DDraceNetwork - bugs

DDraceNetwork

Development / bugs

Report client & server bugs, especially but not limited to release candidates.

Between 2022-05-31 00:00:00Z and 2022-06-01 00:00:00Z

For a map if you use the tile to change the number of jumps a tee can have, that block cannot be toggled with a switch despite being in the switch layer, changing the switch number for it is useless

that's not a bug, that's by design

01:11

switch number is used to select number of jump

Chairn

switch number is used to select number of jump

the delay is used for the number of jumps, the switch number is useless and should be used to toggle it like its used for everything else

that's a switch number 2 that sets jumps to 1 but it doesnt turn on or off from switch it just always is active

Image attachment

@Mr.Gh0s7 do you have more logs where that line came from?

08:14

what were you doing when that happened?

I tried to crash again with debug enabled running from gdb but it didn't crash yet.

08:15

Random moments from what I could tell (DDRace, Fng etc)

Akeno

but im dont using vpn

send me your public ipv4 address (google it) in a DM

08:29

@Mr.Gh0s7 are you building yourself or using our version?

ddnet vanilla from aur and then building with cmake -DCMAKE_BUILD_TYPE=RelWithDebInfo ..

08:31

both crashed

Maybe Arch Linux or your makepkg.conf config enables this assertion and we don't do it in the official build, thus we never noticed

I see

You don't need to run it with gdb btw, after it crashed there should be a coredump

ah right, check coredumpctl

08:34

coredumpctl @Mr.Gh0s7

I use artix linux, which doesn't use systemd

it should still store the core dumps somewhere like /var/lib/coredump or current dir

@Mr.Gh0s7 check the output of cat /proc/sys/kernel/core_pattern

08:42

it tells you where the coredumps go

this gives core

then you should find them in the current directory

08:47

from where you started DDNet

08:47

named core.pid or so

I can't find any

ah, unfortunate

08:49

it's additionally limited by some more configs

08:49

you could check ulimit -c

08:49

it shows the largest possible coredump size

it outputs 0

08:54

I managed to have a /bin/sh coredump

ah, if it outputs 0, it means that coredumps are completely disabled (size limit of 0)

08:56

I guess you'd have to modify that somehow; I don't know how though, sorry

I will look it up thanks

for me, ulimit -c shows unlimited

ulimit -c unlimited sets it to unlimited

only for the current shell, I guess

maybe to it's children also? (edited)

yes

09:00

but not to other shells etc.

09:00

it's nice to have this as a system-wide setting so that you can see why software crashed

I can set it at login time

09:02

Yes it works this way.

09:03

Now if it crashes again I'll have a coredump so we will have more info to go on

backtrace from corefile is this

#0  0x00007fa86c45a36c in ?? () from /usr/lib/libc.so.6
[Current thread is 1 (Thread 0x7fa85cc88000 (LWP 18118))]
(gdb) bt
#0  0x00007fa86c45a36c in  () at /usr/lib/libc.so.6
#1  0x00007fa86c40a838 in raise () at /usr/lib/libc.so.6
#2  0x00007fa86c3f4535 in abort () at /usr/lib/libc.so.6
#3  0x00007fa86c792002 in std::__glibcxx_assert_fail(char const*, int, char const*, char const*)
    (file=<optimized out>, line=<optimized out>, function=<optimized out>, condition=<optimized out>)
    at /usr/src/debug/gcc/libstdc++-v3/src/c++11/debug.cc:60
#4  0x000055ce0cfae200 in CStun::Update() ()
#5  0x000055ce0cfa80c5 in CNetClient::Update() ()
#6  0x000055ce0d042f1b in CClient::PumpNetwork() ()
#7  0x000055ce0d043544 in CClient::Update() ()
#8  0x000055ce0d044b99 in CClient::Run() ()
#9  0x000055ce0cf5b33e in main ()

Skystrife

that's a switch number 2 that sets jumps to 1 but it doesnt turn on or off from switch it just always is active

do you think it would be usefull to make the jump tiles switchable xD

Mr.Gh0s7

backtrace from corefile is this

#0  0x00007fa86c45a36c in ?? () from /usr/lib/libc.so.6
[Current thread is 1 (Thread 0x7fa85cc88000 (LWP 18118))]
(gdb) bt
#0  0x00007fa86c45a36c in  () at /usr/lib/libc.so.6
#1  0x00007fa86c40a838 in raise () at /usr/lib/libc.so.6
#2  0x00007fa86c3f4535 in abort () at /usr/lib/libc.so.6
#3  0x00007fa86c792002 in std::__glibcxx_assert_fail(char const*, int, char const*, char const*)
    (file=<optimized out>, line=<optimized out>, function=<optimized out>, condition=<optimized out>)
    at /usr/src/debug/gcc/libstdc++-v3/src/c++11/debug.cc:60
#4  0x000055ce0cfae200 in CStun::Update() ()
#5  0x000055ce0cfa80c5 in CNetClient::Update() ()
#6  0x000055ce0d042f1b in CClient::PumpNetwork() ()
#7  0x000055ce0d043544 in CClient::Update() ()
#8  0x000055ce0d044b99 in CClient::Run() ()
#9  0x000055ce0cf5b33e in main ()

thats the same trace I already got @heinrich5991 that I also showed you once

seems m_NumUnsuccessfulTries just overflows

10:50

ah no

fixed btw

3

c0d3d3v

do you think it would be usefull to make the jump tiles switchable xD

I am wanting to use it in a map that I’m making but have to use an annoying work around with a door and another jump block

Skystrife

I am wanting to use it in a map that I’m making but have to use an annoying work around with a door and another jump block

First we have to check if any map set the number to something and used the corresponding switch...

c0d3d3v

First we have to check if any map set the number to something and used the corresponding switch...

Oh yeah that’s completely fair I didn’t really think about that

hey i keep getting a bug where i cant click anything on the pause menu

20:46

keyboard and mouseover highlight works just clicking is broken

20:46

clicking in-game also works, just not on menus

Ambidextroid

hey i keep getting a bug where i cant click anything on the pause menu

I've had this before too, I'm not sure what causes it but I have to close and re-open the client to fix it

i have some weird bug where entering some areas of a map will immediately close my lan server (on nightly) (edited)

can you be more specific?

23:44

what map ? Are you triggering a tile or just getting close to a particular zone in the map?

23:44

Can you debug it?

23:45

If you're on windows, there should be a crash report under data/dump or dumps or crash (edited)

23:45

@Ambidextroid @Skystrife what OS are you on?

windows, I'll try to get logs next time I get the bug if it happens again, it hasnt happened in a while

there's probably not much in the logs im afraid

Chairn

what map ? Are you triggering a tile or just getting close to a particular zone in the map?

it was just a gores map, getting anywhere within a certain zone (not triggering any switches, etc) would instantly close the console

23:48

Error occurred on Tuesday, May 31, 2022 at 18:11:18. (with offset 0000000000400000)

DDNet-Server.exe caused an Access Violation at location 000000000042D187 in module DDNet-Server.exe Reading from location 0000000000000408.

sounds like an assertion might be triggered

23:48

ah no, segfault

23:49

is there a crash report available, im not sure we use them for server

yeah that was it basically

23:49

and a lot of other numbers in it

can you send me the map?

im not sure cus its someone elses

well, can't fix the bug if we can't reproduce it

i could try to find the pr or whatnot that caused it

you can try running the server with debugger attached so that we get a stacktrace

23:50

backtrace sorry

23:50

backtrace is often enough to find what's going wrong

23:50

and it tells us where exactly the bug happened

ill try

@louis i just checked, crash reports are enabled for server as well

23:57

can you take a look into data/dumps or %appdata%/teeworlds/dumps ?

Exported 85 message(s)