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 2022-10-08 00:00:00Z and 2022-10-09 00:00:00Z

i just wanna get this done with https://cdn.discordapp.com/attachments/295908390956433410/1027702787041144902/Run_View.demo.mp4 is this run cycle ready to ship or is there anything else i need to attend to

01:11

i -made the body bounce less -fixed the foot going back too much -made the heel touch the ground going to front

sad ending

Looks great to me

murpi

sad ending

daft punk

01:29

anyways ill wait until @Jupstar ✪ reviews this once again

[ddnet-maps:master] 1 new commit

d8b3fdd R 1Hook - ddnet-maps

Looks smooth, but need to test myself xd

What you need to attend to is removing it completely /s

well time to install gentoo

08:45

just gotta remember to not delete my uefi partition this time

dont forget to make benchmarks xdd

i have windows to benchmark against i guess

i want to know the diff between march native, o3, lto and default packaging OS wide

08:46

lol

08:46

windows

how to benchmark

i dunno

xd

pick 2-3 favorite games

08:47

and apps

08:47

and check if they run better

im 2 lazy i just wanna install gentoo rn -.-

08:47

i dont even have steam on nixos

ok

08:47

then at least check how long compiling takes

xd

08:48

ok

that's what makes it most unattractive to me

gentoo?

compiling ROCM from an arch packages took like 1.5h

08:48

yes

lel

self compiling simply slow

What cpu

3800x

hmm

08:49

I have 5600x

yeah gl

08:49

i dunno what packages u use

08:50

but i'd use smth like kde which already takes 1h to compile xD

08:50

firefox slow too

08:50

u should at least estimate 2h i guess

08:51

ohh

08:51

gcc was also insanly slow i think

08:51

llvm is insanly huge too

08:51

and required for mesa

08:51

so maybe better estimate 3h

Kek

The compile time thing is only a problem during the initial setup imo

10:53

I can just compile single updates in the background

10:54

Gcc, llvm, rust are pretty slow though as they usually bootstrap all the way

10:56

Boost is another one that's notoriously slow

Well i hope its doing its thing

11:11

I left it doing the new @ world thing (edited)

11:11

And went out xd

https://en.cppreference.com/w/cpp/utility/expected

cpp/utility/expected

12:42

When can we uwe c++23 ?

12:42

I need this BASED

BASED

@heinrich5991 I end up missing the generated files way too obvious. Any chance we can somehow trigger generation at the beginning of compilation at all times?

13:39

s/ovious/often

13:39

wtf

13:39

somtimes my brain just breaks

Can we get 10x faster snapfinditem pls? @Robyt3 u interested? github.com/teeworlds/teeworlds/commit/e6b8518b49c353963af35efad8465d9066fb819e

14:24

ou can discord do links without https:// ?

no

[ddnet/ddnet] Pull request opened: #5918 Constify snap more (inspired by upstream)

Incorporate const added in those upstream commits: https://github.com/teeworlds/teeworlds/commit/d86d576217eeb843c822568f6b8f97ac852ac0d7 https://github.com/teeworlds/teeworlds/commit/e6b8518b49c353963af35efad8465d9066fb819e

Checklist

[x] Tested the change ingame
[ ] Provided screenshots if it is a visual change
[ ] Tested in combination with possibly related configuration options
[ ] Written a unit test (especially base/) or added coverage to integration test
[ ] Con...

https://github.com/teeworlds/teeworlds/commit/e6b8518b49c353963af35efad8465d9066fb819e here u go then

Optimize CClient::SnapFindItem (10x debug, 3x-4x release) · teeworl...

14:41

go jupstar gotta increase those fps

wont give much

10 TIMES MORE FPS

14:41

pog

14:42

yea watever i managed to crack my pr done without it :D

yeah give me 10 times more fps chillerdragon

i sadly got none :c

100k fps is 1/10th million fps

but according to lordsk you can make findsnap 10 times faster in debug mode

14:44

he has good marketing in his commit message

14:44

i approve

yeah then do it^^

i dont get a single line of dis code

14:45

skill issue

maybe it helps with enough traffic going on

10times faster sounds like it could have impact on a snap spammed server

we will not find out until you do it

i wont

14:46

well not before i got 0.7 working

14:46

im back to 1 step fordward and 2 back :D

Hm, I think we already did a similar optimization

15:05

Atleast I remember reading about it, maybe when trml introduced switch states to the snap?

Most of the changes from https://github.com/teeworlds/teeworlds/pull/2129 are undone again by fixing an issue with demos in https://github.com/teeworlds/teeworlds/pull/2216

ou

@Learath2 how to hide emerge build logs and only show x of y lines etc?

Hm, I never tried it but there is a setting for it. It enables automatically if you enable parallel builds

16:48

I only have my phone with me so can't check, but you should be able to google around for it

16:49

Maybe do man emerge and look for quiet

--quiet-build y

16:50

Works

how far is it

emerge --jobs --load-average 12

16:52

For best right

16:53

Seems like yes

16:53

Im just finishing building kde

lol, why did u even build it

Wanna try kde

but anyway, it builds since more than 3h already isnt it?

I wasnt home

so you stopped it?

And think 20-30 mins passed

ah ok

Whats a good load average to use full resources

16:57

I have 12 threads

16:57

Make.conf has j12

Ryozuki

Whats a good load average to use full resources

Not so easy to figure out tbf, hyperthreading, ram limitations, how well the individual builds parallelize all make it really annoying to figure out

32gb ram

-pipe

Hm, so you have enough ram. If you don’t mind emerge using your whole cpu then load average 12.0 makes sense

17:01

But it will make your computer pretty much unusable while things build

Well i need to build this to hade a desktop

17:01

xd

Oh, for the initial build that’s fine yeah

Jupstar ✪

-pipe

-pipe is nice too yeah, you have the ram for it

Its there already

17:04

Time to buy a threadripper

I don’t think I ever built kde or gnome. Lxde took me about half an hour the time I tried it

17:06

I’m honestly just too used to tiling wms now, don’t think I’ll ever really go back

kde build is bit annoying bcs it doesnt build multiple (independent projects) at once, so u often waste time for the cmake .. command

17:08

would be nicer if they more CPU bound

17:10

but maybe gentoo fixed that ofc, i just used kdesrc-build

Learath2

Hm, so you have enough ram. If you don’t mind emerge using your whole cpu then load average 12.0 makes sense

you can easily add 25% of load

17:11

with 12 cores, -j15 should be the fastest (edited)

Well he said 12 threads. And also tarhetting a load average of 12 in that config will run multiple makes with -j12 each

17:12

Targetting*

i already showed u can easily go much higher

17:13

without real effect

17:13

except maybe more ram usage

17:14

can u also build gentoo with ninja?

No, not all packages build with ninja

17:15

There is one more thing that should be integrated into emerge imo, it should be actively controlling the total amount of jobs instead of total amount of builds

17:15

Launching 3 make -j12's at the same time isn't exactly the sanest of things

in summary

18:14

i lost all day compiling, i got kde without any app

18:14

gg

18:14

tomorrow more

18:14

im gonna play now xD

18:14

maybe i should ditch kde

18:15

kde-apps are 300 things to emerge lol

yeah kde games

18:25

for kids

18:25

compile them

xd

ill remove kde tomorrow

peace in rest

ChillerDragon: you mean your build fails because the generated files aren't there? then some dependency information is apparently wrong

[ddnet/ddnet] Pull request opened: #5919 Tweak the running visuals

Sorry for putting this off for so long. Turns out animating with raw code is harder than you'd think. https://user-images.githubusercontent.com/95713843/194726134-40925aea-6ec4-40ae-9597-b80a28d1b6a2.mp4

Checklist

[x] Tested the change ingame
[x] Provided screenshots if it is a visual change
[ ] Tested in combination with possibly related configuration options
[ ] Written a unit test (especially base/) or added coverage to integration test
[ ] Considered possible...

@Voxel nice, looks much more like running now gj

thanks ^^

now the non running even looks worse at the edge of where the animation changes xD

20:37

the walking generally looks weird now xD

i didnt do the walking tho, matricks did LOL

i know

i think if i change the walking id be twice as flamed

imo if its well done its fine, u can get used to improvements

20:39

i think the running really looks like running now

thats good!

it indeed looks great!

I'm going out for supper, but in the meantime, I want to think about what to do next.

20:45

I want there to be a way where if you, say, scan a folder for images, it will also check its directories if possible. If you have a folder inside your skins folder that isn't accounted for. And the editor's Readd function is kind of useless if you also have folders in mapres. (edited)

Yea and that is a reoccurring pain. And can stay undiscovered during development until the build folder is wiped. I find it quite tricky to then fix those issues months later when cloning the repo on a new machine. (@heinrich5991)

heinrich5991#9827

ChillerDragon: you mean your build fails because the generated files aren't there? then some dependency information is apparently wrong

I've never seen this before

20:51

is your setup different from mine?

20:51

what OS do you use?

Only happend to me in the last few years since I switched to Linux

hm, I only ever compile on linux

20:52

you're talking about ddnet, right?

20:52

what distribution do you use?

But across repos. Seen it in ddnet++ and even in F-DDrace once.

20:53

It’s a issue in the code base

20:53

Not on the os

20:54

If you are curious I can try getting my broken state from today to not work again

20:55

Im not sure if I mix different issues but it’s always the generated code

20:59

I got it in my vanilla protocol debug branch today. I probably messed up some c++ basics and includes but the end result was a missing include of generated protocol. I ended up doing random restructuring until it worked. But during that time I kinda wished that the code would be generated and linked as early as possible to avoid such issues but not sure how bad that would be

[ddnet/ddnet] Issue opened: #5920 Option to disable running/sitting animations

I read heinrich5991's comments on animations PR and why you shouldn't add more config options, but i think this is too big update to just left it as it is. I already posted in #bugs some negative comments by some good players, maybe you received other complaints too, so please consider adding options to disable them.

21:07

[ddnet/ddnet] Pull request opened: #5921 Try http serverbrowser for longer

There are reports from China of timeouts to master1.ddnet.org We don't have a mirror in China unfortunately

Checklist

[ ] Tested the change ingame
[ ] Provided screenshots if it is a visual change
[ ] Tested in combination with possibly related configuration options
[ ] Written a unit test (especially base/) or added coverage to integration test
[ ] Considered possible null pointers and out of bounds array indexing
[ ] Changed no physics that affect existi...

@murpi when you have time could you accept https://github.com/ddnet/ddnet-discordbot/pull/34 and apply the change? heartw

heartw

[ddnet/ddnet-discordbot] Pull request closed: #34 Fix twmap-edit binary call

22:12

[ddnet-discordbot:master] 2 new commits

7fda04f Fix twmap-edit binary call - Patiga 5a01798 Merge pull request #34 from Patiga/patch-1 - def-

I'll try

22:12

Thanks @Patiga

[ddnet/ddnet] Pull request opened: #5922 Update spanish.txt

Checklist

[ ] Tested the change ingame
[ ] Provided screenshots if it is a visual change
[ ] Tested in combination with possibly related configuration options
[ ] Written a unit test (especially base/) or added coverage to integration test
[ ] Considered possible null pointers and out of bounds array indexing
[ ] Changed no physics that affect existing maps
[ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-ad...

[ddnet:master] 2 new commits

b83a1a5 Update spanish.txt - n0Ketchp ca4dd58 Merge #5922 - bors[bot]

22:58

[ddnet/ddnet] Pull request closed: #5922 Update spanish.txt

[ddnet:master] 2 new commits

fc09212 Tweak the running visuals - VoxelDoesCode b7da208 Merge #5919 - bors[bot]

23:20

[ddnet/ddnet] Pull request closed: #5919 Tweak the running visuals

Exported 194 message(s)