Teeworlds - bridge

Teeworlds

IRC / bridge

One-way IRC channel bridge. If you want to be able to send messages to IRC, contact @Dune or @heinrich5991. https://www.teeworlds.com/?page=docs&wiki=rules/irc_rules

Between 2019-08-31 00:00:00Z and 2019-09-01 00:00:00Z

Hey, I am still struggeling with my rainbow. When about 7 players or more are online on the server and everyone has rainbow, the clients get an out of buffer disconnect. Is it possible to fix this in a good way?

06:32

Or can I increase the Buffer?

06:33

(rainbow means sending a skinchange msg to everyone every tick)

no you can't increase the buffer

06:36

but 7 seems odd, maybe some problem in ur code?

if(m_pCharacter && m_pCharacter->m_Rainbow) {		if (!GameServer()->m_apPlayers[SnappingClient] || (!Server()->ClientIngame(SnappingClient) && !GameServer()->m_apPlayers[SnappingClient]->IsDummy()) || Server()->GetClientVersion(SnappingClient) < CGameContext::MIN_SKINCHANGE_CLIENTVERSION)			return; 		CNetMsg_Sv_SkinChange Msg;		Msg.m_ClientID = m_ClientID;	
for (int p = 0; p < NUM_SKINPARTS; p++)		{
Msg.m_apSkinPartNames[p] = m_TeeInfos.m_aaSkinPartNames[p];			Msg.m_aUseCustomColors[p] = 1;			Msg.m_aSkinPartColors[p] = RainbowColor;
}		Server()->SendPackMsg(&Msg, MSGFLAG_VITAL|MSGFLAG_NORECORD, SnappingClient);
}

06:43

This is in the snap, before I had it in Tick, and it also went out of buffer

06:43

so there is no difference

maybe it's called multiple times?

06:49

the buff should be 16k or 32k

How can it be called multiple times

06:50

Espeically when I put it in the Tick

you could check what kinda packets you get on ur client

what would that help

too see why u run out of buffer

but the server disconnects me

doubt it's the rainbow, unless it's called multiple times, client should handle 7 players without problem

But there is nothing else that gets sent every tick

07:08

I will check it in a few days

07:09

I have an idea but that cant be it actually

07:10

@oy, why is sndasyncloading in config vars, and not variables?

07:10

Where is the difference

just checked the recv buffer should be 64k, even when a packet uses max payload of 1400bytes you can't fill the buffer with 7 packets

07:11

there isn't much difference

07:11

the other snd config were in engine/shared so i put it there to have them all together

Ah

07:11

I will try that when I get home

not sure if it's needed to keep track if a user set the default value manually

13:25

https://github.com/teeworlds/teeworlds/issues/2201

Do not register defaults in settings.cfg · Issue #2201 · teeworl...

This lets us change the defaults on future new releases.

how about a new flag in C*VariableData, m_Save. If a value is set on console it'll be set. When writing the config you check m_Save whether to save or not

14:38

Now the issue would be values that are old, e.g. saved by past versions of the client and there are only 2 ways to handle that

14:39

Either this is just useful for all the new config variables, or you allow users to (with a console command or a button) remove default options from their config file

I think not

15:22

(Oy)

Or just leave it as is, it's been broken forever anyway

15:31

Causes a lot of problems when you are releasing a couple versions a month, it's mostly fine for slow releases where the defaults get tested long enough to be unlikely to chang

15:32

s/$/e/

it would be some kinda hassle: if i change a value to a default value, cause i want to use the default value and it should be updated in case

16:15

if we change the default value, it's probably for a reason

16:15

dunno how many want to stick to the old default value

I think it's unnecessary

16:39

06d445a ought to be enough

a default flag for vars, lgtm

17:09

restore default value with some specific command, say unset <varname>

different feature, but seems useful

(or more explicit name, restore_default/reset/to_default etc)

maybe the right course of action would be to simply remove that config option

heinrich5991: when you change defaults?

if we see no reason to not load the sounds asynchronously

17:30

nah, in this case

oh, you are talking about that

there's still a problem with your solution btw, config variables are changed without involvement of the console by the UI, currently, there's no simple way to detect that

oh, the ui doesn't really go through the console, yeah that'd be quite a rewrite

on an old single core machine the loading of the sounds was faster with the not-threaded option

17:48

don't think it hurts to keep the option

lines of code for this particular feature seem to be low (<30), so I guess it really doesn't hurt

Indeed :)

any info on that one https://github.com/teeworlds/teeworlds/issues/2168 ?

master server seems to be vulnerable to whatever · Issue #2168 ·...

it seems that the 0.7 master servers can be influenced in such a way, that individual servers can be blacklisted, seemingly banned and or delisted, I guess, especially masterserver 1 and masterserv...

17:58

otherwise it could be closed

no info known to me

17:59

I guess it can be closed (as "not enough info" or so)

yeah

18:01

heinrich5991: could you look at https://github.com/teeworlds/teeworlds/pull/2194 when you have time?

automatic latest builds/trunk builds for testing new features by A...

Added deployment to appveyor, circleci and travicci to enable automatic trunk builds. os: win64 win32 unix osx (mac) It adds 4 releases (for each os), which are automatically updated and overwrit...

yea, will do

great

about #2168 @jxsl13 might have some info?

re https://github.com/teeworlds/teeworlds/commit/32ee7c774d6e1d20323bcc8df4a500d7e1f81507: cleaning trailing whitespace for r might be better

18:18

with this change, things like sv_name Test server don't work anymore

yeah, you have to use ""

18:21

probably too much a hassle

can be closed

18:44

it seemed to me that thta server is vulnerable to social engineering at most.

Yeah don't break backwards compatibility with previous cfg syntax for that

Taking a look at github actions seems to maybe make things a lot easier. As every operating system can be covered with that.

How are the release tar.gzs crafted for linux?

they can be created using cmake

19:03

use the target "package" or "package_default"

awesome

19:03

ty

(same for the other operating systems btw, package_default probably works better)

package_dmg said 0.7.3 tho

19:04

ah awit#

19:04

i might have messed up my version nvm

19:05

oh GAME_VERSIOn is 0.7.3 and GAME_RELEASE_VERSION is 0.7.3.1?

19:05

https://github.com/teeworlds/teeworlds/blob/master/src/game/version.h

teeworlds/teeworlds

A retro multiplayer shooter. Contribute to teeworlds/teeworlds development by creating an account on GitHub.

Yeah?

19:11

I think the first one is for the server?

wait so server and client have different version numbers?

The version the client sends does I think

quickly grepping through the source code, I think GAME_RELEASE_VERSION is used for the UI and GAME_VERSION for the network

so there can be no network patches?

19:13

well pacth versions that actually get send

19:13

beacuse it seems to be chopped

Not for fixups. Minor versions get sent just fine

Exported 94 message(s)