05:28 -!- TobiX [tobias@zoidberg.org] has joined #mcdevs
05:30 -!- Wulfspider [~Wulf@talk.spout.org] has quit [Ping timeout: 276 seconds]
05:42 -!- eddyb [~eddy@unaffiliated/eddyb] has joined #mcdevs
05:42 -!- Wulfspider [~Wulf@talk.spout.org] has joined #mcdevs
05:50 -!- evil_dan2wik is now known as lua
05:50 -!- Zachoz|Away is now known as Zachoz
05:51 -!- lua is now known as evil_dan2wik
06:08 -!- bobness [~silver@c-24-21-121-12.hsd1.or.comcast.net] has quit [Ping timeout: 256 seconds]
06:35 -!- RainbowDashTable [~cathode@c-76-105-184-52.hsd1.or.comcast.net] has joined #mcdevs
06:40 -!- GameMakerGm [~gamemaker@wikia/Gamemakergm] has joined #mcdevs
07:18 -!- Scootabyte [~Scootabyt@crown-7-41.resnet.ucsc.edu] has quit [Ping timeout: 256 seconds]
07:29 -!- Paprikachu [~Paprikach@178.115.250.80.wireless.dyn.drei.com] has quit [Remote host closed the connection]
07:45 -!- SpaceManiac [~SpaceMani@r74-192-152-131.gtwncmta01.grtntx.tl.dh.suddenlink.net] has quit [Ping timeout: 252 seconds]
07:46 -!- SpaceManiac [~SpaceMani@r74-192-152-131.gtwncmta01.grtntx.tl.dh.suddenlink.net] has joined #mcdevs
07:46 -!- mode/#mcdevs [+v SpaceManiac] by ChanServ
08:07 -!- kahrl [~kahrl@p5790CB0B.dip.t-dialin.net] has joined #mcdevs
08:09 -!- kahrl [~kahrl@p5790CB0B.dip.t-dialin.net] has quit [Client Quit]
08:14 -!- SpaceManiac [~SpaceMani@r74-192-152-131.gtwncmta01.grtntx.tl.dh.suddenlink.net] has quit [Ping timeout: 255 seconds]
08:19 -!- GameMakerGm [~gamemaker@wikia/Gamemakergm] has quit []
08:27 <+md_5> <pbunny> [01:31:59] i don't have mc client
08:27 <+md_5> codes mc server: doesnt own a client
08:27 <+md_5> how do you even know you can log in?
08:28 -!- Paprikachu [~Paprikach@80.120.175.113] has joined #mcdevs
08:28 < dx> maybe he tests with a rcon/query/server list client
08:28 <+md_5> still doesnt let him log in
08:29 <+md_5> also this is the guy who codes in header files and uses epoll/select with 2 threads per connection, and a thread for each antity
08:29 < dx> maybe his mc sever doesn't support the network protocol at all
08:29 <+md_5> and keeps entire map loaded in ram
08:29 < dx> ah lol that guy
08:29 -!- SpaceManiac [~SpaceMani@r74-192-152-131.gtwncmta01.grtntx.tl.dh.suddenlink.net] has joined #mcdevs
08:29 -!- mode/#mcdevs [+v SpaceManiac] by ChanServ
08:29 < dx> yeah, my reaction when i saw his header files... well...
08:29 < dx> let's leave it there.
08:30 < dexter0> Recursion must give him nightmares.
08:30 < dx> don't want to include details about misplaced digestive fluids in this channel
08:31 <+md_5> cant find a good meme for codes mc server: doesnt own a client
08:31 < dx> you probably need to mix a few memes to get a good one
08:34 < dx> md_5: maybe just the [10] guy
08:35 < dx> (the one who's really high)
08:37 <+md_5> dx http://www.quickmeme.com/meme/3t9v73/
08:37 < dx> perfect
08:39 -!- Paprikachu [~Paprikach@80.120.175.113] has quit [Remote host closed the connection]
08:39 -!- Paprikachu [~Paprikach@178.115.248.69.wireless.dyn.drei.com] has joined #mcdevs
08:40 < dx> >trolls.lv:1337
08:40 < dx> not sure if the hostname of his mc server is relevant
08:43 -!- kcj [~casey@unaffiliated/kcj] has quit [Ping timeout: 252 seconds]
08:43 -!- Justasic [~Justasic@unaffiliated/justasic] has quit [Ping timeout: 252 seconds]
08:45 -!- Justasic [~Justasic@unaffiliated/justasic] has joined #mcdevs
09:08 -!- Justasic [~Justasic@unaffiliated/justasic] has quit [Ping timeout: 252 seconds]
09:10 -!- Justasic [~Justasic@unaffiliated/justasic] has joined #mcdevs
09:16 -!- mappum [~mappum@c-67-170-21-29.hsd1.wa.comcast.net] has quit [Ping timeout: 255 seconds]
09:16 -!- conehead [~conehead@unaffiliated/conehead] has quit [Quit: Computer has gone to sleep.]
09:48 -!- MooseElkingtons [~Moose@buttshare.net] has quit [Read error: Operation timed out]
09:49 -!- Moose- [~Moose@buttshare.net] has joined #mcdevs
09:55 -!- Paprikachu [~Paprikach@178.115.248.69.wireless.dyn.drei.com] has quit [Remote host closed the connection]
09:57 -!- Paprikachu [~Paprikach@178.115.249.200.wireless.dyn.drei.com] has joined #mcdevs
10:03 -!- Broken_Syntax [~Nimbus@CPE001310777336-CM602ad07871c0.cpe.net.cable.rogers.com] has quit [Read error: Connection reset by peer]
10:04 -!- Broken_Syntax [~Nimbus@CPE001310777336-CM602ad07871c0.cpe.net.cable.rogers.com] has joined #mcdevs
10:19 -!- Moose- [~Moose@buttshare.net] has quit [Read error: Operation timed out]
10:23 -!- Moose- [~Moose@buttshare.net] has joined #mcdevs
10:23 -!- Moose- is now known as MooseElkingtons
10:27 -!- dx [~dicks@unaffiliated/dxdx] has quit [Ping timeout: 255 seconds]
10:27 -!- bobness [~silver@c-71-236-226-215.hsd1.wa.comcast.net] has joined #mcdevs
10:28 -!- dx [~dicks@unaffiliated/dxdx] has joined #mcdevs
10:52 -!- Xaardas [~tach@p5483C108.dip.t-dialin.net] has joined #mcdevs
11:15 -!- edlothiol [~edlothiol@2a02:810b:80c0:27:76e5:bff:fe22:870a] has joined #mcdevs
11:16 -!- SpaceManiac [~SpaceMani@r74-192-152-131.gtwncmta01.grtntx.tl.dh.suddenlink.net] has quit [Ping timeout: 276 seconds]
11:37 -!- Paprikachu [~Paprikach@178.115.249.200.wireless.dyn.drei.com] has quit [Remote host closed the connection]
11:39 -!- Paprikachu [~Paprikach@77.116.247.97.wireless.dyn.drei.com] has joined #mcdevs
11:52 -!- Scootabyte [~Scootabyt@crown-7-41.resnet.ucsc.edu] has joined #mcdevs
11:52 -!- mode/#mcdevs [+v Scootabyte] by ChanServ
11:55 -!- Justasic2 [~Justasic@174-25-82-237.ptld.qwest.net] has joined #mcdevs
11:56 -!- Justasic [~Justasic@unaffiliated/justasic] has quit [Ping timeout: 256 seconds]
11:57 -!- Justasic2 is now known as Justasic
11:58 -!- Justasic [~Justasic@174-25-82-237.ptld.qwest.net] has quit [Changing host]
11:58 -!- Justasic [~Justasic@unaffiliated/justasic] has joined #mcdevs
12:07 -!- Stormx2 [~Stormx2@cpc18-sotn9-2-0-cust33.15-1.cable.virginmedia.com] has joined #mcdevs
12:13 -!- SpaceManiac [~SpaceMani@r74-192-152-131.gtwncmta01.grtntx.tl.dh.suddenlink.net] has joined #mcdevs
12:13 -!- mode/#mcdevs [+v SpaceManiac] by ChanServ
12:14 -!- Grum [~grum@irc.grum.nl] has quit [Ping timeout: 272 seconds]
12:16 -!- Grum [~grum@irc.grum.nl] has joined #mcdevs
12:16 -!- Scootabyte [~Scootabyt@crown-7-41.resnet.ucsc.edu] has quit [Ping timeout: 245 seconds]
12:16 -!- Scootabyte [Scootabyte@crown-5-224.resnet.ucsc.edu] has joined #mcdevs
12:16 -!- mode/#mcdevs [+v Scootabyte] by ChanServ
12:20 < pbunny> hi, client crashes when i send 'spawn mob' packet
12:20 < pbunny> any suggestions?
12:26 -!- Paprikachu [~Paprikach@77.116.247.97.wireless.dyn.drei.com] has quit [Remote host closed the connection]
12:29 -!- TomyLobo [~foo@91-66-94-221-dynip.superkabel.de] has joined #mcdevs
12:30 -!- Paprikachu [~Paprikach@178.115.250.240.wireless.dyn.drei.com] has joined #mcdevs
12:34 -!- Scootabyte [Scootabyte@crown-5-224.resnet.ucsc.edu] has quit [Quit: Leaving]
12:38 <+md_5> pbunny how would you know? You dont have the client
12:39 <+md_5> unless you pirated
12:39 <+md_5> in which case gtfo
12:40 < pbunny> md_5: a guy tries and gets crash
12:40 <+md_5> you are gonna have fun writing a server without owning the client
12:40 <+md_5> seriously, how the f** did you get this far
12:40 < edk> md_5, by lying
12:40 < pbunny> md_5: http://dpaste.org/IbgPM/raw/
12:40 < pbunny> edk: no.
12:41 < pbunny> md_5: as you see, client actually lists 'Pig' entity in crash dump
12:41 < pbunny> so it parsed it
12:41 < pbunny> why did it crashed?
12:51 < Grum> because it got a NPE
12:53 -!- act4 [1f073cdc@gateway/web/freenode/ip.31.7.60.220] has joined #mcdevs
12:55 < pbunny> Grum: NPE?
12:55 < Grum> NullPointerException
12:55 < edk> (like it says in the error report)
12:55 < pbunny> yeah, and NullPointerException can mean anything
12:55 < pbunny> java sucks totally
12:56 < Grum> not really, you just used a nullpointer :p
12:56 < pbunny> if i don't send 0x18 packet, client does not crash
12:56 < pbunny> Grum: its a native client
12:56 < pbunny> i just sent 0x18 to it
12:56 < Grum> i guess you have to send more
12:56 < Grum> or with proper information
12:58 < pbunny> well it is correct according to http://wiki.vg/Protocol
12:58 < Grum> but the question remains, is THAT correct
12:59 < pbunny> i didn't sent Entity (0x1E) though
12:59 < pbunny> i haven't need it for item entities
12:59 < edk> java does kinda suck
13:00 < Grum> every language sucks
13:00 < edk> yep
13:00 < edk> in their own way, most things are shit
13:00 < pbunny> and is it safe to send 127 in metadata for any mob?
13:00 < edk> the reason java annoys me particularly is it tries to be all not crashy and predictable and compile time and type safe
13:00 < Grum> try and figure out pbunny :D
13:00 < pbunny> still waiting for him to try
13:01 < Grum> edk: erm 'not crashy' and 'predictable' and 'typesafe' are bad? O.o
13:01 < edk> oh, they would be if it worked
13:01 < edk> wouldn't be*
13:01 < edk> I mean, yes, it's type safe
13:01 < edk> but it doesn't make it crash any less, or need less testing
13:01 < edk> so I don't see the point
13:01 < Grum> euuh
13:02 < Grum> *ALL* code needs testing
13:02 < edk> right
13:02 < Grum> no matter what language
13:02 < edk> I agree
13:02 < edk> but that makes me ask the question "why bother having a pessimistic language if you have to test it anyway?"
13:02 < pbunny> 'typesafe' is ridiculous
13:02 < edk> you're ridiculous
13:02 < Grum> euuh wtf?
13:02 < pbunny> in C i can set any pointer to anything
13:03 < edk> I quite like type safety
13:03 < pbunny> no matter the type
13:03 < Grum> awesome nonsense arguments >.>
13:03 < edk> I just don't like the attitude in java
13:03 < Grum> which attitude in java?
13:03 < pbunny> 'predictable' is also BS as it depends on programmer's skill
13:03 < edk> It's hard to pin down. I feel like java has this idea that by making you type a type name enough times, it somehow becomes harder for things to crash
13:03 < pbunny> java is not more 'predictable', java is more 'predictable for noobs'
13:03 < Grum> erm what?
13:04 < Grum> you use an IDE right?
13:04 < Grum> i dont see any problems
13:04 < pbunny> no.
13:04 < pbunny> i use nano
13:04 < edk> yeah, but you're insane
13:04 < Grum> i mean, you wouldnt code java without an ide, if not ... well ..... stop being so daft
13:04 < pbunny> i never code java
13:04 < edk> this is what I don't get
13:04 < Zachoz> Then how can you say it's shit?
13:04 < edk> my complaints about Java may be unreasonable and unfair, but at least I used it before making assertions about it :P
13:05 < Grum> edk: next you are going to argue that a variable named of a single letter is 'much better' than one that describes the meaning properly
13:05 < edk> not so
13:05 < pbunny> edk: i tried to use it
13:05 < edk> but the java way where there's millions of casts everywhere... ugh
13:05 < Grum> erm there is not a single cast anywhere
13:05 < pbunny> when it attempted to javize my brain i threw it away
13:05 < edk> maybe that's a problem with the way specific people write java
13:05 < pbunny> there is actually 'brain java' disease
13:05 < Grum> i've just written ~300 lines of code without having to cast anything
13:05 < pbunny> people start to think like in java, with object and stuff
13:05 < Grum> i think you just suck at writing clean code :)
13:06 < edk> I write plenty of java without many casts in it
13:06 < Grum> or do not understand some of the basic concepts
13:06 < edk> but I see a lot of java that does have them in
13:06 < edk> while we're on the subject of things I don't like about java, fucking generics
13:06 < edk> grrrrr
13:06 < Grum> there are some points where you need casting; not so many
13:06 < pbunny> java is about using 50Mb of RAM for player structure
13:06 < edk> it's not about that
13:06 < Grum> pbunny: its not?
13:06 < pbunny> Grum: official server uses about that
13:07 < pbunny> that's why no 1000+ online server yet
13:07 < edk> one wonders how you worked this out
13:07 < pbunny> not even 100+ online iirc
13:07 < Grum> just for the player structure?.... not really
13:07 < edk> and why you insist on using b for bytes
13:07 < Grum> pbunny: no that is not at all the reason
13:07 < pbunny> edk: because theres also MiB abbrevuature
13:07 < Zachoz> "pbunny: not even 100+ online iirc".. Are you kidding...?
13:07 < pbunny> and nobody needs to concern about bits for at least 20 years
13:07 < edk> however Mb is clearly megabits
13:07 < edk> ...
13:07 < edk> you're retarded
13:07 < pbunny> Zachoz: is there 100+ online server anywhere?
13:08 < pbunny> i mean - a server, not collection of servers
13:08 < edk> fuck yes there are
13:08 < Grum> stop whining about how you captitalize the units god
13:08 < Zachoz> Yes... There's quite a few, actually.
13:08 < Grum> he said 1000+
13:08 < edk> Grum, it's a different fucking unit
13:08 < pbunny> hmh
13:08 < edk> no he didn't
13:08 < Grum> LEARN TO READ Zachoz
13:08 < edk> <pbunny> not even 100+ online iirc
13:08 < Zachoz> ^
13:08 < edk> b and B are not the same thing
13:08 < edk> if it were almost any other letter I wouldn't give a shit
13:08 < Grum> edk: its not, if you mean megabit just fucking type mbit not mb Mb MB mB ...
13:09 < Grum> are you really saying that you ONLY read 'mb' as millibit or millibar?
13:09 < edk> you know what? take that up with SI, not me
13:09 < edk> no, but Mb and MB are equally valid as units of data
13:10 < edk> just one's 8 times the size of the other
13:10 < edk> and I use both units fairly frequently at work
13:10 < Grum> not really; Mega bar is not really a useful measuring tool
13:10 < edk> megabit and megabyte
13:10 < Grum> right lets just stop whining about the capitalization of a letter kk?
13:10 <+Fador> ...you really measure memory usage in bits?
13:11 < edk> who said I measure memory usage in them?
13:11 < Grum> edk: god ....
13:11 < Grum> its just fador who thinks he can be super funny >.>
13:11 < edk> i'm quite happy to stop arguing about it
13:12 < edk> i just get caught up in arguments and then don't want to "lose"
13:12 < Grum> anyhow, pbunny, i'd like to see your 'proof' for the player-structure taking 50mbit/mb
13:12 < edk> i know that's silly
13:12 < edk> yeah, I'm unconvinced by that
13:12 < Grum> i know it is not true
13:12 < Grum> so feel free to prove it otherwise
13:12 < edk> still, a rewritten server would go faster
13:12 < Grum> faster what?
13:12 < edk> but it would go faster if it were rewritten in java, too
13:12 < edk> faster than the current one
13:13 < Grum> define faster
13:13 < edk> *sigh*
13:13 < pbunny> edk: not quite
13:13 < Grum> yes, be fucking specific for once
13:13 < edk> well, let's settle for "take fewer CPU cycles per tick"
13:13 < pbunny> you can't just code something fast in java
13:13 < Grum> faster what? logging in? higher tps? faster ping times?
13:13 < Grum> pbunny: false statement
13:13 < pbunny> because i.e. when you do a=2+5 java actually executes about 500 CPU instructions
13:13 < edk> logging in bottleneck isn't the server, as far as I know, it's doing the authentication dance
13:14 < Grum> pbunny: really?
13:14 < edk> pbunny, consider reading about some of the optimizations the JVM has enjoyed since being invented
13:14 < pbunny> edk: still it uses more than 10 instructions for addition
13:14 < pbunny> i.e. bounds checking
13:14 < Grum> that would mean it is at least 125 slower than ANYTHING out there
13:14 < pbunny> and don't forget about garbage collector!
13:14 < edk> garbage collection is a real weakness of java
13:14 < pbunny> it just stops everything and does its dirty work
13:14 < Grum> which would mean a client not running java but c would have 2500 fps on my machine?
13:14 < edk> however, it's better than a few alternatives
13:14 < Grum> and AT MINIMUM 125fps on ANY machine?
13:15 < edk> Grum, well, unless the bottleneck is the gpu
13:15 < edk> anyway, GC is a pain in the ass, and it is a problem
13:15 < Grum> pbunny: it stops everything? seriously? >.>
13:15 < pbunny> Grum: yeah. it can't check memory during its changing
13:15 < edk> but if the alternative is your horrible C that leaks allocated storage everywhere...
13:15 < pbunny> edk: my C leaks nothing
13:15 < edk> I'll believe that when I see it
13:16 < Grum> then again, you do not have a server written
13:16 < pbunny> and it uses abotu 80Mb for everything
13:16 < Alan> oh boy
13:16 < pbunny> Grum: partially written
13:16 < Grum> without anyone online?
13:16 < Grum> epic
13:16 < Grum> that is even worse than the official server >.>
13:16 < Alan> what a fun channel this is
13:16 < pbunny> Grum: it uses about 300 bytes per player online
13:16 < Grum> pbunny: then its not a minecraft server
13:16 < pbunny> ...
13:16 < Alan> half a proof-of-concept that isn't feature complete isn't really comparable...
13:17 < Grum> because it is 100% impossible to have that little memory usage for an online player *AND* actually have a proper minecraft server
13:17 < edk> well
13:17 < Grum> s/proper/featureful/
13:17 < Grum> s/featureful/feature complete/
13:17 < Grum> lets be more exact :)
13:17 < pbunny> Fador: can you please show me code that sends 0x18 (Spawn Mob) packet?
13:18 < edk> I'm trying to think about what you actually need to store for a player
13:18 < pbunny> Grum: why?
13:18 < pbunny> Grum: i store player's location, inventory, network stuff
13:18 < Grum> pbunny: because each player has to have at least 19*16x19*16x256 blocks loaded around itself
13:18 < pbunny> LOL
13:18 < edk> I think he's discounting that
13:18 < Grum> i know he is
13:18 < pbunny> now that's a java thinking
13:18 < edk> you can have single inventory items that take more than 300B though
13:18 < Grum> pbunny: no its not
13:18 < edk> pbunny, that's stupid
13:18 < pbunny> Grum: why the hell should i store the map in player's structure????
13:19 < edk> that's just someone interpreting the question different from you
13:19 < Grum> pbunny: quote me where i say that you store that IN the player structure?
13:19 < pbunny> Grum: if 1000 players will be standing on the same place - i must store the map around them 1000 times?
13:19 < pbunny> oh god why
13:19 <+Fador> pbunny: https://github.com/fador/mineserver/blob/master/include/protocol.h#L111
13:19 < Grum> quote me.
13:19 < Grum> because you are just making shit up
13:20 < edk> I wouldn't bother waiting for that quote :P
13:20 < pbunny> 13:37:23         Grum | pbunny: because each player has to have at least 19*16x19*16x256 blocks loaded around itself
13:20 < Grum> i'm saying it should be loaded, not saying WHERE it should be stored/loaded
13:20 < edk> pbunny, although a couple of the problems you have listed in java are actually real problems
13:20 < Grum> yes, loaded around itself, i'm not saying anything about WHERE to store that
13:20 < edk> you seem to completely misunderstand their significance and when they apply
13:20 < edk> so I'm inclined to say you don't actually know anything about java at all
13:20 < pbunny> Grum: why would server keep a copy of map that he loaded?
13:20 < pbunny> why would it take memory on server?
13:20 < Grum> because that is what a feature complete minecraft server does
13:20 < Grum> but you'll figure that out when you grow up
13:20 < pbunny> storing a map for every player?
13:20 <+Fador> pbunny: metadata should be just (int8_t)127 if no metadata is present
13:20 < Grum> no ....
13:21 < Grum> loading the world-data around players in memory
13:21 < pbunny> Fador: well i send 127
13:21 < edk> I have to go
13:21 < edk> have fun
13:21 < edk> pbunny, try not to hurt your brain
13:22 < Grum> pbunny: you do far too many assumptions
13:22 < pbunny> Fador: << (int8_t)0 << (int8_t)0 << metadata;
13:22 < Grum> but you'll find out soon enough why its not trivial to be a feature complete minecraft server
13:22 < pbunny> what are these 2 zeros there? o.O
13:22 < Grum> *AND* be fast
13:22 < pbunny> Grum: map is already loaded on server...
13:22 < pbunny> it is always kept there loaded......
13:22 < Grum> the whole map?
13:22 < pbunny> yeah
13:22 < Grum> lol ...
13:22 <+Fador> pbunny: not sure..it might still be using some older protocol version because I haven't updated everything =/
13:23 < Grum> you do realize that people have a 1.5gb map very often?
13:23 < pbunny> Fador: i will try these zeros
13:23 < Grum> or even far far bigger
13:23 < pbunny> Grum: do you realize my server has 8Gb of RAM?
13:23 < pbunny> and it will have 256Gb
13:23 < Grum> >.>
13:23 < pbunny> 256Gb is quite enough for everything
13:23 < jast> Grum: we've done all this before
13:23 < Grum> will be nice startup times then for you lol :p
13:23 < pbunny> Grum: nice performance on other hand
13:23 < jast> if you're not in it for the entertainment, you can't back out too early
13:23 < Grum> loading the whole world into ram, epic startup wait >.>
13:24 < pbunny> Grum: so?
13:24 < pbunny> good performance after it for everything
13:24 < Grum> pbunny: yeah because waiting 20 minutes before your whole world is loaded is worth it >.>
13:24 < pbunny> and world will be a true world, not 'combined singleplayer' 
13:24 < Grum> what?
13:24 < pbunny> mobs will be independent etc
13:24 < Grum> that makes no sense at all
13:24 < pbunny> think about it some more then
13:24 < Grum> 'true world', 'combined singleplayer'
13:24 < Grum> those terms mean nothing
13:25 < jast> he's explained it before
13:25 < pbunny> yeah, currently mc servers offer 'combined singleplayer'
13:25 < pbunny> no players -> nothing happens
13:25 < Grum> what? O.o
13:25 < Grum> euuh
13:25 < pbunny> world is saved and stopped
13:25 < Grum> no stuff happens in the areas of the world that is loaded
13:25 < pbunny> until player comes
13:25 < jast> the idea is, apparently, that supposedly it's 5000% better if the world still updates everywhere even when nobody is near (or even logged in)
13:25 < Grum> lol
13:25 < Grum> LOL
13:25 < jast> yes
13:25 < jast> exactly
13:25 < Grum> good luck running out of cpu on the best of machines
13:25 < pbunny> jast: yes, it allows for better realism
13:25 < Grum> there is a good reason we unload parts of the world
13:26 < jast> Grum: oh, but what you don't know is that pbunny is a self-certified genius and master of C
13:26 < pbunny> Grum: CPU usage is about 1% for processing world and physics for 1000 entities
13:26 < Grum> only 1000 yes
13:26 < Grum> but entities would keep spawning
13:26 < pbunny> and its when running via valgrind compiled with debug
13:26 < jast> so for 100000 entities it's 100% :)
13:26 < Grum> because the world is loaded
13:26 < pbunny> jast: if running via valgrind
13:26 < jast> yeah yeah
13:26 < jast> we haven't considered a bigger world yet
13:26 < Grum> so yeah, having a couple of million entities and dont forget all the other stuff the server has to do
13:26 < pbunny> Grum: sure, no problem
13:26 < Grum> really?
13:26 < Grum> and how fast do you think to finish this server?
13:26 < jast> if the surface of the world increases by a factor of 2, the RAM required increases by a factor of up to 512
13:27 < Grum> that 'has no problem' having thousands of people online while keeping any map that fits in memory loaded forever?
13:27 < jast> in short, web scale!
13:27 < Grum> because i'd like to see it :)
13:27 < Grum> oh, and obviously do *everything* the current minecraft does logicwise
13:27 < jast> you will, Real Soon Now(tm)
13:28 < jast> did we mention that pbunny does not have a minecraft client?
13:28 < Grum> oh also, how will it get to this size world?
13:28 < pbunny> jast: what do you mean by "surface of the world increases"?
13:28 < Grum> pbunny: people explore
13:28 < pbunny> what exactly increases by factor of 2?
13:28 < pbunny> Z size? X size?
13:28 < Grum> or do you plan to pre-gen a 60million*60million world?
13:28 < jast> surface area
13:28 < pbunny> Grum: no.
13:28 < Grum> so how do you plan to do worldgeneration?
13:28 < jast> surfaces are two-dimensional
13:28 < pbunny> Grum: i already done that
13:28 < Grum> did you also do the lighting of them?
13:28 -!- act4 [1f073cdc@gateway/web/freenode/ip.31.7.60.220] has quit [Ping timeout: 245 seconds]
13:29 < pbunny> just call the relevant function of needed world prior to using this chunk
13:29 < jast> oh please, lighting is for free! didn't you know?
13:29 < Grum> also i really wouldnt want to see this code in c
13:29 < Grum> i mean it must have locks all over the place :)
13:30 -!- Zachoz is now known as Zachoz|Away
13:30 < Grum> beceause the only way you'll get remotely close to what you claim to achieve performance wise is to use every single core on a system
13:30 < Grum> and good fucking luck with the locking :)
13:30 < pbunny> Grum: http://dpaste.org/P7mog/ - here you go
13:31 < pbunny> common world loop, some physics stuff, handling of dumping entities to players
13:31 < Grum> and no threading
13:31 < Grum> i'd love to see your lighting code
13:31 < Grum> assuming it is feature complete and correct
13:31 < pbunny> no threading?
13:31 < jast> because it's so cheap, let's implement radiosity, too
13:31 < Grum> wait; there is threading
13:31 < Grum> nevermind :)
13:32 < Grum> the fact that 95% of the code is on the righthand side of my page kinda raises a huuuuuge red flag btw
13:34 < jast> hey, m*n loop on entities? awesome
13:34 < pbunny> Grum: http://dpaste.org/MQNyD/raw/
13:34 < pbunny> boo!
13:34 < jast> and it's even symmetrical
13:34 < Grum> but i'm really curious how you did your block and skylighting calculations
13:35 < jast> oh, actually it's an m*n*p loop in some situations
13:35 < pbunny> Grum: search for "chunk_set_block" there
13:35 < pbunny> or chunk_set_field / chunk_set_hfield
13:35 < jast> i.e. one billion iterations for 1000 entities
13:35 < Grum> that doesn't set any light pbunny
13:35 < Grum> you just set the heightmap
13:35 < pbunny> jast: i will optimize that
13:36 < pbunny> split entities by quadrants
13:36 < pbunny> so only nearby ones will be concerned
13:36 < pbunny> Grum: i haven't implemented lighting yet, but you can set it with chunk_set_hfield
13:36 < Grum> pbunny: i'm missing the skylight and blocklight calculations
13:36 < jast> oh see, he just hasn't implemented it yet
13:36 < Grum> you cant just 'set' a light value without doing propagation
13:37 < jast> but of course we all know how easy it is to do that
13:37 < pbunny> jast: easier than making client not crash on pig entity packet receive
13:37 < pbunny> :|
13:37 < jast> if you say so
13:37 < Grum> also, all of your macros are broken, unless c++ automatically puts ()'s around your arguments :)
13:38 < pbunny> no.
13:38 < pbunny> ( ) are not required
13:38 < pbunny> go learn some C
13:38 < jast> well, they aren't if he happens to be using variable names whenever he passes stuff
13:38 < pbunny> right
13:38 < jast> if he doesn't... boom
13:38 < Grum> erm they are pbunny? if you define a macro as: foo(i) { return i*2; } you better make that return (i)*2
13:39 < pbunny> Grum: not quite
13:39 < jast> oh, but you never pass something more complex than 'foo' to macros. everyone knows that!
13:39 < Grum> because when you'd ever be passing in: 1+1 as argument for i
13:39 < jast> err, bad example word :)
13:39 < Grum> it expands to: return 1+1*2 ... which is NOT 2 :P
13:39 < pbunny> Grum: if i would need to pass 1+1, i will use ( )
13:39 < jast> foo(1+1) == 3! yay!
13:39 < pbunny> don't worry about that
13:39 < Grum> pbunny: oooh wait; so when you use the macro with a composed argument you will remember to fix up your macro?
13:39 < jast> so you always have to remember which things you're calling are macros and which aren't. you must really hate yourself...
13:40 < pbunny> Grum: of course
13:40 < Grum> i mean, that is not at all a way that will give you unexpected bugs at any time
13:40 < pbunny> jast: everything are macros there
13:40 < pbunny> except for main() and thread functions
13:40 < jast> yeah, because you hate yourself, right
13:40 < pbunny> and some init functions that are called only once
13:40 < pbunny> jast: what is wrong with defines?
13:40 < jast> and because the inline keyword doesn't exist
13:40 < pbunny> i don't get it
13:40 < pbunny> jast: its not the same
13:41 < Grum> oh nothing; as long as you write them properly ;)
13:41 < jast> the code gets even less readable
13:41 < Grum> these arent really robust
13:41 < pbunny> jast: its 100% readable by me
13:41 < Grum> if this isnt robust .. that says quite some about the rest of the code
13:41 < pbunny> what do you mean by "not robust"?
13:41 < Grum> 'it breaks when you put shit in it depending on what you put in it'
13:41 < Grum> aka: fragile :P
13:41 < pbunny> Grum: this applies to any code
13:41 < pbunny> you shouldn't put shit there
13:42 < pbunny> well except for java which is shit itself
13:42 < Grum> erm
13:42 < pbunny> so it isn't suffering
13:42 < Grum> foo(1+1) is not 'shit' :)
13:42 < Grum> so tempting to find a bug in your code right now where you failed to put ()'s in your macro ;)
13:42 < Grum> but i dont think you have implemented enough to easily find one
13:43 < pbunny> Grum: i don't spawn bugs.
13:43 < Grum> actually, you do
13:43 < pbunny> and you are right, shitty developers spend 80% of their time to 'implement enough' to find bugs easier
13:44 -!- SpaceManiac [~SpaceMani@r74-192-152-131.gtwncmta01.grtntx.tl.dh.suddenlink.net] has quit [Ping timeout: 245 seconds]
13:44 < Grum> anyhow, pbunny, is this project of yours opensource?
13:44 < jast> nope
13:44 < jast> he's IP-paranoid and stuff
13:45 < pbunny> ip-paranoid?
13:45 < Grum> just let him answer jast gawd
13:45 < jast> whatever. have fun.
13:45 < pbunny> Grum: no, maybe in the future
13:45 < pbunny> do you want to participate?
13:46 < Grum> i'd like to see you solve some problems
13:46 < pbunny> what problems?
13:46 < Grum> lighting for example
13:47 < pbunny> don't see problem there
13:47 -!- SpaceManiac [~SpaceMani@r74-192-152-131.gtwncmta01.grtntx.tl.dh.suddenlink.net] has joined #mcdevs
13:47 -!- mode/#mcdevs [+v SpaceManiac] by ChanServ
13:47 < pbunny> what seems hard for you?
13:47 < Grum> what do you consider lighting?
13:47 < pbunny> light and skylight portion of chunk
13:48 < Grum> right, so that code is easily to do efficiently? :)
13:48 < pbunny> depends on algorithm
13:48 < pbunny> and requirements
13:48 < Grum> the algorithm is: you are one lighter than your brightest neighbor; max 15; min 0;
13:49 < Grum> blocks can be lightsources of different strengths and allow light to pass through 
13:49 < Stormx2> Grum, you can't win an argument with pbunny. He brings you down to his level then beats you with experience
13:49 < Grum> some blocks let light through but they diminish it with more than 1, like water
13:50 < pbunny> Grum: most probably i will create a special type of metaentity - lightsource
13:50 < Grum> so the furthest light will go from a single source is 15 blocks
13:50 < pbunny> and will iterate through them in entities loop
13:50 < pbunny> in world.c
13:50 < pbunny> they will update light around them if needed
13:50 < pbunny> or simply on every iteration
13:50 < Grum> but you will check every gametick if it is needed?
13:50 < Grum> for all lights?
13:50 < Grum> for a single floating torch you'll end up checking ~4k blocks around it
13:50 < pbunny> Grum: don't see any problems of updating ~30 bytes per entity per tick
13:51 < pbunny> why 4k blocks?
13:51 < Grum> because that is how far the light shines?
13:51 -!- Justasic2 [~Justasic@174-25-126-72.ptld.qwest.net] has joined #mcdevs
13:52 < pbunny> well implement some tracking of light changes
13:52 < pbunny> so light will be updated only when needed
13:52 < pbunny> i.e. when torch is placed
13:52 < Stormx2> How do you track which block actions affect which lights?
13:52 < Stormx2> e.g. a piston moving a block
13:52 < Grum> pbunny: block placement, block breaking indeed
13:53 < Grum> or rather, on delta-block :)
13:53 < pbunny> Stormx2: any block movement inside light-change-trackable area will trigger light update
13:53 < Grum> but that is not really the issue
13:53 < pbunny> block change *
13:53 < pbunny> i don't see a problem
13:53 < Grum> the issue is the sheer volume of blocks you have to calculate the lighting for
13:54 -!- Justasic [~Justasic@unaffiliated/justasic] has quit [Ping timeout: 250 seconds]
13:54 -!- Justasic2 is now known as Justasic
13:54 < Grum> and because light *can* decay with more than 1 you are limited in the amounts of shortcuts you can take
13:54 -!- Justasic [~Justasic@174-25-126-72.ptld.qwest.net] has quit [Changing host]
13:54 -!- Justasic [~Justasic@unaffiliated/justasic] has joined #mcdevs
13:54 < Grum> but please do, i'm really interested if you find a good & fast way
13:56 < pbunny> i don't see the problem...
13:56 < pbunny> it's like 200 bytes write (at most) for light update
13:56 < pbunny> maybe 400 bytes read
13:56 < Stormx2> pbunny, the issue isn't the number of bytes you're writing...
13:56 < pbunny> RAM can handle that
13:57 -!- zz_dav1d [dav1d@static.82.162.46.78.clients.your-server.de] has quit [Quit: nothing to see]
13:58 -!- dav1d [dav1d@unaffiliated/dav1d] has joined #mcdevs
14:03 < pbunny> Stormx2: what is the issue?
14:05 < pbunny> can somebody join trolls.lv:1337 and test the freaking pig?
14:07 < pbunny> can client crash if pig entity is (theoretically possible) inside block?
14:37 < Stormx2> no
14:38 < pbunny> what can be the problem then
14:38 < Stormx2> pbunny, the issue is cpu cycles as lighting is O(a^n)
14:39 < pbunny> so what
14:39 < pbunny> lightning won't change every tick
14:39 < pbunny> lighting *
14:39 < Stormx2> so an implementation written by people who understand what big-O is is slow
14:40 < jast> have fun at dusk and dawn, then
14:40 < pbunny> dusk and dawn doesn't happen every tick
14:40 < Stormx2> jast, that's client-side
14:40 < Stormx2> the lighting changes anyway
14:40 < pbunny> oh?
14:40 < Stormx2> server still sends the same skylight
14:41 < Stormx2> yeah, it does it based on time irrc
14:41 < jast> right. magic.
14:41 < jast> but still you'd have to recalculate light
14:41 < jast> after all you need to figure out the correct light level for places that used to be lit by sun and now aren't
14:42 < jast> I suppose you could cheat and do it just once
14:42 < dav1d> meh I am bored
14:42 < jast> (per dusk and per dawn)
14:42 < dav1d> sitiing at university "learnling" java -.-
14:42 < jast> dav1d: feel free to join in yet another iteration of this pointless discussion
14:42 < Stormx2> not true. the client treats blocklight and skylight separately. The *actual* skylight it computes is a product of the skylight sent in chunks, and the time-of-day
14:42 < Stormx2> i.e. you don't need to do any new light calculations for night-time
14:42 < dav1d> atm Hello World, line for line explanation
14:43 < dav1d> jast: light calculations?
14:43 < jast> so the server doesn't have to decide where it can and can't spawn mobs?
14:43 < jast> dav1d: that's just a sub-topic
14:43 < Stormx2> Oh, shit.
14:43 < dav1d> if you come up with a good and fast way, let me know, gotta implement it for BraLa some day
14:43 < Stormx2> Hm.
14:43 < dav1d> ah ok
14:43 < jast> the overarching story is a certain person's super server that does everything
14:43 < Stormx2> I'm sure I recall someone telling me it moved client-side but perhaps I'm wrong
14:43 < dav1d> and they use a mac at university -.-
14:43 < dav1d> they should know better
14:44 < jast> well, I suppose there's no point in sending the dusk/dawn-related light updates to the clients, but you still have to calculate it on the server
14:45 < pbunny> jast: my server won't "spawn" mobs
14:45 < pbunny> mobs will reproduce
14:46 < jast> I know, I know
14:46 < jast> but you still need to know about lighting if you want to, say, determine correct growth for plants
14:46 < pbunny> i.e. pigs will infest players with larvae which after some time will kill player and several new pigs will be born
14:46 < jast> oh, just like in real life
14:46 < pbunny> yeah
14:47 < pbunny> plants will eat players and other mobs and grow after that
14:47 < pbunny> they will need nutrients
14:48 < jast> I can see it now... a patch of high grass chasing and eating players
14:49 < pbunny> no, they won't move
14:49 < pbunny> but they can digest anything that will step on them, slowly
14:51 < pbunny> i will need light calculations for UFO appearing
14:51 < dx> god dammit guys
14:51 < dx> you talk too much
14:52 < pbunny> btw, is it possible to surrond some mob with darkness (set all light to 0)? how will he look like?
14:52 < dx> and the worst part is that i was going to grab popcorn to continue reading, but then you went on to say 'java is shit'. nobody cares if java is shit.
14:52 < pbunny> dx: but there are many that need to be told the truth
14:52 < pbunny> i.e. dav1d 
14:53 < dx> i don't know what you're talking about but i've already saved my popcorn in the fridge for another day
14:53 < dav1d> pbunny: mh?
14:53 < dav1d> didn't follow the conversation
14:54 < dav1d> pbunny: what?
14:54 < dav1d> is it my fault that I have tests in java to finish university?
14:56 < dx> ha
14:57 < dx> yeah, dav1d joined in the middle of this, only mentioned the university thing, and then pbunny attacks him randomly for having mentioned java. lol.
14:57 < dx> okay, okay, i'll grab the popcorn again and continue reading this conversation from the beginning
14:57 < dx> but it better be good!
14:58 < pbunny> dav1d: this is your fault that you attend useless timewaste
14:58 < dav1d> haha
14:58 < dav1d> what a funny guy
14:58 < pbunny> its just a business of making money from parents that are 100% sure educations is needed now as it was 50 years ago
14:58 < dx> dav1d: i agree!
14:59 < dav1d> dx: thanks
15:00 < pbunny> education now serves 2 things
15:01 < pbunny> first, its a business. a bunch of 'professors' earn money by wasting time of students that aren't aware that they can just google all they need like 100 times faster
15:02 < pbunny> second, its a government structures for brainwashing the youth of society
15:02 < pbunny> i.e. they enforce m$ products and position linux to students like some exotic stuff
15:02 < dav1d> wrong
15:03 < dav1d> university pcs run linux, profs use all OS mixed
15:03 < pbunny> so students will use m$ products and can be controlled by the system via microsoft
15:03 < dx> meh
15:03 < pbunny> dav1d: there are exceptions
15:03 < pbunny> 90% of univercities position m$ as primary OS simply because m$ 'sponsors' them
15:03 < pbunny> also they teach C#, ASP.NET and other ridiculous bs
15:04 < dx> we already know you never had formal studies, pbunny, it was obvious from the beginning
15:04 < pbunny> i had until i left
15:04 <+sadimusi> by choice? :D
15:04 < pbunny> yeah
15:04 < pbunny> sadimusi: i noticed they try to brainwash me
15:04 < pbunny> it was too dangerous to stay
15:05 < dx> maybe brainwashing isn't always a bad thing after all
15:05 < pbunny> they only reason i went there at all was because of parents persuaded me
15:05 < dav1d> nope, no c#
15:05 < pbunny> as always happens
15:05 < dav1d> racket and java
15:05 < pbunny> dx: it is bad for mental freedom, obviously
15:05 < pbunny> and for individuality
15:05 <+sadimusi> how so?
15:06 < dav1d> bbl going home
15:06 < dx> some people make you consider that mental freedom leads to awful results
15:06 < pbunny> what people?
15:06 < dx> stupid people
15:06 < pbunny> only a very few people are totally free
15:06 -!- dav1d is now known as zz_dav1d
15:06 < pbunny> and they are at the top of all of this
15:07 <+Fador> they teach C++ at my university =)
15:07 < dx> haskell / C at mine
15:07 <+sadimusi> we can code in whatever language we choose, the examples are mostly in java
15:08 < pbunny> still i don't see how google is less efficient than sitting there for years
15:08 <+sadimusi> look at your code and you'll know
15:08 < pbunny> nothing wrong there :)
15:08 < dx> would you decide to google spontaneously "how to measure algorithmic complexity"?
15:08 < pbunny> it's just not what you are used to
15:08 < dx> because you don't seem to know what that is
15:09 < pbunny> dx: yes, if i will need to measure it
15:09 < dx> lol
15:09 < dx> that answer...
15:09 < dx> i'll take it as a "no"
15:09 < dx> since you'll never realize when you need it if you don't know it exists
15:10 < pbunny> because i'm fine without it
15:10 < pbunny> and it exists only to garbage the minds of students some more
15:10 < dx> lol
15:10 <+sadimusi> :D
15:10 < dx> computer science theory is garbage, got it
15:10 < pbunny> there are tons of principles, standards, etc that do nothing useful
15:10 < pbunny> abstractions...
15:10 < pbunny> dx: yeah, theory is garbage if it isn't specific
15:10 < dx> yep, that seems to match the way you code perfectly
15:10 < dx> i do believe you're being honest
15:10 < pbunny> well my code works and is efficient
15:11 < pbunny> i don't need to know freaking hundreds of 'theories' to code well
15:11 < Stormx2> Can someone please ban this turd
15:11 < Stormx2> I'm sick of this shit
15:11 < pbunny> lol
15:11 < Stormx2> Stop taking pride in being uneducated
15:11 < dx> Stormx2: sorry for continuing it :(
15:11 < dx> i just find him very entertaining
15:12 < pbunny> Stormx2: education is not important in programming
15:12 < pbunny> ability to learn and invent is important
15:12 < Stormx2> Yes it fucking is you tard. This is why your code sucks.
15:12 < Stormx2> Every time we've got a snippet of your code it has been total garbage
15:12 < Stormx2> And you can't even admit to it.
15:13 < pbunny> because it isn't what you are used to in univercity?
15:13 < dx> fun fact: education is important for many things, including realizing how your code is shit!
15:13 < pbunny> dx: education can be different
15:13 < Stormx2> Everyone else
15:13 < Stormx2> Opinions please
15:13 < pbunny> univercities perform the "useless shit brainwashing"-style education
15:13 < pbunny> to make education periods longer and grab more money
15:14 < Stormx2> pbunny, they also teach you important concepts like big-O, which I still haven't heard your rationale for not accepting
15:14 -!- SpaceManiac [~SpaceMani@r74-192-152-131.gtwncmta01.grtntx.tl.dh.suddenlink.net] has quit [Read error: Connection reset by peer]
15:14 < pbunny> Stormx2: can you give any example of when i will need this big-O ?
15:14 <+sadimusi> Stormx2: I think it's obvious that nobody agrees with him on anything
15:14 < pbunny> and anyway i can just google it in 5 minutes
15:14 < dx> what sadimusi said
15:14 < Stormx2> sadimusi, you can stop this
15:14 < Stormx2> Make #mcdevs-1989
15:14 < Stormx2> send pbunny there
15:14 < pbunny> lol.
15:14 < pbunny> don't be so mad
15:15 < Stormx2> anyone who wants to follow him and write servers with pthreads and mutexes
15:15 < Stormx2> and not abstract datatypes
15:15 < Stormx2> or any of that shit anyone sensible does
15:15 < pbunny> i have abstract datatypes
15:15 < pbunny> i.e. item
15:15 < dx> what i'm really worried about is, am i the only one who isn't 'affected negatively' by his presence?
15:15 < pbunny> dx: obviously univercities inject a program to hate enemies of education system
15:15 < dx> loling
15:15 < jast> what do you mean by 'affected negatively'
15:16 < Stormx2> This is a channel I enjoy talking about technical parts of MC in. And I really value the opinions of almost everyone here.
15:16 < dx> jast: he's amusing, that's what i mean
15:16 < jast> oh, definitely
15:16 < dx> cool.
15:16 < Stormx2> But I hate some 15yo chipping in with "lol java is slow"
15:16 < Stormx2> or some shit
15:16 < Stormx2> It ruins the channel.
15:16 < pbunny> lol.
15:16 < jast> let's face it, java *is* slow (lol)
15:16 < pbunny> ++
15:16 < dx> yeah, java is really slow, REALLY slow
15:16 < Stormx2> Java being slow is not why minecraft is slow though.
15:16 < dx> but the issue might be that i have 32mb of ram
15:17 < Stormx2> Not entirely, not even majority.
15:17 < dx> and a 2gb swap partition mounted over sshfs
15:17 < pbunny> Stormx2: yeah, that's only one of reasons
15:17 -!- SpaceManiac [~SpaceMani@r74-192-152-131.gtwncmta01.grtntx.tl.dh.suddenlink.net] has joined #mcdevs
15:17 -!- mode/#mcdevs [+v SpaceManiac] by ChanServ
15:17 < pbunny> the other reason is - java is a memory hog
15:17 < pbunny> that's why they had to offload world regions
15:17 < pbunny> that's why additional overhead was introduced just to make server work at all
15:17 < Stormx2> world regions are stored as byte or nibble arrays iirc.
15:18 < pbunny> yeah, to disk
15:18 < Stormx2> I run my server off a ramdisk
15:18 < Stormx2> So, no disk.
15:18 < jast> the worst thing about java is that many of the biggest projects that use it are written badly and solve non-problems, and everyone jumps on those and integrates them into anything they, in turn, write in java
15:18 < pbunny> in C i can easily keep 2500x2500x256 world in 4Gb of RAM
15:18 < pbunny> with java i doubt 64Gb would be enough
15:18 < Stormx2> what makes you say that?
15:18 < pbunny> i did calculations.
15:18 < Stormx2> java stores byte arrays just like C does
15:18 < pbunny> ah
15:18 < pbunny> how about array sizes?
15:19 < dx> Stormx2: this guy stores the whole world in memory all the time
15:19 < pbunny> hmm
15:19 < Stormx2> They're byte[], it's not a python list or anything
15:19 < pbunny> dx: yeah, and i can keep 2500x2500 world that way
15:19 < Stormx2> dx, so do I but I don't need a new server to do it :P
15:19 < dx> Stormx2: don't bother arguing his shit, he knows about this
15:19 < pbunny> Stormx2: well maybe problem lies in another reason
15:19 < pbunny> the brain java of mojang devs?
15:19 < pbunny> the wish to OOP anything as much as they can
15:19 <+sadimusi> it's starting to get annoying, is anybody still having fun and wants to keep him around?
15:20 < dx> i do, but don't mind if someone kicks him
15:20 < jast> what's annoying is that everyone thinks they can educate him
15:20 < Stormx2> I could go either way.
15:20 < pbunny> omg, you can just ask me to stop this topic
15:20 < Stormx2> Maybe see if he keeps it up?
15:20 < jast> only slightly, though. it's kind of funny, too, at the same time
15:21 <+sadimusi> I'll just disable notifications for this channel for a while then :)
15:22 < dx> actually i think the conversation is over
15:22 < jast> your face is over
15:22 < jast> !!
15:22 < dx> :(
15:22 < jast> oh wait, no, there it is. my mistake.
15:22 < dx> oh cool
15:23 < dx> i'm going to write a minecraft server in pypy, they say it's faster than C so it's even better than java right?
15:23 < Stormx2> I'm going to write a minecraft server on a lisp machine because I hear they're all the rage
15:24 < dx> i'll just use python generators instead of arrays, python generators are just code and don't use any memory at all
15:24 < dx> i'd use the term "heap memory" but i've never heard it before
15:24 < dx> i can still use it - i just have to google it
15:24 < pbunny> are you serious?
15:25 < pbunny> you shouldn't use python at the first place
15:25 < dx> are YOU serious?
15:25 < pbunny> of course
15:25 < dx> oh god dammit sorry i'm continuing it
15:26 < jast> from now on I'll write all my code in Ada 83
15:28 < pbunny> Stormx2: what is your version about why mojang decided to not keep world in ram?
15:29 < pbunny> why did they go to all these complications and limitations (no offline mobs etc)?
15:29 < dx> dude
15:29 < dx> the answer to that question is trivial
15:29 < Alan> ugh
15:29 < Alan> this channel
15:29 < Alan> is full of people who haven't got a fucking clue, and people raging/sighing at people who don't have a clue
15:29 < Alan> who suggested I should come here?
15:29 < Alan> was it eddyb ?
15:29 < dx> Alan: uh, sorry
15:30 < dx> Alan: my pypy comment was sarcastic btw
15:30 < Alan> dx: I could tell you were being sarcastic
15:30 < dx> oh ok
15:30 < Alan> just, in general
15:30 < dx> Alan: then i don't know what other people you're talking about
15:30 < dx> Alan: it's just pbunny being derp...
15:30 < eddyb> Alan: yeah, well, I'm sorry if it turned out worse :(
15:30 < Alan> mainly pdelvo 
15:30 < Alan> erm
15:30 < Alan> pbunny: 
15:30 < dx> Alan: it's not usually like this
15:31 < Alan> yeah, it seems to usually be dead instead
15:31 < Alan> :P
15:31 < dx> ..well, yeah, lately yes
15:31 < pbunny> dx: trivial? can you sound it?
15:31 < dx> it's mostly about mc protocol and nobody interested happened with the protocol in a long time
15:31 < dx> pbunny: infinite world
15:32 < pbunny> dx: are you serious about somebody can ever need infinite world?
15:32 -!- Stormx2 [~Stormx2@cpc18-sotn9-2-0-cust33.15-1.cable.virginmedia.com] has quit [Ping timeout: 248 seconds]
15:32 < pbunny> isn't 2500x2500 already big enough?
15:32 < pbunny> or 15000x15000 (256Gb RAM) ?
15:32 < Alan> infinite world = very easy for players to accidentally make it use all the RAM
15:32 < dx> lol
15:32 < pbunny> Alan ++
15:32 < pbunny> dx: world is still not infinite
15:32 <+Fador> some run servers on raspberry pi with limited RAM ;D
15:32 < Alan> without regard for what the hardware specs are
15:32 < Alan> loading only what's in use lets memory usage scale with player count, not world size
15:33 < Alan> and lets you still have your effectively unlimited worlds
15:33 < Alan> how is that so hard to grasp?
15:33 < pbunny> why not limit world to 15000x15000 and loop it as planet?
15:33 < pbunny> it's a HUGE world
15:33 < dx> pbunny: my feed the beast server is hosted on a 512mb openvz - it runs flawlessly and the generated chunks extend up to 9000
15:33 < Alan> because that's not Minecraft
15:33 < dx> pbunny: i pay $30 per year for this server
15:33 < pbunny> dx: yes, but it doesn't iterate over everything
15:33 < dx> pbunny: to have a server with 32gb ram, you have to pay $100 per month usually
15:33 < dx> pbunny: there are no servers with 256gb ram in this planet
15:33 < pbunny> it's not a real independent world
15:34 < pbunny> dx: LOL.
15:34 < pbunny> there is
15:34 < dx> ..
15:34 < pbunny> dx: http://www.newegg.com/Product/Product.aspx?Item=N82E16813131643
15:34 < Alan> dx: try telling my workplace that :P
15:34 < pbunny> you can insert 256Gb of RAM there
15:34 < dx> fug
15:34 < dx> oh dammit
15:34 < Alan> we're about to get 2 new servers: a 32-core 256GB RAM, and a 64-core 512GB RAM
15:34 < pbunny> i will get one in autumn maybe
15:34 < dx> actually there was a 48tb supercomputer
15:35 < Alan> but big servers are expensive
15:35 < dx> god dammit, i'm the stupid one now ;_;
15:35 < dx> okay
15:35 < dx> how about
15:35 < Alan> and your "average" group of minecrafters does not have the money needed to throw at servers that big
15:36 < Alan> and yet again, putting a really high lowerbound on amount of memory needed by making the usage scale with world size is horrible
15:36 < dx> pbunny: if you can tell me about a dedicated non-collocated service that provides 256gb ram, i'd love to know its price
15:36 < pbunny> dx: why not just buy one?
15:36 < pbunny> it's cheaper and better privacy
15:36 < dx> ..
15:37 < dx> okay i'm bored of this
15:37 < dx> bye
15:37 < pbunny> dx: imagine a hosting provider will read the emails and passwords of users of mc, then hacks their emails (as 99% of people use the same password everywhere)
15:37 < pbunny> and they all will blame you for that
15:37 < pbunny> how can you just trust your data to somebody you never saw
15:37 < dx> please don't ping me anymore, thank you.
15:38 < pbunny> also he can steal your code and run his own server, etc
15:38 < pbunny> ok
15:38 < jast> that's very simple: you make an agreement with the hoster to not do that. if they do it anyway, you sue them.
15:38 < Alan> "steal your code"
15:38 < pbunny> jast: they can hide all traces
15:38 < pbunny> easily
15:38 < jast> it's also illegal in the first place
15:38 < Alan> yeah, now you're going into the realms of "breach of contract", "illegal" etc
15:38 < pbunny> jast: you can't monitor your hosting provider
15:38 < Alan> running your server in a datacentre already puts you at risk
15:38 < jast> well
15:38 < pbunny> he can do whatever he wants, you can only guess
15:38 < Alan> even colocated
15:39 < pbunny> Alan: sure
15:39 < Alan> physical access = you lose
15:39 < pbunny> that's why i keep my server at home
15:39 < Alan> enjoy dat latency
15:39 < jast> it's fairly difficult to read stuff on a server without powering it off if you don't have the password
15:39 < pbunny> i have 1Gbps connection there
15:39 < jast> you can't do that with a screwdriver and a hammer
15:39 < Alan> you have a very broken/skewed view of what an average minecraft server looks like or should look like
15:39 < pbunny> jast: um, you can even read from RAM on running server with special devices
15:39 < Alan> jast: who says you can't power it off?
15:39 < jast> yes, with special devices
15:40 < Alan> "oops, power interruption"
15:40 < pbunny> jast: hosting provider can have these
15:40 < jast> any hoster who has these special devices will be viewed with, shall we say, a small amount of suspicion
15:40 < jast> Alan: o hi encrypted hard disk
15:40 < Alan> jast: and how do you suggest ever being able to reboot it?
15:40 < jast> (yeah yeah, not perfectly secure, I know)
15:41 < jast> though I suppose you could put the bootloader in a hardware dongle
15:41 < Alan> any method that you can use to unlock the encryption can be intercepted
15:41 < pbunny> you can encrypt all important partitions, but hosting provider can still sniff your password when you attempt to decrypt it (i.e. via ssh)
15:41 < jast> that should make it pretty safe
15:41 < Alan> ... the hardware dongle that's in the datacentre?
15:41 < jast> yes
15:41 < Alan> right, so how does any of this protect you against the datacentre staff?
15:41 < jast> e.g. hardware dongle has a private key baked into it
15:41 < pbunny> my root fs is unencrypted, so in case of reboot i can ssh into it and mount all encrypted partitions
15:42 < Alan> jast: in which case they've got the private key
15:42 < pbunny> but hosting provider would just intercept ssh and get your password
15:42 < Alan> it's almost like people don't grasp information theory.
15:42 < jast> you underestimate the difficulty involved in reverse-engineering hardware
15:42 < jast> especially if you don't want to damage said hardware
15:42 < Alan> why would you need to reverse-engineer it?
15:43 < jast> because the hardware dongle isn't going to tell you the secret key if you just ask nicely
15:43 < pbunny> jast: hosting provider can just use hotpluggable RAID in your server
15:43 < pbunny> then it can unplug 1 hdd and read it
15:43 < pbunny> then plug it back
15:43 < pbunny> etc
15:43 < pbunny> limited opportunities
15:43 < jast> it's still encrypted o/~
15:43 < Alan> jast: ... the bit of code that reads the private key from the dongle can't be encrypted with the key that's in the dongle
15:43 < pbunny> jast: encryption password is intercepted
15:43 < Alan> otherwise it couldn't run
15:43 < jast> the private key is _never_ read from the dongle
15:43 < jast> the dongle uses it ONLY to authenticate you
15:43 < Alan> or whoever has the machine
15:44 < pbunny> jast: can you explain in more detail how will you mount encrypted partitions/disks ?
15:44 < jast> it also contains another secret key used for disk decryption, sent out to the OS after a system integrity check (and after you have authenticated yourself)
15:44 < Alan> jast: the authentication can still be intercepted
15:44 < jast> yes, but since the secret key is unknown, that doesn't matter
15:44 < jast> asymmetric cryptography, you know
15:44 -!- zz_dav1d is now known as dav1d
15:44 < jast> challenge-response
15:45 < jast> ever heard of any of those?
15:45 < pbunny> jast: hosting provider can do mitm as he still has all the keys
15:45 < jast> where from?
15:45 < pbunny> unless you drove to him physically and configured server by yourself
15:45 < jast> yes, that's called colocation
15:46 < jast> obviously you can't do any of this if you use hardware provided by the provider
15:46 -!- levifig [~levifig@spwn.co] has quit [Excess Flood]
15:47 -!- levifig [~levifig@spwn.co] has joined #mcdevs
15:47 < jast> and now, if you've set up a secure server box (with tampering sensors), things should be pretty good
15:48 < jast> I still wouldn't give critical data to anyone else, but we're talking about minecraft servers here, for which pretty much anything is overkill
15:48 < Alan> so, now that we've thoroughly explored fantasy-land...
15:48 < Alan> you don't co-located critically sensitive information
15:48 < Alan> minecraft servers are not critically sensitive information
15:48 < Alan> oh
15:48 < Alan> haha
15:48 < Alan> 2 people saying the same thing, must be true
15:48 < jast> yeah, that's how it works
15:49 < Alan> I hope not
15:49 < Alan> otherwise we're all screwed
15:49 < jast> yes we are
15:49 < jast> boom
15:49 < Alan> anyway, pretty sure my life isn't going to be enriched by staying here
15:49 < Alan> good luck with whatever things you attempt to do
15:49 < jast> took you long enough to figure that out :}
15:50 < Alan> no, i just stuck around for the rage
15:50 -!- Alan [alan@unaffiliated/alan] has left #mcdevs []
15:53 < jast> dx: EC2 offers an instance type with 244 GB RAM for "just" $3.50/h :}
15:54 < dx> jast: nice
15:54 < dav1d> 244GB ram, not bad
15:54 < jast> also 88 bogocores
15:54 < dx> jast: i actually needed to do a heavy video encoding job once, grabbed one server with 48gb ram for 40 cents an hour or so
15:55 < jast> yeah, I did something similar once
15:55 < dx> never knew what the provider was and i paid the guy in bitcoins lol
15:55 < dx> kinda shady but it was little money for what i had to do
15:55 < dx> and it worked!
15:56 <+Fador> dx: and now they have a copy of your videos? =D
15:56 < dx> ...well
15:56 < dx> it was just anime. lol.
15:57 <+Fador> ^_^
15:57 <+Fador> -!- Fador [fador@hentai.fi]
15:57 < dx> lewd
15:57 < dx> a friend had to send his only PC to repair something, and it took a really long while, like two months
15:57 < dx> he had a PS3 but it doesn't support the usual codecs
15:58 <+Fador> like..HEVC ;)
15:58 < dx> so i offered to encode all the episodes he missed from the season, thought it would be fun
15:58 < dx> lol hevc
15:58 < dx> well, actually, no softsubbed mkv and no 10bit. really common situation.
15:59 < dx> nobody supports 10bit ever
15:59 -!- BizarreCake [~BizarreCa@46.121.251.157] has joined #mcdevs
15:59 <+Fador> http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6324420 <- I was writing a paper about HEVC
15:59 < dx> whoa.
16:00 < dx> ...but no thanks, i'll pass
16:00 < dx> i barely know how to get decent results with the x264 command line tool :P
16:00 <+Fador> http://www.cs.tut.fi/~moncef/publications/complexity-analysis-ISCAS-2012.pdf my other paper about it that you can actually read without paying ;D
16:00 <+Fador> hehe, sure, HEVC was standardized just few months ago
16:01 < dx> and fansubbers are already considering using it as soon as possible
16:02 <+Fador> well it does reduce the size 50% with the same quality..=D
16:03 < dx> at the cost of a lot more people not being able to watch it!
16:03 <+Fador> sure!
16:03 <+Fador> they can watch it in the *future* ;D
16:03 < dx> actually they did that already with hi10p, and the hardware decoders haven't caught up
16:04 < dx> (they probably don't care about it)
16:05 <+Fador> the switch to 10bit was quite fast after x264 started supporting it
16:07 -!- clonejo|offline is now known as clonejo
16:08 < dx> in hindsight, the only decent ways to handle it would be either adopting it ASAP like they did, or not adopting it at all
16:09 -!- Stormx2 [~Stormx2@2001:630:d0:f110:4cfb:5a63:110a:4de1] has joined #mcdevs
16:09 < dx> as in, there was no reason to delay it if stuff that isn't ffmpeg isn't planning to implement it too
16:09 < dx> (ffmpeg as a generalization because i'm too lazy to check alternate software only implementations)
16:10 < dx> (we all know ffmpeg has a monopoly on the codec market anyway)
16:12 < Stormx2> dx, libav? :P
16:12 < jast> libav is pretty much the same thing
16:12 < dx> Stormx2: same shit
16:12 < jast> except it appears to be a pointless fork