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 2021-01-11 00:00:00Z and 2021-01-12 00:00:00Z

Is Zwelf around on irc?

00:29

I think I've seen him speaking from the bridge once or twice :P

is online on freenode

00:31

along with me and twbridge

I just noticed we basically reimplement std::future for our sql stuff. Was wondering if he looked into it at all

and std::future actually is C++11

Yep

00:32

For our engine jobs too btw

what does it do?

I for one didn't realize it

00:32

that I was reimplementing something C++ already has

I didn't know about it at all until I listened to a talk yesterday

00:33

Promises are things that will happen in the future

00:34

You can wait on them, they evaluate either to a result or error out

00:34

Which is what we do with our sql result thing

how can I poll this thing?

It has the benefit of resolving properly in all cases, including exceptions and thread exit

oh ok ty

is it guaranteed to just poll if I pass a time of 0 to wait_for?

@heinrich5991 I have no idea, never used them. The guy was talking about asynchronous features of boost and just mentioned c++11 has futures aswell

if we can't poll them, they're not useful for our use case

I'd guess wait for with 0 would return immediately. I don't see why it wouldn't

00:39

I will check the standard tomorrow

00:43

Can it really ever guarantee no blocking at all though? The scheduler can just drop us if it wants, no?

it could in some way guarantee that we're not dropping to the kernel to check if the result is there

00:44

I guess that's hard to phrase in standard legalese

00:44

(our solution currently does that)

Ah, maybe some lockfree programming magic?

00:44

Oh do we use atomics?

yes

00:45

an atomic bool, to check if the result is there

I like the api of a future btw. Maybe we could adopt a similar idea for the sql stuff. I don't like us having to allocate two objects separately

good night

suggestion: add search in demos page in client

suggestion: add voicechat proximity

[ddnet-maps:master] 1 new commit

dc24dc3 A Juma - ddnet-maps

@Learath2 @heinrich5991 what about the actor model, is it any useful here? https://en.wikipedia.org/wiki/Actor_model

The actor model in computer science is a mathematical model of concurrent computation that treats actor as the universal primitive of concurrent computation. In response to a message it receives, an actor can: make local decisions, create more actors, send more messages, and determine how to respond to the next message received. Actors may modif...

08:24

https://actor-framework.org/

CAF - C++ Actor Framework

08:24

i found about this thanks to actix.rs

I only sort of like the idea. It's a nice design but I find it overkill for most systems

08:30

@Ryozuki what problem are you thinking of solving with it?

@Learath2 i thought about putting the db behind that

08:32

here is an exampel of the pattern implemented in a chat app, in rust tho https://github.com/actix/actix/tree/master/examples/chat

Actor framework for Rust. Contribute to actix/actix development by creating an account on GitHub.

Ah, I see. I think it's a rather massive solution for the simplest of concurrency problems

xd

08:33

i mean

08:33

if u have it already implemented by someone

08:33

it doesnt rly matter if its "massive"

We don't reall need any of the great properties of the actors

but yeah i guess

Ryozuki

it doesnt rly matter if its "massive"

Well these libraries are usually generic so you pay in compile time cost, it's also one more lib to link and it adds another pattern most people here aren't familiar with so it introduces some thinking cost

08:35

If we needed massive scalability and two way arbitrary messaging, it would work great

yeah

08:36

CAF allows you to transparently connect actors running on different machines and OSes via the network.

08:36

this is cool tho

08:36

intercommunication between servers

Yeah great stuff especially if you like microservice architectures

[ddnet/ddnet] Pull request opened: #3493 Keep showing last checkpoint after kill

Some maps kill you quickly and you can't see how good your last run was, for example "Buckle Up".

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 if it works standalone, system.c especially
[ ] Considered possible null pointers and out of bounds array indexing
[ ] Changed no physics that affect existing maps
[ ] Tested the cha...

[ddnet:master] 2 new commits

4520ea3 Reset practice after team saved and got killed - def- c8b4ce9 Merge #3465 - bors[bot]

13:19

[ddnet/ddnet] Pull request closed: #3465 Reset practice after team saved and got killed

[ddnet/ddnet] Pull request opened: #3494 Remove unused onChangeTeamState

This function was never used and just disrupts the reading flow a bit.

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 if it works standalone, system.c especially
[ ] Considered possible null pointers and out of bounds array indexing
[ ] Changed no physics that affect existing maps
[ ] Tested the change with [ASan+UBSan or valgri...

@Learath I didn't know about std::future in C++11 yet

hello,anybody?

[ddnet:master] 2 new commits

87d422d Remove unused onChangeTeamState - Zwelf 460f219 Merge #3494 - bors[bot]

13:48

[ddnet/ddnet] Pull request closed: #3494 Remove unused onChangeTeamState

hi yiyi

can anyone tell me, which engine does ddnetwork developped by?

DDNet is based on Teeworlds game, we use SDL2 and pure OpenGL for graphics

yiyi: oh, and the reason you don't see us in irc channel is because most of us are on Discord, which is bridged to this IRC channel

aha, i see, thank you!

@Lady Saavik or someone speaking polish, can you please check https://github.com/ddnet/ddnet/pull/3433 ?

Update Polish translation by def- · Pull Request #3433 · ddnet/ddnet

Anonymously contributed, will need someone with Polish language to check Checklist Tested the change ingame Provided screenshots if it is a visual change Tested in combination with possibly re...

[ddnet-web:master] 1 new commit

d7a313b Another css-halloween stats fix - def-

[ddnet/ddnet] Pull request opened: #3495 Send logout packet after using timeout protection (fixes #3490)

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 if it works standalone, system.c especially
[ ] 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-addresssanitizer--u...

[ddnet/ddnet] Pull request opened: #3496 Freeview mouse moves faster when zoomed out (fixes #3459)

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 if it works standalone, system.c especially
[ ] 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-addresssanitizer--u...

What is the file naming convention here? We have "allwordstogether" convention for the most of files, and snake_case convention for some hundred files here and there (engine/shared, game/client/components, etc). There are also DDRace files in game/server/gamemodes (which I'm actually going to rename).

[ddnet/ddnet] Pull request opened: #3497 No network clip in client prediction

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 if it works standalone, system.c especially
[ ] 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-addresssanitizer--u...

17:28

[ddnet/ddnet] Pull request opened: #3498 Make netclipping of entities and events respect shown distance (fixes #3420)

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 if it works standalone, system.c especially
[ ] 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-addresssanitizer--u...

@Kaffeine feels like "allwordstogether" except_when_it_gets_too_long_to_stay_readable

[ddnet/ddnet] Issue closed: #3453 queue system when server is full

17:31

[ddnet/ddnet] Issue closed: #3428 Open News tab automatically

17:32

[ddnet/ddnet] Issue closed: #3379 client suddenly takes 100% cpu and starts lagging

[ddnet/ddnet] Issue reopened: #3428 Open News tab automatically

suggestions

18:20

also we need a blocker server for NA, every time i go to SA shithole blmap i get violated intimately by spikes/stutters there is always a number of ddnet CAN maps that have 0 players on at all times, could we convert some into blocker servers?

18:20

thoughts?

@cummy DDNet USA - BLOCKER exists

[ddnet/ddnet] Issue opened: #3499 Race Points please

As suggested by... everyone on discord: 3 6 12 18 32 If you don't have an option to make reward points amount whatever you like - it's just a great time to implement this, cuz it's not that hard and really needed for balance. (I'm planning to make a voting for maps difficulty in the future. Could be useful to set exact points amount instead of only stars)

32 for 5* is crazy af lol

19:41

should be 3 6 12 15 18

19:41

or 3 6 9 12 15

Cristian

32 for 5* is crazy af lol

no 15 points is not much for 5* there are some rly hard 5* race

19:44

maybe not 32 but at least more than 25

mm i think 25 for 5* is the limit

19:46

more than that would be crazy

19:48

and rerate race server would be a nice idea cuz players will play it so much

19:48

for pointa

still not much

20:09

imagine playing antibuguse and gettins 25 points

20:09

it doesnt seems like much

20:09

while 25 points would be fair in other 5* race maps

20:12

maybe we could get some kind of time based points, so that it would be about average time too and not only about difficulty

Does anyone brave enough to check a commit? https://github.com/ddnet/ddnet/commit/af843600e73cf9b2aff28b78d66a075e9261f745 The commit introduces a Reset() method and what is the ugliest is that the Reset() is called twice — one time from the CGameControllerDDRace() constructor which constructs a CGameTeams member, the constructor of which in its turn calls the Reset(). And the second time on the next line: ((CGameControllerDDRace *)m_pController)->m_Teams.Reset(); I see no reasons to keep this. (I'm OK with Reset() method but not OK with the ugly second call) (edited)

[ddnet/ddnet] Issue closed: #3459 Freeview mouse should move faster when zoomed out

20:21

[ddnet:master] 2 new commits

9033968 Freeview mouse moves faster when zoomed out (fixes #3459) - def- 3bc3a80 Merge #3496 - bors[bot]

20:21

[ddnet/ddnet] Pull request closed: #3496 Freeview mouse moves faster when zoomed out (fixes #3459)

@Kaffeine if that truly does nothing, there's no problem in removing it

20:28

just make sure it's actually doing nothing

20:30

refactoring to avoid the cast is also fine, obviously ^^

[ddnet/ddnet] Pull request closed: #3497 No network clip in client prediction

20:30

[ddnet/ddnet] Pull request closed: #3493 Keep showing last checkpoint after kill

20:30

[ddnet:master] 3 new commits

95e4418 Keep showing last checkpoint after kill - def- f9aae5d No network clip in client prediction - def- d0cdf23 Merge #3493 #3497 - bors[bot]

heinrich5991

refactoring to avoid the cast is also fine, obviously ^^

I'm going to surprise you.

heinrich5991

@Kaffeine if that truly does nothing, there's no problem in removing it

Nothing in this code is doing obviously nothing

21:15

There should be some shitty case

[ddnet/ddnet] Pull request opened: #3500 Add Discord rich presence

Continued the work of @edg-l. Supersedes #3472.

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 if it works standalone, system.c especially
[ ] 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://gi...

Deleted User

okay thanks! will try this

update: the spike lags are still appearing in my game and gave up on finding new solutions. im gonna wait for next update now

[ddnet/ddnet] Issue closed: #3420 Distance when zooming out not used for effects

22:31

[ddnet:master] 2 new commits

cbda5b7 Make netclipping of entities and events respect shown distance (fixes #3420) - def- 5d45f06 Merge #3498 - bors[bot]

22:31

[ddnet/ddnet] Pull request closed: #3498 Make netclipping of entities and events respect shown distance (fixes #3420)

Deleted User

update: the spike lags are still appearing in my game and gave up on finding new solutions. im gonna wait for next update now

You can try nightly version already to see if that helps: https://ddnet.tw/news/nightly-build-revived/

whats the difference in nightly build? (i still have the spikes) i felt the game more tough

do u get lagspikes on older clients

23:05

nightly build just has the new features still in development

yes i do

then it might not be a problem with the client and rather a problem with your specific computer or your internet

[ddnet/ddnet] Pull request opened: #3501 RFC: Refactor (from the upstream and own custom)

Port some cleanup and refactoring from the upstream and add some custom refactoring. Changes overview:

Backported angle() and direction() functions (upstream commit https://github.com/ddnet/ddnet/commit/7735557aa5c6fb81daa724cce3caa109d1c76c11)
Backported vector math cleanup (remove some pointless templates, refactor other functions; upstream commit https://github.com/ddnet/ddnet/commit/bdaabc10bcb04e28cf650bb489b87fe8a0e03dee)
ddracechat.h, used by both client and server, moved...

[ddnet/ddnet] Pull request opened: #3502 Small (mainly backported) refactor

Port some cleanup and refactoring from the upstream and add some custom refactoring. Changes overview:

Backported angle() and direction() functions (upstream commit https://github.com/ddnet/ddnet/commit/7735557aa5c6fb81daa724cce3caa109d1c76c11)
Backported vector math cleanup (remove some pointless templates, refactor other functions; upstream commit https://github.com/ddnet/ddnet/commit/bdaabc10bcb04e28cf650bb489b87fe8a0e03dee)
ddracechat.h (used by both client and server) moved...

23:52

[ddnet/ddnet] Pull request closed: #3501 RFC: Refactor (from the upstream and own custom)

Exported 129 message(s)