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 2020-05-01 00:00:00Z and 2020-05-02 00:00:00Z

@jao I've temporarily given up on it

00:02

@jao I've temporarily given up on it

kk

Kk

[ddnet-libs:master] 1 new commit

573d8a9 Rebuild mac libs (MacOSX 10.13.4 SDK) - def-

@Learath2 thanks, now it built. Can you test please? https://ddnet.tw/downloads/DDNet-13.1-osx.dmg

[ddnet-web:master] 1 new commit

13f4f17 Version 13.1 - def-

07:58

[ddnet:master] 1 new commit

a723299 Update ddnet-libs (rebuild mac libs again) - def-

07:59

[ddnet:13.1] 2 new commits

c9dfcb2 Fix Mac OS X notifications linking - def- a723299 Update ddnet-libs (rebuild mac libs again) - def-

[ddnet/ddnet7] Issue opened: #95 too weak connection: out of buffer

On Vanilla client when voting for a different server type, noticed on far away servers like Chile.

[ddnet:master] 1 new commit

df278aa Make the github workflow badge link to appropriate page - def-

08:32

[ddnet/ddnet] Pull request opened: #2158 WIP: Try to fix macos action

@deen works great

08:39

@noby you should give it a test too, you are on older macOS

how

https://ddnet.tw/downloads/DDNet-13.1-osx.dmg

test what

whether it works

i can open it

open the package contents, run the actual executable to see the termina

08:42

l

08:42

check if curl is giving an error

[2020-05-01 03:42:34][http]: libcurl version 7.69.1 (compiled = 7.55.0-DEV)

08:42

thats the only result for curl

great

08:43

ship it

08:44

I think the github vms are built wrong tbh, you don't need command line tools anymore, xcode includes that nowadays

08:48

@deen you should also try passing VERBOSE=1 to make so we can see the commands executed, it'd be nice to see the -isysroot and -isystem values

can we just fix up the github vm in a preparatory step

08:51

?

08:51

@Learath2 finally

Hm, we could try

The github action logs are the worst thing I've ever seen

08:53

Half of the time they don't load all the text for me, sometimes nothing, and ctrl-f is totally broken

yeah, they are really meh

How difficult can it be to just show a text box?

Well I doubt they are all stupid, I'm sure there is more to it

I guess they tried to make it lazy-load because some have dozens of MBs of logs

(I don't see how they fail to do it either though, it should just be simple polling for the logs (or html5 sse))

I recently wrote Javascript at work instead of C++/Python btw

08:57

We needed some fancy userscript to integrate Jira and Bugzilla

@deen can we run commands before the run? It'd be nice if we could see xcode-select -p and make VERBOSE=1

08:57

@deen javascript isn't too bad as long as you know the pitfalls, the pitfalls are like dozens of meters deep

08:58

I mostly use typescript nowadays

yeah, so far it's survivable, but I'm basically just moving data from A to B, so nothing too hard

09:00

xcode-select -p gave me /Applications/Xcode_11.4.1.app/Contents/Developer

okay that's pointing to the correct place

09:02

hm, I wonder how it ends up including the command line tools version of the headers, I thought cmake used xcrun

@deen I think there's a raw log somewhere in github actions

09:02

that one is usable

-isysroot /Library/Developer/CommandLineTools/SDKs/MacOSX.sdk -g -isysroot /Applications/Xcode_11.4.1.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.15.sdk

that doesn't sound right

ah, raw logs are available once the run is finished

09:08

I see isysroot twice

09:08

oh, you saw it too

try not touching the sysroot

09:09

xcode-select thinks we are using xcodes version of the sdk as apparent from the xcode-select -p

09:15

homebrew is a joke of a package manager, I wonder how no one came up with a better version yet

09:19

I see, -isystem /Library/Developer/CommandLineTools/SDKs/MacOSX10.15.sdk/usr/include is what's breaking the build

09:20

I wish github would just let people ssh into their containers like circle used to do

I mean you can hack that together

@Learath2 hints for how to get rid of it?

I'm not sure where it's coming from who decides the -isystem?

oh, i think I am?

Yep you are

09:25

Don't use anything from commandlinetools, just let clang decide those, lets see what it decides to pull in

09:25

@heinrich5991 would be an interesting project, I wonder if the containers can access the outside world

they can access the internet, so yes

Try running DDNet servers on the containers

09:26

and you can start new ones on demand

09:26

no need for servers anymore, check mate

heh, if only they didn't time out

just need some way to migrate servers

https://github.com/mxschmitt/action-tmate/blob/master/README.md

mxschmitt/action-tmate

Debug your GitHub Actions via SSH by using tmate to get access to the runner system itself. - mxschmitt/action-tmate

09:27

Apparently someone already hacked the ssh thing together

i think I'm out of ideas

Did you just let it run with verbose so we can check whats happening in the first place?

yes, last push

-I/Library/Developer/CommandLineTools/SDKs/MacOSX10.15.sdk/usr/include

09:34

now we just need to find where this is coming from

09:34

@heinrich5991 do you know of a way to trace why cmake is adding this flag?

09:34

or whether it is the one adding it to begin with?

09:35

if it's any help, it's only present when compiling the client, so it's one of the clients dependencies thinking that's the correct include dir

09:36

tbh, we could just nuke CommandLineTools while preparing, but I'd rather find the issue

09:37

@deen can you add CMakeCache.txt as an artifact so we can download it?

09:38

or just cat it to the console before the make

@Learath2 no idea where this is coming from

CMakeCache should tell us why

09:56

I'll take a look when I'm in front of a computer I guess

heh

10:56

PC_CURL_INCLUDEDIR:INTERNAL=/Library/Developer/CommandLineTools/SDKs/MacOSX10.15.sdk/usr/include

10:57

@heinrich5991 I found the problem, can you gimme a hand with the cmake part please?

@fokkonaut oh btw i deleted whatsapp just that u know

why we dont have blocks that are half hookable and half unhookable

11:09

or even 1/4 of block

11:09

we could do hard maps with that feature

no one wants such pixel perfect shit

okay, pkg-config in homebrew installs the curl pkg config file, which includes the wrong sdk root

yep it's homebrews fault, I guess we try nuking the commandline tools directory

that fixes it

@Smetanolub sounds like you just want the blocks to be smaller

@deen I think you can merge #2518 now

yay

12:25

I like the way you fixed it @Learath2

this is what happens when you half-ass bolt on a package manager onto an OS hostile to anything except itself

12:26

https://github.com/Homebrew/brew/blob/master/Library/Homebrew/os/mac/pkgconfig/10.15/libcurl.pc

The missing package manager for macOS (or Linux) - Homebrew/brew

pacman for life

12:27

actually i just use yay

I use aurutils on archlinux

yay is the best

Brew is supposed to be able to handle the situation if (XCode || CLT), it can't if the sdkdir is hardcoded...

macos sounds like a pain in the ass

@ChillerDragon I invited the user you needed

https://twitter.com/mxcl/status/608682016205344768?lang=en

Max Howell (@mxcl)

Google: 90% of our engineers use the software you wrote (Homebrew), but you can’t invert a binary tree on a whiteboard so fuck off.

Retweets

7456

Likes

12045

12:33

No wonder google didn't hire this guy

12:36

I'm shit with algorithms, even I could invert a binary tree, but this guy is on another level, he pretends the packages aren't in any structure, they are all laid out flat, then he annotates the packages with dependencies to create a fake graph

@ChillerDragon how do you want me to contact you now

https://leetcode.com/problems/invert-binary-tree/

Level up your coding skills and quickly land a job. This is the best place to expand your knowledge and get prepared for your next interview.

12:41

haha

12:43

greenthing

Image attachment

swap nodes, head down, keep going until nowhere to go down

Image attachment

exactly

can i work at google now?

12:46

poggers

well you didn't do it on a whiteboard

12:47

I do agree that algorithms are not a great measure of a programmer, but still failing this one, I wouldn't hire him either

yea

Oh, we should do that golf thing again sometime

12:48

it was fun

ye but it would be better if it had better challenges

yeah, it did get repetitive and meh after a while

i would love to do versus challenges with https://www.codewars.com/

Codewars: Train your coding skills

Codewars is where developers achieve code mastery through challenge. Train on kata in the dojo and reach your highest potential.

12:49

best website for this kind of thigns

12:49

to register u need to solve a challenge xd

12:50

(a rly ez one tho)

I bet I'd fail it, I'm great at failing easy things

@heinrich5991 bors is still choking on something

13:42

ah, interactions are not allowed to touch files inside .github

13:42

integrations*

13:43

hm, but it was choking on stuff earlier too

[ddnet/ddnet] Pull request closed: #2158 Fix macos github action

13:43

[ddnet:master] 2 new commits

6f040e4 Try to fix macos action - def- 52d9c18 Merge pull request #2158 from def-/pr-macos-action - Learath2

13:45

[ddnet:master] 1 new commit

6d433fc Revert "Allow macOS CI to fail for merging" - Learath2

Finally the green tick is back

brownbear

2

14:38

@jao are you around?

@Learath2 awesome!

@Learath2 somewhat

Is it hard to run a copy of the ddnet bot locally?

@Learath2 of bors?

no of @DDNet

you need to setup postgres rest is easy

15:30

i think it only works on linux btw

@jao how do I choose the db it uses?

you create a .pgpass file at your home dir

15:35

https://wiki.postgresql.org/wiki/Pgpass

nice, it doesn't care about .pgpass

15:46

passing the dsn manually to create_pool does seem to work

why doesnt it work

troll

1

no idea didn't even read .pgpass

15:50

what are you using the database for btw?

menus branch is the up to date one btw

15:51

for $profile $points etc

15:52

you can see data/schema.psql

hm am I supposed to be the one to create the tables btw?

y

15:54

cant u run the entire file in psql or smth

yeah, btw you have a syntax error there, a . instead of a ,

oh no

do u need the cronjobs too

16:10

?

nope, just doing a learning exercise

nice

@heinrich5991 do you happen to know why regex \b would limit the matches to 1?

figured it out, new question, are regex lookaheads bad for performance?

16:44

I feel they aren't that bad because you don't need backtracking

@jao what do you use when coding in python?

vsc

huh, the autocompletions are shit for me

17:12

does it complete something like pat = re.compile(r".*"); pat.?

17:15

type inference for pat should be absolutely trivial

ye autocompletion works

17:34

you can use pycharm

17:34

it’s better for that kinda stuff

@Learath2 install python extension

I'm not retarded, ofc I have the python extension

what are you even making

@Learath2

18:56

how to fix lags

🤦‍♂️

3

1

19:00

redstripe

ctrl shift q

@heinrich5991 https://rust-analyzer.github.io/ seems to do a way better job than RLS

rust-analyzer: building a better Rust IDE

19:24

they just released their first release on 20 apr

19:24

https://rust-analyzer.github.io/blog/2020/04/20/first-release.html

19:25

i agree

Image attachment

19:26

https://github.com/rust-lang/rfcs/pull/2912

Transition to rust-analyzer as our official LSP (Language Server Pr...

The RFC proposes a plan to adopt rust-analyzer as Rust's official LSP (Language Server Protocol) implementation. The transition to rust-analyzer will take place in a few stages: Feedback -...

[ddnet/ddnet] Pull request opened: #2159 Make the dummy always press fire when cl_dummy_hammer is active

Hello, i have a very weird bug ingame (GER Server). I played the map Topsy Turvy with my dummy yesterday and saved at a part which couldnt be done with dummy. Now i loaded the savegame with a partner who did it and saved again. And here comes the bug. When i load the save i can play normally except for that one of the tees cant use weapon. Not like that he hit a hammer off tile, he just cant use the weapon, pressing mouse1 has no effect. Th

21:26

Before loading and after saving his weapons work again

the first message was cut off after "no effect. Th"

I tried to change the server but the same bug happened again, also any other tee always the same bug for exactly the same tee

21:27

The other tee normally can

which weapon btw? or multiple don't work?

i have no extra weapons, only hammer and pistol. none of them, works

21:32

with antiping it "tries" activity but with antiping off just nothing happens

so you can select the weapons but not shoot. sounds like m_Ammo is 0 or maybe the regeneration counter is set to a wrong time

ah

21:33

it's the fix we recently pushed

21:33

from fokkonaut

21:33

it probably didn't consider saves

21:33

damn it, I missed that

21:33

deen: ^

@deen ^

https://github.com/ddnet/ddnet/pull/2086 this one? I remember you commenting "fingers crossed" :D

Remove setting ammo for freeze to 0 by fokkonaut · Pull Request #20...

what was the intended use of this fix^^ And will my save work again when its refixed or is it dead now?

heinrich5991: since you already reviewed, any idea how to fix?

Konsti: will work again after this is fixed

thanks for help so fast

my guess without looking at the source: disregard ammo while loading

21:36

and set it to -1 if we have the weapon and 0 if we don't

for all weapons?

yes

21:37

but let me look at the saving code

why don't we care about it anymore?

21:37

it's easy to do in save code, i can make a pr

needs to be the load code

yeah, i figured

because we don't track freeze by "no ammo" anymore

21:38

(which was a hack that I'm happy about getting rid of)

[ddnet/ddnet] Pull request opened: #2160 Don't load m_Ammo to fix loads (following #2086)

As reported by Konsti, thanks to heinrich5991 for remembering #2086

Is there any commit that adds the ddnet ingame updater?

21:40

I would like to apply it for F-Client

nope

21:40

we had to change lots of stuff about it

21:40

but the code is relatively modular

21:41

see updater.cpp/h and all uses of its functions I guess

21:41

and be prepared for lots of fiddling and failing updates :D

thanks, I will try it :)

and on the server side you'll need something like https://update4.ddnet.tw/update.json

uhm, what? :D

to specify what to update

when will the commit be merged?

when I finish reviewing it

21:44

and then it needs to be deployed to the servers

ok thanks

@deen what does all of it mean? and do i need to do it manually?

Konsti: yes, manually

konsti? :D

whether to update server executable, client exec, libs, files. and you'll need to provide them from your web server as well

deen: I think you need to set it to m_aWeapons[i].m_Got ? -1 : 0

@heinrich5991 it can be -1, it doesnt matter

ah, so we don't ignore it totally

if you dont have the weapon, it doesnt matter whether you have 0 or -1 ammo

@deen or remove that variable completely, that would work as well

in m_aWeapons you mean?

.m_Ammo*

in CCharacter

I wouldnt remove it tho

21:47

So mods have it easier to work with it

it's non-functional for us

true

we don't have code to decrement it

@deen, if i just want to update the executable i just set client = true, and if i need to update some other files i need to set them in download: ?

yes

and does the server need anything more?

only the update.json and the files in correct location, but you'll see the client error messages if they're not there

okay

21:49

how do you keep track of what changed?

21:49

do you go back commit by commit before a release?

by looking at https://github.com/ddnet/ddnet/compare/13.0.2...13.1

DDraceNetwork, a cooperative racing mod of Teeworlds - ddnet/ddnet

deen: but you do need to set it to -1 if you got the weapon, otherwise it won't fix Konsti's problem

if something in data or ddnet-libs changed -> need to put it on server and add it to update.json

21:50

heinrich5991: yeah, I noticed, so always -1 is the easiest fix I guess

@deen okay, can you give me an example of how the server structure has to look like? so where to put the json, where the other files?

deen: yes

Konsti: wanna give it a try?

21:52

huh, only one empty ger server

21:52

on ger2, none on ger

21:54

ok, tested with konsti, works

21:54

it's in the queu

21:54

e

thanks to u both

i'll rebuild servers once it's submitted

21:55

you're welcome Konsti

I'm really not pleased about the updater, I wrote it but apparently I wasn't really thinking while making it

heinrich5991: bors is a bit stupid about the queue isn't it? If the builds were green already and nothing was merged inbetween, then no need to rerun them

yes @deen I think there's an issue about it

anyone else having problems with borderless window? sometimes the border appears

[ddnet/ddnet] Pull request closed: #2160 Don't load m_Ammo to fix loads (following #2086)

21:59

[ddnet:master] 2 new commits

7ae8988 Don't load m_Ammo to fix loads (following #2086) - def- 2067a1a Merge #2160 - bors[bot]

usually when im tabbed out for a while or when pc is sleeping and i repoen

@deen https://github.com/bors-ng/bors-ng/issues/185 didn't have movement for three years though

Take advantage of pull request status information when available · ...

Consider the situation where: Pull requests are built the same way as staging/trying. The tests pass on a pull request. The pull request has master as an ancestor. There is nothing else in bors&amp...

@deen oh no, the updater requires http.h, which is not in 0.7

22:10

so I guess I need that too :/? Uhm... I think this wont happen if no one PRs it :D

@Learath2 i need your help/opinion... I am in a conflict with myself regarding 128 players. I got it to a point where it works pretty good, also with dummy and so on. I made it so that players with the same IP get the same faked IDs in the map. Vanilla clients 0.7 works fine of course, and I made an adjustment in my 0.7 F-Client for some things for the dummy (because everything is different, e.g. team changes, client enter, drop etc. everything is a netmsg now). So F-Client with dummy works flawless too.

23:23

DDNet clients work fine too, also with dummy. Expect one point: whisper. When whispering to the dummy that is not active at the moment, it will of course say the name of the player with the fake id of the current player's map. I could fix that by sending RECV whispers as server messages for DDNet clients.

23:24

But generally, I am not 100% sure whether I should keep the whole thing or go back to 64 players. I mean, it definitely is a nice thing to have, since for example today I had 80 players at a time. Usually tho, it is around 60-70

@Learath2 did you figure out why bors chokes on some PRs?

[ddnet:master] 2 new commits

bcf9e00 Make dummy hold fire with cl_dummy_hammer 1 - trml 65284f3 Merge #2159 - bors[bot]

23:26

[ddnet/ddnet] Pull request closed: #2159 Make the dummy always press fire when cl_dummy_hammer is active

@heinrich5991 well it was choking on that one because it was modifying a file in .github/workflows, apparently integrations are not allowed to make changes to that

ah yea: https://github.community/t5/GitHub-Actions/GitHub-action-Resource-not-accessible-by-integration/m-p/19950/highlight/true#M75

Re: GitHub action: Resource not accessible by integration

Not sure why they did not reply in this thread, but I did get a confirmation from GitHub support via mail: OAuth tokens and GitHub Apps are restricted from editing the main.workflow file. This means it isn't possible to merge a Pull Request via an Action if it updates the m...

@fokkonaut I'd stick to 64 players, the translation patches were never nice, I think 64 is more then enough

ty

@Learath2 well, BlmapChill is a big map tho, and 128 is working fine besides that

23:31

but i also think the algorithm is a bit laggy

this is all fixable but it'll take a lot of work, to get the algorithms all optimized

23:33

I don't get your whisper issue, the current players id map should match the dummies exactly

No, of course it doesnt on the serverside

23:34

if the dummy is in another area where other tees are

23:34

they have different id's on dummy/main

yeah but the first part is the same

its about WHO sent the message

oh, do we send a client id and not an actual name?

yes, thats how ddnet whisper works.

23:35

so as i said, i could fix it by sending a servermsg to all clients with the same IP as the receipiant

why all clients? you as the server always know the correct recipient

Yes, but if I send it to the dummy only, the main wont receive it

23:36

only whisper and team messages are received when they go to the udmmy

this all sounds very meh

yea. but okay, another thing. Lets imagine I remove 128 support. F-Client still has some things in the client, for example a complete set of m_aClients[] for thedummy, so that the dummy can have different infos for each client, that was of course used for 128 players, as the info is not in snaps anymore. But if I remove that, also team changes for specific tees to specific tees only wouldnt work in 0.7 dummy anymore. My mod does team changing to team blue to hide players from minigames.

23:38

i am completely overwhelmed right now xD

23:39

i hate 0.7 for moving stuff out of snaps

with 64 clients you should never be lacking info

yes, but if my dummy joins a minigame, it gets teamchanges to team blue for everybody thats not in the same minigame, when swapping to the main again, the others are of course still in team blue, which is false.

23:42

The second m_aCleints array fixes that

23:42

but i inititally did that for 128 players, so i think i would need to keep that too in the client, even if i remove 128

having different sets information for dummy and main sounds logical to me

yea

23:43

This is a point where I really notice how much better the Snap thing is

@heinrich5991 why? the dummy and the main client exist in the same game world

But for cases like the team changes it is neccessary

@Learath2 in every server implementation, currently

23:44

but if I write my own tw server, there's no guarantee I send the same IDs to every player

23:45

e.g. I could choose to send cid=0 for the client's own tee

Well I don't like these kinds of "many mods in one mod" mods to begin with tbh

23:46

@heinrich5991 I don't think this is a good idea

why not?

@Learath2 well, teeworlds adventure mod also does this, they have multiworld and also send faked ids

adding 50 layers of translation is just not a good idea

23:47

then I don't like the "teeworlds adventure mod" either

XD

@heinrich5991 why would you want inconsistent ids anyway, it would only give you more players in one server, or support some mod in mod situations

because I don't want to leak server state e.g.

dummy is only a mod too, it should work as a real new client i think

no harm to leaking the actual client ids

23:50

only more complexity with designing the translation layer that'll just add overhead

"runtime overhead"? meh (edited)

at the very best case I can imagine (a swap from your cid to 0 and put the guy at 0 in your position (at that point why bother you are still leaking state)) it's still a constant overhead of an extra branch instruction

you shouldn't worry about that extra instruction

23:52

you should worry about making a proper collision algorithm first

23:52

(that is not O(n^2). that costs more instructions)

At no gain, it's still an extra instruction

say you want to swap tees in 0.7

23:53

then you need your ID swapping feature anyway

why?

ah, you could copy over the tee data, fair enough

any useful translation table and you are looking at some very bad performance characteristics

why?

at the very least while generating the table

23:56

if you want to support more then 64 clients, you are going to have to sort the tees based on distance, that's O(logn) with a good algorithm

23:57

If you want just a 64 -> 64 mapping, it'll have a [64][64] memory cost

which is 4KB btw

besides, put aside the cost, we don't seem to care about that much in 2020, it doesn't bring anything to the table, I don't see a good use for it

but okay, I see that you don't need it, so it's unnecessary overhead

23:58

but I do think that you worry about the wrong costs if you're worried about one extra mispredicted branch per network packet

this is how these things add up, we have all become too lax, started relying on the processors raw power to take care of things, fancy branch predicting algorithms in cpus, speculative execution

Exported 367 message(s)