DDraceNetwork - developer

heinrich5991

[ddnet/ddnet] Pull request opened: #5093 Fix /practice chat responses and a few others

Fixes #5087.

bors[bot]

[ddnet:master] 2 new commits

ed03945 do not assert on entities without snap id - C0D3D3V e4ed71f Merge #5088 - bors[bot]

07:49

bors[bot]

[ddnet/ddnet] Pull request closed: #5088 do not assert on entities without snap id

07:49

bors[bot]

[ddnet/ddnet] Issue closed: #5054 Crash in SnapNewItem

bors[bot]

[ddnet/ddnet] Issue closed: #5087 No notification in team 0 if you try to use /practice

08:11

bors[bot]

[ddnet:master] 2 new commits

0c2bb5a Fix /practice chat responses and a few others - heinrich5991 bab2def Merge #5093 - bors[bot]

08:11

bors[bot]

[ddnet/ddnet] Pull request closed: #5093 Fix /practice chat responses and a few others

bors[bot]

[ddnet:master] 2 new commits

b6a9389 Render tees in skin selector with default emote - sjrc6 d87edb9 Merge #5089 - bors[bot]

08:37

bors[bot]

[ddnet/ddnet] Pull request closed: #5089 Render tees in skin selector with default emote

@Learath2 I did it

TsFreddie

@Learath2 I did it

Now backport it to teeworlds ❤️

I have some question about the licensing. I see no License notice on https://github.com/ddnet/ddnet-maps, does that mean all maps are copyright by default? also, what is the license for the default image assets and teeworlds / ddnet code?

Default assets are all CC-BY-SA

09:29

https://raw.githubusercontent.com/ddnet/ddnet/master/license.txt

09:29

09:30

Since the mapres are just in the wild with most of their authors long gone, it's really hard to get all maps to be CC-BY-SA now :/

define "gone"

3

not reachable

Best case: Doesn't play ddnet anymore but we can identify and email them

09:31

Worst case: Unknown author

~~or dead~~

1

or dead

OK thanks. So if I understand correctly I technically cant distribute a map from ddnet-maps on an open-source repo, but in practice it's fine to do so?

Maps in data/maps are all CC-BY-SA, for the rest of the maps you need to check their header for their license and 99% of them don't have any license at all

1

k2d222

OK thanks. So if I understand correctly I technically cant distribute a map from ddnet-maps on an open-source repo, but in practice it's fine to do so?

I am not a lawyer, but I think there was some verbal agreement at least between us and the map authors that their maps would be public. I doubt that also extends to secondary parties, so I guess legally you probably can't distribute them

09:36

I doubt anyone would actually sue you over it. I'm not even sure if these authors can even assert any copyright over the maps given the mapres don't belong and aren't licensed to them

1

09:38

I did try to get actual legal advice about this, but the lawyer I asked quoted me about 300-400 euros per hour of his time iirc. So I just dropped the question instead

don't sell them and you'll be fine (probably) (not legal advice) (please sign the waver so you don't sue me if you get sued)

ez 400€

I'm not very knowledgeable about copyright laws, but I find the licensing on the teeworld-related repos a bit loose considering it's a quite popular game with a lot of contributors. I think copyright / licensing it is a crucial skill that most developers me included lack of.

except for the maps it should be fine

09:41

where do you see a problem except for the maps?

Actual teeworlds and ddnet license are pretty clear actually

09:42

Do whatever you want, don't claim you made it, mark your altered versions as altered, don't remove the license

09:42

Oh and we are not liable if you run it and it launches a nuclear warhead

I believe non-liability clauses have limits

https://discord.com/channels/252358080522747904/293493549758939136/723645849996886126 scanned the maps for licenses some time ago, 10 maps had that field set https://discord.com/channels/252358080522747904/293493549758939136/723645849996886126

1

09:43

hm double link

heinrich5991

I believe non-liability clauses have limits

There very well could be, not something I've looked into

the liability clause should be in all caps

09:44

makes it more effective

1

Thanks everyone for the clarifications! I know I'm not really in legal issues but I want to make things right and I would like to learn more about licensing overall.

@heinrich5991 if you have a moment can you take a look at #5092 and why it's not linking? It linked just fine locally

https://github.com/ddnet/ddnet/pull/5092

Rework HTTP to use curl-multi interface (DEPENDS ON #5069) by Leara...

This PR tries to rework the http interface to utilize the curl-multi interface. In the process we lost the IJob interface. I experimented with getting rid of the std::unordered_map using CURLOPT_PR...

Here's something I haven't had happen before (screenshot with phone camera for obvious reasons)

09:53

well u dont have space in ur hard drive

Yep

btw df -h is better

1

tf is that screen

Cipy29

Here's something I haven't had happen before (screenshot with phone camera for obvious reasons)

Happens to me regularly on the macbook

10:01

Also happened a lot on my server while I was also archiving teehistorian files

@Learath2 need to update our curl stub

You really get to see which software is built well when space runs out

heinrich5991

@Learath2 need to update our curl stub

Oh, hm, do we have some sort of documentation on the process? Is it something I should be doing?

run scripts/generate_fake_curl.py on linux (edited)

10:02

ddnet-libs documents how the libcurl.so is generated IIRC

10:03

see ddnet-libs/curl/linux/lib32/VERSION or ddnet-libs/curl/linux/lib64/VERSION

Learath2

You really get to see which software is built well when space runs out

I also got to see how much useless stuff I had downloaded justatest

frontend guy sent this for review.

3

1

smh not even creative enough for a fourth variant of the =

checkout that parse tho

aye, nice code

@heinrich5991 how do you handle changes in ddnet-libs? Should I do a PR on ddnet-libs too?

@k2d222 I agree btw that licensing is sadly not handled very well with maps, mostly because it was never really started I suppse

just push there if you're somewhat sure you're doing the right thing @Learath2

10:08

should I generate the curl libraries for you @Learath2 (if you don't have a linux handy)

I'm on linux, generated them, will push

10:09

python3 scripts/generate_fake_curl.py --link-args=-m32 is enough for getting the lib32 one, right?

yes

10:09

I mean try file on it

It does generate an 32-bit ELF file, just wasn't sure if I also needed to be on 32bit for some reason to generate it properly

no

Learath2

[ddnet-libs:master] 1 new commit

b0fca23 Update curl to add curl-multi functions - Learath2

Hopefully I didn't break it

looks good to me

@heinrich5991 I'm not sure how to handle CURLcode without including curl/curl.h in http.h. I tried to assume it's unsigned int and static_asserted for it in http.cpp, turns out it's not unsigned int on windows

heinrich5991

where do you see a problem except for the maps?

Skins ^^

I'd need something like std::has_greater_range_than<unsigned int, std::underlying_type<CURLcode>>

10:32

I got an excellent idea from the C++ irc

perhaps just do sizeof?

Does anyone have knowledge of sound-layers? First of all, what does the PAN option do? And secondly directional sound isn't possible, or is it? (edited)

was it somehow planned to have like dynamic tiles? like on KoG, having dynamic freeze or solid tiles

11:47

it could be cool for new maps

maybe @Patiga knows something about sound layers

11:58

@Learath2 added the suggestion in the http PR

I am pretty sure I will add a feature request (and maybe even a PR if I can make sense of the code)

it'd be nice if you could add an issue first

12:01

so effort isn't wasted in case we don't like the design or something

heinrich5991

@Learath2 added the suggestion in the http PR

I have no issues with anything else, I was a little on the fence about the unique_ptrs but I see why you'd want them

unique_ptrs where?

12:03

as return type?

yeah

12:03

in http.h

@Ryozuki told me that's the way to go, and I thought it kinda makes sense

12:04

we return the ownership, and they're easily convertible to raw pointers or shared_ptrs

Can't really return raw pointers without allocation anyway

12:05

So the unique_ptr is better as it takes care of allocating and de-allocating

AssassinTee

I am pretty sure I will add a feature request (and maybe even a PR if I can make sense of the code)

the answer is yes and PAN is actually the setting for it <.<

@heinrich5991 #5092 still won't link :/

https://github.com/ddnet/ddnet/pull/5092

Rework HTTP to use curl-multi interface (DEPENDS ON #5069) by Leara...

This PR tries to rework the http interface to utilize the curl-multi interface. In the process we lost the IJob interface. I experimented with getting rid of the std::unordered_map using CURLOPT_PR...

I gtg for a while, small things to do

def-

[ddnet-web:master] 2 new commits

c25e165 Add NLD funding - def- ada69b4 Add 5 € donation by metlix - def-

https://pyscript.net/

Pyscript.net

Run Python code in your HTML.

13:36

ow

13:36

finally something to stop javashit

yes, let's do interpreted languages on top of another interpreted language on top of a browser sandbox (edited)

wasm and js are not really interpreted anymore

oh my bad it actually uses pyodide := cpython ported to wasm not that bad in that case!

AssassinTee

[ddnet/ddnet] Issue opened: #5094

Vulkan sometimes renders clipped quads

Bug

I have multiple clipped quad regions and one quad is always rendered in the wrong region. This only happens ingame and with Vulkan Ingame (Vulkan): One blue stars appears between all red stars !screenshot_2022-05-13_15-41-52 Editor(Vulkan): No blue star ![screenshot_2022-05-13_15-41-39](https://user-images.githubusercontent.com/821401/168303057-85fc28d6-dc0c-4cfe-8cf1...

Sorry for breaking the game again

@AssassinTee ?

14:24

AMD beta-testing Ryzen 7000 on Linux, unfortunately they didn't select DDNet benchmark

https://web.archive.org/web/20220509071950/https://openbenchmarking.org/result/2205052-NE-MAY42016271

When people join a community, they don't just adopt the values of the community, but they also come with pre-existing values. As a community grows, it will adapt to the people who are joining it, and this may alienate the early-adopters who joined for a different reason than the people joining during the exponential or diminishing returns phases of logistic growth. This doesn't mean the community is "ruined" but from the perspective of early adopters, they might describe it that way.

14:27

https://news.ycombinator.com/item?id=31363153

techsin101

Ask HN: Is it true that any community that grows big enough, gets ruined?

Have you ever seen an example where anything became better than before when it grew big. There is almost a law that a company, a group, a community, a game or anything once it grows past its core audience it just gets ruined.I have seen a game with millions of players become crappy once they became mainstream and started to attract very young an...

14:28

interesting discussion, that may apply to ddnet at some level i guess

14:28

xd

deen

AMD beta-testing Ryzen 7000 on Linux, unfortunately they didn't select DDNet benchmark

https://web.archive.org/web/20220509071950/https://openbenchmarking.org/result/2205052-NE-MAY42016271

5.21 ghz and 8 cores damn

hus3h

[ddnet/ddnet] Issue opened: #5095 Can't type into editor inputs when entering the editor using 'cl_editor 1'

Reproduction: you have to be in a server, in F1 or using a bind execute cl_editor 1, try using inputs (Save As for example), doesn't take any keys except backspace and delete (enter to submit also works)
Applies to all inputs including layer options and server commands
Entering the editor through the UI icon and using ctrl+shift+e works fine
Pressing ctrl+shift+e, even if you don't input anything and immediately close the editor, fixes it till client is closed (makes you ...

scam

stop posting that

HappyFray

[ddnet/ddnet] Issue opened: #5096 no output while using specific commands

No output while writing /rules nor /cmdlist Neither console/chat or external console.

Ryozuki

interesting discussion, that may apply to ddnet at some level i guess

aka Eternal September https://en.wikipedia.org/wiki/Eternal_September

Eternal September

Eternal September or the September that never ended is Usenet slang for a period beginning around 1993 when Internet service providers began offering Usenet access to many new users. The flood of new users overwhelmed the existing culture for online forums and the ability to enforce existing norms. AOL followed with their Usenet gateway service...

xD

15:22

i never used usenet nor rly know what it is

15:22

i guess im too zoomer

deen

AMD beta-testing Ryzen 7000 on Linux, unfortunately they didn't select DDNet benchmark

https://web.archive.org/web/20220509071950/https://openbenchmarking.org/result/2205052-NE-MAY42016271

is also pretty slow apparently

well slower than intel

15:29

so must be a test sample

what's slow

nobody cares about irc

xd

15:30

will check from Discord

@Jupstar ✪ it looks like the nvidia open source thing is not rly mergeable into the linux tree like amd drivers rn

15:31

and its mostly function calls to private firmware stuff

1 million lines of codes i read 4 million AMD

xd

but anyway, its a start

15:32

and they got positive response

maybe it tells driver devs some hints

maybe that helps them to do more in future

https://news.ycombinator.com/item?id=31365307

palebluedot

Twitter Deal Temporarily on Hold

But it may help Nouveau:

15:35

In the meantime, published source code serves as a reference to help improve the Nouveau driver. Nouveau can leverage the same firmware used by the NVIDIA driver, exposing many GPU functionalities, such as clock management and thermal management, bringing new features to the in-tree Nouveau driver.

15:35

https://developer.nvidia.com/blog/nvidia-releases-open-source-gpu-kernel-modules/

Ram Cherukuri

NVIDIA Releases Open-Source GPU Kernel Modules | NVIDIA Technical Blog

The first open-source release of GPU kernel modules for the Linux community help improve NVIDIA GPU driver quality and security.

yeah but its more a PR move

yeah maybe

i dont think so

15:36

i think its pressure from google and other big players

Why Google?

they want custom code and amd gives it to them

k2d222

oh my bad it actually uses pyodide := cpython ported to wasm not that bad in that case!

wait what

well google and other big players

15:36

everyone wants GPUs nowadays

15:37

thats also why its "datacenter" code

15:37

as nvidia claims itself

i always thought this, but its actually insane companies gives so much preference to windows, you actually give your lifeline to microsoft, companies should actually be interested in making linux mainstream, i think valve knows this and thats why they dont wanna depend on microsoft

Don't they make TPUs?

Ryozuki

i always thought this, but its actually insane companies gives so much preference to windows, you actually give your lifeline to microsoft, companies should actually be interested in making linux mainstream, i think valve knows this and thats why they dont wanna depend on microsoft

windows is really only on desktop market

15:38

and germany xd

yeah

15:38

still

15:39

im also sure knowing the inner workings of linux source code could help big AAA game engines perform better

its probably a dead horse in the long run tho. Valve is not a small player remember that

on windows they are limited to reverse engineering maybe

the gaming market is one of the biggest argument for microsoft

but its stupid