DDraceNetwork - developer

DDraceNetwork

Development / developer

Development discussion. Logged to https://ddnet.tw/irclogs/ Connected with DDNet's IRC channel, Matrix room and GitHub repositories — IRC: #ddnet on Quakenet | Matrix: #ddnet-developer:matrix.org GitHub: https://github.com/ddnet

Between 2017-09-21 00:00:00Z and 2017-09-22 00:00:00Z

[ddnet-maps:master] 1 new commit

980c078 A Guhimbarwa - def-

[ddnet:master] 1 new commit

94c7c8d Use C++14 - def-

<Learath2> deen: skipping 11 are we? :P

:D

what a huge jump

13:12

maybe c++17 ?

to new i thing

<Learath2> should slowly start getting rid of base/tl stuff

13:13

<Learath2> poor matricks and his nicely optimized memory structure

13:13

wouldn't be fast to rewrite teeworlds haha

<Ryozuki> c++ has lambdas

13:14

<Learath2> only people who want to show off seriously use lambdas :P

https://www.youtube.com/watch?v=OLH3L285EiY&t=1270s

LAMBDA Functions: Powerful And Elegant Abstractions

<Ryozuki> i like them :o

lecture about lambdas

<Learath2> Ryozuki: you should try Scala or haskell

13:15

<Ryozuki> i never see videos about programming, i don't like watching, reading is better

it's not about programming

13:16

it's about lamdas

<Ryozuki> ah

in programming either, but generally

13:16

about lambdas

<Learath2> needs a pitch change for her voice :/

13:16

<Ryozuki> Learath2, which one you recommend first? is scala functional also?

<Learath2> there is a specific pitch that some girls speak with, it totally distracts me from everything including what they are saying

Im sorry : (

<Learath2> Ryozuki: I reccommend neither, functional languages feel like more of a toy for me tbh

i would tell her that

<Learath2> Ryozuki: well they do give you a nice new perspective

13:18

<Learath2> i'd try haskell I guess, more people using it and it's hip nowadays

13:19

<Ryozuki> i also try to learn nim but i give up easily

13:19

<Learath2> @Savander, not your fault her voice has one specific pitch that annoys me :P

13:19

<Ryozuki> one thing i remember is that you can make your own operands, i found that cool

Why does every dev type with with irclink but savander doesnt

<Learath2> Ryozuki: nim is nice, python with the performance of C and very nice meta programming facilities

13:20

<Learath2> also really nice that you can run nim code at compile time

13:20

<Ryozuki> thats nice

13:21

<Ryozuki> ill retake it :P

13:21

<Learath2> @Ezy Savander is speaking in discord, we are talking in irc

Oh

<heinrich5991> Learath2: you should learn a functional language as well, it changes the way you think :)

13:21

<Ryozuki> hehe

13:21

<Ryozuki> i was learning nasm these lasts 2 days

13:22

<Ryozuki> (linux)

13:22

<Learath2> I've had the pleasure of working with Racket last year

13:22

<heinrich5991> deen: why C++14 as opposed to C++11? ^^

13:22

<Learath2> functional languages and their obsession with no side effects only pure functions nearly drove me insane

13:23

<heinrich5991> try to embrace it :) don't work against it, try to understand how they work

13:23

<Learath2> i understand how they work and i don't like it

13:23

<heinrich5991> heh, that's a bold statement

13:24

<Learath2> the extremes of everything is stupid, the obsession with OOP in java is retarded, the obsession with having no state in functional languages is also stupid

13:25

<heinrich5991> I don't think you understand how they work; I don't think I understand how they work

13:27

<Learath2> heinrich5991: what is there not to understand? they define functions in the mathematical sense

13:27

<heinrich5991> to understand how you would write programs in it

13:28

<Learath2> makes them good for mathematical concepts but also inconvenient to code on without changing your entire perspective + are not very indicative of how cpu's compute stuff

13:28

<heinrich5991> "inconvenient to code on without changing your entire perspective"

13:28

<heinrich5991> that's what you should try ^^

13:28

<Learath2> i

13:28

<Learath2> i'd rather keep my perspective == the cpus perspective

13:29

<heinrich5991> you can keep that as well

13:29

<heinrich5991> but broaden your horizon ^

13:29

<heinrich5991> ^

13:30

<Learath2> I mean if you have the time, sure they are nice toys to play around with

13:31

<heinrich5991> don't think about them like that

13:31

<Learath2> don't expect them to perform well or be convenient for using as a general language

13:31

<heinrich5991> try to approach new things without prejudice ^^

Guys, any decisions about randomness ?

<heinrich5991> not yet, I think

13:33

<Learath2> heinrich5991: i approached it without prejucide when i used racket, i formed some conclusions which are "they are not performant, not memory efficient, not good at writing general programs, are nice to express mathematical constructs in"

13:34

<heinrich5991> I'd say 3 is wrong, they're good at writing general programs

13:34

<Learath2> either haskell or rust is next on my list so it's not like i scratched them off completely

13:36

<heinrich5991> haskell is also at most one category slower than C

13:40

<Learath2> heinrich5991: it's not haskells fault that it's slower, recursions are not performant and hard to optimize, functional languages use a lot of it

13:40

<heinrich5991> haskell can optimize them away

13:41

<Learath2> some of them surely

13:42

<heinrich5991> also, function calls aren't that expensive

13:42

<heinrich5991> they compile to a direct jump :) that can be predicted

13:43

<Learath2> they compile to a call unless the compiler does some clever optimizing, calls are atleast 2x worse then jmps

13:45

<Learath2> think llvm/gcc can do tail recursion nowadays

13:46

<heinrich5991> on cpu level, call is just a push and a jmp if you want to talk in asm terms

13:46

<heinrich5991> but it's a direct jump so it can be predicted

13:47

<heinrich5991> the worst thing for modern processors is mispredicted jumps AFAIK

13:49

<Learath2> if it's tail recursion it just gets optimized to a loop anyways

13:49

<Learath2> non-tail recursion should cause lots of mispredicted jumps

13:50

<heinrich5991> no

13:50

<heinrich5991> whether tail or non-tail recursion doesn't matter for mispredictions

13:50

<heinrich5991> the if branches are what matters

13:51

<Learath2> true nvm

13:52

<Learath2> could run out of cpu cache \o/

13:53

<Learath2> the point is it's hard to think what would be performant while thinking in terms of a functional language

13:56

<Learath2> my turn to ask questions: heinrich5991 why do you think that they are good for writing general programs while also claiming you don't understand much about them?

13:58

<heinrich5991> because I know people who are very fluent in haskell

14:01

<Learath2> and do these people use haskell day to day for everything?

14:01

<heinrich5991> this one person actually only uses haskell

14:01

<heinrich5991> and I try to get them to learn a bit more about C, but just like you, they think they don't need it ^^

14:02

<Learath2> can he match my claim of "I can do anything the CPU allows in C"?

14:04

<heinrich5991> can he match the claim you tried to make up to be unmatchable?

14:04

<heinrich5991> but probably yes

14:04

<Learath2> apparently haskell can now do asm and your friend is a compiler

14:05

<heinrich5991> asm isn't in C, do you want to use compiler extensions?

14:06

<Learath2> I'm pretty sure I can hack up something in C to move the EIP around

14:08

<Learath2> I'm fairly confident in my claim that nothing matches C that is higher level then C in flexibility thus generality

14:08

<heinrich5991> anyway, I'm sure he can similarly make up a claim that tries to be unmatchable with C

14:08

<heinrich5991> try rust

14:08

<heinrich5991> try c++

Guys, can you tell me how to test my hard drive disc ?

<Ryozuki> elixir (?)

it's getting slower

14:10

:/

<heinrich5991> Savander: os?

w10

14:11

i mean, it worked fine, even over last year

<Learath2> rust has the most obnoxious memory-safety rules, thus less flexible. c++ is basically c with classes that decided it no longer wanted to be c with classes so got their own compiler, but i'll give you that one

but that drive has over 4-5 years

<Ryozuki> syntastic doesn't support nim? :(

@Savander read out the S.M.A.R.T data

(y)

everything you need to know I guess: https://en.wikipedia.org/wiki/S.M.A.R.T.

S.M.A.R.T. (Self-Monitoring, Analysis and Reporting Technology; often written as SMART) is a monitoring system included in computer hard disk drives (HDDs) and solid-state drives (SSDs) that detects and reports on various indicators of drive reli...

<heinrich5991> Learath2: what about c++ or rust in response to your question? ^^

14:28

<Learath2> gave response about rust and c++

14:28

<Learath2> missed C++ b/c it used to be C :P

14:28

<heinrich5991> rust can do the same things if you want

14:28

<heinrich5991> you have unsafe{}

14:28

<heinrich5991> :)

14:28

<Learath2> besides i'm fairly sure your friend can't make a claim I can't fulfil in C and he can in Haskell

14:29

<Learath2> given the haskell compiler is written in C

14:29

<heinrich5991> that is indeed not an argument ^^

14:30

<heinrich5991> you can write a c compiler in brainfuck ^^

14:30

<heinrich5991> but you can't do random numbers in brainfuck

14:30

<heinrich5991> or open files, or anything

14:30

<heinrich5991> it would only matter if it compiled to C

14:31

<Learath2> k, but i already made the claim that you can do anything the processor can do in C so at the very best he can match my claim

14:32

<Learath2> wait how would you even write a c compiler in brainfuck?

14:33

<heinrich5991> get source from stdin, output binary to stdout

14:34

<Learath2> thought so... you know that's not what I mean when I say the GHC is written in C

14:35

<heinrich5991> ok, waht do you mean?

Just question, what do you to do in so low level ?

<heinrich5991> you mean it's compiled to C?

14:35

<heinrich5991> I don't think that's true anymore

14:35

<Learath2> a piece of C code can take Haskell code in and output LLVM IR

14:36

<heinrich5991> a piece of haskell code can take C code and output ELF binaries

14:36

<heinrich5991> (or LLVM IR if you want)

14:37

<Ryozuki> endless war :D

14:37

<heinrich5991> why war? ^^

14:37

<Learath2> you'd still need the GHC to compile said haskell code

14:37

<Ryozuki> it's a way to say it -.-

14:37

<Learath2> anyways they should just write the GHC in haskell already

14:39

<Ryozuki> 1 + 2

14:39

<Ryozuki> (ignore my last msg)

4

<Learath2> i do not believe it is possible to craft a problem that is sovable in Haskell and not in C, feel free to prove otherwise

14:39

<Learath2> s/Haskell/<Insert higher level language then C>/

14:40

<heinrich5991> i do not believe it is possible to craft a problem that is sovable [sic :P] in C and not in Haskell, feel free to prove otherwise

14:40

<heinrich5991> problem a bit underspecified though

14:41

<Ryozuki> someone here uses weechat?

14:41

<Learath2> Ryozuki: i do

14:41

<Ryozuki> i'm new with it, struggling with... all of it :D

14:42

<Ryozuki> is there a way to change what channel you are viewing without /join ?

14:42

<Learath2> Ryozuki: alt-number

14:43

<Learath2> if you have the mouse extensions enabled you can also click on the channel

14:43

<Ryozuki> finally :D

14:43

<Ryozuki> how to install it?

14:44

<Learath2> heinrich5991: yeah problem doesn't seem to be the word i'm looking for

14:45

<Learath2> but feel free to show me a haskell function that takes a pointer and modifies the pointed-to object

14:46

<Ryozuki> (and how to "scroll up"?)

14:46

<Ryozuki> ah, re pag

14:47

<Learath2> Ryozuki: can't haz mouse extensions if your distro doesn't have it compiled in their package and your terminal doesn't

14:47

<Learath2> i'd check those

14:47

<Learath2> if not you might need to compile weechat yourself

14:47

<Learath2> but you better get used to the shortcuts, they are really nice :D

14:48

<Ryozuki> yeah :P

14:49

<heinrich5991> Learath2: https://hackage.haskell.org/package/base-4.5.1.0/docs/Foreign-Storable.html#v:poke

14:49

<Learath2> Ryozuki: ever gave a tiling window manager a try?

14:49

<Ryozuki> what is that? c:

14:50

<heinrich5991> Ryozuki: windows always fill up the whole screen

14:50

<Ryozuki> i have 2 screens, never faced that prob

14:51

<Learath2> heinrich5991: i stand corrected, haskell does allow non pure functions then

14:51

<heinrich5991> which problem?

14:51

<Ryozuki> having a terminal fill whole screen

14:51

<heinrich5991> it is an IO action, it's a pure function

14:51

<Ryozuki> or u refer another thing?

14:51

<heinrich5991> look at the return type ^^

14:51

<Learath2> Ryozuki: it's not a problem :D instead of having windows around your windows are tiled to your screen

14:51

<Ryozuki> :o

14:52

<Ryozuki> which one you recommend?

14:52

<heinrich5991> what's your OS?

14:52

<Ryozuki> debian 9

14:52

<heinrich5991> i3 works for me

14:52

<Learath2> https://learath2.info/i3.png

Embedded image

14:52

<Learath2> like this f.e.

14:54

<Ryozuki> :o

14:55

<Learath2> heinrich5991: no i mean, your argument that "it is not possible to craft..." is completely correct in this case because haskell does allow non-pure functions

14:55

<Ryozuki> that looks good

14:55

<heinrich5991> the function is pure though

14:55

<Learath2> Ryozuki: i thought so too but it takes some time to get used to having a billion shortcuts

14:56

<heinrich5991> because it returns an IO ()

14:56

<Ryozuki> hmmm

14:56

<Ryozuki> "apt-get install i3 suckless-tools" funny package

14:56

<Learath2> switched between windowed and tiled dms for a while until i got used to it

14:56

<Ryozuki> heinrich5991: you are also using i3?

14:56

<heinrich5991> yes

14:56

<heinrich5991> why do you want to install suckless-tools?

14:57

<Ryozuki> and you recomend it more than gnome?

14:57

<Ryozuki> https://wiki.debian.org/i3 tells me to do so

14:57

<Learath2> heinrich5991: that function might be but haskell does allow non pure functions which means it's probably impossible to craft a problem such that it's solvable in C but not in haskell

14:57

<heinrich5991> dunno, depends on what you want, Ryozuki

14:57

<Ryozuki> "suckless-tools should also be installed to enable dmenu and i3status for application launching and system information respectively. "

14:58

<Learath2> Maybe it is but I do not possess another quality of haskell that limits it which i can use to craft said problem

14:59

<Learath2> Ryozuki: well using a tiled wm means you have no desktop and they work better on bigger screens

14:59

<Ryozuki> i have a 24' and a 21' screen

15:00

<Learath2> i reccommend it more then gnome but i'd reccommend anything more then gnome :P

15:00

<Ryozuki> hmm

15:00

<heinrich5991> gnome is also fine

15:00

<heinrich5991> I also know a lot of people who use that

15:00

<Learath2> yeah my issues with it is purely from a principle standpoint

15:01

<Learath2> i don't like the bloat and i don't like that it's helping systemd take over the world

15:01

<Learath2> kde is acceptable, openbox and lxde look cute but i've never used them personally

15:02

<Learath2> Ryozuki: you need to try a lot of them before you choose :P

15:02

<Ryozuki> Learath2: i see you dislike a lot of things xD

15:03

<Learath2> I like a lot of things aswell so they balance out to make me a normal human being :P

Learath just like to spread his negative word instead of positive :V

15:09

;D

<Learath2> the thing is I don't like some things that are common nowadays, e.g. gnome-systemd. So it seems like I am a bitter cunt who doesn't like anything new

15:17

<Learath2> and sometimes i'm quick to judge things so I like to discuss them over with other people to make sure I'm not missing something

<deen> Always skip c++11, 14 was just minor fixes and it's supported in gcc, msvc and clang already

15:38

<heinrich5991> https://msdn.microsoft.com/en-us/library/hh567368.aspx

15:38

<heinrich5991> that doesn't look like c++14 is well-supported

15:39

<deen> only sfinae is missing, right?

15:39

<deen> fits my experience from work

15:39

<deen> ah, that's 11^^

15:40

<heinrich5991> do you want to drop support for anything but VS2013?

15:40

<heinrich5991> (just out of interest)

15:40

<deen> meh

15:40

<heinrich5991> VS2015 I mean

15:40

<deen> then just use the stuff from 14 that keeps building on windows

15:40

<deen> we have the ci

15:46

<heinrich5991> sorry to bother you, but do we need 14? I'd personally prefer using the smaller standard

15:46

<heinrich5991> it seems that c++11 has threading stuff already

15:47

<deen> I'd prefer using the same version everywhere so I don't have to look up what features I can use :D

15:47

<heinrich5991> at work/tw?

15:47

<deen> anyway, just make it 11 if you want to

15:47

<deen> yeah

15:48

<deen> also both tw and work don't use the stl :D

15:49

<heinrich5991> why at work?

15:50

<deen> we have our own implementation using allocators

15:50

<heinrich5991> ah

15:50

<heinrich5991> the stuff does support allocators though I think

15:50

<heinrich5991> the stl

15:50

<deen> probably not back when our replacement was written :D

15:50

<heinrich5991> http://en.cppreference.com/w/cpp/container/vector

15:51

<deen> and we explicitly forbid everything without allocators

15:51

<deen> otherwise how would you enforce it?

15:52

<deen> and there are lots of weird corner cases about copying and moving with respect to allocators

15:52

<heinrich5991> move constructors ._. ^^

the CI has VS2015

<heinrich5991> yep, can change to VS2013

16:05

<deen> Learath2: why do you think functions should always have a name, even if you use them only once?

I don'

16:06

don't

<deen> I thought you said you don'T like lambdas

I said only people who want to show off use lambdas iirc, and i was just fooling around then

<deen> ok

16:07

<heinrich5991> lambdas are useful :)

16:07

<deen> i think it's nice when you want to sort by something specific just one time, then you just make an anon. function

16:07

<deen> and it's clear for everyone that it's not used anywhere else

16:08

<deen> saying lambdas are bad is like saying rvalues are bad and everything should be named^^

1

all for anonymous functions

16:11

does c++11/14 have foreach with lambdas?

<deen> Learath2: Want to read something by your favorite author

16:11

<deen> ?

who is my favourite author?

<deen> https://lwn.net/Articles/250967/

What every programmer should know about memory, Part 1

Ulrich Drepper recently approached us asking if we would be interested in publishing a lengthy document he had written on how memory and software interact. We did not have to look at the text for long to realize that it would be of interest to many LWN readers. Memory usage is often the determining factor in how software performs, but good information on how to avoid memory bottlenecks is hard to find. This series of articles should change that situation. Click below (subscribers only) for the first installment.

looks like a nice read

<heinrich5991> Learath2: c++ foreach is without lambdas

why is he my fav author tho?

<heinrich5991> for(auto &x : vec)

16:13

<deen> I think Learath2 means this: http://en.cppreference.com/w/cpp/algorithm/for_each

16:13

<heinrich5991> ah

16:13

<deen> that seems to support functions since forever and policy since c++17

16:14

<deen> and you can use lambdas anywhere you would use a function I guess

ah yes that's it

<heinrich5991> yes, and more

16:14

<heinrich5991> because you can capture the environment

well can you really? there aren't many functions that take a function as an arg

<deen> you can do that with a function too, heinrich5991

16:14

<deen> Just pass everything manually :D

16:15

<deen> Learath2: "higher-order functions"

16:15

<heinrich5991> you can emulate it, but not with the same syntax, and not typesafe

16:15

<heinrich5991> we always pass void *pUser for that

16:15

<heinrich5991> but that's the opposite of type-safety ^^

16:16

<deen> Learath2: well, you can always write your own!

@heinrich5991 how do lambdas carry the environment around? what are their scope?

hmm, no i dea what's going on. butt

yesterday i enabled cl_predict 1

16:17

then i disabled it

16:18

and ehm, my game looks laggy idk

16:18

wiht delays

<deen> then enable it again

weird stuff

<heinrich5991> it's on by default

it surely needs to carry something like pUser around in the background?

<heinrich5991> you don't want to disable it

16:18

<heinrich5991> Learath2: yes. but in a typesafe way :)

16:18

<heinrich5991> you declare what it carries around in the [] of lambdas

hmm, i thought it was from anti ping

16:18

it isn't ?

<heinrich5991> no

16:19

<heinrich5991> cl_predict is vanilla

really ?

16:19

im stupid

<heinrich5991> deen: does that fall under "users even disable stuff in the console and don't know what's wrong?" :P

16:20

<heinrich5991> you said that the other day ^^

god knows how they implemented that. Sounds scary

what happens if the variable goed out of scope?

<deen> yes

16:21

<Ryozuki> i'm sure the one who implemented that knows and not god ^^

16:21

<heinrich5991> Learath2: same as with other parameters. if you pass it by reference, it's bad, if you pass it by value, then not

had a guy enable gfx_noclip and wonder why clipping caused problem

16:21

s

disabled

<heinrich5991> ohhh fuck

he disabled that thing

16:22

or

<heinrich5991> I think that must have been the problem with my teeworlds installation many years ago

16:22

<heinrich5991> a disabled gfx_noclip ^^

16:22

<deen> Learath2: Ooops, mixed up one unpopular linux dev with another one :D

16:22

<heinrich5991> ulrich drepper: author of glibc btw

16:22

<heinrich5991> knows his stuff, sometimes makes unpopular decisions :)

16:24

<Ryozuki> http://en.cppreference.com/w/cpp/utility/functional/function

btw, what was purpose on gfx_noclip ?

testing i'd guess

16:24

deen who did you think it was?

<heinrich5991> lennart poettering probably

lennart poettering is not smart enough to write any article

<heinrich5991> talking about prejudices :D

16:25

<Ryozuki> https://de.wikipedia.org/wiki/Ulrich_Drepper yay, only in german

Ulrich Drepper war der Chefentwickler und Verwalter des „GNU C standard library“-Projektes, glibc. Drepper erlangte sein Diplom im Fach Informatik an der Universität Karlsruhe. Ab 1996 arbeitete er für die Firma Cygnus Solutions. Von 1999 b...

16:26

<deen> after my audio has problems on windows and android I'm not that unhappy with pulseaudio anymore

16:26

<heinrich5991> how dare you not insult poettering's software

16:26

<heinrich5991> you must be a bad person :P

16:26

<heinrich5991> /s

16:26

<Ryozuki> is pulseaudio bad?

poettering is the worst thing to happen to linux in a century

<heinrich5991> ask Learath2 if you want to hear yes

16:27

<Ryozuki> Learath2, i expected u to say that :D

16:27

<heinrich5991> :)

i'll say it's bad and heinrich and deen will prolly say its a blessing

<Ryozuki> Learath2, is there someone who did something good for linux?

16:27

<Ryozuki> in this century

Linus torvalds obv

<Ryozuki> :D

16:28

<Ryozuki> they are all red hat developers

16:28

<Ryozuki> Learath2, u consider red hat bad for linux?

have not made my mind up about red hat yet

<deen> since I can't even remember these guys' names I think it's pretty obvious I don't care about who to like and who not :D

because I don't like to judge stuff I know nothing about

<heinrich5991> you don't know linus torvalds?

16:29

<Ryozuki> (i do)

16:29

<deen> i didn't mix that one up

16:31

<Ryozuki> if tux the penguin is the mascot of linux, what's windows mascot?

16:32

<heinrich5991> the flag is the logo of windows

16:32

<Ryozuki> a flag with 4 colors

16:32

<Ryozuki> so creative :/

look at poetterings face, it's the perfect punch me face

<heinrich5991> not anymore

16:32

<heinrich5991> it's only blue now

16:32

<Ryozuki> "modern"

16:34

<deen> Learath2: if you don't like his software, go join a project that develops an alternative. if you don't like him, don't go on a date with him

can't possibly compete with something loved by so many people

16:34

when fanboyism gets involved people don't think logically anymore

<heinrich5991> are you sure it's only fanboys who aren't opposed to systemd?

16:35

<deen> I don't think he has fanboys, always just read negative stuff about him

there are ofc some people who aren't fanboys

16:35

those can be converted but they are there for convenience

16:36

nothing i can make beats PnP systemd

16:36

even if i do i'd only get some of the convenience people because the rest just dont know how to linux

<Ryozuki> what means "how to linux"?

I mean the people who are here to see the "Year of the linux desktop"

16:38

they just care if it works

<Ryozuki> usually competitions involves making better things to win

usually

16:40

in this case poettering has people that likes him in ubuntu and debian so it's more of a leverage game

16:41

and it seems that wherever debian/ubuntu goes linux follows nowadays

16:42

so yes, get me friends who will push my software in board decisions, some fanbase with pitchforks preferably, and i'll fork OpenRC and beat systemd

16:43

also need a development team to help me dig out the remnants of the software systemd absorbed from that mess of lines they call code

i guess noone wants to have the systemd debate tonight

<heinrich5991> ddnet has 7.5k commits

17:25

<heinrich5991> exactly :)

nice

A biiiig project

do you know how many there was before refactor trunk to master or w/e

i saw, thjast linux kernel has 500k commits ;oll

17:26

:ooo

[ddnet/ddnet] Pull request opened: #872 Change to C++11 in `CMakeLists.txt`, `bam.lua`

also should we start phasing out base/tl?

<heinrich5991> deen has 2271 commits

17:29

<heinrich5991> dunno

17:30

<heinrich5991> I think proper threading would be a priority

proper as in using c++11 instead of pthreads?

<heinrich5991> proper as in we don't have to fiddle around with named semaphores on macos

17:31

<heinrich5991> c++11 is also more widely supported, windows as well ^^

@heinrich5991 does c++11 have semaphores?

<heinrich5991> good question

17:32

<heinrich5991> I don't see them

17:32

<heinrich5991> I only see condition variables and mutexes

17:32

<heinrich5991> but atomics!

17:32

<heinrich5991> :)

heinrich5991: also how do you think we should handle system.c?

<heinrich5991> we could add system.cpp :)

17:36

<heinrich5991> what did we use semaphores for?

to avoid spinlocks in the queues we have

17:38

fetcher and CJobs

<heinrich5991> for queues, ok

well there has to be another good way or else more people would complain on osx

<heinrich5991> yea, can probably implement it using condition variables

well atomics guarantee complete read writes right?

<heinrich5991> yes

couldn't we just use an atomic int as a semaphore?

<heinrich5991> how do you wait for an atomic int to change?

17:41

<heinrich5991> can't, really

oh nvm what i thought about gives you a spinlock again

<Ryozuki> Savander, it actually has 705k

o.O

https://github.com/torvalds/linux

linux - Linux kernel source tree

http://www.wasz.ovh/img/GFc9m6yl87131

Screenshot_GFc9m6yl87131

2 view(s)

19:04

:/

19:05

no errors in SMART, no damaged blocks

19:05

;/

<eeeee> i think concurrent queues might get added to the current c++ concurrency TS

after having a short look it should be no problem to replace our semaphores with condition_variables, we might want to think about some wrappers or classes we want for threading tho

<eeeee> might appear in c++20, so we'd be able to use it around 2025. if we move fast maybe we'll already start working on teeworlds 0.8 by then!

so implement their api ?

<Learath2> our use of semaphores can be completely replaced by condition_variables and a way to keep track of the number of entries in queue

19:14

<Learath2> I'll also implement the fetch task as a CJob in a couple of days so we can get rid of the unnecessary queue in CFetcher

finally

@Savander what is this tool lol

?

19:50

HD Tune

never heard of it

19:52

did you only pay attention to the damaged blocks?

i checked SMART as well

19:55

im not good in diagnostics

19:55

;/

[ddnet:master] 1 new commit

a9b80e8 Update to Portuguese translations (by RockuS) - def-

∞ contributors (linux kernel)

21:04

@Ryozuki more than 700k commits, they created a new git repo at some point

xD

and it's mirror

how many commits did you make @Ryozuki

to the kernel? none

over 100k

21:07

@Ryozuki god of commits

?

to ddnet?

i didnt do 100k commits

21:08

14 commits / 177 ++ / 121 --

21:08

nearly nothing

:<

21:09

lier

if you are curious https://github.com/Ryozuki/ddnet/commits?author=Ryozuki xD

ddnet - DDraceNetwork, a mod of Teeworlds

<heinrich5991> deen seems to be the only one who has debloated the code :o

21:12

<heinrich5991> I on the other hand, added >25k lines :(

GreYFoX 735 commits / 49,977 ++ / 1,372,346 --

21:13

rm -R

21:13

i guess

deleting the old ddrace code I guess?

its lines or characters?

<heinrich5991> lines

are you sure? 1.5mln lines of code ?

21:16

of ddrac e

<heinrich5991> yes, I'm sure that it's lines

I thought you did more @Ryozuki

contribution can't be measured in mere LOCs or commit count

how else did you contribute?

talking, ideas

21:31

discussions

21:31

reporting bugs

you can contribute in many ways

debugging if neccessary

@heinrich5991 does your teehistorian will save names as well?

<heinrich5991> yes

it won't be problem in terms of law?

<heinrich5991> saving pseudonyms? interesting question

Yes, because you can identify for example me

22:03

by my nickname

<heinrich5991> nah, could be anyone else as well

it's not like my problem

<heinrich5991> but we also do that already, so I probably don't introduce something new

22:03

<heinrich5991> see ranks on ddnet.tw

I can ask my lawyer if you want

<heinrich5991> if you want

22:04

<heinrich5991> although as indicated above, we already save names

but it's more like saving every move of player

22:06

right, ranks etc. Im curious how f.e google solve problems, about saving every your move

22:06

for adverts for example

<heinrich5991> they have ToS

we don't have right?

we don't have one

so maybe we should have one, like "We are storing names to serve ranking system" etc.

22:11

but, meh

22:11

then they should accept

doesn't sound desirable

I don't know

22:12

im not sure to be honest

I don't think anyone cares if you save their nickname

You would be surprised

ghost

2

Exported 505 message(s)