00:00 < laxa> np :) 00:13 <@deen> DDNet client/server 6.3 released 00:13 <@deen> the menus at the left side on http://ddnet.tw are getting bigger and bigger 00:14 <@deen> Should I just make the menu move with the rest of the site? 00:14 <@deen> or keep increasing the necessary screen height 00:22 < eeeee> i tried coming up with a better layout for the website but it seems i have no design powers whatsoever :( 00:23 <@deen> me neither, but thanks for trying! 00:23 < eeeee> probably have to read a book on that or something 00:24 < eeeee> or i just get a designer friend drunk an make him come up with some ideas 00:37 < xrothx> deen: 00:38 <@deen> hi xrothx 00:38 < xrothx> I get a player telling that ddnet is infected 00:38 < xrothx> he uploaded the mod to virustotal 00:38 < xrothx> & he is watching ddnet for over 4months now? 00:38 < xrothx> Who is the compiler of the mod before its published? 00:38 <@deen> I am 00:39 < xrothx> aight, why would it be scanned as virus? 00:39 <@deen> many programs do that when compiled with gcc for windows 00:39 <@deen> virus scanners are shit 00:39 < xrothx> ok :P 00:39 <@deen> it's usually some generic detection 00:39 <@deen> "something looks suspicious" 00:40 <@deen> and of course ddnet client is suspicious, it has an autoupdater that can load any code from the internet 00:40 * eeeee pretty sure that's the autoupdater part 00:40 <@deen> I think Tsin was flaming about DDNet client being malware before we had the autoupdater 00:40 <@deen> I explained to some guy how I compile it with gcc 00:40 <@deen> He did the same and got the same virus results 00:44 < eeeee> https://www.virustotal.com/en/file/d96ccfd7deec47dee2bb68d3bcb098dbb0d21ad57039ea8833c3bf4ae88d5673/analysis/1416959048/ 00:46 < xrothx> !twp xRoThx 00:46 < Nimda> xRoThx is currently playing 1TeeMap on server : DDNet GER2 - Oldschool [DDraceNetwork]. 00:46 <@deen> eeeee: yeah, always looks something like that 00:47 < eeeee> dafuq is ByteHero 00:47 <@deen> haha 00:47 < xrothx> deen: 00:47 < xrothx> ger2 tested & secure :) 00:48 <@deen> xrothx: i see ddos still incoming 00:48 <@deen> 100 mbit/s make it through 00:48 <@deen> xrothx: how strong was the attack? 00:48 <@deen> Everyone seems to know ByteHero just from virustotal^^ 00:48 <@deen> and it seems to have many false positives 00:49 <@deen> chinese anti virus software I'd say 00:50 < laxa> !fr-en reconnaissance 00:50 < Nimda> Translation: recognition 00:51 < eeeee> that reminds me of old windows xp days when it was more like a real fight against viruses 00:52 <@deen> Or win95/98, when your system would be infected by connecting it to the internet for a few minutes^^ 00:53 < laxa> there was a worm on win XP 00:53 < laxa> I came back from holiday 00:53 < eeeee> i once had to fix friend's computer after a virus got onto it. friend had "mcafee" antivirus which somehow couldn't prevent infection and in turn got infected and somehow deleted itself leaving the computer without any antivirus whatsoever 00:53 < laxa> it was restarting computer every 1 minute 00:53 <@deen> I sent ByteHero a report about the false positive, lets see what happens 00:53 < laxa> and you were gotting it just by internet connection iirc 00:54 < eeeee> had to fight the virus with bare hands after that. it was restarting way too fast after being killed from task manager so you couldn't delete the binary, but i wrote a batch file to kill and delete it in loop and it lost eventually. 00:54 <@deen> eeeee: that's what the virus wanted you to believe!!! 00:54 < eeeee> oh shi- 00:55 <@deen> haha 00:55 < laxa> !fr-de beaux-frère 00:55 < Nimda> Translation: Feine Bruder 01:08 < xrothx> deen: Got 3 people today that I know for a couple years telling me that ddnet is not good 01:08 <@deen> xrothx: why? 01:08 < xrothx> master pw in it, ? 01:08 <@deen> what? 01:08 < xrothx> in the mod 01:08 <@deen> i don't understand the question 01:09 < xrothx> Can you check the latest update 01:09 < xrothx> pretty weird all those people 01:09 <@deen> hm? 01:09 <@deen> There is no master pw 01:09 <@deen> I can't control other people's servers and clients, if that's what you say 01:18 <@deen> eeeee: I like simple webdesign, but people said the ddnet.tw website looked too simple 01:18 <@deen> so I made it into the monster it is now 01:21 <@deen> Like my personal website: http://felsin9.de/nnis/#teaching-research (Just uploaded my paper!) 01:23 <@deen> xrothx: no report on the ddos size? 01:24 <@deen> xrothx: and i still didn't understand why ddnet is bad according to your old guys 01:24 <@deen> the usualy argument is that we kill off teeworlds because we're too big 01:43 < Maple> can somebody tell me what Naiß means 01:43 <@deen> "nice" 01:43 < Maple> ok phew 01:43 < Maple> i was thinking it meant noob 01:44 <@deen> when you pronounce it in German it sounds exactly like "nice" 01:44 < Maple> xd 01:44 < Maple> !twp Fudgyking 01:44 < Nimda> No Fudgyking found online. 01:44 < Maple> !twp Broken 01:44 < Nimda> Broken is currently playing TakeUrLuggage on server : DDNet USA - Moderate [DDraceNetwork]. 01:44 < Maple> !twp mikey12\ 01:44 < Nimda> No mikey12\ found online. 01:44 < Maple> !twp mikey12 01:44 < Nimda> No mikey12 found online. 01:46 < Maple> deen the name faking prevention still doesnt work 01:46 <@deen> Maple: really? 01:46 < Maple> yea im in server with broken 01:46 < Maple> and i have his name 01:46 <@deen> let me check 01:47 <@deen> That server wasn't restarted yet 01:56 <@deen> Maple: any better idea than just regarding whitespaces as nonexistant? 01:57 < Maple> well 01:57 < Maple> in vanilla servers 01:57 < Maple> my faking doesnt work 01:57 <@deen> in vanilla server only ascii characters work 01:58 < Maple> http://en.wikipedia.org/wiki/Whitespace_character 01:58 <@deen> the only negative effect i can imagine is if someone is called "B r o k e n", and Broken enters, he will get named "(1)Broken" 01:58 < Maple> i use U+2060 : word joiner 01:58 <@deen> yeah, i know 01:59 <@deen> it's recognized, just not in the middle of a name^^ 01:59 <@deen> only at start and end 01:59 < Maple> well that one is not even visible 01:59 < Maple> so the person using it is faking for sure 01:59 < Maple> for the visual whitespace, you can make exceptions 02:00 <@deen> yeah, but i don't want to blacklist every single whitespace that may be too small to be visible 02:00 < laxa> deen: Wtf, only ddrace use utf-8 chars for nickname ? 02:00 <@deen> laxa: it's a change in 0.6.2 for vanilla i think 02:00 < laxa> ok 02:00 <@deen> most mods have kept utf8 names (probably because they're at 0.6.1) 02:00 < laxa> cause otherwise, I would have said, completely remove utf8 for nickname and just use ascii 02:00 <@deen> yeah, great 02:01 < laxa> it's complete bullshit to me to be honest 02:01 <@deen> let's force russians to use english names! 02:01 < laxa> hum 02:01 <@deen> everyone has to have an english name! 02:01 < laxa> just normal chars I mean 02:01 <@deen> fuck other cultures! 02:01 <@deen> yes, "normal chars" are english :P 02:01 < laxa> just alpha-chars 02:01 <@deen> that's the english alphabet 02:01 < laxa> String.Alpha.NoCulture() 02:01 < laxa> better ? :D 02:01 < Maple> *grabs popcorn* 02:02 < laxa> Maple: fuck off :D 02:02 < laxa> Just normal alpha chars in every culture should be allowed 02:02 < Maple> tbh, faking should be considered as admiration 02:02 < laxa> that's what I mean 02:02 < laxa> Maple: I am not talking about faking at all anyway 02:02 < laxa> just about nickname 02:02 < Maple> i think its deen's issue 02:03 < laxa> faking or nickname ? 02:03 < Maple> faking 02:03 < laxa> nickame isn't deen's problem, he has to follow the rules of teeworlds 02:03 < laxa> teeworlds is like the first game I ever seen accepting every char from utf8 02:03 < laxa> ok, gotta go sleep, anyway, nice talking, bye bye :) 02:04 <@deen> bye 02:06 < ddnet-commits> [ddnet] def- pushed 1 new commit to DDRace64: http://git.io/M47V3Q 02:06 < ddnet-commits> ddnet/DDRace64 771eb1a def: Disregard whitespaces in confusable name comparison 02:06 <@deen> There you go, Maple 02:06 <@deen> follow me again 02:06 < Maple> Naiß 02:06 < Maple> xD 02:29 < Maple> !twp Fudgyking 02:29 < Nimda> No Fudgyking found online. 02:52 < Maple> !twp Fudgyking 02:52 < Nimda> No Fudgyking found online. 03:26 < eeeee> nice webpage. i reloaded a couple times because i thought css didn't load :) 03:31 <@deen> haha 03:31 <@deen> it even has a css: http://felsin9.de/nnis/css.css 03:32 <@deen> Especially nice: background-color: #eeeeee; 03:32 * eeeee approves 03:46 < eeeee> so if you wanted to simplify the website i guess you could use some tools like google analytics to understand how do people use your site and then optimize for that 03:46 < Nimda> Xabier by Waschlappen & Broken just released on Moderate at 2014-11-26 03:43 03:47 <@deen> good night 03:47 < eeeee> nite 14:33 <@EastByte> good afternoon 14:38 <@EastByte> deen: I want to test stuff 14:45 < laxadedi> LL 17:33 <@deen> EastByte: what stuff? 19:22 < BeaR> sup? 19:34 <@deen> hi BeaR 19:35 <@deen> There's elevator music at the start of Xabier 19:35 <@deen> released tonight 19:35 < BeaR> ah cool 19:35 <@deen> apparently still no one can compress opus files properly^^ 19:35 < BeaR> hm lol 19:35 <@deen> I've had to reencode for all maps with sound 19:36 < BeaR> there are more ?: 19:36 <@deen> Little Forest by Kintaro 19:36 <@deen> The halloween moderate map by Ama 19:36 <@deen> and now Xabier 19:36 < BeaR> k 19:36 <@deen> I've made a category for maps with sound in the votes^^ 19:37 < BeaR> you changed sth at the soundsystem? (I didnt follow the commit history in the last time (: ) 19:37 <@deen> I reduced the max drift, but I think you saw that 19:37 <@deen> from 2 seconds to 0.2 or something 19:37 < BeaR> ah k 19:37 <@deen> because phacrum & 645654 are working on Justice 2 and they always use every single feature 19:38 <@deen> at the start of the map they have a huge clock and it's ticking 19:38 <@deen> analog clock that shows the actual game time^^ 19:38 < BeaR> lol 19:38 <@deen> But they've been working on the map for a year, so don't expect a release soon 19:38 < BeaR> sounds fun with the envelope editor only :> 19:39 <@deen> I did some bugfixes yesterday, but couldn't figure out the sound bugs we still have 19:39 < BeaR> there are more expect that cutoff? 19:39 <@deen> today i'm busy with grading lambda calculus homework 19:40 <@deen> only what's here: https://github.com/def-/ddnet/issues 19:40 <@deen> oh yeah, so it's just that 1 + more complicated sounds as a feature^^ 19:40 < BeaR> have to write a paper so currently im lacking time :< 19:40 <@deen> about what? 19:40 < BeaR> API comparisons 19:40 < BeaR> not interesting tbh 19:41 < BeaR> that complicated sound stuff would high probably break compatibility with older clients /: 19:41 <@deen> ah, i fixed HD flag for sound 19:41 <@deen> with older clients and with older maps? 19:42 < BeaR> can't remember if I added some versioning for all the sound stuff 19:42 < BeaR> older clients 19:42 <@deen> if no sound plays on older client that's fine 19:42 <@deen> people who stay at old client can't expect that new features work 19:42 <@deen> but map should still be playable 19:42 < BeaR> nah, will it will probably fuck up the serialization /: 19:42 <@deen> ok, shit 19:42 < BeaR> - will 19:43 <@deen> that's too bad 19:43 <@deen> round sound is often not what mappers want apparently 19:43 < BeaR> hm, I fixed that cutoff thingie already, dunno where the source is 19:43 < BeaR> gonna tackle it tomorrow 19:44 <@deen> 2 of 3 mappers who used sound have asked for a way to do "same sound level in an entire room, nowhere else" 19:44 < BeaR> hm? 19:44 < BeaR> no falloff? 19:44 <@deen> exactly 19:44 <@deen> sound should just play in room 19:44 <@deen> or maybe rectangular falloff 19:45 <@deen> I wrote a small twping tool and am now recording my ping to the 3 european ddnet servers: http://ddnet.tw/ping/ 19:45 < BeaR> (: 19:46 < Axomar> Deen, r u there. Have u a minute for me pls 19:46 <@deen> hi Axomar 19:46 < Axomar> hi :D 19:47 < Axomar> give it other chatroom to talk alone? :D 19:47 < Axomar> s 19:47 < BeaR> hm there is an easy workaround for that compatibility break maybe (: 19:47 <@deen> BeaR: nice =) 19:48 < BeaR> deprecating the old LAYERTYPE_SOUNDS and add sth like LAYERTYPE_SOUNDS_V2 would probably work 19:48 <@deen> yes 19:49 <@deen> how do you plan to extend sounds? 19:53 < BeaR> basically extend the CSoundSource, add some m_DoFalloff bool and index for a list of shapes 19:54 < BeaR> shape would probably be some sort of tagged union with Sphere/Rectangle/(?) 19:55 <@deen> maybe make the falloff a float between 0 and 1 19:55 <@deen> or an int from 0..255, or sth like that 19:55 <@deen> so you can set earlier dropoff 19:55 <@deen> 255 would mean dropoff starts in the middle 19:55 <@deen> 0 means no dropoff 19:55 < BeaR> true, probably more flexiable (: 19:56 < BeaR> flexible* 19:56 <@deen> Sphere and Rectangle are also the only shapes I can think of that would make sense 19:56 <@deen> but it's still a problem for rooms in weird shapes 19:56 <@deen> what if two rectangles overlap? 19:56 <@deen> ah, that's what you mean by "tagged union"? 19:56 < BeaR> huh? 19:57 <@deen> ok, not^^ 19:57 <@deen> so, what would happen if they overlap? 19:57 <@deen> Sound plays louder, right? 19:57 < BeaR> y 19:57 < BeaR> tagged union is some kind of algebraic data type 19:58 < BeaR> sth like data shape = Rectangle | Sphere 19:58 <@deen> I'd prefer to imagine all shapes as a union, so that you can create a more complex shape by combining multiple 19:58 < BeaR> hm would be pretty complex :D 19:58 <@deen> BeaR: syntax error, should be Shape ;) 19:58 < BeaR> :p 19:59 < BeaR> I actually learn it tomorrow :> 19:59 <@deen> I taught it on Monday 19:59 <@deen> But we've finished Haskell now anyway 20:00 < BeaR> ah lol 20:00 <@deen> now lambda calculus, type inference, prolog, scala, x11, mpi, compilers :P 20:00 <@deen> it's a weird course 20:00 < BeaR> x11 ? 20:00 < Maple> xorg 20:00 < BeaR> but why 20:01 <@deen> the terrible x11 language 20:01 <@deen> oh wait, it's x10^^ 20:01 <@deen> http://x10-lang.org/ 20:01 < Maple> i started seeing assembly today 20:01 <@deen> I don't like it anyway, always joke about how long compiling programs takes 20:01 < BeaR> heh also want to learn Scale some day 20:02 <@deen> Maple: hopefully not the java bytecode we do^^ 20:02 < BeaR> all hail SIMD vector math \o/ 20:02 < Maple> we see the basics using WinSimul 20:02 < Maple> more lke the logic 20:03 <@deen> BeaR: I should be writing a paper for a conference using that, but can't access the machine i need for that^^ 20:03 < BeaR> :P 20:04 <@deen> Maple: cool, now you can optimize DDNet server to work well with 512 players 20:04 < Maple> xD! 20:04 < BeaR> concerning the complex shapes, I believe everything beside polygons is pretty much out of scope tbh 20:05 < BeaR> MAX_CLIENTS = 512 :> 20:06 <@deen> BeaR: then let's just do circles and rectangles, but have to make sure they work with the grid 20:06 <@deen> then it should be fine to build bigger shapes with 0 falloff (i hope) 20:06 < BeaR> y 20:25 < Learath2> deen: what happens when i try to autoupdate with a client 3 versions older or smth ? 20:33 <@deen> Learath2: it should work! 20:33 <@deen> Learath2: it should make a list of what has changed since that version 20:33 <@deen> and download and update all these files 20:34 < Learath2> just verifying couldnt really read what the old one was doing 20:34 <@deen> I didn't write it! 20:34 <@deen> I think I tried around some time ago and it seemed to work and I even understood why 20:35 < Learath2> but you probably know how it behaves and its easier to ask you :) 20:48 <@deen> !ddnetpeak 20:48 < Nimda> Current players on DDNet : 472 20:48 < Nimda> Current DDNet peak : 612 users online at 2014-10-26 20:15:01 20:48 <@deen> looks good 20:49 <@deen> there was a bug today on ddnet: when you changed server type from solo to non-solo it would stay solo^^ 20:54 < o_be_one> hi:) 20:54 <@deen> hi o_be_one 20:54 < o_be_one> how are you deen ? 20:56 <@deen> half-dead 20:56 < eeeee> be optimistic! half-alive 20:57 <@deen> shouldn't have bicycled 100 km right after donating blood^^ 20:57 < eeeee> lol you're hardcore 20:59 <@deen> o_be_one: what are you up to? I've seen middle americans on CAN^^ 21:01 < o_be_one> lol eeeee :D 21:01 < o_be_one> ahah deen yes CAN is getting more regular players, good to see :) and funny to see some US players ... 21:01 <@deen> eeeee: i don't know how to autogenerate what characters are rendered as empty 21:02 < o_be_one> Broken, Pepe and other come often too lol :o 21:02 < o_be_one> Hard Candy x) 21:02 < o_be_one> finally i think you can state that it was a good idea ^^ 21:03 < o_be_one> even if there is only like 10 players, its a new way to get a better ping for many of them 21:03 < o_be_one> ive just 23ms, Mapple 20, seen some with 15ms ....... 21:03 <@deen> well, i don't mind. but some people complain that we have too many servers 21:03 <@deen> citing CAN, CHN and Brazil as main examples 21:03 < o_be_one> yes thats a problem cause so there is players on all servers 21:03 < o_be_one> but not all on only one 21:04 < Learath2> you do have alot of servers but not sure how that is a problem 21:04 < eeeee> deen: i thought you were going to render each character to bitmap and compare those to find the ones which look alike 21:04 < o_be_one> but i take advantage of that, cause all servers are very customisable by voting 21:04 < eeeee> can't you just also note which ones look like space? 21:04 < o_be_one> we can set a server on the mod and on the map we want in only 2 clics ... 21:04 < o_be_one> all servers can be like "private" easily, its not a bad thing 21:05 <@deen> eeeee: I've just used the unicode confusables, seemed easier 21:05 <@heinrich5991> deen: you could specify that you may only use characters from a single language 21:05 <@heinrich5991> I believe unicode has some properties for that 21:05 <@deen> heinrich5991: yeah, I've read about that in the confusables document 21:05 <@deen> heinrich5991: but that would break many existing names 21:05 <@deen> which I would like to avoid 21:06 <@deen> (even if I hate that kind of name) 21:06 < eeeee> oh okay then. the existing function seems to be okay as well. 21:06 <@heinrich5991> deen: you could whitelist all the characters that are in use today 21:06 <@deen> eeeee: actually I had to add a space character there 21:06 <@deen> recently 21:06 <@deen> heinrich5991: so much work^^ 21:06 <@deen> heinrich5991: i think the current solution is ok 21:06 <@heinrich5991> deen: do you keep logs? 21:06 <@deen> no 21:06 <@deen> they would be huge 21:06 < eeeee> D: 21:06 <@deen> don't even have the space to store them 21:07 <@heinrich5991> but wait 21:07 <@heinrich5991> you keep records 21:07 <@heinrich5991> enumerate the chars from there 21:07 <@deen> then that's unfair 21:07 <@deen> some characters are allowed, some not. not very transparent 21:08 <@deen> Logs are deleted whenever server is restarted 21:08 <@deen> still GER has 15 GBs of logs right now, even though I restarted a few times in last days 21:08 < Learath2> sounds like you need logrotate and some compression 21:08 <@deen> oh wait, most of that is a single log file 21:08 <@deen> looks like a potential attack^^ 21:09 <@deen> indeed! 21:09 <@deen> 14 GB of [14-11-26 03:43:03][]: connection less packet too small, 4 21:09 < eeeee> oh man, such a classic 21:09 <@deen> attack was by "Hard *Candy*₡" it seems 21:09 <@deen> tonight at 4 am 21:09 < Learath2> BAN 21:10 <@deen> I shouldn't log that 21:10 <@deen> too easy to create 14 GB of log 21:10 < eeeee> yeah comment out that dbg_msg first 21:10 < Learath2> then BAN 21:10 < eeeee> i thought i had that in ddrace64 already 21:10 < ddnet-commits> [ddnet] def- pushed 1 new commit to DDRace64: http://git.io/c3KuwA 21:10 < ddnet-commits> ddnet/DDRace64 9d3cb79 def: Remove log that happens too often 21:10 <@deen> but now that i commented it out, people will discover this^^ 21:10 <@deen> apparently attacks read ddnet git commits 21:11 <@deen> attackers* 21:11 <@heinrich5991> deen: that's no problem after commenting it out 21:12 < Learath2> not deployed yet tho 21:13 <@deen> exactly 21:13 <@deen> and i can't restart already full servers 21:14 <@heinrich5991> what about not pushing until deployed? 21:14 < eeeee> well hopefully impact is not very big as we have async logging 21:14 <@deen> yeah, that would've been smart 21:14 <@deen> eeeee: except if they get more traffic than disk can write 21:14 < eeeee> it was more of an issue in old times when logging was sync 21:14 <@deen> my async logging code is pretty stupid 21:15 <@deen> it works well in normal scenarios, not sure if it's attack proof 21:15 <@deen> but currently I deploy by pulling from github^^ 21:17 < ddnet-commits> [ddnet] def- pushed 1 new commit to DDRace64: http://git.io/HJwaZQ 21:17 < ddnet-commits> ddnet/DDRace64 8fdd6da def: Keep retrying for FifoFile even if it's temporarily unavailable 21:18 < ddnet-commits> [ddnet] def- pushed 1 new commit to DDRace64: http://git.io/C1G76A 21:18 < ddnet-commits> ddnet/DDRace64 d2573e7 def: comment 21:25 <@deen> Now I have nice visual proof how terrible ping on GER is for me: http://ddnet.tw/ping/ 21:25 <@deen> ^^ 21:34 < ddnet-commits> [ddnet] def- pushed 1 new commit to DDRace64: http://git.io/RA2Rpw 21:34 < ddnet-commits> ddnet/DDRace64 e4c4ba2 def: Remove another bad log 22:42 < xrothx> deen: 22:42 < xrothx> deen: did you choose? 22:44 <@deen> xrothx: no, but I'm graphing my ping to locations now: http://ddnet.tw/ping/ 22:45 <@deen> you can see the problems on GER quite well there 22:46 < xrothx> maybe because ger has 200+ players at the peak moments 22:46 <@deen> no 22:46 <@deen> that has nothing to do with it 22:46 < xrothx> ok 22:46 < xrothx> So you ll most likely switch? :D 22:46 <@deen> no 22:46 <@deen> the performance of GER is great 22:46 <@deen> and the ddos protection as well 22:47 < xrothx> But..? 22:47 <@deen> I would want another solution, but I guess I need eeeee's help for that, because I don't know how to set it up 22:47 < xrothx> Ok 22:47 < xrothx> deen: I had an idea 22:47 <@deen> but the ping jumps to 120 for many players 22:47 <@deen> it also happens from Britain and Poland sometimes 22:47 < xrothx> I can buy a cheap dedicated server with 64gb etc 22:47 < xrothx> really good, 22:47 < xrothx> but what if we take a small vps with nfo 22:47 < xrothx> and make everything go through it? 22:47 < xrothx> :) 22:47 < xrothx> great host, 22:48 < xrothx> great protection 22:48 < xrothx> kind of vpn style, but for hosting 22:48 <@deen> yeah, that's a possibility, but in reverse 22:48 <@deen> hm 22:48 <@deen> our approach would be superior 22:49 <@deen> also, nfo has the ping problems, which your idea doesn'T solve 22:49 <@deen> the idea would be this: 22:49 <@deen> keep GER running as it is, use GER2 as a proxy to GER, as you suggested 22:49 <@deen> and people can connect to either IP and end up playing on the same server 22:50 <@deen> then we could also add some client magic so that client automatically sends packets to both servers (or the best one) 22:50 < xrothx> Will it not be their ping + ping of GER2-GER 22:50 <@deen> yes, it will be 22:50 < xrothx> higher ping 22:50 <@deen> but ping of GER2-GER is 0.8ms 22:50 < xrothx> ok 22:50 <@deen> on GER I have 120 sometimes, on GER2 mostly 20 22:50 < xrothx> Kind of cloud creation of teeworlds hosting ^^ 22:50 <@deen> 21 wouldn't be a problem either^^ 22:51 < xrothx> ok 22:51 < xrothx> Should try 22:51 < xrothx> ofcourse ^^ 22:51 < xrothx> But why not taking GER2 alone then? 22:51 < xrothx> GER is useless in this situation then 22:51 <@deen> GER2 has worse performance and probably worse ddos protection 22:51 < xrothx> upgrade GER2 a little bigger & it will be 0.8 ping less? 22:52 <@deen> also GER2 has bad ping for some people... 22:52 < xrothx> lol? 22:52 < xrothx> conclusion: 22:52 < xrothx> You will Never find the perfect server 22:52 <@deen> yeah, that's why the new solution would be great 22:53 <@deen> combines all the advantages of both servers 22:53 < xrothx> Aight 22:53 < xrothx> nice 22:53 < xrothx> I have to go now, Goodluck with it ;) 22:53 <@deen> bye 23:01 < Nimda> Peaks by Ama just released on Moderate at 2014-11-26 23:00 23:15 <@deen> is there a small and simple library for checking that updated files are signed? would be nice for the autoupdater in ddnet 23:20 <@deen> since Learath2 wants to use libcurl (I think), maybe we should just use https with our own hard-coded CA 23:22 <@deen> all i could find was gnupg, openssl/libressl, libtomcrypt, crypto++ 23:24 <@deen> and libsodium 23:40 < Learath2> deen did that before with ddwar 23:41 < Learath2> https://github.com/Learath2/teeworlds/tree/fetcher1 23:42 < Learath2> I did finish the libcurl part 23:42 <@deen> nice 23:43 < Learath2> just need to make a pretty parser for the upd file 23:43 <@deen> you could also change the update file 23:43 <@deen> the format 23:43 <@deen> EastByte likes json, so we have a json parser in the source now^^ 23:44 < eeeee> yeah hardcoded custom CA worked fine but in the end we used a real CA because we needed it for the website anyway (and you can't host two different https certs on same ip) 23:44 <@deen> i don't trust real CAs but ok =/ 23:45 <@deen> for example: what if someone steals ddnet.tw domain. then they can get a certificate and push malware updates 23:47 < eeeee> hmm maybe you can still have a cert signed by real CA but make the client accept only that particular cert (not any valid one) 23:47 <@deen> yeah 23:47 <@deen> i think we could just use a different certificate for update.ddnet.tw than for ddnet.tw 23:47 < eeeee> yeah i though that too but we only had one ip 23:48 < eeeee> could use a different port for update though 23:48 < eeeee> also note how using different certs doesn't actually solve the malware problem 23:49 <@deen> at least it solves it in ddnet client 23:49 < eeeee> if someone steals ddnet.tw he will just drop some OMG DOWNLOAD THIS NEW UPDATE DDNET CLIENT v. 9001 on the front page 23:49 <@deen> can't fix web browsers