DDraceNetwork - developer

DDraceNetwork

Development / developer

Development discussion. Logged to https://ddnet.org/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 2024-03-07 00:00:00Z and 2024-03-08 00:00:00Z

Matodor

Click to see attachment 🖼️

can you check for me?

pie

the entire snap chunk payload:

\x9A\x51\x9B\x51\xB3\x92\xAD\xB9\x0B\x9F\x01\x00\x03\x00\x06\x00\x00\x00\x00\x00\x14\x00\x00\x01\x0B\x00\xDA\xC8\xF0\x91\x02\xCC\xB0\xD0\xB1\x02\xDA\xE8\xD8\xF0\x0B\xFF\xFF\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFF\xFB\xF7\x0F\x40\xDE\xE4\xD0\xB1\x03\xFF\xAD\x98\xA1\x01\xFF\xFD\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFF\xFB\xF7\x0F\x00\x80\xFE\x07\x80\xFE\x07\x0A\x00\x01\x00\x00\x00\x00

only the snap payload (content after size bytes):

\x00\x03\x00\x06\x00\x00\x00\x00\x00\x14\x00\x00\x01\x0B\x00\xDA\xC8\xF0\x91\x02\xCC\xB0\xD0\xB1\x02\xDA\xE8\xD8\xF0\x0B\xFF\xFF\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFF\xFB\xF7\x0F\x40\xDE\xE4\xD0\xB1\x03\xFF\xAD\x98\xA1\x01\xFF\xFD\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFF\xFB\xF7\x0F\x00\x80\xFE\x07\x80\xFE\x07\x0A\x00\x01\x00\x00\x00\x00

00:04

since teeframe is currently only a server implementation, it only encodes snaps, it cant decode, so I think I cant help much more than this (edited)

Miguilim

the entire snap chunk payload:

\x9A\x51\x9B\x51\xB3\x92\xAD\xB9\x0B\x9F\x01\x00\x03\x00\x06\x00\x00\x00\x00\x00\x14\x00\x00\x01\x0B\x00\xDA\xC8\xF0\x91\x02\xCC\xB0\xD0\xB1\x02\xDA\xE8\xD8\xF0\x0B\xFF\xFF\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFF\xFB\xF7\x0F\x40\xDE\xE4\xD0\xB1\x03\xFF\xAD\x98\xA1\x01\xFF\xFD\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFF\xFB\xF7\x0F\x00\x80\xFE\x07\x80\xFE\x07\x0A\x00\x01\x00\x00\x00\x00

only the snap payload (content after size bytes):

\x00\x03\x00\x06\x00\x00\x00\x00\x00\x14\x00\x00\x01\x0B\x00\xDA\xC8\xF0\x91\x02\xCC\xB0\xD0\xB1\x02\xDA\xE8\xD8\xF0\x0B\xFF\xFF\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFF\xFB\xF7\x0F\x40\xDE\xE4\xD0\xB1\x03\xFF\xAD\x98\xA1\x01\xFF\xFD\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFD\xFB\xF7\x0F\xFF\xFF\xFB\xF7\x0F\x00\x80\xFE\x07\x80\xFE\x07\x0A\x00\x01\x00\x00\x00\x00

since its a snap single and not a slice, you sure have the removed items, updated items, and unused 00 at the beginning

00:04

so its 0 removed items

00:05

3 updated items

00:05

00 - unused

00:05

since there is no removed items, all snap items sent have it item id, id, and payload

00:06

first item is 6 - obj_game_info

00:06

id is 0 - as obj_game_info id is always zero

Miguilim

deltaTick 5211

since delta tick in this case is current tick - (-1), there is no confirmed snap known by the client (delta snap) (edited)

00:09

so this snap single payload contains all snaps items and game context the client should be aware of

Miguilim

since delta tick in this case is current tick - (-1), there is no confirmed snap known by the client (delta snap) (edited)

and because of that, all values are the real values and not the diff from the last delta snap

[ddnet/ddnet] Pull request closed: #8072 List all practice commands when enabling /practice

00:32

github-merge-queue[bot]

[ddnet:master] 2 new commits

209df98 List all practice commands when enabling /practice - furo321 1afdf47 Merge pull request #8072 from furo321/practice-helptext - heinrich5991

[ddnet-maps:master] 1 new commit

e09bdfa A A - ddnet-maps

5

1

A

chillerdragon: how can i rebase my teeworlds branch to latest community, am little dum =]

Image attachment

07:45

its llvm 18 time

Ryozuki

its llvm 18 time

owo

xar

@Jupstar ✪ i know ur afk from discord but i enabled lto on gentoo

08:51

i forgot to do so

08:51

i also enabled o3 on some packages

now u have a fast

oh i just realized thats a bit annoying since its a pr upstream as well then you break your teeworlds pr if you merge it into community

Replying to @MilkeeyCat chillerdragon: how can i rebase my teeworlds branch to latest community,…

08:59

Maybe just pr against community directly from a latest version.

git remote add communism https://github.com/teeworlds-community/teeworlds/
git fetch communism
git checkout community
git checkout -b your_epic_branch_name

then add changes there and pull directly here https://github.com/teeworlds-community/teeworlds/pulls

Pull requests · teeworlds-community/teeworlds

A community fork of teeworlds. Trying to have more active maintainers and merge prs faster. - Pull requests · teeworlds-community/teeworlds

08:59

i might have the mirror bot do that actually to fix the pipeline problem for all branches

09:00

so annoying that the merge queue has to pass the pipeline before merge. Does not make much sense to me. The only relevant state is after merge.

09:02

btw you dont have to copy paste changes you can just run this git cherry-pick b855a9db2cf0cd8fa6e5f56c68c78a0667d579d5 on the new branch and push

do i have to be on teeworld or community-teeworlds? justatest

justatest

what is "be on"

09:03

you can do that from your teeworlds/teeworlds fork its fine

ok

you can do inter fork cross entanglement pullrequests

oh interesting so my parser falsely detected it as map download hmm thanks a lot

Replying to @Miguilim second chunk is a snap single , so no sequence number

10:00

no time for that now and will forget later lol. But you can pr or issue all my infra is open source this should be located at zillyhuhn/server-config on github

Replying to @Matodor ChillerDragon: when any allow-origin? […]

10:00

pls open an issue if you found a bug they get lost in this chat

Replying to @zhn ChillerDragon: two bugs we found

<ChillerDragon> > can you give an example of when you'd use multiple cursors?

11:14

<ChillerDragon> @heinrich5991 idk where to start so many times i need multiple cursors

11:14

<ChillerDragon> for example editing this piece of code

11:14

<ChillerDragon> https://zillyhuhn.com/cs/.1709810001.png

Embedded image

11:15

<ChillerDragon> or copy pasting a wall of text and transforming its structure

What would u want to edit in these lines?

<ChillerDragon> or doing this

11:16

<ChillerDragon> https://zillyhuhn.com/cs/.1709810175.png

Embedded image

11:17

<ChillerDragon> @MilkeeyCat doesn't matter usually if you want to edit something in those lines its the same for every line. So if you can do word jumps and have a separate clipboard and word boundaries for everyline its amazing

11:18

<ChillerDragon> turn yaml to json etc

ws-client

<ChillerDragon> > can you give an example of when you'd use multiple cursors?

Ws Client broken again, reply didnt Work

<ChillerDragon> what didnt? work?

Up is working - down is Not

Image attachment

<ChillerDragon> one is a reply from matrix the other is from irc which has no replies

11:25

<ChillerDragon> @MilkeeyCat https://tube.zillyhuhn.com/video.php?t=vscode_multi_cursor.mp4&u=chiller if the function signature changes this could be an example

11:26

<ChillerDragon> with multicursor anything tedious becomes fun

11:26

<ChillerDragon> s/anything/manything/ yes thats a word

ws-client

<ChillerDragon> https://zillyhuhn.com/cs/.1709810175.png

11:31

chillerdragon

<ChillerDragon> no doubt you can do every multicursor vscode action in vim

11:33

<ChillerDragon> but it looks more complex to type and also more complex to reason about

11:34

<ChillerDragon> id claim vscode is fast because it requires less keystrokes and less mental capacity

11:34

<ChillerDragon> another common use case is this one @heinrich5991

11:34

<ChillerDragon> https://tube.zillyhuhn.com/video.php?t=vscode_multi_cursor2.mp4&u=chiller

I kinda like writing macros

11:35

Its fun xd

11:35

But not fun when ut doesn't work first try

<ChillerDragon> yes multicursor is very forgiving when it comes to mistakes

11:36

<ChillerDragon> undo and corrections and tinkering is smooth

11:37

<ChillerDragon> if its only 2 or 3 lines its so nice because the overhead of multicursor is non existant

<ChillerDragon> mostly there is a few twists so one fix macro usually doesnt catch all edge cases

11:46

<ChillerDragon> but a few different multi cursors do

11:47

<ChillerDragon> json to yaml

11:47

<ChillerDragon> https://tube.zillyhuhn.com/video.php?t=vscode_multi_cursor3.mp4&u=chiller

11:47

<ChillerDragon> i can do this all day

Send json

santatrollet

<ChillerDragon> {"name":"John","age":30,"car":null,"foo":"bar","baz":"bang","nob":"cat","skill":420}

11:52

<ChillerDragon> @MilkeeyCat we should compete in some codewars site just reformatting code xd

11:52

<ChillerDragon> is that a thing?

if its not, it can become a thing xd

<ChillerDragon> for sure i could see this being fun for 5mins getting random snippets pvp'ing a friend first to refactor every tool allowed

chillerdragon in the 0.7 coding competition (he won because hes the only person who ever codes 0.7)

<ChillerDragon> pastes into chatgpt5

<ChillerDragon> lmao jq

11:56

<ChillerDragon> nice one

11:56

<ChillerDragon> again prime enjoyer detected hehe

Its pretty useful thing

<ChillerDragon> nice subst

11:58

<ChillerDragon> ye u pro

11:58

<ChillerDragon> but id still claim im faster :p

11:58

<ChillerDragon> i guess we never know

You are faster because you are pro

12:00

And me nub

<ChillerDragon> skill issue?

I dont do that kind of refactoring often

12:00

But yea, skill issue

<ChillerDragon> but using ! is bit chot

12:03

<ChillerDragon> lemme do speedrun using ! xd

U can use bang in vscode

<ChillerDragon> https://tube.zillyhuhn.com/video.php?t=nvim_moment.mp4&u=chiller

U can do this in vim

<ChillerDragon> if u need ! to be effective in vim you are basically bash enjoyer not vim enjoyer axaxax

why would i spend 20 mins writing ugly ass regex when i can just use ! thonk

thonk

<ChillerDragon> xd

12:08

<ChillerDragon> agree

12:08

<ChillerDragon> i am big bash enjoyer and its no secret

12:08

<ChillerDragon> in bash you dont even need !

I saw Minecraft server in bash vid

12:09

From prime

12:09

kek

<ChillerDragon> ye me too

That thing is so cursed justatest

justatest

<ChillerDragon> made me think about minetee

12:09

<ChillerDragon> i should learn mc protocol before discussing w heinrich again :D

ws-client

<ChillerDragon> https://zillyhuhn.com/cs/.1709810175.png

ah. in vim, I usually use macros for these

I remember they use 9 digits after . For position in one block

<ChillerDragon> in prime vid the blog mentioned 4 bit per block xd. And the protocol we plan in #7666 has 41 BYTES per tile LMAOOOO

12:11

<ChillerDragon> https://github.com/ddnet/ddnet/issues/7666

Add netmessage to update tiles in the map · Issue #7666 · ddnet/ddn...

Not sure how exactly it should look like. I quickly came up with a poc that looks like this NetMessageEx("Sv_ModifyTile", "modify-tile@netmsg.ddnet.org", [ NetIntAny("m_X&q...

12:11

<ChillerDragon> if a minetee creeper explodes the removed tiles are basically ddos xd

ws-client

<ChillerDragon> https://tube.zillyhuhn.com/video.php?t=vscode_multi_cursor2.mp4&u=chiller

ah, looks nice. thanks for the example

I didn't find a website where u can compete with others by refactoring code, and it goes into my todo project list

<ChillerDragon> nice

But before that there're two ddnet related projects i have to finish first XD

MilkeeyCat

But before that there're two ddnet related projects i have to finish first XD

demo webviewer first pls

Matodor

demo webviewer first pls

Did u see that giga cool laggy map? xd

1

12:48

I made a feature to render object containers aaand it just silently doesn't work

12:48

No errors, just empty screen

12:48

Poggers

Matodor

demo webviewer first pls

https://ddnet.github.io (edited)

Oh ye, my demo viewer is completely useless

archimede67

https://ddnet.github.io (edited)

12:50

need pure js project

Lmao

monkalaugh

✖️ client wrapper ✖️

Then add server with ws and send requests from frontend

wow

12:54

there's a web demo render now??

bamcane

there's a web demo render now??

just a client ported to web

archimede67

https://ddnet.github.io (edited)

@Lumpy ◐ω◑ i did it

hm

12:55

i couldn't load it, there's only a white screen be displayed

MilkeeyCat

@Lumpy ◐ω◑ i did it

what is it thonk

thonk

Lumpy ◐ω◑

what is it thonk

thonk

Cant u read? ._.

MilkeeyCat

Cant u read? ._.

ye I read ddnet..

MilkeeyCat

I made a feature to render object containers aaand it just silently doesn't work

writing types for webgl calls from typless language is so much fun pepeW

pepeW

@Learath2 c question

15:20

how do u benchmark a function

15:21

clock_t begin = clock();

/* here, do your time-consuming job */

clock_t end = clock();
double time_spent = (double)(end - begin) / CLOCKS_PER_SEC;

15:21

is tihs good enough?

15:21

time.h

That should be fine but you probably want a high precision time rather than just any time

15:22

I would also do it a couple thousand times in a loop and divide

15:23

Maybe do it a dozen times or so before you start timing to help the branch predictor. Idk it all depends on how accurate you need to get

@Learath2 how can u make sure it doesnt get optimized out

15:23

the fn call

15:23

does it optimize extern fns?

Put it in a separate TU and it won't be optimized indeed

tu?

15:25

ah

15:25

the extern fn is from a static lib

15:25

my lib compiled from my lang

There is also a compiler attribute you can mark your function with but I don't remember

Ryozuki

the extern fn is from a static lib

It it's extern from a static lib you are fine. It can't be optimized out completely unless you use LTO

15:32

This sort of micro benchmark on the function level is not very useful unless you already have a theory as to how it will turn out btw

15:34

Because of emergent behaviour of compilers it's really hard to extrapolate from microbenchmarks to what happens in an overall program

ye i know

15:37

xd

15:38

@Learath2 tbh this is for the lang im making at work

15:38

the boss wants a microbench

kek

1

Oh btw, careful clock() iirc is the cpu clock. If there is any io or if your cores are contended you will get bad results

15:41

You want a quiet system, and you want to use a different source of timing if there is any io involved at all

gm

Or if one solution involves offloading the computation to something that is not the CPU

@Learath2 well its more like a regression bench

15:48

to know if we lose perf

15:48

i guess

Ryozuki

to know if we lose perf

The issue I was trying to highlight is that even if you do lose performance at this level, it might not change anything in the overall performance of your program. That's all, but I'm sure your boss already knows that.

yeye

Learath2

The issue I was trying to highlight is that even if you do lose performance at this level, it might not change anything in the overall performance of your program. That's all, but I'm sure your boss already knows that.

some stuff is more done for marketing than practical uses

16:01

too

#include <assert.h>
#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
#include <time.h>

extern uint64_t edlang_function(uint64_t n);
extern uint64_t rust_function(uint64_t n);

struct timespec timer_start() {
  struct timespec start_time;
  clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &start_time);
  return start_time;
}

long timer_end(struct timespec start_time) {
  struct timespec end_time;
  clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &end_time);
  long diffInNanos = (end_time.tv_sec - start_time.tv_sec) * (long)1e9 +
                     (end_time.tv_nsec - start_time.tv_nsec);
  return diffInNanos;
}

int main(int argc, const char **argv) {
  if (argc < 2) {
    fprintf(stderr, "missing iteration arguments\n");
    return 1;
  }

  if (argc < 3) {
    fprintf(stderr, "missing input number argument\n");
    return 1;
  }

  int num_iters = atoi(argv[1]);
  int input = atoi(argv[2]);

  uint64_t result_edlang;
  uint64_t result_rust;

  // warmup + sanity check
  for (size_t i = 0; i < 3; ++i) {
    assert(edlang_function(input) == rust_function(input));
  }

  printf("Running %d iterations\n", num_iters);
  printf("Using input value:\t%d\n", input);

  {
    struct timespec vartime = timer_start();
    for (size_t i = 0; i < num_iters; ++i) {
      result_edlang = edlang_function(input);
    }
    long time_elapsed_nanos = timer_end(vartime);
    printf("edlang Result =\t%lu\t\tTime taken : %.2Lf ms\n", result_concrete,
           (long double)time_elapsed_nanos / 1000000.0L);
  }

  {
    struct timespec vartime = timer_start();
    for (size_t i = 0; i < num_iters; ++i) {
      result_rust = rust_function(input);
    }
    long time_elapsed_nanos = timer_end(vartime);
    printf("Rust Result =\t\t%lu\t\tTime taken : %.2Lf ms\n", result_rust,
           (long double)time_elapsed_nanos / 1000000.0L);
  }

  assert(result_edlang == result_rust);

  return 0;
}

16:55

@Learath2 behold my C program

https://github.com/llvm/llvm-project/pull/83992

[MLIR] Add llvm (debug) attributes to CAPI by edg-l · Pull Request ...

This PR adds the following to the mlir c api: The disctinct mlir builtin attribute. LLVM attributes (mostly debug related ones)

17:14

another pr merged into llvm

17:15

this one way bigger

17:15

got already more pr ideas

Ryozuki

@Learath2 behold my C program

not enough ub

linux enjoyers, pls send help, why my sed not working Zzz_7062057890308424505

Zzz_7062057890308424505

paul@ubuntu-hp:~/test2$ cat random2.cfg | grep password
  password: "superduperpass"
paul@ubuntu-hp:~/test2$ sed -i 's/password:.{0,}/password:asd/g' random2.cfg
paul@ubuntu-hp:~/test2$ cat random2.cfg | grep password
  password: "superduperpass"

troll

2

try without -i

18:00

to see the result

18:00

without cat

18:00

ah well u grepping

18:00

hm try with -E idk

18:00

its been long

18:01

.{0,}

18:01

are u sure this is matching

18:01

just do .*

18:02

❯ sed 's/password:.*/password: asd/' a
password: asd

18:02

0 matches only 0 times

Ryozuki

just do .*

lol, worked smh

18:02

regex test website showed .{0,} works as well (edited)

18:03

thanks a lot

yes

18:03

it works

18:03

with -E

18:03

as i said

18:03

❯ sed -E 's/password:.{0,}/password: asd/' a
password: asd

18:03

-E enables fancier regex

18:03

sed is ancient

Ryozuki

-E enables fancier regex

justatest

not listed on man sed

18:03

oh ok it does

❯ sed --help | rg E
  -E, -r, --regexp-extended
                 (for portability use POSIX -E).
E-mail bug reports to: <bug-sed@gnu.org>.

xd

sed doesn’t usually work in regex mode

18:29

ex

18:29

regexex

Ewan

sed doesn’t usually work in regex mode

idk why i read it as

sex doesn’t usually work in regex mode

also true

xd

omg, that's so gooood

19:18

sadly Jupstar can't see all these fps feelsbadman

feelsbadman

MilkeeyCat

omg, that's so gooood

where map

no map

20:09

but there many many tiles

Image attachment

i'm trying to cross compile from fedora to windows, you guys ever got this error?

CMake Error at cmake/FindSSP.cmake:12 (message):
  could not find ssp paths
Call Stack (most recent call first):
  CMakeLists.txt:544 (find_package)

Cipy29

i'm trying to cross compile from fedora to windows, you guys ever got this error?

CMake Error at cmake/FindSSP.cmake:12 (message):
  could not find ssp paths
Call Stack (most recent call first):
  CMakeLists.txt:544 (find_package)

On Windows this error message usually means ddnet-libs are missing or not updated. I don't know how to get the ssp lib for cross compiling though.

I got it here /usr/x86_64-w64-mingw32/sys-root/mingw/bin/libssp-0.dll but it seems like it doesn't see it

Take a look at FindSSP.cmake

ye I tried fixit it like this (without comments)

#set(SSP_COPY_FILES
    #"/usr/x86_64-w64-mingw32/sys-root/mingw/bin/libssp-0.dll"

but it didnt do it

21:14

i was hoping someone got it before and someone knew quick fix xd, but if not, imma do more gpt4-ing xd

Did you try cloning ddnet-libs already and building with precompiled libs or does that not work?

nope, haven't tried, imma try

Robyt3

Did you try cloning ddnet-libs already and building with precompiled libs or does that not work?

haha, worked, thanks ❤️

troll

1

CMake Error (dev) at debug/googletest-src/googletest/cmake/internal_utils.cmake:239 (find_package):
  Policy CMP0148 is not set: The FindPythonInterp and FindPythonLibs modules
  are removed.  Run "cmake --help-policy CMP0148" for policy details.  Use
  the cmake_policy command to set the policy and suppress this warning.

Call Stack (most recent call first):
  debug/googletest-src/googletest/CMakeLists.txt:84 (include)
This error is for project developers. Use -Wno-error=dev to suppress it.

21:39

Any idea how to fix this in GH actions?

fokkonaut

CMake Error (dev) at debug/googletest-src/googletest/cmake/internal_utils.cmake:239 (find_package):
  Policy CMP0148 is not set: The FindPythonInterp and FindPythonLibs modules
  are removed.  Run "cmake --help-policy CMP0148" for policy details.  Use
  the cmake_policy command to set the policy and suppress this warning.

Call Stack (most recent call first):
  debug/googletest-src/googletest/CMakeLists.txt:84 (include)
This error is for project developers. Use -Wno-error=dev to suppress it.

#6913

https://github.com/ddnet/ddnet/pull/6913

Fix CMP0148 by heinrich5991 · Pull Request #6913 · ddnet/ddnet

https://cmake.org/cmake/help/latest/policy/CMP0148.html Checklist Tested the change ingame Provided screenshots if it is a visual change Tested in combination with possibly related configuratio...

thanks

[ddnet/ddnet] Pull request opened: #8073 Add validation for StrToInts and IntsToStr, move and rename

Rename StrToInts function to str_to_int32 and add strict validation. Because this function should only used with trusted internal strings, assertions are added to ensure that no invalid UTF-8 is being encoded as integers and that the string is not truncated. Some buffer sizes are adjusted accordingly, so truncation cannot happen. Rename IntsToStr function to int32_to_str and add lenient validation. An additional argument specifying the size of the output buffer is added to assert t...

@zhn you're such a chad sometimes

Image attachment

justatest

2

Image attachment

2024-03-07 22:51:10 I assert: /home/deen/isos/ddnet/ddnet-source/src/base/system.cpp(4587): Invalid UTF-8 passed to windows_utf8_to_wide
2024-03-07 22:51:10 I assert: /home/deen/isos/ddnet/ddnet-source/src/base/system.cpp(4587): Invalid UTF-8 passed to windows_utf8_to_wide

this one's interesting, what could've caused this? - our map has a graphical issue, but i dont think its related is it?:

2024-03-07 22:51:10 I Warning: The width of texture "embedded: 5" is not divisible by 16, or the height is not divisible by 16, which might cause visual bugs.

22:30

i got this sent to by a player, he said joining other servers/maps works fine - and we currently have 20 players without any issues

22:32

update: redownloading the map helps, but im still curious as to why that happens/happened

This was a bug a while ago where loading the map into the editor and then saving it would corrupt the images right?

meloƞ

update: redownloading the map helps, but im still curious as to why that happens/happened

Are you on the newest version?

oh it didnt happen for me, it was a player on our server, and redownloading the map fixed the issue - he was on latest release

Hmm okay

22:34

Weird

ye i couldnt connect the dots on this one :D

Maybe his network gave up on him a ripped a few packets outta there

meloƞ

oh it didnt happen for me, it was a player on our server, and redownloading the map fixed the issue - he was on latest release

If it happens again, copy the corrupted map and send it please. Invalid UTF-8 in image/sound names should already be checked with latest client though. Checks for invalid layer/group/envelope names will be added by the PR I just opened.

22:40

Did this happen on an official DDNet server?

it happened on a teeworlds mod with latest ddnet

Maybe it was the skin name related crash, which should be fixed on the client side by the PR above and on the server-side on DDNet and KoG servers already

hmm maybe, i'll drop the stuff in here when it happens again and i cant figure it out myself!

If you use Windows you should have a crash log (.RTP file), which should tell where the invalid UTF-8 comes from

Exported 260 message(s)