> statistics
17:46 < techkid6> ?
17:55 < MrARM> fixed it
17:55 < MrARM> techkid6: yup
17:56 < techkid6> MrARM: thats a shame, what I'd like to do is have a way for the "Achievement" GUI to become a notification GUI, with achievements being only part of it
17:59 < MrARM> anyway, if anyone is wondering what I did wrong, I created an buffer that was being deallocated and my app was reading memory that got already modified
18:01 <+Amaranth> MrARM: Should have written it in Rust ;)
18:02 < MrARM> uh no
18:03 < barneygale> hacker news reader detected
18:03 <+Amaranth> barneygale: Rust user, actually
18:03 < MrARM> I’m actually coding it in C++ for a reason
18:03 < MrARM> as I’m gonna plug this into MCPE without too much hassle
18:04 < barneygale> Amaranth, playing with rust is about #45346 on my to-do list. 
18:04 <+Amaranth> heh
18:04 < barneygale> I need to find some good projects to contribute to
18:05 <+Amaranth> I'm one of the maintainers of the noise-rs library, we've got what I'm willing to bet is one of the fastest CPU implementations of perlin noise
18:06 <+Amaranth> At least for 2D noise
18:06 <+Amaranth> I'm not sure the 3D is up to par on speed yet
18:06 < shoghicp> hmm
18:07 <+Amaranth> I'm going to experiment with a SIMD implementation too but the compiler is autovectorizing the lerps already so I'm not sure how much gain there would be
18:07 < MrARM> also, in PlayerAbilitiesPacket, there’s a ’previous integer value divided by 250’
18:07 < MrARM> there’s no previous integer
18:08 < MrARM> so is it an float or it is an int?
18:45 < MrARM> k
18:45 < MrARM> I did all client->server packets :D
18:47 < Bibl> thnx boss
18:47 < Bibl> wait i dont see it MrARM 
18:47 < Bibl> it just says client or server
19:08 < Deaygo> Amaranth: I enjoy the * on the website.    * In theory. Rust is a work-in-progress and may do anything it likes up to and including eating your laundry.
19:09 <+Amaranth> :D
20:12 < TkTech> Notifico and any other services that depended on my DNS should now be back up. Sorry again for the downtime.
20:16 < Bibl> lol look what i just found in my old code
20:16 < Bibl> 								braceFor: /* impact lmao */for (; i < braceIndex; i++) {
20:16 < Bibl> 									LexerToken t = tokens[i];
20:18 < Xor_Boole> yey I broke models again http://puu.sh/eqwdL/c6182d297f.png
20:24 < Bibl> i like the magenta-black block
20:24 < Bibl> what is it
20:24 < Weloxux> probably one of them new stone types
20:24 < Weloxux> there´s hundreds
20:34 < dx> vaguely relevant to this channel: https://github.com/apache/thrift/blob/master/lib/java/src/org/apache/thrift/protocol/TCompactProtocol.java#L452-L458
20:34 < dx> "Convert n into a zigzag int. This allows negative numbers to be represented compactly as a varint."
21:21 < Xor_Boole> dx "zigzag int"
--- Log closed lun. janv. 12 21:27:29 2015
--- Log opened lun. janv. 12 21:33:46 2015
21:33 -!- Irssi: #mcdevs: Total of 149 nicks [1 ops, 0 halfops, 9 voices, 139 normal]
21:37 -!- Irssi: Join to #mcdevs was synced in 223 secs
--- Day changed mar. janv. 13 2015
05:09 < Not-f18a> [Glowstone] SpaceManiac pushed 3 commits [+2/-0/±8] http://git.io/MIrVGw
05:09 < Not-f18a> [Glowstone] ChioriGreene ce8b93b - Implemented door placement and interaction (see #119).
05:09 < Not-f18a> [Glowstone] jimmikaelkael 2634e8d - Added better game rule map and default values (see #455).
05:09 < Not-f18a> [Glowstone] SpaceManiac e75792e - Implemented doDaylightCycle, doTileDrops, and reducedDebugInfo.
16:18 < MrARM> hmm
16:18 < MrARM> http://wiki.vg/Protocol#Entity_Equipment
16:18 < MrARM> is that not an byte instead of an short?
16:26 < MrARM> nv,
16:26 < MrARM> *nvm
16:33 < MrARM> hmm, my code fails in reading a slot data, with reading the nbt data length short
16:33 < MrARM> only 1 byte is left, no two
21:05 <+SpaceManiac> MrARM: NBT data length isn't length-prefixed anymore
21:05 <+SpaceManiac> it's either a zero byte (no data) or the actual data (with no prefix)
21:08 < MrARM> Thanks 
21:09 < MrARM> isn't that 0 a TAG_END? 
21:22 < Fenhl> I guess you could call it that
22:40 < Xor_Boole> when are skin blobs going to require an yggdrasil signature?
22:40 < Xor_Boole> as in, is this actually going to happen?
23:03 < Thinkofdeath> Xor_Boole: they already do, its just the extra checks disabled
23:03 < Thinkofdeath> (excluding skulls)
23:03 < MrARM> are you ill, freenode? 
--- Day changed mer. janv. 14 2015
01:14 < Not-f18a> [Glowstone] SpaceManiac pushed 3 commits [+0/-0/±7] http://git.io/kywP9A
01:14 < Not-f18a> [Glowstone] SpaceManiac ffd04c4 - Implemented help.yml loading (closes #90).
01:14 < Not-f18a> [Glowstone] SpaceManiac 96e6bbe - Improved potion effect support (fixes #588).
01:14 < Not-f18a> [Glowstone] SpaceManiac 5303fee - Added player notification for difficulty changes.
--- Log closed mer. janv. 14 03:51:02 2015
--- Log opened mer. janv. 14 04:15:39 2015
04:15 -!- Irssi: #mcdevs: Total of 152 nicks [1 ops, 0 halfops, 9 voices, 142 normal]
04:19 -!- Irssi: Join to #mcdevs was synced in 217 secs
20:23 < MrARM> K, wrote my simple C++ nbt lib
--- Day changed jeu. janv. 15 2015
03:24 < Not-f18a> [Glowstone] SpaceManiac pushed 2 commits [+0/-1/±3] http://git.io/u0c8Cw
03:24 < Not-f18a> [Glowstone] ZephireNZ 1be5df0 - Provide a view for getOnlinePlayers().
03:24 < Not-f18a> [Glowstone] SpaceManiac a4e35c3 - Added better bind error messages (closes #586).
15:51 < MrARM> In http://wiki.vg/Object_Data, Falling Block data, how do I decode the block id/meta?
15:52 < Dinnerbone> It says in the "notes" column
15:55 < Dinnerbone> (Well, it says how it's encoded, you can reverse that of course)
16:32 < shoghicp> I know of someone that has not reviewed a pull request I made ;D
16:33 < Weloxux> Is it me? Is it me? :D *jumps up and down excitedly*
16:34 < shoghicp> NO!
16:35 < Dinnerbone> I've been busy with company workshop and hugging people and fixing computers and stuff! :(
16:35 < MrARM> dammit
16:35 < MrARM> I crashed xcode
16:35 < Dinnerbone> I did take a quick glance and do have comments though, you only do it for the query not ping too
16:36  * Weloxux starts sobbing
16:36 < shoghicp> well, I do it for the UDP case
16:36 < shoghicp> as ping starts a TCP socket
16:36 < shoghicp> it doesn't matter anymore
16:36 < shoghicp> but on the UDP case, it'll resolve the DNS on each packet sent
16:36 < dx> heh, pull requests to dinnerbro, without context that sounds weird
16:36 < shoghicp> so it's important to cache it, it can change or have a slow resolution
16:37 < Dinnerbone> True! I didn't consider that. That's why I only took a quick glance and didn't investigate it yet :D
16:40 < shoghicp> I added extra info to the PR
16:45 <+sadimusi> does this thing handle SRV records at all?
16:46 < shoghicp> I think that step is done before that code
16:46 < shoghicp> and that's for the port
16:47 <+sadimusi> not only the port, the host can be completely different as well
16:47 < shoghicp> yep
16:47 < shoghicp> D: only TCP
16:47 <+sadimusi> but yeah I found it
16:47 < shoghicp> https://github.com/Dinnerbone/mcstatus/blob/master/mcstatus/server.py#L22-L31
16:47 <+sadimusi> I was just wondering if you're using something more clever than python's own dns module since it tends to behave differently than java in some cases
16:48 < shoghicp> dnspython, I think
16:54 < Dinnerbone> It's better than the one used in Minecraft!
16:55 < Dinnerbone> Which doesn't work on all systems and sometimes gets it wrong. :D
17:01 < Dinnerbone> I pulled it in, shoghicp, thank you <3 I'll deploy a new version later.
17:01 < shoghicp> :D
17:02 < shoghicp> ping me once it is deployed so I can test it :)
17:45 < Paprikachu> hey Dinnerbone, could you please make packet headers fixed length? :3
17:46 < dx> Paprikachu: why
17:47 < Paprikachu> it makes stuff much easier
17:47 < Paprikachu> and there isn't really a reason to use varints there anyways.
17:47 < dx> i'm sure the easiest thing for mojang is to not change it
17:48 < Paprikachu> well, not changing anything is always the easiest thing, obviously
17:48 < dx> also you'd be dealing with backwards compatibility issues with the status part of the protocol
17:48 < dx> s/you/they/
17:48 < Paprikachu> could prob add a switch after handhshake.
17:48 < shoghicp> more code :D
17:48 < dx> and way uglier
17:49 < Paprikachu> welp
17:49 < dx> and if the best reason you have is that you're lazy...
17:49 < Paprikachu> not really
17:49 < Paprikachu> generating packets is a huge pita
17:49 < dx> 'generating'?
17:50 < Paprikachu> you know... putting the data together before sending it
17:50 < dx> if it's a huge pita then you should consider refactoring your code so that it's not
17:50 < Paprikachu> it fundamentally is
17:50 < dx> layers and stuff
17:50 < Paprikachu> yes, more inefficient layers :)
17:50 < Paprikachu> just think about it
17:51 < Paprikachu> if you want to keep the data in a single buffer, the following process has to happen
17:51 < Paprikachu> 1) leave enough space for the maximum possible header size
17:51 < Paprikachu> 2) encode the data at that location
17:51 < Paprikachu> 3) variably encode the type
17:51 < Paprikachu> 4) variably encode the length of that + data length
17:52 < Paprikachu> and that's pretty damn annoying, because you have to construct packets in reverse.
17:52 < Paprikachu> and inefficient too.
17:53 < dx> uhhh are you doing malloc()s manually in a way that it hurts you every time you do it?
17:54 < Paprikachu> and considering that the protocol uses json in some places, saving a single byte as opposed to having a 2 byte header length and a 1 or 2 byte type seems like a very questionable gain.
17:55 < Paprikachu> when you're trying to write a high performance server, cycles start to count.
17:55 < dx> considering that you'll have to encode json in some places, mallocs don't seem like a big issue
17:55 < Paprikachu> i'm not saying the protocol is a bottleneck, but i will say that there is a lot of cycles wasted for very little gain.
17:55 < dx> also this converstion is going nowhere
17:56 < Paprikachu> actually, json encoding doesn't have to use many allocations.
17:56 < Paprikachu> in fact, you can probably do with a single one, or none at all if you have a strategy for keeping buffers around (which i do)
17:56 < dx> omg all this microoptimization
17:57 < Paprikachu> i dont see any micro optimization
17:57 < Paprikachu> since im not performing any of those optimizations yet
17:57 < Paprikachu> im just saying it can be done.
17:57 < dx> "(which i do)"
17:57 < Paprikachu> i have such an optimization for receive buffers, yet
17:57 < Paprikachu> *yes
17:58 < Paprikachu> when you're aiming for 10k users, stuff like that counts.
17:59 < dx> so i assume you do profiling and benchmarking and have a decent test suite to prove all this?
17:59 < Paprikachu> i spend a fair amount of time making sure that my code is not designed in a way that is fundamentally inefficient.
17:59 < Paprikachu> other than that, i do what works and will optimize when i get there later.
18:00 < dx> http://blog.codinghorror.com/the-sad-tragedy-of-micro-optimization-theater/
18:00 < dx> read that, highly relevant to the way you're thinking about problems
18:00 < Paprikachu> heck i dont have a working protocol implementation yet, what the hell
18:00 < Paprikachu> no, i'm not. you
18:00 < Paprikachu> you are misunderstanding.
18:01 < dx> you don't even have a protocol implementation and you're asking about microoptimizing malloc()s
18:01 < dx> the post is relevant
18:01 < Paprikachu> do you not see how packet parsing is singnificantly worse with varints
18:01 < Paprikachu> i *already* use a single buffer for packets, there is nothing else to do for now.
18:02 < dx> there is no proof that it 'significantly', it might be in theory, but the theory you have in your mind doesn't necessarily match the way the compiler/operating system/cpu will handle your code
18:02 < dx> that it's*
18:03 < dx> read the post.
18:03 < Paprikachu> it doesn't change my original concern, which is that generating packets is significantly harder with next to no gain.
18:03 < Paprikachu> the argument "use library X" is a lazy excuse.
18:04 < dx> what? how are libraries involved?
18:04 < Paprikachu> oh i figured you would say that
18:04 < dx> i didn't even think of that
18:04 < Paprikachu> doing X is hard? use lib Y!
18:05 < Paprikachu> someone somewhere down the road parses those varints
18:05 < dx> uhhh no, using a library won't fix the way you think about optimization
18:05 < Paprikachu> he's gonna cry.
18:06 < Paprikachu> well sorry to disappoint, apparently *thinking* about optimizations is really bad.
18:06 < dx> read the damn post!
18:06 < Paprikachu> how about no
18:06 < dx> why?
18:06 < Paprikachu> because it doesnt fucking apply?
18:06 < dx> did you read it to reach that conclusion?
18:06 < Paprikachu> i am NOT optimizing ANYTHING at ALL
18:07 < Paprikachu> how did you reach the conclusion i am doing that
18:07 < Paprikachu> just because im having some thoughts?
18:07 < dx> yes, like i said, this is about the way you're thinking about the problem
18:07 < dx> it's not about any particular optimization
18:07 < dx> it's not even against optimizing
18:08 < Paprikachu> what im thinking is: i will do the slow thing first, and if that is too slow, i will do what i thought of before
18:08 < Paprikachu> i dont see whats fundamentally wrong with that
18:08 < Paprikachu> you're suggesting im optimizing without evidence, which i'm clearly not
18:08 < dx> i'm referring to what you said after
18:09 < dx> 13:50 < Paprikachu> if you want to keep the data in a single buffer, the following process has to happen
18:09 < dx> that way of thinking about problems is what the post addresses
18:09 < Paprikachu> except it's not
18:09 < dx> it's not?
18:09 < Paprikachu> i can add 1+1 you know
18:10 < dx> what
18:10 < Paprikachu> some "optimizations" aren't "optimizations" because they're just so fucking obvious that you don't even need to measure it
18:10 < dx> uh
18:11 < dx> okay i was actually trying to help, but since you keep refusing to read that post, i see you don't actually see what i'm doing as help
18:11 < dx> so i give up
18:11 < dx> i guess i already said that this conversation is going nowhere fast.
18:11 < Paprikachu> if you're handling 200k packets per second, you don't want any superficial allocations and copying, it's really easy
18:14 < Paprikachu> oh and another thing, varints prevent packets from being sent in batches
18:17 < dx> the current implementation limits the size of varints to 5 bytes, hope that helps
18:17 < Paprikachu> i already knew that, and no it doesn't
18:17 < dx> ok
18:17 < Paprikachu> the size of a varint(T) is (sizeof(T) * 8 + 6) / 7
18:18 < Paprikachu> you really dont seem to understand that copying is the issue
18:19 < dx> i was only pointing out a fact
18:36 < Paprikachu> actually, you know what, i have an idea that could actually work
18:36 < Paprikachu> though it doesn't solve the complicated varint problem
18:37 < Paprikachu> it should be possible, given a packet, to calculate its wire length in advance.
19:26 <+Amaranth> Most packets are like that, sure
19:26 <+Amaranth> They have a fixed structure, the only thing that changes is the values in them. Some of them also change the number of values with pretty large swings in size
19:38 < MrARM> in the Block Action packet, is Block Type same as Block Id in other packets?
19:50 <+SpaceManiac> MrARM: No, in Block Action it's still the old type (no data)
19:50 < MrARM> so what it actually is? Only block id?
19:51 <+SpaceManiac> yes
19:51 < MrARM> k, thanks
19:52 < MrARM> hmm, why is block id even sent here? :o
19:52 <+SpaceManiac> it's used to determine which category of action is being played (note block, piston, chest)
19:53 < MrARM> well, client could just get it from map… maybe it’s to prevent some stuff, like if map went desync
22:30 < gurun> Paprikachu, the best optimization so far has turned out to be caching of already serialized packets, ready to distribute.
22:32 < Paprikachu> gurun: and what would you cache
22:32 < gurun> chuks
22:32 < gurun> s/chuks/chunks
22:32 < Paprikachu> well sure, if you use compression
22:33 < Paprikachu> i'm thinking of just not compressing anything
22:33 < shoghicp> D:
22:33 < shoghicp> compress at least chunks!
22:33 < shoghicp> or packets with a size higher than the MTU
22:33 < Paprikachu> compression uses cpu
22:33 < Paprikachu> a *lot*
22:33 < shoghicp> the default value is too low, takes less time to send a packet ncompressed
22:33 < Paprikachu> not sure if worth
22:33 < shoghicp> but having to send mor epackets = bad
22:33 < gurun> well
22:34 < gurun> when you get to really high numbers of connectivity, i run out of bandwidth way before i run out of memory and CPU.
22:34 < Paprikachu> if you're aiming as high as me (10k concurrent players), stuff liek that matters
22:34 < shoghicp> 64 is... too low. 1470 - 1500 would be a good value
22:34 < shoghicp> was the default 64 or 128?
22:34 < Paprikachu> gurun: that might be true, i'd have to do the math
22:35 < Paprikachu> but since compression is optional, it can wait
22:35 < gurun> at 80 players on the MCPE protocol, i was running at 2Gbit/s
22:35 < gurun> 20Giga that is
22:35 < gurun> fuck
22:35 < shoghicp> I can tell you that sending a 256x256 block area took a lot of time (and CPU time) without compression (23MB), compression helped a lot
22:35 < gurun> 20mbit :-P
22:35 < Paprikachu> 20 mbit isnt an issue
22:35 < shoghicp> on zlib, I found levels 6-7 to be the best taking into account CPU usage
22:35 < Paprikachu> good servers have 100 mbit and up
22:35 < gurun> that was just 80 players
22:36 < Paprikachu> my mc server is not supposed to run on home servers
22:36 < gurun> but if you are playing the PC league, i guess you have TCP .. so you will be better off.
22:36 < Paprikachu> it's for large servers
22:37 < Paprikachu> my ram limit is somewhere around 16-32 gb (may use more if available) and bandwidth 200mbit+
22:38 < Paprikachu> it's a design choice to optimize for cpu usage
22:38 < Paprikachu> if i can do without compression, i will
22:41 <+SpaceManiac> shoghicp: the default is 256
22:41 < shoghicp> alright
22:42 < shoghicp> but as Minecraft sends packets directly, without bundling them in bigger packets
22:42 < shoghicp> having it compress only if bigger than MTU would be a good idea :)
22:43 < Paprikachu> MTU depends on who you
22:43 < Paprikachu> oops
22:43 < Paprikachu> MTU depends on who you're asking though
22:43 < shoghicp> yep
22:44 < Paprikachu> i'd just go with something like 1024
22:44 < shoghicp> I work with Minecraft: PE, there you need to guess the MTU
22:44 < shoghicp> the overhead of compressing is bigger than sending a packet :)
22:45 < Paprikachu> makes sense
22:49 < gurun> i think that comes down to where you want and can scle
22:50 < gurun> scale.
22:50 < gurun> cpu, memory or band
22:51 < Paprikachu> yeah well, i just don't think memory and bandwidth are an issue for big servers
22:52 < Paprikachu> from my own experience of hosting servers, going above 100 people is hard one a single instance
22:52 < Paprikachu> *on
22:54 < gurun> i thought you where talking about 10k
22:54 < shoghicp> he is
22:54 < Paprikachu> 10k is my goal for this server
22:54 < Paprikachu> i meant 100 players for a bukkit/spigot server
22:55 < gurun> ok, maybe the discussion should be around this .. how many players in line of sight do you imagine supporting?
22:55 < Paprikachu> no clue :D
22:55 < Paprikachu> that's probably just bandswith limited
22:55 < Paprikachu> bandwidth
22:56 < gurun> i don't think in neither MiNET or PM it would be a great issue suporting a LOT of players, as long as they don't "group up" too much.
22:56 < Paprikachu> not sure what either of those are
22:56 < gurun> it's just two different servers
22:57 < Paprikachu> what are they written in?