DDraceNetwork

I just finished this map in 37mins and now it shows I did in 37s

same bug

03:29

I finish a solo map in 2:02

03:29

bug shows 0:02

03:29

wtf

we're just too pro

yeah hahhahah

same bug

@koko_awa

It seems like all the maps have become like this

04:02

I'm fixing it right now

1

I was rushing a map, and the time it took changed to the time you see there in the screenshoot

05:36

Here is the map, my real record is the 17 mins

it's due to this, the change has already been reverted https://github.com/ddnet/ddnet/pull/8221 (edited)

05:38

but your rank will still probably stay that way until an admin gets to it

05:39

ideally they can just filter out every rank from this window and then reverse the operation but i'm worried that the data loss would be too bad

05:39

you may just lose your rank

Ewan

you may just lose your rank

will the ranks be reset?

no idea

05:40

most likely

05:40

it's the safest way to gaurantee no fraudulent ranks lol

No, ranks are not being reset, all ranks will be multiplied by 50

nice

deen

No, ranks are not being reset, all ranks will be multiplied by 50

good job!

map : umbrella

11:15

14:52

Found a really weird bug. Not sure if it's known. When you change the map the tee seems to have some sort of "invincibility frames". If you change the map to the solo map "I will be back" for example, you won't get frozen by the freeze at the start and can just move freely while falling. In that map it would lead to a huge skip because you get rocket right at the start or more like BEFORE you even start (edited)

14:53

Doesn't seem to happen if you kill or join the map, only if you change the map

14:54

Also a friend figured out that it had something to do with the renderer. Apparently only happens with the Vulkan renderer

Pennervomland

Also a friend figured out that it had something to do with the renderer. Apparently only happens with the Vulkan renderer

tested and it works on opengl

Cipy29

tested and it works on opengl

friend had opengl first and changed it to vulkan. with opengl it didn't work for him, with vulkan it did. Weird that it works for you xD but maybe doesn't have anything to do with the renderer after all

Pennervomland

Also a friend figured out that it had something to do with the renderer. Apparently only happens with the Vulkan renderer

maybe one renderer simply initializes the map more quick but generally rendering code is unrelated

@Robyt3 Could this be similar to https://github.com/ddnet/ddnet/pull/8201 but for server code?

Fix tee briefly appearing at previous position when joining by Roby...

Previously, when connecting to servers repeatedly, the local tee would appear at the position it had on the previous server for a short time when joining. This (and potentially other bugs) are fixe...

deen

@Robyt3 Could this be similar to https://github.com/ddnet/ddnet/pull/8201 but for server code?

problem is he claims u can cheat with it

15:26

so it must be server side

15:26

if true, that would invalidate basically all of ddnet lol

Hmm, maybe the game world is not reset immediately when changing maps, I can take a look later

upon further investigation, i found out that this bug doesn't actually work, false alarm

7

15:36

on an unrelated note, i got really good at speedrunning

15:36

7

@admin time to shutdown all servers

15:51

15:53

@Zwelf would that be easy to detect ? (https://discord.com/channels/252358080522747904/757720336274948198/1229806559765205002)

15:53

or is teehistorian potentially also unaware

hopefully i didnt add this problem, i remember i once changed smth about the init stuff

yes, okayish easy to detect. (needs physics implementation to detect behavior) but I wanted to improve my tooling anyway xd

The bug seems to depend on if the client sends NETMSGTYPE_CL_STARTINFO with the MSGFLAG_FLUSH flag, if the client doesn't send it with that flag they won't be able to reproduce the bug. I git bisected it to https://github.com/ddnet/ddnet/pull/6823.

furo

The bug seems to depend on if the client sends NETMSGTYPE_CL_STARTINFO with the MSGFLAG_FLUSH flag, if the client doesn't send it with that flag they won't be able to reproduce the bug. I git bisected it to https://github.com/ddnet/ddnet/pull/6823.

but it cant be client side xD