DDraceNetwork - developer

One "pID" can be used only for one ui element

4:57AM

multi-touch handling be like

ddnet-maps

[ddnet-maps:master] 1 new commit

8cc28ce A Purple Peaks 2 - ddnet-maps

TsFreddie

Click to see attachment 🖼️

pain

6:44AM

I tried to add multi-touch to one of my games once (edited)

6:45AM

Then I got the conclussion of literally never making a mobile game

6:45AM

atleast in the game engine I'm used to use

2023-11-09 10:41:21 I assert: /ddnet-source/src/game/editor/editor.cpp(2386): Unhandled entities image for explanations

8:44AM

I think I now what's causing this but I am not sure.

8:45AM

Here's the entities I use.

8:48AM

I have added more tiles (specifically more spike tiles and flag goal tiles) to the entitiesdfngv2.png which is located under editor/entities/

8:49AM

Hmm I forgot to test with nightly doing so now...

8:50AM

Although looking at the code the result should be the same

                EExplanation Explanation = EExplanation::DDNET;
                if(m_SelectEntitiesImage == "DDNet")
                    Explanation = EExplanation::DDNET;
                else if(m_SelectEntitiesImage == "FNG")
                    Explanation = EExplanation::FNG;
                else if(m_SelectEntitiesImage == "Race")
                    Explanation = EExplanation::RACE;
                else if(m_SelectEntitiesImage == "Vanilla")
                    Explanation = EExplanation::VANILLA;
                else if(m_SelectEntitiesImage == "blockworlds")
                    Explanation = EExplanation::BLOCKWORLDS;
                else
                    dbg_assert(false, "Unhandled entities image for explanations");

Same thing on nightly 2023-11-09 10:56:56 I assert: ddnet/src/game/editor/editor.cpp(2564): Unhandled entities image for explanations (edited)

I made by patching ddnet but I am not sure if my approach is considered correct/if it would cause problems. I basically added OTHER value to the EExplanation enum (at editor.h::(958)) and changed the else statment to Explanation = EExplanation::OTHER; (at editor.cpp::(2564)). Lastly I added

case EExplanation::OTHER:
    return nullptr; // TODO: Explanations for Other?

(at explanations.cpp::(682)).

9:10AM

*made it work by patching...

Mr.Gh0s7

Here's the entities I use.

oh wait hold on fng with teleport and live freeze? and JETPACK?

1:08PM

ok. this might be peak

Voxel

oh wait hold on fng with teleport and live freeze? and JETPACK?

I think you've already played at dfng xD

oh wait i haaaave

1:23PM

i need to play more of that

1

1:23PM

i just didnt know these features

Mr.Gh0s7

Here's the entities I use.

is that a 10 team fng

no more

1:54PM

63 plus team 0 = 64 teams fng

Mr.Gh0s7

63 plus team 0 = 64 teams fng

thats just called solofng LOL

2:56PM

or if theres 128 max server itd be like duos from fortnite

https://blog.rust-lang.org/2023/11/09/parallel-rustc.html

Faster compilation with the parallel front-end in nightly | Rust Blog

Empowering everyone to build reliable and efficient software.

4:20PM

@Jupstar ✪

4:22PM

@Jupstar ✪ can u time

4:22PM

RUSTFLAGS="-Z threads=8" cargo build --release

4:22PM

change 8 to ur cores

4:22PM

read the blog post, it halved the compile time if i read correctly

4:22PM

u need nightly

4:23PM

When the parallel front-end is run in multi-threaded mode with -Z threads=8, our measurements on real-world code show that compile times can be reduced by up to 50%, though the effects vary widely and depend on the characteristics of the code and its build configuration. For example, dev builds are likely to see bigger improvements than release builds because release builds usually spend more time doing optimizations in the back-end. A small number of cases compile more slowly in multi-threaded mode than single-threaded mode. These are mostly tiny programs that already compile quickly. We recommend eight threads because this is the configuration we have tested the most and it is known to give good results. Values lower than eight will see smaller benefits. Values greater than eight will give diminishing returns and may even give worse performance.

4:23PM

ok try with 8

4:24PM

https://en.wikipedia.org/wiki/Amdahl%27s_law

Amdahl's law

In computer architecture, Amdahl's law (or Amdahl's argument) is a formula which gives the theoretical speedup in latency of the execution of a task at fixed workload that can be expected of a system whose resources are improved. It states that "the overall performance improvement gained by optimizing a single part of a system is limited by the ...

Ryozuki

RUSTFLAGS="-Z threads=8" cargo build --release

without cargo build --release 894,20s user 59,23s system 1001% cpu 1:35,22 total with RUSTFLAGS="-Z threads=40" cargo build --release 920,27s user 71,94s system 1040% cpu 1:35,38 total was slower for some reason xd or is it especially important to use exactly your core count and not logical threads?

5:10PM

but i can try with 8.. but sounds stupid on 16 core pc xD

@Jupstar ✪ bro

5:13PM

use 8

5:13PM

not 40

5:13PM

read the article

5:13PM

to understand why

5:13PM

cargo already calls rustc paralel

5:14PM

but they also made rustc itself paralel

the article is 1000 characters above my attention span

they tested and said with 8 on a 24 core worked best

5:14PM

try 8

Ryozuki

try 8

RUSTFLAGS="-Z threads=8" cargo build --release 1035,68s user 98,75s system 1156% cpu 1:38,06 total

5:14PM

it's even slower

lmao

5:14PM

with clean?

yes

how much cores u got

or is the blog post for the nightly tomorrow or smth? xDD

try 4

16

or 2

5:15PM

or 6

so to use one AMD processor group basically??

cargo doesnt fully saturate when calling multiple rustcs iirc

5:16PM

so paralelozing a bit rustc may

5:16PM

the article explains it idk

5:16PM

i also over read it a bit

5:17PM

but they got 50% gain

ok i ask gpt for a summary

though the effects vary widely and depend on the characteristics of the code and its build configuration. For example, dev builds are likely to see bigger improvements than release builds because release builds usually spend more time doing optimizations in the back-end.

5:17PM

try with debug mode too

5:18PM

A small number of cases compile more slowly in multi-threaded mode than single-threaded mode. These are mostly tiny programs that already compile quickly.

mhh

5:20PM

or they simply did synthetic benchmarks

5:21PM

rerun without these flags.. defs faster in release cargo build --release 888,18s user 57,35s system 1007% cpu 1:33,88 total now i try debug

@Jupstar ✪ using 40 threads is a bit much when cargo itself spawns 40 rustcs

heinrich5991

@Jupstar ✪ using 40 threads is a bit much when cargo itself spawns 40 rustcs

i didnt read the article xD

5:23PM

i see 8, i put 40

it means your CPU would run 1600 threads in parallel, likely diminishing the returns of parallelizationj

5:23PM

btw 40 was faster xD

xddd

5:23PM

did u rustup up

5:24PM

to get latest nightly

5:24PM

just in case

debug: cargo build 506,87s user 73,18s system 1678% cpu 34,567 total with: RUSTFLAGS="-Z threads=8" cargo build 674,66s user 120,29s system 1794% cpu 44,299 total

% rustc --version                                                                                                                                 git:playground
rustc 1.75.0-nightly (fdaaaf9f9 2023-11-08)

5:24PM

i dunno, it's simply slower for me

5:24PM

or is it for the nightly tomorrow? XD

5:25PM

but even then it would be weird that the flags make a difference i guess

Jupstar ✪

i dunno, it's simply slower for me

Slower than normally compiling? Now that's an absolute failure

Learath2

Slower than normally compiling? Now that's an absolute failure

well atleast for the project i try with

5:26PM

they apparently tried many libs

5:26PM

or they have some NASA pc

Learath2

Slower than normally compiling? Now that's an absolute failure

its a experimental nightly feature

You should try one of their test projects

but in their reports they get good results

If you cant reproduce that either now that would be interesting

@Jupstar ✪ ur crate is big right

Learath2

You should try one of their test projects

at least cranelift is part of my compilation too

Ryozuki

@Jupstar ✪ ur crate is big right

it's huge, but mostly bcs the deps are huge xD

5:27PM

cranelift compiler alone takes so fcking long xD

Ew imagine having deps and not rolling everything on your own

yeah imagine not living in the 80'

Ew

EwE

eWe

WeW

serde in debug: cargo build 7,80s user 1,26s system 292% cpu 3,095 total vs RUSTFLAGS="-Z threads=8" cargo build 11,94s user 2,20s system 485% cpu 2,910 total

5:32PM

so also slower for me

seems worthy of a report (edited)

also tried with 2 and 4 as threads.. still slower. but actually a tiny bit faster than with 8

5:33PM

now with 40

5:33PM

40 is slow xdd ~18 secs

5:34PM

with threads=1 it's around the same speed as without, but still bit slower

have you tried with 1?

5:34PM

ok

xd

5:34PM

odd

Ryozuki

odd

have u tried it?

5:36PM

i mean would be worth a try.. maybe it's some AMD thing that it's slower

i havent

5:41PM

im not home rn

5:41PM

i also got amd btw

yes

but i can try with m1

ok

imagine running a model on gpu to determine fastest compile parameters xd

the EU law just reinforces free software btw

5:53PM

the ability to modify the source code and compile allows u to not abide by the stupid shit

5:53PM

gov CAs

5:53PM

#stallmanwasright

5:53PM

5:55PM

https://wiki.mozilla.org/CA/Changing_Trust_Settings#Deleting_a_Root_Certificate

Changing Trust Settings

what

the EU law just reinforces free software btw

is it even proposed already?

6:09PM

"EU law" sounds like it's in place already

Robyt3

[ddnet/ddnet] Pull request opened: #7428 Fix editor crash when showing explanation for custom entities

The failing assertion is removed because users can add custom entities images besides the predefined ones. Reported by Mr.Gh0s7 on Discord.

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 (especially base/) or added coverage to integration test
[ ] Considered possible null pointers and out of bounds array indexing
[ ] Changed no ph...

github-merge-queue[bot]

[ddnet:master] 2 new commits

df66426 Fix editor crash when showing explanation for custom entities - Robyt3 962379d Merge pull request #7428 from Robyt3/Editor-Entities-Explain-Assertion-Fix - heinrich5991

7:00PM

heinrich5991

[ddnet/ddnet] Pull request closed: #7428 Fix editor crash when showing explanation for custom entities

One question why does ger10 laag for some spefic players since yesterday ?

7:04PM

Is this some network stuff or is there some kind of firewall "protection" something to slow them down?

@Ryozuki did u see steam deck oled

Normaly ger10 ~25 ping, and now ~100 ping, while ger2, pol everything else is same as before just ger10

Are you ingame? In the browser you might have a cached bad ping

Learath2

The regular version will be defined properly by cmake. That's why it's grey

bruh

Sedonya

bruh

Did you modify the cmake project to change the version aswell?

Learath2

Did you modify the cmake project to change the version aswell?

I have a private fork, I just upgraded from 17.2.1 to 17.3

Did you modify CMakeLists

7:23PM

That's the question

Learath2

Did you modify CMakeLists

no

7:24PM

I didn't, but when I renewed it, I noticed it had been cheated on by someone else (edited)

7:24PM

and

Mr.Gh0s7

2023-11-09 10:41:21 I assert: /ddnet-source/src/game/editor/editor.cpp(2386): Unhandled entities image for explanations

Should I make an issue/pull request here this seems to be lost

Ok then the behaviour makes absolutely no sense to me. GAME_RELEASE_VERSION is defined by cmake. Or if not by version.h. Neither would define it as 2.1.1 unless you changed something

Learath2

Ok then the behaviour makes absolutely no sense to me. GAME_RELEASE_VERSION is defined by cmake. Or if not by version.h. Neither would define it as 2.1.1 unless you changed something

I can add you to the repository if it's more convenient for you, you can see the commits for yourself

Uhhhhhh I'm on a phone. I'd rather not tbf

1

Sedonya

and

Everything that has changed in CMake since the last working version is in the screenshot. The rest is just adding files to the compilation

Did you try a clean recompile?

7:28PM

Like delete the entire build dir

no

7:29PM

I'm gonna give it a try

Learath2

Did you modify CMakeLists

@Sedonya I think Learath2 means if you have modified CMakeLists.txt at all

7:31PM

and the answer is probably yes

7:32PM

(not just since it broke, but in all of your custom client history)

a

7:32PM

yes

ReiTW

@Ryozuki did u see steam deck oled

no

heinrich5991

(not just since it broke, but in all of your custom client history)

But I was just adding the name of the files to compile, no more than that

@lerath2 sry for the late response its not just browser, if u go into any ger10 server there are few players runnning with 100 ping u can ask everyone the same question, they lag since yesterday

7:37PM

ps. "everyone the same question, they lag since yesterday", and this also only on ger10 btw

7:40PM

btw the minute i wrote it everyone, got better ping now, so the problem is solved for everyone

7:40PM

if someone did something thanks alot

@Learath2

7:41PM

I have Wnodev in the release, so no complaints

Sedonya

and

This part of the code was introduced 3 years ago: https://github.com/ddnet/ddnet/commit/1a17feff12e03e8c95c4ca0199f0e55612fc8b1a then I tried to remove it, and had to restore again: https://github.com/ddnet/ddnet/pull/7310

Kaffeine

This part of the code was introduced 3 years ago: https://github.com/ddnet/ddnet/commit/1a17feff12e03e8c95c4ca0199f0e55612fc8b1a then I tried to remove it, and had to restore again: https://github.com/ddnet/ddnet/pull/7310

Got it, but how do I fix my code? Undoing your changes is not the best option I guess

So, this code is not something new and you have issues due to other changes in CMakeLists. CMake and the current packaging are not as flexible as we'd like it to. GAME_RELEASE_VERSION is a generic name for the project version.

9:12PM

VERSION and GAME_RELEASE_VERSION are used to indicated the packaged version. Maybe in your case you'd want to define (via CMake target_compile_definitions) SCLIENT_RELEASE_VERSION instead of the GAME_RELEASE_VERSION (and probably modify the version.h for that).

1

dobrykafe

[ddnet/ddnet] Pull request opened: #7429 Hide "Show directory" when "All combined" selected

closes #7424

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 (especially base/) or added coverage to integration test
[ ] 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/...

does ddnet use libpng or pnglite?

libpnglite

9:49PM

xd

9:49PM

no, libpng nowadays (edited)

hmm

9:52PM

so the pnglite lib ty

wat

9:56PM

pnglite gone

hmm xD

9:56PM

libpng is bloated no?

it's libpng, ignore my past me

Mr.Gh0s7

libpng is bloated no?

why?

9:57PM

it's the reference implementation

idk I am asking

it does what it should xd

9:57PM

loading pngs

Ok that's what I am using then :)

ah for your project?

9:58PM

i'd say better use an easier to use lib

Jupstar ✪

ah for your project?

yeah

but pnglite can only load certain png images

9:58PM

but robyte made a pr for vanilla tw

9:58PM

using some png lib

9:58PM

maybe u can use that

https://github.com/teeworlds/teeworlds/pull/2934

Proof of concept: replace pnglite with stb_image/stb_image_write by...

Using stb_image decreases loading times for images by around 25% when compared to pnglite for me, but this should be tested with other compilers, computers and images. This would also allow using o...

9:59PM

LodePNG is also easy to use for small projects

9:59PM

https://lodev.org/lodepng/