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 2018-02-21 00:00:00Z and 2018-02-22 00:00:00Z

[ddnet/ddnet] Pull request closed: #1044 Remove the blanket rules for files without extension

06:38

[ddnet:master] 2 new commits

2d55136 Remove the blanket rules for files without exte... - heinrich5991 ca75ab3 Merge pull request #1044 from heinrich5991/pr_d... - def-

06:38

[ddnet/ddnet] Pull request closed: #1045 Some CMake improvements

06:38

[ddnet:master] 3 new commits

ac7e92d Rerun wavpack version detection when the librar... - heinrich5991 1578e27 Ignore old generated files if still present - heinrich5991 6c41e95 Merge pull request #1045 from heinrich5991/pr_d... - def-

[freenode] <nemo> if factorio was idle, it was probably a better candidate than an active process

17:02

[freenode] <nemo> 4530344kB is a hell of a lot tho

17:03

[freenode] <nemo> ditto 1431628 in actual memory. 1.4 gigs

17:03

[freenode] <nemo> sounds like a good candidate

17:07

[freenode] <nemo> if you really want to let factorio run wild, you might want to adjust its oom score to protect it

17:08

[freenode] <nemo> looks like a score of -17 is total immunisation

17:08

[freenode] <nemo> I guess downside of that would be you could kill your system

17:08

[freenode] <nemo> https://backdrift.org/oom-killer-how-to-create-oom-exclusions-in-linux

OOM Killer - How To Create OOM Exclusions in Linux - Backdrift

When a linux machine runs low on memory the kernel OOM Killer will kill processes to free up ram. This article explains how to protect processes from OOM Killer.

yo im having trouble with special chars on ubuntu and with my searchengine xd

17:18

how to write things like [ { ] } @ etc in ubuntu or search them in searchengines because they get ignored there .-.

[freenode] <nemo> quote them?

i tried

[freenode] <nemo> shrugs

17:40

[freenode] <nemo> is this some integration of search into gnome3 or unity or something?

?

are you talking about sth like google

duckduckgo

17:43

ye

[freenode] <nemo> is this in browser or in your desktop environment

yep

17:44

nobody using ubumtu here?

[freenode] <nemo> lol @ "yep"

17:44

[freenode] <nemo> to a question w/ 2 choices

17:45

[freenode] <nemo> I'm currently using ubuntu on a few computers. Mate + 14.04 LTS

how do type all these special chars?

17:45

o.O

[freenode] <nemo> everything else is devuan or gentoo, apart from one experiment w/ mint

17:45

[freenode] <nemo> so. again ChillerDragon - I gotta ask. is this in your browser, or your desktop environment

yes

17:45

as i said already

[freenode] <nemo> …

ah

17:45

xd

17:46

xxxD

17:46

i have two probs

17:46

the one not being able to type it

[freenode] <nemo> entering @ sign in browser works fine w/ DDG and Google

and the other one not being able to search for a solution because entering "type { ubuntu" in duckduckgo gives random results

[freenode] <nemo> well. DDG's search may be crap is all

17:47

[freenode] <nemo> but I gotta say. that search string is not great

well nvm just tell me how to write them on my keyboard .-.

[freenode] <nemo> I wouldn't have any idea what you were looking for either

i want to type [ { @

[freenode] <nemo> well on a standard QWERTY american english kbd, that would be <shift> + [

17:48

[freenode] <nemo> no idea what your kbd is

german

[freenode] <nemo> welp. lemme ask Wuzzy in #hedgewars

17:48

[freenode] <nemo> he's german

c:

alt-gr 8

17:48

on the german keyboard

[freenode] <nemo> ah

17:48

[freenode] <nemo> yay. more germans

shift ö on the english layout with a german keyboard IIRC

great

17:49

ty @heinrich5991 working

[freenode] <nemo> FWIW he said alt-gr 7

17:49

[freenode] <nemo> I guess that's { and }

ye all these numbers

17:49

yy

17:49

altgr is key

@ChillerDragon don't you have those printed on your keyboard? ^^

i have .-.

17:49

but i tried all keys except altgr

17:49

i didnt know i have this key

17:49

xd

nemo: for reference: https://upload.wikimedia.org/wikipedia/commons/3/36/KB_Germany.svg

[freenode] <nemo> I never remap kbd off qwerty - always leads to tears when on a random computer - but I guess if one is in germany QWERTZ is the standard

17:50

[freenode] <nemo> for french I use XCompose

yes, qwertz is standard for computers here

[freenode] <nemo> € → compose + e=

ąçȩįų

[freenode] <nemo> huh. no ẞß ? is it too rare?

nemo: top right

[freenode] <nemo> oh

17:51

[freenode] <nemo> there it is

17:51

[freenode] <nemo> yeh

Image attachment

17:52

whats going on?

read is a special command from bash

17:52

system can only call executables

isnt system === bash?

17:52

oh LOOL

17:52

ty

17:52

so i have to call bash

17:52

xd

[freenode] <nemo> heh

maybe don't use system

system rocks .-.

17:53

hwo do i pause my programms then?

std::getline(std::cin) or sth

grpf

every time you call system you're spawning another program, which is pretty expensive on windows e.g.

[freenode] <nemo> not to mention incompatible possibly

hrmpf

17:54

?

17:54

u mean win to linux compatible

[freenode] <nemo> remote env may not have bash

17:54

[freenode] <nemo> or bsd or android or whatever

#ifdef WIN32 system("pause >NUL"); #elif (_linux) || (APPLE) system("read -n1"); #endif

17:55

ifdef rocks

[freenode] <nemo> O_o

17:55

[freenode] <nemo> that's... unique...

^

[freenode] <nemo> so heinrich5991 - does ChillerDragon have commit access to ddnet?

xxxxD

[quakenet] <heinrich5991> no

17:55

[freenode] <nemo> thank god

17:55

[quakenet] <heinrich5991> but he sent PRs already

hehe

[quakenet] <heinrich5991> nemo: but I'm trying to not have a thing like "commit" access anyway

17:57

[quakenet] <heinrich5991> rather a couple of people who can merge pull requests

17:57

[quakenet] <heinrich5991> and everyone makes PRs when they code sth

getchar() works great btw

17:57

maybe im gonna use it instead

[quakenet] <heinrich5991> getchar is even a standard C function :) +1

ye f3

17:58

prs are great

17:58

ye c > c++

17:58

this looks weird

17:58

xd

[quakenet] <heinrich5991> that's undefined behavior

17:58

[freenode] <nemo> heinrich5991: we're too lazy for that

17:58

[quakenet] <heinrich5991> ;)

we have bors

[freenode] <nemo> heinrich5991: it's easier to just give someone access if they are fairly responsible than review/merge everything

the spammer

17:58

xd

[freenode] <nemo> heinrich5991: that's how I ended up w/ access - unc0rr got tired of merging my stuff

17:59

[quakenet] <heinrich5991> :)

18:00

[freenode] <nemo> oh. ChillerDragon - there's a silly thing from the IOCCC that I use in shells from time to time that might entertain you

the word IOCCC already entertains me

18:00

tf is that

[freenode] <nemo> alias ccstdin='g++ -pedantic -ansi -Wall -o ~/.stdin+ ~/.stdin.cc -lm && ~/.stdin+'

18:00

[freenode] <nemo> alias cstdin='gcc -pedantic -ansi -Wall -o ~/.stdin ~/.stdin.c -lm && ~/.stdin'

18:00

[quakenet] <heinrich5991> international obfuscated C contest

18:00

[quakenet] <heinrich5991> c

18:00

[freenode] <nemo> https://www.ioccc.org/

18:01

[freenode] <nemo> http://m8y.org/tmp/stdin.c

18:01

[freenode] <nemo> http://m8y.org/tmp/stdin.cc

18:01

[freenode] <nemo> $ cstdin

18:01

[freenode] <nemo> printf("Hello World\n");

18:01

[freenode] <nemo> Hello World

this is some weird stuff

[freenode] <nemo> your approach to C reminded me of this

18:02

[quakenet] <heinrich5991> why so many flags to gcc?

18:02

[freenode] <nemo> heinrich5991: they accumulated no doubt, over the past 20 years

18:02

[freenode] <nemo> heinrich5991: mostly was trying to be yelled at for whatever I was testing

18:02

[quakenet] <heinrich5991> -ansi = please give me the oldest version of C/C++ you have available

18:02

[quakenet] <heinrich5991> :(

18:03

[freenode] <nemo> hehe

18:03

[freenode] <nemo> no doubt I wanted "safest subset"

18:08

[quakenet] <heinrich5991> nemo: what were the languages again? pascal and haskell?

18:12

[freenode] <nemo> heinrich5991: oh. for hedgewars?

18:12

[quakenet] <heinrich5991> yes

18:13

[freenode] <nemo> heinrich5991: C++/QT5 frontend (iOS and Android have separate frontends ofc), Haskell gameserver (there's a new Rust one unc0rr has started), Pascal engine

18:13

[freenode] <nemo> heinrich5991: rust for engine might happen. dunno. qt5 is higher priority tho

18:13

[quakenet] <heinrich5991> rewrite it in rust?

18:13

[freenode] <nemo> yeh

18:13

[quakenet] <heinrich5991> mh

18:13

[freenode] <nemo> well... more like... unc0rr adapts his pas2c converter to pas2rust

18:13

[freenode] <nemo> then we clean up the result

18:13

[quakenet] <heinrich5991> :D

18:14

[quakenet] <heinrich5991> incremental rewrites might be easier :)

18:14

[freenode] <nemo> eh. the pas2c works already

18:14

[freenode] <nemo> mostly would be changing concepts

18:14

[freenode] <nemo> all the tangled up global state

18:14

[quakenet] <heinrich5991> yea

18:14

[quakenet] <heinrich5991> you might run into quite a few conceptual problems on the way

18:14

[freenode] <nemo> possibly. we've been looking into it a bit

18:15

[freenode] <nemo> the biggest issue, IMO, is the main "gear" iterator - the linkedl ist of struct + function callback for all the things that need to do things in the game

18:15

[freenode] <nemo> but. I think we can address that reasonably

18:15

[freenode] <nemo> the other one that's guaranteed to be an issue is the AI - but we've already decided we'll use unsafe for that - screw rust's paranoia

18:15

[quakenet] <heinrich5991> …

18:16

[freenode] <nemo> (the AI has direct access to the land array from a separate thread, and we do not, in fact, care if it mutates out from under us)

18:16

[freenode] <nemo> so locking is just silly

18:16

[freenode] <nemo> we'll just unsafe take a refererence to it. moving on.

18:17

[freenode] <nemo> say, have you seen what the output of his pas2c looks like?

18:17

[freenode] <nemo> http://m8y.org/tmp/uGears.c

18:18

[quakenet] <heinrich5991> make the world's inner members atomics

18:18

[quakenet] <heinrich5991> then accessing it from different threads is no longer unsafe

18:18

[freenode] <nemo> yeah. I know. that's what #rust said too

18:19

[freenode] <nemo> not gonna do it

18:19

[freenode] <nemo> we don't need it

18:19

[freenode] <nemo> checkout line, oh, 1527 of that c file

18:21

[quakenet] <heinrich5991> nemo: what's wrong with using atomics? :)

18:21

[freenode] <nemo> heinrich5991: https://hg.hedgewars.org/hedgewars/file/09edbd8a311d/hedgewars/uGearsHandlersMess.pas#l1714 the equivalent line of pascal

18:21

[freenode] <nemo> heinrich5991: it's not necessary is all

18:21

[quakenet] <heinrich5991> with Ordering::Relaxed, it should compile down to the same machine code, although you won't invoke UB by modifying stuff while reading it

18:21

[freenode] <nemo> uh

18:21

[quakenet] <heinrich5991> (on x86)

18:22

[freenode] <nemo> how on earth could atomic access ever compile down to same as non-atomic?

18:22

[freenode] <nemo> but, if it is zero cost, I'm open to it

18:22

[freenode] <nemo> it seriously doesn't matter tho

18:22

[freenode] <nemo> the AI only needs a fuzzy view of the Land

18:27

[quakenet] <heinrich5991> nemo: https://www.cl.cam.ac.uk/~pes20/cpp/cpp0xmappings.html

18:27

[quakenet] <heinrich5991> Load Relaxed: MOV

18:27

[quakenet] <heinrich5991> Store Relaxed: MOV

18:29

[freenode] <nemo> and how does that guarantee atomicity?

18:29

[freenode] <nemo> isn't really clear to me at all

18:29

[freenode] <nemo> but again, if it compiles down to 'sact same code as an unsafe read, w/ no extra costs imposed at all, then fine.

18:30

[quakenet] <heinrich5991> the guarantees you get from relaxed atomics are the ones that intel provides you anyway

18:30

[quakenet] <heinrich5991> so no extra cost for you

18:30

[freenode] <nemo> https://stackoverflow.com/questions/31214983/does-loadstdmemory-order-relaxed-cost-same-as-reading-unatomic-variable seems so yeh

Does .load(std::memory_order_relaxed) cost same as reading unatomic variable?

I have 64 bits which I need to read extremely quickly before an event and then after the event perform a compare-and-exchange. I was thinking I could load(std::memory_order_relaxed) before the eve...

18:30

[freenode] <nemo> aight. fair 'nuff

18:30

[quakenet] <heinrich5991> atomicity at this level means just "you can only observe the value before it was written and afterwards"

18:30

[freenode] <nemo> and. sorry. "intel guarantee" is funny

18:30

[freenode] <nemo> can't help but giggle these days

18:30

[quakenet] <heinrich5991> which is how programmers usually model memory anyway :P

18:30

[quakenet] <heinrich5991> ^^

18:31

[freenode] <nemo> after meltdown and meltdownprime

18:31

[freenode] <nemo> from response, seems to be architecture dependent

18:31

[quakenet] <heinrich5991> yes

18:31

[freenode] <nemo> I wonder what the situation is on arm

18:32

[freenode] <nemo> it would matter a lot more there

18:32

[quakenet] <heinrich5991> would you handle complete garbage values ok in the AI?

18:32

[quakenet] <heinrich5991> like not these that were there before or are there afterwards

18:32

[freenode] <nemo> On strongly-ordered systems (x86, SPARC TSO, IBM mainframe), release-acquire ordering is automatic for the majority of operations. No additional CPU instructions are issued for this synchronization mode, only certain compiler optimizations are affected (e.g. the compiler is prohibited from moving non-atomic stores past the atomic store-release or perform non-atomic loads earlier than the atomic load-acquire).

18:32

[freenode] <nemo> On weakly-ordered systems (ARM, Itanium, PowerPC), special CPU load or memory fence instructions have to be used.

18:33

[freenode] <nemo> yeah. complete garbage values are totally fine

18:33

[freenode] <nemo> so long as they aren't, oh, more than 1% of the values

18:33

[freenode] <nemo> which they probably won't be

18:33

[quakenet] <heinrich5991> …

18:33

[freenode] <nemo> basically the only thing that matters is "is there collision or not"

18:33

[freenode] <nemo> most of the time

18:34

[freenode] <nemo> that is, is the value non-zero

18:34

[freenode] <nemo> and. it has to actually be checking that cell

18:34

[freenode] <nemo> out of 8+ million others

18:35

[freenode] <nemo> of which the vast majority never change after initial load of the map

18:35

[freenode] <nemo> and, even if it gets wrong value, odds are it wouldn't impact the AI's strategy in any way

18:36

[freenode] <nemo> and even if it impacted it, most of time it wouldn't impact it in a way that actually has any significance

18:36

[freenode] <nemo> anyway. would have to see what the cost on ARM is. can always strip it

Embedded image

18:37

not even a response

[freenode] <nemo> lol

18:38

[freenode] <nemo> so this jxsl13 person - did they actually contribute anything useful?

18:38

[freenode] <nemo> 'cause that would be best way to get trusted with commits

18:39

[freenode] <nemo> or were they just there to complain

18:39

[quakenet] <heinrich5991> not in the code base, but in the forums

thats not the point of my screenshot

18:39

read https://www.teeworlds.com/forum/viewtopic.php?id=12115

[quakenet] <heinrich5991> nemo: https://www.cl.cam.ac.uk/~pes20/cpp/cpp0xmappings.html

18:39

[quakenet] <heinrich5991> Load Relaxed: ldr

18:39

[quakenet] <heinrich5991> Store Relaxed: str

18:39

[quakenet] <heinrich5991> still 0-cost

nemo, it recommends other people ot be added as "admin" in the repo, such as heinrich

18:39

to*

18:39

but oy clossed it without even saying a thing xD

18:40

the most funny thing, is that oy clossed the issue in 10 mins but didn't respond to a pr from 2013 yet

[freenode] <nemo> "

18:40

[freenode] <nemo> Note: the only way to get atomic 64-bit memory accesses on ARM without LPAE (the Large Physical Address Extensions) is to use ldrex/strex with a loop (ldrd and strd instructions are not guaranteed to appear atomic). Processors with LPAE have single-copy-atomic LDRD and STRD instructions to normal memory. "

18:41

[freenode] <nemo> that is proooobably not gonna be a problem

18:41

[quakenet] <heinrich5991> where do you get this from

18:41

[freenode] <nemo> your link

18:41

[freenode] <nemo> Ryozuki_: but, heinrich5991 already is a core dev no?

18:41

[quakenet] <heinrich5991> nemo: not for teeworlds

18:42

[freenode] <nemo> ah...

18:42

[quakenet] <heinrich5991> oy is currently the only one, and refuses to extend it to other ppl

thats what i mean :)

18:42

and thats how oy killed teeworlds dev

[freenode] <nemo> oh well. ddnet seems to be doing fine

18:42

[freenode] <nemo> unc0rr was much more open w/ his baby

ye ddnet responds to prs even in the same day

[freenode] <nemo> he just complained a lot about all the weird stuff we were doing

18:43

[quakenet] <heinrich5991> ^^

and this is oy response "When I started playing Teeworlds, there were usually just 2 servers with people playing - that was fine. If it goes back to that - still fine. It was never the aim to get a lot of players, it's a niche product, even though it attracted lots of players during time. People start and stop playing games, that's life. Games die and people move on.s."

18:43

with such thinking, dev is for sure dead xD

[quakenet] <heinrich5991> wait, nemo: do you need 64-bit atomics?

18:43

[quakenet] <heinrich5991> the thing you quoted

18:44

[freenode] <nemo> that's why I said "probably not"

18:44

[freenode] <nemo> so, we technically support some truly insane map sizes

18:45

[quakenet] <heinrich5991> more than 4M in each direction?

"Oy wrote: It's not turning into an anti-Teeworlds like ddnt or become mod-friendly etc."

[freenode] <nemo> heinrich5991: for the LandPixels we use 4 bytes for pixel (true colour). 2 bytes for the collision

18:45

[freenode] <nemo> heinrich5991: we do support maps as large as libpng will load

18:46

[quakenet] <heinrich5991> well

18:46

[quakenet] <heinrich5991> you need to access the 2 bytes for collision

18:46

[quakenet] <heinrich5991> that's a 16-bit atomic load

18:46

[freenode] <nemo> I'm not saying it's a sane thing to do

18:46

[freenode] <nemo> oh. I misunderstood

18:46

[freenode] <nemo> I thought it was about addressable size

18:47

[freenode] <nemo> then yeah. not a problem. only 64 bit type is hwfloat and isn't used for mem access

18:49

[quakenet] <heinrich5991> "hwfloat"

18:49

[quakenet] <heinrich5991> intel hwfloat is 80 bit

18:50

[freenode] <nemo> hw = hedgewars

18:50

[quakenet] <heinrich5991> ah ;)

18:51

[freenode] <nemo> it's not actually float either

18:51

[quakenet] <heinrich5991> fixed point? <3

18:51

[freenode] <nemo> yep

18:51

[quakenet] <heinrich5991> cool

18:51

[quakenet] <heinrich5991> teeworlds deals with floats :(

18:51

[freenode] <nemo> basically unc0rr ran into the same inconsistency problems that the poor TA:Spring guys were running into

18:51

[freenode] <nemo> where they were endlessly playing with compiler variations and flags to avoid desyncs on different hardware/OSes

18:51

[quakenet] <heinrich5991> yes

18:52

[quakenet] <heinrich5991> we had that recently

18:52

[freenode] <nemo> I think they ended up resolving it w/ streflop - but prior to that they were super precise about gcc version

18:52

[freenode] <nemo> you guys aren't deterministic lockstep tho

18:52

[freenode] <nemo> is death for poor spring/hedgewars.

18:52

[quakenet] <heinrich5991> true

18:52

[freenode] <nemo> anyway. he fixed it w/ a fixed point number - he called it float since it was dropin

18:52

[freenode] <nemo> I think I have an old pdf of the Spring guys testing this

18:52

[quakenet] <heinrich5991> who are the "spring guys"?

18:53

[freenode] <nemo> http://m8y.org/spring/sync_matrix.pdf

18:53

[freenode] <nemo> https://springrts.com/

18:53

[freenode] <nemo> I got interested in it due to Total Annihilation - it was intended as a FOSS variant - Total Annihilation being an RTS far ahead of its time - in like 1996 had true 3d terrain and ballistics

18:54

[freenode] <nemo> later in 2007 Supreme Commander was written - it had an engine suprisingly similar in terms of features to what the Spring guys had come up w/ ðŸ˜ƒ

18:55

[freenode] <nemo> http://m8y.org/hw/fp.html this one might entertain you too

i got supreme commander and its dlc

[freenode] <nemo> I never ended up buying that one, was too much into Spring at the time

18:55

[freenode] <nemo> I did buy Planetary Annihilation tho

i have it too

18:56

is good

[freenode] <nemo> oh. cool. we should play sometime

tho i dont play lot rts rn

[freenode] <nemo> well. I don't have much time to play much of anything

18:56

[quakenet] <heinrich5991> we're now using -ffloat-store on some gcc versions

18:56

[freenode] <nemo> I did just finish Starbound - but only way I got away with that is my 5 year old daughter enjoys watching me play

18:56

[freenode] <nemo> she'd make up stories as to the various creatures I encountered

18:57

[freenode] <nemo> at one point she drew on a bunch of pieces of paper a doodle book of absolutely everything I did that evening in starbound

18:57

[quakenet] <heinrich5991> is your daughter playing sth? :)

18:57

[quakenet] <heinrich5991> nice :o

18:57

[freenode] <nemo> she plays... Twenty, DragonBox 5+

18:57

[freenode] <nemo> and some simple phone games

18:57

[freenode] <nemo> I highly recommend Twenty BTW if you want a fun simple game for your phone

18:58

[freenode] <nemo> is free for the basic game

18:58

[freenode] <nemo> I've gotten up to... 13x20 - my coworker is up to 17

18:58

[freenode] <nemo> he's got a 2 thumb technique tho

18:58

[freenode] <nemo> I just can't manage that

18:58

[freenode] <nemo> my daughter has gotten up to like 17 or 18 or so

18:58

[freenode] <nemo> (not 20)

yo why is teeworlds folder hidden under linux?

[freenode] <nemo> ??

19:01

[quakenet] <heinrich5991> because it has a leading dot in the name

19:01

[freenode] <nemo> heh

is there any reason? to use .teeworlds instead of teeworlds

19:01

ikr

[freenode] <nemo> that's standard for config folders under linux

19:01

[quakenet] <heinrich5991> everything with a leading dot is hidden by convention

19:01

[freenode] <nemo> to avoid clutter

19:01

[freenode] <nemo> the only improvement would be to do ~/.local/config/teeworlds

hm

[quakenet] <heinrich5991> if you want, you could support the xdg stuff and put it into .config/

19:01

[freenode] <nemo> which, we haven't bothered doing yet

19:01

[quakenet] <heinrich5991> we neither

19:01

[freenode] <nemo> it's a hassle

hmkay it does make sense

[freenode] <nemo> gotta handle migration/legacy

ty c:

[freenode] <nemo> since none of us care enough, hasn't happened

why isnt it hidden on macOS then?

[quakenet] <heinrich5991> nemo: fun stuff, same in rust: https://github.com/rust-lang/rfcs/pull/1615

Let Cargo put data into platform-specific directories by tbu- · Pull Request #1615 · rust-lang/rfcs

(rendered)

so macOS convention shows config folders?

I think we put it into a special path on macOS, like on windows

19:02

let me chck

application support

[freenode] <nemo> I assume you guys use Library/Application Support/Teeworlds

classical thing

19:03

ya

[quakenet] <heinrich5991> yes

19:03

[freenode] <nemo> yeah. MacOS determines how that is displayed

allright

[freenode] <nemo> if you open home folder in macos - how is Library displayed?

19:03

[quakenet] <heinrich5991> supporting different OSs is fun, right? :)

yes!

19:03

c:

19:04

what do you mean?

[freenode] <nemo> under Windows we use My Documents/Hedgewars

19:04

[freenode] <nemo> SHGetFolderPathA(0, CSIDL_PERSONAL, NULL, 0, path)

19:04

[quakenet] <heinrich5991> we use %APPDATA%\Teeworlds

19:04

[freenode] <nemo> we used to do that

19:04

[freenode] <nemo> Smaxx (our former windows guy) said it wasn't intuitive for windows users

19:04

[freenode] <nemo> they wouldn't be able to easily find it

19:04

[quakenet] <heinrich5991> mh...

i like appdata c:

[quakenet] <heinrich5991> we ship a config_directory.bat

19:04

[quakenet] <heinrich5991> that opens the folder :D

its best

19:05

nemo uhm idk Library seems to be hidden

[freenode] <nemo> https://hg.hedgewars.org/hedgewars/rev/560e2766c445

Image attachment

19:06

its hidden in gui

19:06

but not in ls

[freenode] <nemo> yeah, it's just arbitrary conventions. I dunno. Doesn't suprise me apple hides it tho

19:06

[freenode] <nemo> APPDATA is hidden in windows no?

19:06

[quakenet] <heinrich5991> yes

y

[freenode] <nemo> unless you set Show Hidden Files and Folders

19:06

[quakenet] <heinrich5991> yes, but even then

so its somehow hidden in macOS too then

[freenode] <nemo> "* Hedgewars will now use a sub directory of "My Documents" like other games instead of "%userprofile%/.hedgewars" under Windows"

but without .teeworlds

[quakenet] <heinrich5991> you still have to navigate through C: Users your name

just in a weird hidden folder

[freenode] <nemo> yeah, no idea why Smaxx thought My Documents was better

19:07

[freenode] <nemo> but he was windows guy, so didn't argue

19:07

[quakenet] <heinrich5991> backward compatibility strategy:

19:07

[quakenet] <heinrich5991> * Due to those changes old files/settings/teams will be ignored

19:07

[quakenet] <heinrich5991> great :P

19:08

[freenode] <nemo> we did add interface buttons to open the ~/.hedgewars/Data folder (and other locations) so that people who were unfamiliar w/ linux commandline could actually find it

19:08

[freenode] <nemo> (or OSX users

)

only the entry folder has to be hidden am i right?

19:08

all sub directorys can be listed normal i guess

[freenode] <nemo> yeah

cool

[freenode] <nemo> heinrich5991: heh. yeah, yay breaking changes.

19:08

[freenode] <nemo> heinrich5991: I think we still have unicode problems under windows w/ team names

19:09

[quakenet] <heinrich5991> rust solves these <3

19:09

[freenode] <nemo> \o/

19:09

[freenode] <nemo> hm

19:09

[freenode] <nemo> well

19:09

[quakenet] <heinrich5991> they did unicode right as well as with interfacing with windows

19:09

[freenode] <nemo> if we change engine to rust that doesn't fix frontend

19:09

[freenode] <nemo> http://utf8everywhere.org/

frontend sux xd

19:09

uninstall it

[freenode] <nemo> qml rewrite of frontend is on the todo

19:10

[freenode] <nemo> but. dunno if that'll fix the team naming

19:10

[freenode] <nemo> we probably just need to be more aggressive on escaping

19:11

[quakenet] <heinrich5991> escaping?

19:11

[freenode] <nemo> special characters in the filenames

19:11

[quakenet] <heinrich5991> ah yea

19:11

[freenode] <nemo> what happens right now is that if someone uses, um, non-BMP unicode, it loads fine in linux

19:11

[freenode] <nemo> but gets mangled in windows

19:11

[quakenet] <heinrich5991> do you save the client from arbitrary file overwrites? :)

19:12

[freenode] <nemo> er?

19:12

[freenode] <nemo> not sure what you mean

19:12

[quakenet] <heinrich5991> do you have a "map download to file" function or similar?

yo nemo how did u even get into this irc? xd

19:13

have u ever played teewoods?

[freenode] <nemo> I played w/ heinrich5991 !

ah

[freenode] <nemo> and. I joined because we were looking for help from you guys w/ our SDL2 migration

great!

[freenode] <nemo> which you helped w/

19:13

[freenode] <nemo> I played teeworlds a long long time ago

19:13

[freenode] <nemo> didnt' care for it

19:13

[freenode] <nemo> but ddnet is better

.-.

[freenode] <nemo> I like the puzzle aspects

19:14

[freenode] <nemo> even if I kinda sucked at it

19:14

[freenode] <nemo> some things were a bit weird graphically. the ice effect esp

19:14

[quakenet] <heinrich5991> yea, we don't even have graphics/prediction for our main gameplay element :<

not needed

19:17

its all perfect c:

19:18

getchar(); btw sux bcs it only reacts on enter .-.

[freenode] <nemo> heh

19:20

[freenode] <nemo> ChillerDragon: soooo I hope you seriously aren't bothered by that?

19:20

[freenode] <nemo> 'cause toggling buffered/unbuffered in a crossplatform way is kind of a PITA

19:20

[quakenet] <heinrich5991> that's why he was trying system

hehe

19:21

ye sys rocks

[freenode] <nemo> oh. was that the underlying reason. well. you aren't going to solve cross-platform that way

.-.

[quakenet] <heinrich5991> I saw you were already using ncurses?

19:21

[quakenet] <heinrich5991> it probably has a function for htat

yerp

19:21

hm

19:21

lemme check

[freenode] <nemo> oh. yeah. if he has an ncurses dependency already...

19:22

[freenode] <nemo> https://linux.die.net/man/3/cbreak ?

AHAHAHHAHA

19:22

searched for a ncusrses solution

19:22

2nd answer is using system

[freenode] <nemo> …

https://stackoverflow.com/questions/1449324/how-to-simulate-press-any-key-to-continue#1449355

How to simulate "Press any key to continue?"

I am trying to write a C++ program in which when user enter any character from keyboard and it should move to next line of code. Here is my code: char c; cin>>c; cout<<"Something"&l...

[freenode] <nemo> oh. eh

19:23

[freenode] <nemo> that's not curses

i searched for ncurses -.-

[freenode] <nemo> http://tldp.org/HOWTO/NCURSES-Programming-HOWTO/init.html yeah. ncurses does same as curses

19:25

[freenode] <nemo> I would hope that cbreak() is implemented on whatever windows version of the lib is there

19:25

[freenode] <nemo> last time I built an ncurses anything for windows was w/ a cygwin dependency tho

well i gues im going to revive my #ifdef agian to seperate oses

[quakenet] <heinrich5991> curses seems to have a getch function

ya

19:27

i used it

19:27

but it didnt stopped somehting yet

[freenode] <nemo> er. why not use this cbreak() thing?

19:27

[freenode] <nemo> seems exactly what you want?

19:28

[freenode] <nemo> and nocbreak to turn buffering back on

hum

19:29

im too noob to use it

19:29

it turns buffering off

19:29

so it doesnt pause yet am i right?

19:30

so by combinding it with getchar() it should work fine?

[freenode] <nemo> yeah

doesnt work .-.

[freenode] <nemo> whaaat

19:30

[freenode] <nemo> pastebin?

ye it still lets me write hughe textts in menu

19:31

pastbin of my cod?

19:31

e

yes

https://github.com/ChillerDragon/chidraqul6/master/src/client/base.cpp

19:31

gg

19:31

xd

19:31

sec

[freenode] <nemo> heh

19:32

[freenode] <nemo> not to mention a stripped down testcase would be appreciated

forhot the blob

19:32

https://github.com/ChillerDragon/chidraqul6/blob/master/src/client/base.cpp

ChillerDragon/chidraqul6

chidraqul6 the console game in the 6th generation designed to be cross platform

19:32

line 12

19:33

well ait

19:34

i didnt push yet xd

19:34

haha fok me

[freenode] <nemo> I gotta think a pastebin of a testcase would be a lot faster

idk what do you mean by that

[quakenet] <heinrich5991> not only for nemo


    cbreak();
    getchar();
    nocbreak();

[freenode] <nemo> just would like a mini program similar to your "system" example above

these lines i use

[quakenet] <heinrich5991> it might be helpful for yourself if you try it out on your own in a small testcase

19:36

[freenode] <nemo> ok... and in a small testcase, does it not work as expected?

no it works as getchar() without any changes

19:36

i can press all keys and entering a message

19:36

it waits for the enter

[freenode] <nemo> you just can't be bothered to send us your reduced testcase

19:37

[freenode] <nemo> btw you are testing in a reduced testcase right?

nope

19:37

well i can try to crate a testcase for u c:

[freenode] <nemo> \o/

19:38

[quakenet] <heinrich5991> not only for nemo, it's also good for yourself to create a reduced test case

19:38

[quakenet] <heinrich5991> faster testing :)

19:38

[quakenet] <heinrich5991> the knowledge that the proble doesn't lie elsewhere

in this case...

19:43

https://pastebin.com/raw/GB97N2ea

[freenode] <nemo> I like that you added a license

ikr xd

#include <curses.h>
int main() {
    initscr();
    noecho();
    cbreak();
    getch();
    nocbreak();
    echo();
    return 0;
}

19:47

gcc a.c -lncurses (edited)

19:47

./a.out

19:47

works for me

[freenode] <nemo> yeah. he was lacking the initscr();

19:48

[freenode] <nemo> the noecho isn't necessary ofc

19:48

[freenode] <nemo> hm. how was an ncurses program working without that?

19:48

[freenode] <nemo> I thought that was like fundamental

19:48

[quakenet] <heinrich5991> ChillerDragon: do you use any other functions from ncurses?

hmm idk i have some initscr all over the place in the src

19:49

i think so

[quakenet] <heinrich5991> it's the noecho you're missing I believ

19:49

[quakenet] <heinrich5991> if it's "there is still character output when I type"

it works great in the testcase btw

19:50

but it fucks my game lul

19:50

like totally

19:50

you just createt a whole new game

19:51

this is escalating way too hard im going to stick to the getchar() for now lul

19:51

but ty for ur support guiis c:

Exported 550 message(s)