01:11 < eeeee> Learath2: so why do you have to do a separate request just to check if map exists?
01:14 < eeeee> somehow programming in prolog and haskell doesn't sound like a very practical paradigm ;\)
01:16 < eeeee> and is it easier to learn about compilers using java bytecode compared to using llvm bytecode?
01:25 <@deen> ehm, no
01:25 <@deen> but java seems to be the "cool thing" at universities
02:15 <@heinrich5991> deen: https://www.teeworlds.com/forum/viewtopic.php?id=11259
02:55 <@deen> heinrich5991: yes?
02:55 <@heinrich5991> maybe you could add your experience there? :)
02:56 <@deen> it depends a lot on the mod
02:56 <@deen> and no idea^^
02:57 <@deen> GER has 51 GB/day
03:10 <@deen> hope this is fine: https://www.teeworlds.com/forum/viewtopic.php?pid=112846#p112846
03:11 <@heinrich5991> thanks :)
03:16 <@deen> RUS always loses players randomly and I don't understand it
03:16 <@deen> zoom into today, players around ~40, then sudden drop and 2 hours later back
03:16 <@deen> I ask the players if there are any lags or problems but haven't heard of any
04:32 <@deen> there are people who play with the name "/kill"
04:32 <@heinrich5991> mh
04:32 <@heinrich5991> why?
04:33 <@deen> we have "kill protection"
04:33 <@deen> so after 20 minutes of gametime killbind doesn't work anymore and instead you have to explicitly type /kill
04:33 <@deen> not sure how to solve this
04:33 <@deen> I've always been annoyed by people's names starting with a slash
04:35 <@heinrich5991> deen: fix the problem with accidently killing oneself or with the people's names?
04:35 <@deen> well, we fixed the problem with accidentally killing oneself
04:36 <@deen> but now we have the new problem of people making others accidentally kill themselves using exactly this kill prevention^^
06:13 <@heinrich5991> deen: what is patafix secret sauce to faster server browser reload?
06:13 <@heinrich5991> patafix's
06:14 <@deen> retrying servers that don't answer
06:15 <@heinrich5991> :)
06:15 <@heinrich5991> that might not work for 64p servers
06:15 <@deen> indeed
07:14 < Teetime> deen: you could check if something is written after /kill and then ignore it
08:32 < GoJEGrEEN> hi
08:38 < GoJEGrEEN> deen: i found 2 bug ( i think maybe not ) in ddnet , one in client another in server
12:52 < KinG_> hi all
12:52 < KinG_> deen: u here>
12:52 < KinG_> ?
12:54 < KinG_> EastByte: hi u know where deen is?
14:12 <@deen> KinG_: hi
14:12 < KinG_> deen: hey
14:13 < KinG_> deen: some problems came in
14:21  * GoJEGrEEN slaps KinG_ around a bit with a large fishbot
14:24 <@deen> ^^
14:24 <@deen> afk
14:24 < GoJEGrEEN> well deen here finaly :)
14:24 < GoJEGrEEN> hi
14:24 < GoJEGrEEN> ops :
14:36 < KinG_> hey Savander :D
14:37 < Savander> Hey ! : )
14:41 < Nimda> DDNet Brazil went down!
14:42 < Nimda> DDNet Brazil went back online!
14:57 < Nimda> DDNet Brazil went down!
14:58 < Nimda> DDNet Brazil went back online!
15:00 < KinG_> anyone knows what does this mean
15:00 < KinG_> : /lib/ld-linux.so.2: bad ELF interpre                           ter: No such file or directory
15:47 < KinG_> laxa: u here?
15:52 < laxadedi> KinG_: hu; M
15:53 < KinG_> laxadedi: to run a -pl file in linux i should do chmod +x first then ./ ?
15:54 < laxadedi> I guess so
15:54 < laxadedi> maybe there is another way for perl to do that too
15:54 < laxadedi> but if you have your shebang and you give +x to your file, you should be able to do ./
15:57 < KinG_> hmm
15:58 < KinG_> so to run ts3server_linux_amd64-pl what ways can i try?
16:00 < KinG_> laxadedi: ?what is shebang?:D just tell me how can i run that file
16:00 < KinG_> it's so important
16:01 < laxadedi> google
16:01 < laxadedi> I am working, no time
16:03 < KinG_> someone help meeee
16:05  * GoJEGrEEN_ slaps GoJEGrEEN around a bit with a large fishbot
17:20 < GoJEGrEEN> deen: 
17:20 < GoJEGrEEN> still afk :?
17:25 < GoJEGrEEN> i must go tell him i found 2 bug in client and server ( i think )
17:25 < GoJEGrEEN> bb
17:58 < Maple> nobody will test my map
18:17 < GoJEGrEEN> zZZ
18:27 < Learath2> eeeee: i could probably stop on the header check then continue but that would require me to do some really ugly stuff 
18:53 < GoJEGrEEN> deen: are u there ?
18:56 <@EastByte> GoJEGrEEN: what kind of bug did you find?
18:59 < GoJEGrEEN> hi EastByte 
18:59 <@EastByte> hi
19:00 < GoJEGrEEN> in client and server
19:00 <@EastByte> what is it about?
19:00 < GoJEGrEEN> in server
19:00 < GoJEGrEEN> when you hook player and go pause
19:00 < GoJEGrEEN> when you do resume 
19:00 < GoJEGrEEN> your hook still release
19:01 < GoJEGrEEN> do you understand me ?
19:01 < GoJEGrEEN> :D im very bad in english :P
19:02 <@EastByte> when you press hook and /pause the hookstate will stay the same
19:02 < GoJEGrEEN> go in one server with sv_pausable 1
19:02 < GoJEGrEEN> and test it
19:02 <@EastByte> after that you'll need to press hook twice
19:03 < GoJEGrEEN> i know but
19:03 < GoJEGrEEN> some ppl
19:03 < GoJEGrEEN> for fun do this :/
19:03 <@EastByte> what do they do?
19:03 < GoJEGrEEN> they hook me and they go to pause
19:04 <@EastByte> do you mean /pause or /spec ?
19:04 < GoJEGrEEN>  i mean /pause
19:05 < GoJEGrEEN> give me ip of server with sv_pauseable 1
19:05 <@EastByte> 84.200.43.133:8303
19:05 < GoJEGrEEN> i show u what happen exactly
19:05 < GoJEGrEEN> k
19:12 < GoJEGrEEN> EastByte: bb and dont forgot tell it to deen im bad in english and thanks for your time
19:12 <@EastByte> no problem
19:13 <@EastByte> cya
19:24 <@EastByte> deen: the jetpack prediction isn't non-ddnet compatible
19:25 <@EastByte> the strength is 400 by default and there is no gametype check
19:46 <@EastByte> deen: and the bug GoJEGrEEN talked about already was fixed fe86eef7d8d524bc86044907fcbe170ebae4e1b5
21:06 <@deen> EastByte: other servers have jetpack?
21:07 <@deen> or what do you mean?
21:15 <@deen> and yes, gojegreen's bug should indeed be fixed. was it an old server version?
21:15 <@deen> i think I even released a new version just because of that
21:16 <@EastByte> deen: the jetpack prediction is triggered on vanilla servers for example
21:16 <@deen> wow, it is?
21:16 <@EastByte> jetpack strength is 400 except ddnetsrv tells the client otherwise
21:16 <@EastByte> if I understand correctly
21:16 <@deen> oh god...
21:16 <@deen> why did no one say anything?!
21:16 <@EastByte> hmmm
21:16 <@deen> I'm sure I tested this at some time
21:16 <@EastByte> they use stupid bots for vanilla? :D
21:16 <@deen> must have gone wrong in between
21:17 <@deen> ah, with low ping it's not thaaat noticable
21:17 <@deen> still, needs to be fixed quickly
21:17 <@EastByte> haha
21:17 <@EastByte> you should higher your ping for testing^^
21:17 <@deen> actually with low ping it looks cool
21:17 <@deen> it looks like recoil
21:18 <@EastByte> I know^^
21:18 <@EastByte> I thought it was a feature
21:18 <@EastByte> but it's horible
21:18 <@deen> yeah, with 200 it's terrible
21:18 <@EastByte> the screen is always shaking when you shoot the gun no matter of speed
21:18 <@deen> yeah
21:18 <@deen> so, how to fix it?
21:19 <@EastByte> it looks like the prediction is always enabled but depends on the strength
21:19 <@EastByte> so a strength of 0 like disabled
21:19 <@EastByte> so we could set the strength to 0 by default
21:19 <@EastByte> or we check for the gametype in gamecore
21:20 <@deen> Not sure where to set it to 0 best
21:21 <@EastByte> in some init function when checking for gametype?
21:49 <@deen> we should do it for every game type
21:49 <@deen> i think ddnet server sends a tune packet anyway
21:55 < eeeee> Learath2: i didn't really have time to look at your code but you're most likely doing it wrong if you need two requests
21:57 <@deen> yes, should really only be 1 request
21:57 <@deen> and if the request returns 404 in the header, return immediately and start regular map download instead
21:58 <@deen> otherwise download the map
21:59 < Learath2> deen: it somehow needs to be a blocking but not blocking request
21:59 <@deen> ^^
21:59 <@deen> it needs to be blocking but with a callback
21:59 <@deen> and the callback can stop the block
21:59 <@deen> (I've never used curl)
22:05 < Learath2> deen: there is only one thing you should know about libcurl and that is curl_easy_perform blocks 
22:05 < Learath2> and thats it 
22:06 < Learath2> i really couldnt devise a plan to do it in one go
22:07 <@heinrich5991> maybe the `easy` in `curl_easy_perform` is the problem, it sounds like it's a shortcut for many other functions
22:11 < Learath2> heinrich5991: i checked it isnt there is easy or multi
22:12 <@heinrich5991> Learath2: there aren't other functions, there's just easy and multi?
22:12 < Learath2> there is curl_easy and curl_multi group of functions
22:18 < ddnet-commits> [ddnet] def- pushed 1 new commit to DDRace64: http://git.io/5f09sg
22:18 < ddnet-commits> ddnet/DDRace64 66d3e5e def: Fix: Not jetpack prediction on DDNet incompatible servers
23:15 < eeeee> Learath2: so can you explain again what's the problem with http codes >=400 ?
23:16 < eeeee> what happens if you try to download the file in one request and it's not there?
23:21 < Learath2> eeeee: should fail with an error pTask->State == STATE_ERROR 
23:24 < eeeee> so can't you just catch that error and figure that update is probably not there
23:28 < eeeee> Learath2: also y u no use CJob
23:32 < eeeee> like take a look at https://github.com/def-/ddnet/blob/251af2fd49fd0325033cd2875c67f3359a9a0ee1/src/engine/shared/jobs.cpp it does pretty much the same thing as your code
23:40 < Learath2> eeeee: iguess the only different thing about CJob is that it doesn't kill off the thread if the queue is empty
23:42 < Learath2> I'll be busy for the next 4 or 5 days feel free to improve on it and comment on it and ill fix it all when im back
23:50 < eeeee> Learath2: yeah and you probably shouldn't kill the thread
23:50 < Learath2> eeeee: y ?
23:50 < eeeee> because why would you do that
23:51 < eeeee> also spinning up a new thread requires allocations and stuff - might take a bit of time
23:51 < Learath2> its just going to sit there until something else comes by
23:51 < eeeee> yeah but it will only consume a few kb of mem
23:55 < Learath2> eeeee: really dont feel like making it all into CJob ill just not kill the thread and make it sleep instead 
23:58 < eeeee> but why
23:58 < eeeee> it shouldn't be too much work
23:59 < eeeee> sleeping would increase the latency further i.e. make it even slower compared to native download
23:59 < Learath2> thats what CJob does tho
23:59 < Learath2> i dont see how it could be better then what i did if im doing the exact same thing
23:59 < eeeee> yeah if you use signals like cjob does then it wouldnt have latency