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 2020-09-24 00:00:00Z and 2020-09-25 00:00:00Z

why does ctrl+shift+d make u move

@louis you cant open chat?

no actually i cant

thanks to whoever shortened the search/exclude/server address fields in the server browser greenthing

greenthing

(edited)

@752947895 update viruses definition list

05:20

a.k.a. viruses database

05:21

and update the client too >.>

@murpi you're welcome

darkLove

3

[ddnet-web:master] 1 new commit

e395c3e Add Umbra to Vanilla mod - def-

[ddnet:master] 1 new commit

1c3244a Update ddnet-libs - def-

07:54

[ddnet/ddnet] Pull request opened: #2915 HTTP: Set final m_State after running completion function

Ensures that we don't delete a file that is being used by ingame map downloader already. As reported by Jupeyy in #2901

07:58

[ddnet/ddnet] Pull request opened: #2916 Make sure turning dyncam off in settings works

Even if players have cl_mouse_followfactor turned on Alternative fix for #2761 since players keep reporting confusion about this

07:59

[ddnet/ddnet] Pull request closed: #2871 Allow unmasked use of entities if wanted by the server (closes #2870)

[ddnet:master] 2 new commits

84cad65 Add new renderer(as streamed vertices) for editor - Jupeyy ceefdac Merge #2894 - bors[bot]

08:15

[ddnet/ddnet] Pull request closed: #2894 Add new renderer(as streamed vertices) for editor

[ddnet-web:master] 1 new commit

df705f6 Fix 15.0.5 release date - def-

I've heard about a skins DB cleanup recently. I understand that right now people can choose any skins from ddnet skins DB and use that on any server or mod. So for example if someone will upload a skin that's only appropriate for his mod (like a skin of a tank) that could lead to everyone will use that in DDRace and other mod and it will look inconsistent and lame. What I propose is adding to a skin a special flag (that could be a part of skin name for simplicity) that will ensure that this skin could only be used on server which want it. For example we can start those skins from underscore like _tank. I am sure it will be super easy to add validator to client info parser that will reject that skin and default it to something for servers that don't have allowed those type of skins allowed explicitly on their servers. Also that will be a simple patch for old mod servers owners that they could apply if interested in filtering. I wanted to add some special skins to be used in my mod that will be 100% inappropriate in other mods like: door, car, platform to use them as player interacatble tiles. But I think that right now those skins will not be accepted for the reason players will use that in ddrace and it will looks crappy. So I suggest this workaround. (edited)

08:42

Thanks @Soreu for telling me about a cleanup

08:43

sorry for editing here. Can't stand re-reading my lang errors

I'd suggest the prefix to be x_ like it is done with x_ninja & x_spec, so for INF mod that could be x_zombie1

08:45

and such flag for not autodownloading skins would be good idea in general, as discussed already at the beginning of cleanup plan

to be honest I am not much interested in filtering at all as my players are not allowed to choose their skins as it will break the game. Server assigns skins to them now based on a game class

08:46

*on a player class, e.g. Soldier, Zombie etc

08:47

Do we have an issue opened on that?

@Deleted User in the end your server decides which skins are allowed. You can modify the server to forbid some skin names

and such flag for not autodownloading skins would be good idea in genera

client will download a skin once he will see another character in a game. But only if server allows that character to set that skin at the first place

The client will respect what the server tells it

Oh, that is proposed and already is on roadmap, @deen ?

08:49

*was

why? we don't need it for ddnet-server

08:49

you can add it in the code of your own mod

I can make anything on my server right. But first the skin should appear in skin db.

08:50

The problem is that skin will not make it there if it could be abused on ddrace servers.

08:50

I have filters already (I don't allow client skins at all, but I want to be able to choose skins serverside from ddnet skin DB to assign to characters)

08:51

But ddrace servers don't have any filtering system AFAIA

I see, true

08:51

do we need to forbid soldier and zombie? do they look bad?

No, I am about adding newer skins to the DB that could be inappropriate, like skins of door, of visual FX of a platform, of a tank

08:52

I don't think you will like people running with a skin of a wall on ddrace servers

08:53

or even with a skin of a tank or a part of a tank even

Hm, easiest solution would be for server to tell client where to download skins, but we probably don't want that

08:53

I'm not sure how it works with the x_ skins

That will be awesome soluation. But I understand it will require MUCH more effort

x_ right now is excluded in client level, not server-side

08:55

The problem is if we do it server-side, many run old servers

08:56

so we'd end up with doors on all other mods and KoG etc

It will work like that:

server is loading and have something sv_allow_x_skins
if he sees a client wants to set skin to a skins which body starts with x_ it should not allow that setting a default skin instead of sv_allow_x_skins is not set to 1
players on server with svallow_x_skins enabled will be able to set skin and will download x\ skins when they will see a character with a skin set (is that a way skin downloading works now?)
players on a server without svallow_x_skins set will not be allowed to set x\ skins and thus will not see any characters with such skins set on the server and so they will not download it

and I don't want to maintain a list of skins allowed by mod in client

I think that after adding that to ddnet server code (client skin validation) the only additional thing you will do is to tell a person pushing a skin to db to rename it as it is too wild to be used freely.

08:58

In a case the skin is not appropriate ofc

08:58

For all server owners who don't change anything it will just be as if there are no x_ skins

08:58

which is like now, because "wild" skins are not accepted to a db I think

08:59

but server owners who want to use special skins will be able to do that

that doesn't work since the server is not the one forbidding x_ skins, it's the client

I am proposing the server to do that. I can give you lines of code for that. I've already prohibited all client skins in my server as only server should assign them. But I can make it conditional.

09:00

(in zombie mods server assigns skins)

But we can't modify all the old servers running

hmm...

09:00

you're right

09:00

I've got this. Let me think about it

I guess we could create a new network message that the server sends to client to inform it that the server will do skin filtering

09:01

if no such msg is sent, as on old servers, the client does the filtering. otherwise client allows any skin server sends

oh... we should add a server info flag that server will do skin validation

yeah, something like that

and disallow client to use x_ skins on a server without it

09:02

though it could be easily hacked

hacked?

as everything that's done clientside

09:02

abuser could remove clientside validation

09:02

and server will accept x_ skin

09:02

a server without serverside validation I mean, old one

yeah, but that's fine since the server sent the flag to indicate that it will do the validation on its own

09:03

I would say that's a case of "works as intended"

but we want to protect old servers without validation mechanism and a flag

old server would never send that flag

a hacked client will ignore absense of that flag on old servers

hacked client? I don't follow, sorry

09:04

How are we getting hacked?

I mean that a player could remove client side validation that checks for a server flag and will hardcode to allow x_ skins on any server

09:04

so absense of such a flag will mean nothing as old server will accept any skin

By modifying the client source code and recompiling?

yes

Of course everyone can do that

09:05

how is that hacking?

09:05

You can modify the source code right now to allow x_

Ok let's call it abusing

I don't see what that would abuse

09:05

you just see more skins on your own client, nothing else

09:05

no one else would see the skins

We want to protect old servers without serverside skin validation, right?

not protect, there is no attack

And we do that by adding to a client a check that will search for a server info flag

I think it will be much easier for me to just implement this than explain to you why it's no problem, sorry

Ok, if you don't care that a door could be used as a skin on old ddrace servers then it's perfect for me

No, that can not happen...

or if you think it will not be abused

it can't be abused with what I proposed

Hm...

the client only decided which skins itself can see, not which other clients see

09:07

Do you want to implement the feature or should I?

I expect you to do that for many reasons.

ok

I just still don't get why this is not a problem

09:09

1. We add x_ to inappropriate skins like a dunno penis 2. Old servers will accept any skins client will request 3. An old client which can set any skins (for example he downloaded a skins manually from database) sets x_penis 4. He logs into an old server and sets his skin 5. Newer clients will see his skin as they will download it and nothing will stop them from doing that

09:09

am I getting it wrong?

No, old client can not set any skin

09:10

Old clients ignore all skins with x_

Wdym? Is x_ filtering is already there?

new clients will ignore the x_ skin too if the old server doesn't say that it does any filtering

09:10

of course it's there already

09:11

Then it's much simpler

09:11

Thanks for explaining wasn't aware of that

but the x_ filtering is in the wrong place so that's why we need the update path

09:11

to move it from client to server

It should stay both on server and both on client I think

Didn't read it all, but client has set download link for the db

once you will release any client with no x_ filtering there will be no way back

as I explained a few times, client should only check x_ if server didn't say that it will do the check

could the link contain also exclude of specific skinpack?

@Deleted User what do you mean by no way back?

@deen ah, yes, Riht

sorry, but this discussion is leading nowhere

then we could just set skinpack to i.e. MOD

@Soreu that's not flexible enough. We want server to be able to set any skin

1

server yes, but not the client

09:13

and skins with prefix of x_ are not selectable by the player

what do you mean by no way back?

hm... I just liked how x_ magically appeared to be alreay in a client. If it wasn't there we would have a bigger problem But I realised it's not a problem as client side validation will reman there just will become conditional

x_zombie in skinpack MOD

@Soreu yes, they are selectable with f1, player_skin. and they are accepted by server. But in the end the c lient doesn't render them but replaces wiht default instead

so that is kind of a solution though

09:15

not best one ofc

~~can't wait to make a game of penises fighting tanks~~

lol

Can't wait to give a real transport to my mod's players

09:16

and destroyable walls not made of lasers too (edited)

09:16

also maybe to use zombie skins from that pack too

@heinrich5991 help, GameInfoFlags can't be extended any more

Thanks for listening me, explaining and finally accepting this @deen

09:17

by a chance will you add allow zoom flag to it?

we have ALLOW_ZOOM already

right now ifcity devs are enforcing clients to use k-client just to use zoom

they are doing that because they want people to use their client, not because DDNet client doesn't support it

as they don't want to rename their mod to contain "race" in their game type

what their reason is for enforcing their client I don't know

As far as I know vanilla client checks if Race is in the name of a game type

that is not required, they just have to set ALLOW_ZOOM bit in GameInfoFlags

and then decides if it should allow zooming

no, that is just the fallback for old mods that no one updates

09:19

new mods can use ALLOW_ZOOM, as I just said

hm... so they could just set that flag serverside? only ddnet supports this, right? not 0.6 and 0.7 vanilla clients?

09:19

(which is fine too)

yes

Nice

09:19

I will advice my players to tell ifcity devs to set that flags then

09:20

and thay they don't really need k-client

I'm sure sure they don't even know K-Client is often autobanned

I think they don't care

09:21

or are seeing this as a holy fight against bans

09:21

*bot bans

Why is my ddnet getting deleted randomly? (edited)

@Deleted User it's the gameinfo@netobj.ddnet.tw net object they need to implement

1

09:22

@Kingston probably your antivirus solution mistakenly auto-detecting it as malware, happend to many

09:22

please report to your AV vendor that that is a false positive, and add an exception

09:22

@Learath2 if I need more flags, can I just make m_Flags2 for GameInfoEx, right?

Maybe my WS detects it as virus or smth

09:23

I have no AV atm

@deen GameInfo is one of those versioned netobjects, so yes you add an m_Flags2 and you bump version/add a size check

great

actually don't even need to do anything with the size, we already disabled size checks for that one

@Kingston AV = Antivirus

yeah, i noticed

@deen what are you adding?

I will advice my players to tell ifcity devs to set that flags then

@Deleted User kclient has probs some backdoor or some shit

Support for server to allow client to use x_ skins

then mod-specific skins can be added to DDNet's skin DB

We can also just hardcode support for zooming out in city mods, because I don't like them enforcing their client for no good reason

won't hardcoded solution no longer work if they happen to add like more letters to the name? xd

09:28

to mod name*

sure, they could do that, but then it's obvious they're doing something shady to enforce their own client

09:29

and are not just too lazy to implement the proper net object

what is their own client?

k-client

Ah, already checked that

09:30

No bitcoin miner

haha, would be super easy to hide in a game btw

09:30

since many games run at high cpu load for no good reason anyway

yea

so no one would wonder why cpu is at 100% while gaming

You could probably even run it on the gpu with a game like tw

well maybe they use the gpu

Image attachment

09:31

monkalaugh

or maybe you just set no vsync and it's rendering 100 frames and only 1 gets shown

im at 60-80 fps

I'm sure we have a combination of settings that gets our cpu usage more sane, I couldn't figure it out

black desert online has rly demanding graphics

09:33

xD

@deen do you want me to open an issue to keep track of it?

@Deleted User nah, implementation is nearly done

twinbop

2

09:34

@Learath2 cl_refresh_rate 480

best dev team ever

is what I play with, ~20 CPU usage

By the way, do you think that allowing a server to set position for motd will be hard?

09:40

Will it require breaking changes or could be implemented easier way?

09:41

I want to make an onboarding (interactive help/tutorials) system for my mod. Movable MOTD window would be sufficient, no need to make any complext and separate netgui.

09:42

Broadcasts are not well suitable for this as they're not very readable and are at the center too.

09:42

Chat is even less visible.

[ddnet/ddnet] Pull request opened: #2917 Create ALLOW_X_SKINS game info flag

To allow server to set any x_ prefixed skin and client won't filter it out. As requested by Pure_luck for his mod to show players as tanks, walls, etc. Won't be enabled on DDNet-Servers, thus such skins can be added where a server modification wants to fine-control what skins are allowed and can enforce such skins.

1

1

1

I think that's how my players feel when they request something to find out that I am about to deploy changes already

09:47

@deen I like that this solution allow a server owner to decide how skins filtering system will be implemented.

[ddnet/ddnet] Pull request opened: #2918 Allow specifying -DGAME_RELEASE_VERSION

to be used for nightlies and release candidates to be able to tell them from regular releases more easily. For example: CPPFLAGS="-DGAME_RELEASE_VERSION=\"15.0.5-rc1\"" cmake .

ohhh

@Ravie @louis Just so you know: #2917

https://github.com/ddnet/ddnet/pull/2917

Create ALLOW_X_SKINS game info flag by def- · Pull Request #2917 · ...

To allow server to set any x_ prefixed skin and client won't filter it out. As requested by Pure_luck for his mod to show players as tanks, walls, etc. Won't be enabled on DDNet-Ser...

[ddnet-scripts:master] 1 new commit

6395fb3 build-and-deploy script - def-

I can also build the official releases/rcs/nightlies on my normal computer instead of my home server now. 8 min instead of 34 min is much more comfortable. rcs and nightlies will also have a proper version number showing like "15.0.5-rc1" or "nightly20200924"

3

poggers

[ddnet-scripts:master] 1 new commit

79d1e3d More build fixes - def-

@heinrich5991 I'm confused. I can't get CPPFLAGS to work in my Debian 6 chroot

14:07

works everywhere else

14:07

cmake 3.9.2

have you checked that whether the flags get to the actual compilation lines?

they don't get there

14:31

upgrading cmake also didn't work

14:31

I'm really confused why cmake is not looking at CPPFLAGS on Debian 6

why do you use debian 6?

have u tried CXXFLAGS

yes, same problem

15:18

I'll do it another way now

15:18

I use Debian 6 because building a portable executable for Linux is the ugliest thing in the universe

do we need such downward compability for linux

You need to choose a distribution with super-low libc and libstdc++ version to stay compatible with all the weird ones

15:19

I think centos used some 12 year old libc on a still-supported version last time I checked

ah, thats sad then xd

They made progress! Centos now only supports 9 year old glibc!

justatest

debian wiki only says that CPPFLAGS isnt supported https://wiki.debian.org/Hardening

15:21

Notes for packages using CMake

ah no, RedHat 5 from 2007 is still in support and uses glibc 2.5 from 2006. So we're at 14 years we need to support atm

I wonder how this stayed this horrible for this long

but why should we support redhat 5 users from 2007?

oh well, I use Debian 6 with glibc 2.11 for building, so RedHat 5 never worked

15:25

because some people run that on their servers and can't recompile on their own

i find that to be a rly weird and rare case

15:25

did someone complain in the past?

yes

Not really rare

well its just like windows xp

every time I updated glibc version

damn

We had someone here a month or so ago running on a container where you can't replace the kernel so you can't get above a certain glibc version

and why they dont update their server

and we also can't use musl, since we have c++ additionally

15:26

some weird-ass hosters living in 199X I guess

OpenVZ doesn't let you touch the kernel

15:26

OpenVZ is much cheaper for the hosters

oh right, and without new kernel you often can't update userspace

how to know what my vps uses?

Capitalism dictates that upsell shitty cheap stuff

15:27

@Ryozuki your kernel is usually a giveaway

what does that mean

15:27

i use hetzner and u can upload ur own distro iirc

Hetzner doesn't do OpenVZ

they use something kvm-based

15:28

i don't think i've seen openvz used for years actually

The latest OpenVZ kernel is something like 2.6

so much images

Image attachment

15:28

monkalaugh

You can install anything on a hetzner vps/dedi. I use debootstrap to install devuan from the rescue system

ah yes, ur battle against systemd

15:29

i wanna try nixos

15:29

some day

I think I also bootstrapped netbsd in the oddest way possible. Writing the live usb image to a hdd

15:30

Why would you say systemd on such a nice day? Now I need to go sleep it off

monkalaugh

15:31

@Learath2 is there a alternative to runit

There are many, how simple do you want it to be?

is there one made with rust?

15:31

monkalaugh

Probably not, but given PID1 is supposed to be dead simple, you can be the first one to write one

:O

https://www.fpcomplete.com/rust/pid1/ see this is how you not do it

Implementing pid1 with Rust and async/await

This post covers a simple Rust application to explore topics like async/await, unsafe Rust, signal handling, and error handling.

might be a interesting project to do

come on, what's wrong with systemd?

@Comrade dont start!

15:34

this wont end good

ok

justatest

i know

minit, runit, busybox init are all good examples of what one should be doing in pid1

Since 2015, the majority of Linux distributions have adopted systemd, having replaced other systems such as the UNIX System V and BSD init systems. systemd has faced mixed reception from Linux users, with arguments that systemd suffers from mission creep and bloat, as well as criticism over software (such as the GNOME desktop) adding dependencies on systemd—frustrating compatibility with other Unix-like operating systems

15:36

i guess this is the main complain

15:36

unix philosophy about doing 1 thing only and doing it well

15:37

@Learath2 what about OpenRC

Systemd is a monolith, a blackbox absorbing projects at a staggering rate. It's sole intention is to make it so impossibly hard to replace it that you just have to use it

15:37

The person that invented it even said so

15:38

I've used OpenRC for quite some time, it works splendidly

15:38

For a desktop you really do want something more comprehensive than minit/runit

Image attachment

15:39

xd

15:39

aw apparently nixos uses sytemd

15:39

i rly want to try a non systemd next

It's mind boggling how someone could think fragmentation is bad in this context

15:40

You want more options, not less

15:41

And for the absolute troglodytes that argue systemd helps with compatibility across distros. How you initialize your system and what you have running as PID1 has squat to do with compatibility

@Learath2 did u join this creed https://systemd-free.artixlinux.org/

Archlinux, systemd-free

A comprehensive, step-by-step guide for migrating from systemd to OpenRC in Arch Linux, with a proof-of-concept video.

15:41

?

15:41

xD

It's systemd absorbing other projects that actually help with this and create a vendor lock-in to create this illusion

15:42

I'm only using arch so people can't use the "you never used systemd, you can't have any opinion" on me anymore

15:42

I'll be going back to gentoo when I'm back in italy

why gentoo?

Learath2 and I should preorder some CPU fans for inevitable replacements

I love that I'm not locked into any software with gentoo. It also has a great way to patch software I'm installing with ebuilds

I've been test-compiling ddnet all-day because of this stupid debian6 issue

I also love that I can strip features that I don't need from software

justatest

15:44

@deen what cpu u have btw?

i7 6700k

How long does it take to compile with a 6700k?

can we time it?

15:44

i want to know how well my ryzen does

15:45

its 100€ cheaper xD

all of ddnet from scratch?

cmake -GNinja -DWEBSOCKETS=OFF -DMYSQL=ON -DCMAKE_BUILD_TYPE=Debug -DVIDEORECORDER=ON -DAUTOUPDATE=OFF -DANTIBOT=OFF -DUPNP=OFF -DPREFER_BUNDLED_LIBS=ON; time ninja
ninja  89.44s user 8.97s system 693% cpu 14.190 total

I've been looking into getting a ryzen

debug build though

How has it been @Ryozuki?

@Learath2 i love it

15:47

@deen i get errors compiling using bundled

Let's see how well a macbook with a 7700hq does

ok, let's switch to unbundled

15:48

13.288

❯ time ninja [215/215] Linking CXX executable DDNet-Server real 0m11.242s user 1m42.917s sys 0m10.609s

15:48

11?

yeah

my ryzen has 6 cores

15:48

12 threads

linking probably takes a lot of time, you used debug?

i used ur cmake line

15:48

with bunblded off

15:48

bundled*

15:49

AMD Ryzen 5 2600 (12) @ 3.400GHz

15:49

did u do it in release?

nope, debug

real 0m17.601s in release

19.754

:o

15:51

btw when i tried building with bundled it was mainly this errors

15:51

./obj/mysys_ssl/../../mysql-8.0.15/mysys_ssl/my_aes_openssl.cc:245: undefined reference to `EVP_CIPHER_iv_length'

15:51

EVP_

15:51

everything

eh, we should probably just delete the mysql bundle

15:51

it doesn't even work i think

I can't build with sql at all

./obj/libmysql/../../mysql-8.0.15/sql/auth/sha2_password_common.cc:153: undefined reference to `EVP_MD_CTX_destroy' /usr/bin/ld: /usr/bin/ld: DWARF error: could not find variable specification at offset 19a9 /usr/bin/ld: DWARF error: could not find variable specification at offset 1a09 /usr/bin/ld: DWARF error: could not find variable specification at offset 1a74

15:51

what is this

15:51

xD

15:52

dwarf error?

15:52

xD

One day we will get a nice way to compile for old glibc and nice package management for C/C++

15:53

Atleast I hope

the package management for c/c++ are distro package managers

fully static build?

lol

@Learath2 for glibc there is a nice way I think, but stdlibc++ is harder

15:54

@Comrade glibc might also depend on kernel versions, not sure

Executed in 24.77 secs

executed?

sounds like mac

yep mac

everyone has at least 3.2 kernel today, so don't worry

mac cross-compiling is much slower to me

@Comrade just a month or so ago we had someone with 2.6, stuck there because OpenVZ

im on 5.8.10

OpenVZ is broken thing

living on the edge

nobody should use that

There are still many hosters selling OpenVZ vpss and many users that just don't know any better buying them

15:55

@Ryozuki does that have wireguard built in yet?

how do i know that

15:56

xD

maybe it's rude, but it's only their problem if they pay for some OpenVZ trash

wirefilter?

Ah they added it in 5.6 so you have it

15:56

wireguard is high performance low level tunnels

[ddnet-scripts:master] 1 new commit

10ae233 Clean up build.sh - def-

@Comrade sure, but we still try to help them as well as they can

15:57

and for Linux in my opinion you should use your package manager or build yourself

yes

I never download prebuilt binaries online

1

yeah thats a rare thing to do on linux

This is why I don't really like flatpak

15:58

Well that and it's part of the global conspiracy to turn linux into hacker windows

@Learath2 r u familiar with ddnet editor code?

@heinrich5991 @Learath2 could you review https://github.com/ddnet/ddnet/pull/2918 please? I need it for my automatic builds script changes

Allow specifying -DGAME_RELEASE_VERSION by def- · Pull Request #291...

to be used for nightlies and release candidates to be able to tell them from regular releases more easily. For example: CPPFLAGS="-DGAME_RELEASE_VERSION="15.0.5-rc1"" cmake .

monkalaugh

15:59

@deen what about including a small git hash if u built it on ur own

16:00

commit*

16:00

by default

[ddnet:master] 2 new commits

2be54a3 HTTP: Set final m_State after running completion function - def- 56e996c Merge #2915 - bors[bot]

16:00

[ddnet/ddnet] Pull request closed: #2915 HTTP: Set final m_State after running completion function

@Ryozuki you could do that, I spent enough time already on this thing

ok maybe ill try after this gets merged

Is there any concern with the version string getting too long?

CMake supports only major.minor.patch.tweak

yeah, for cmake I still used the one from version.h

It's still in that format, we just add a suffix for our own version string, as far as I can see

@Learath2 nah, you can also replace it totally with -DVERSION=nightly20200924

and is cmake okay with that?

yes

how does it obtain these https://cmake.org/cmake/help/latest/variable/PROJECT_VERSION_MAJOR.html ?

16:07

if you use "nightly20200924" instead of "1.2.3.4"

It still reads them from src/game/version.h

16:08

so those will still be 15.0.5

16:09

Hm, I guess I could also switch to only allowing a suffix

16:09

and use 15.0.5-nightly20200924

16:09

is that better?

yes

1

i dont think it makes sense

16:10

to use version there

16:10

i mean the "15.0.5"

The only thing that would give more info would be replacing the date with a commit hash

its probs easier for your RC to allow fully change the version

@Learath2 but not so easy to understand for average players. they'll understand 20200924 is the date

definitely

@Deleted User i think for rc appending -rc1 etc is enough

@Learath2 https://github.com/ddnet/ddnet/pull/2884#issuecomment-696691153 do u know how to prevent scrolling from affecting the editor zoom when used on a scrollbar?

16:11

xD

@Ryozuki in the editor zoom location check if a popup is hot/active/something like that

^^

agh

16:11

ok

16:12

15.0.5-20200924

16:12

shorter and better imo

I really want SnapReplaceItem but it's a marathon to code that and I don't really feel motivated enough

nightly is redundant

[ddnet:master] 5 new commits

eb725f8 Allow specifying -DGAME_RELEASE_VERSION - def- ebb4e3d Center update info text, add -DINFORM_UPDATE flag - def- 1a17fef Use VERSION variable in cmake to specify version number more cleanly - def- d6edc78 Use steam info in cmake more, so i need fewer ugly shell scripts - def- 75dcbb8 Merge #2918 - bors[bot]

16:14

[ddnet/ddnet] Pull request closed: #2918 Allow specifying -DGAME_RELEASE_VERSION

[ddnet-scripts:master] 1 new commit

521fcef Nicer nightly version - def-

@Ryozuki are you looking into the git hash? Otherwise I'll create an issue I guess

im fixing the scrollbar rn

Hm, dissecting interleaved connections is soooo hard

16:30

Not sure if it's supposed to be dissectable actually

https://github.com/ddnet/ddnet/pull/2884

Add scrollbar to auto rules list, fixes #2514 by edg-l · Pull Reque...

It doesn't show a scrollbar when there aren't more than 5 rules.

16:53

fixed

16:54

it should also fix any other scrollbar scrolling we add in the future xd

16:54

(in the editor)

[ddnet:master] 2 new commits

f7cf854 Make sure turning dyncam off in settings works - def- ab8ad77 Merge #2916 - bors[bot]

17:01

[ddnet/ddnet] Pull request closed: #2916 Make sure turning dyncam off in settings works

17:01

[ddnet/ddnet] Issue closed: #2750 Entities are hidden in case if we didnt received server type

17:04

[ddnet/ddnet] Pull request closed: #2913 Center update info text, add -DINFORM_UPDATE flag

17:06

[ddnet/ddnet] Pull request closed: #2904 Determine binary dir correctly (fixes 2891)

17:08

[ddnet/ddnet] Issue closed: #2514 Editor: Auto rules sometimes don't fit on screen

17:08

[ddnet:master] 7 new commits

2587328 add scrollbar to auto rules list, fixes #2514 - edg-l ffade19 don't show scrollbar if not needed on auto rules - edg-l 9cdaba8 fix style - edg-l af29d8d remove unnecessary condition - edg-l 76f3dbf fix scrolling triggering zoom - edg-l fb3b409 fix style script - edg-l 7000987 Merge #2884 - bors[bot]

17:08

[ddnet/ddnet] Pull request closed: #2884 Add scrollbar to auto rules list, fixes #2514

17:12

[ddnet/ddnet] Pull request closed: #2877 "Always show start screen" option on the start screen

i wonder how should i indicate/know "built on ur own"

17:12

u already can know the hash with GIT_SHORTREV_HASH

17:12

we can shorten it to 7 chars tho

17:13

tee_thinking

Image attachment

Does anybody know why ddnet and teeworlds are setting display mode even when I am on my native display resolution and the game is set to use it too? It makes my display turn black and blink a led for 2 seconds. That's annoying and seems to be unnecessary.

17:14

(in fullscreen mode)

linux?

Yes

DE?

gnome

@Deleted User different color mode?

wdym?

17:15

aren't any program use 24-bit nowadays?

@deen he means the display turns black like when u change resolution

17:15

no?

yes

Also I am still annoyed to get 5 second freeze on exit. But that's another story.

17:16

That's DDNet only

debug it?

17:16

compile your own debug build, run with gdb

17:16

and check where ith angs for the 5 seconds

What signal I should send to it?

how would u check where it hangs with gdb?

no signal, run inside of gdb, when you hit quit and it hangs, press ctrl-c in gdb

17:17

then check backtrace

I will try

oh yea

But still that resolution change annoys me too.

17:17

It was there for ages I guess. That's the reason I use borderless window. But it has lower performance.

17:17

Much lower

https://stackoverflow.com/questions/19636832/find-out-where-my-c-program-hangs-with-gdb

Find out where my C program hangs with GDB?

I am trying to debug my C program with GDB to find out where it hangs. As a newbie to GDB, I am able to run it, but have no idea how to make GDB stop and tell me where the program hangs. I tried...

did u try gfx_resizable 0 @Pure_luck?

oh... sorry

17:19

I feel so stupid

oh, we should definitely ignore resizing events when fullscreen is enabled

It uses the same resolution for fullscren and for windowed mode

17:19

when I set it to 1600x1200 it tries to set it in fullscreen mode too

17:19

though I'd prefer it to use native resolution in fullscreen mode as lower resoluations don't make sense

17:20

I sometimes set it to windowed to be able to read some server debug while testing

still black screen for you with native res?

no, it's gone

17:20

sorry for that

I still get it, weird

Well, I have black screen for ~300 ms but that's nothing

17:21

it's ok

17:21

not 2 seconds

depends on your screen probably

or CPU maybe

17:21

I have rather recent one

17:22

so I don't notice it now

is this a gnome thing?

I don't have problems now with screen resolution change

@deen why do the news dont use the full space? e.g the last bulletpoint

Image attachment

it wasn't native

17:23

I cannot debug in GDB as it's hanging entire DE for 5 seconds on exit

is it intentional?

17:23

I cannot debug in GDB as it's hanging entire DE for 5 seconds on exit

@Deleted User monkalaugh

monkalaugh

maybe I should send it segstop idk

17:24

I will make a console script that will print which process is using cpu on ddnet exit

ah, just a problem with my window manager

17:24

@Ryozuki so it fits in 5:4

tee_thinking

eh, got crash just during game (edited)

try to remember what u did

nothing, i guess

17:28

just played

4Head

17:29

r u using a official release or did u compile from master

I've made a simple script to monitor processes in background. No need to debug DDNet. It's Xorg and gnome-shell processes going 100% on DDNet exit. It have started after last system update. I cannot tell if that's DDNet code fault. Just no other program is affected. (edited)

17:35

6168_Kim_Jong_Un

@Ryozuki steam client

I think DDNet is calling X/WM something and then is waiting for it to complete something.

Dance

1

17:35

It was instant before, but not now.

ghost_sad_but_cute

1

17:36

I am glad I am not epileptic but that emoticon is still annoying.

17:36

Too bad it doesn't deletes on user block.

17:36

kids that just got discord nitro use it

17:37

Oh.. I still haven't sent a SIGSTOP to gdb'ed DDNet with a background script

17:37

I might know what it is calling

17:39

Does it help?

Image attachment

17:40

Turning off V-Sync does not help. It happens in windowed mode too.

17:41

It's that 5 seconds delay with 100% cpu on Xorg and gnome-session before DDNet exits.

Guys, I have problem with one player - @Saiborg gets his client frozen every single time we're at the end of [School] Hook and he falls in speedups

17:41

Image attachment

17:42

his client is frozen, he can't do a thing, but only hears sounds

17:42

his console shows nothing though https://media.discordapp.net/attachments/757166272122454076/758745033720463410/Stuck_3.jpg?width=908&height=536

Embedded image

maybe this map contain tiles his client doesn't support? is he using ddnet?

17:42

I mean recent one

DDNet, latest, from Steam

Yeah

17:42

DDNet

Try to zoom in

what will happen if he will spectate to that point of a map from the start?

he can go thru the part

I still have the game open, cant do anything

problem starts when he falls in & gets dragged by speedups

17:43

it's already 4th time, his friend had same problem two days ago when we played together

17:43

same map, same place

17:45

there's not a single unused tile in this place, just checked in editor

17:45

also, somehow I don't get this problem

17:45

@Saiborg I just saw you hitting 3 times before leaving

17:46

so seems like something worked at the end

Nope xD The thing is the game doesnt freeze, i can hear the sounds and everything, but my screen gets stuck

17:46

It was stuck at this screen, but i could hear all the sounds

Image attachment

17:47

and move as well probably because i heard it

well, I was seeing you at the beginning on part

17:47

of*

probably because i died here when it "froze" and my controls are working, and i hear everything, but my screen is locked to this image

version platform old gpu? opengl3 on?

17:49

xd

17:49

discord slow damn

17:49

version platform old gpu? opengl3 on?

17:49

discord slow damn

Image attachment

@Saiborg Settings -> Graphics -> just paste screenshot as that will be easiest xd

Image attachment

Embedded image

17:54

^ his btw

is it exactly that part on the screen?

What do you mean Jupstar?

18:12

If you mean if its that part i get stuck on yeah.. I have a screenshot for when i got stuck the fist time as well.. let me send the pic 1min

18:13

Image attachment

18:13

This was the first time

henlo

hi

this channel is much more active

18:16

full of discord doods

why you have to type /notice Nimda I am not a bot to join

when do you return from wild true-ascetic life

oh

18:21

ok its boring now i should go bye

troll

2

i valgrinded, we have quite some invalid reads and writes but only in the server browser,

18:36

also cant reprod any frame drops

18:36

@Saiborg can you send me ur ddnet_settings.cfg privatly

@Deleted User I happened 4 times in total, with Pak & his friend, alway in this part, and always only to the player that felt in the speedups

18:42

though to make it weird, their screen freezes right before that

almost sounds like a bug in prediction xd

on the screen from them they are in air, so looks like out of nowhere screen freezes in ~middle of this part, and due to this they fall in speedups

do they all have background maps active?

18:43

and entites

but I have totally no idea why or how, since I'm not affected when playing with them nor when doing it alone

18:44

For sure entities, but on first ocassion Pak didn't have custom entities bg, I explained him later how to do that

18:44

(I think)

18:45

idk about his friend Drakii, because he played that map only once

18:45

but I doubt not, since he knows even less xd

18:46

but I doubt he could have it somehow*

18:46

sorry, brain half dead rn

ok, so many possibilities .D

18:49

could be a NaN or some overflow

https://developers.redhat.com/blog/2020/09/24/new-c-features-in-gcc-10/

New C++ features in GCC 10 - Red Hat Developer

Explore the front end-based C++ features that C++ application programmers care most about in GCC 10.1 (G++ 10.1), which include many C++20 proposals.

18:54

monkalaugh

@Deleted User but then how does it happen every time to Pak, but not a single time one to me? :c

18:59

I'd bet something with drivers, since that's lately hot topic xD

maybe his setup, his CPU, his GPU

18:59

could be anything

the only weirdness is that we have played several different maps, and it didn't occur on any of them

19:00

only on [School] Hook and only in the last part ._.

yeah thats why i think its something with floating points

19:00

maybe it needs some position on the screen to happen (edited)

19:00

but really hard to tell

19:05

xD

19:05

@deen i cant type any passwort in passwort protected servers anymore

oh, about passwords and servers... if I have password set in the config, could it just connect to the server automatically without the need to still click ok to join? The pw is autofilled, just still requires confirmation to join

yes that was made bcs of security reasons

19:26

e.g. passwort steal

@Deleted User where do i find that file?

@Saiborg DDNet client directory -> double click config_directory.bat -> will be in directory that will pop up

i have a file called settings_ddnet

19:27

is it that one?

yes ty

19:29

enabled floating point exceptions now... instant crash xD

absolutly cant find anything, but the x coordinate on both pictures is the same

20:31

thats insanly suspecious

1

sendhelp

1

1

@Deleted User I saw some mesa bugs about this, for example https://forums.factorio.com/viewtopic.php?t=56377

20:39

@Deleted User what do you mean by typing pw? in f2?

no

20:39

in the password field

where is that field?

go on password protected server

ah, when connecting

hola

20:41

alguien juega

20:41

someone play?

can server's ip become an admin on join?

20:42

im about situation when you are testing stuff and reload it quite often, so, you are hosting it on localhost

20:42

@Quixe #general

absolutly cant find anything, but the x coordinate on both pictures is the same

@Deleted User oh lol, indeed, didn't even notice o.0

20:52

@Quixe ask in #general, not in developer's channel ^^

soooooOOOOOOOOOOOOOrrY

@gerdoe F2 reload doesn't log you out

20:59

also, if you find it helpful: on my local server I set rcon pw to be single letter/number, so it's not annoying to login often, and you can also do bind with rcon_login <pw>

[ddnet/ddnet] Issue opened: #2919 Can't enter password anymore on pw-protected server

As reported by @Jupeyy on Discord. @TsFreddie I'll take a quick look if I can fix it, otherwise please you take a look since your change caused this: ``` $ git bisect bad 387bc53030695b9b81e56278bb49a19f203971e8 is the first bad commit commit 387bc53030695b9b81e56278bb49a19f203971e8 Author: TsFreddie Date: Tue Sep 22 17:01:13 2020 +0100 ime window position & ime input handling ddnet-libs | 2 +- src/engine/client/input.cpp | 34...

@gerdoe if you connect to 127.0.0.1 and are logged into rcon map reloads automatically when you save it in editor

@Learath2 i find it funny, ##C guys citing the standard like someone cites the bible

Image attachment

22:27

ah, i must be rly bored to be reading this chat at 1230

I can cite the standard like that too

@Learath2 i see ur a man of culture too

lmao

Exported 651 message(s)