DDraceNetwork - developer

heinrich5991

[ddnet/ddnet] Pull request closed: #870 Made int packing functions safe

00:36

heinrich5991

[ddnet:master] 1 new commit

43f246f Made int packing functions safe - Redix

def-

[ddnet-web:master] 1 new commit

b62836f upd - def-

anyone know how to open a .demo file outside of tw

12:42

like program or anything to open it

I don't think there's another program that can open them

12:42

what do you want to do?

yt

you want to put a demo onto youtube?

y

your best bet is to open the demo in the client and record it

12:43

I think

ok

(maybe there was a client that could actually output a video file, but I don't remember it)

hmh made a client that can convert demo to mp4

12:47

https://github.com/ddnet/ddnet/pull/546

Demo to video converter by H-M-H · Pull Request #546 · ddnet/ddnet

This adds a way to convert demos into mp4 (h.264) videos using ffmpeg-libraries. This feature is optional and defaults to off. It can be included by setting ffmpeg.value = true in config.lua. Demos...

12:51

why wasn't this merged btw

read the discussion I guess

there is no reason :/

sound didn't work, apparantly

12:52

from my reading

ah okay, kinda letdown tho

12:59

can't vote/change maps on test server, neither with rcon nor with vote call (edited)

I just closed it because nothing happened for many months

14:53

but I'd still accept that PR once it works fine

maybe build that tool as a headless client or sth, I feel like we have a shit ton of libraries now

1

heinrich5991

[ddnet/ddnet] Pull request opened: #871 Teehistorian

teehistorian records all inputs from the players as well as the player positions in each tick. It stores this info in a highly compressible output format (I've achived 5x compression using xz or bz2).

coool

22:08

have you recorded anything ?

I have recorded me playing Kobra 3 Solo for 10 minutes

22:09

that yielded a conservative approximation of max ~2GB per day on all DDNet servers together

nice

22:09

have to open/run it ?

22:09

from client or ?

we concluded that this would be feasible to store

22:09

I don't understand

how *

oh

22:10

the server records the stuff

22:10

"sv_teehistorian 1" into the config

22:10

might need to create the directory ~/.teeworlds/teehistorian

22:10

(for client side stuff, you have the ghosts)

data is saved similar to ghost/demo? Or rather like csv ?

22:11

im just curious

it's stored in a binary format that I just invented :/

oh

it was necessary to bring down the file size

understandable

https://github.com/ddnet/ddnet/pull/871#issuecomment-330098227 make ddnet teleporter deterministic again!

Teehistorian by heinrich5991 · Pull Request #871 · ddnet/ddnet

teehistorian records all inputs from the players as well as the player positions in each tick. It stores this info in a highly compressible output format (I've achived 5x compression using xz or bz2).

22:24

please comment with your ideas

22:24

@Savander ^

I don't know if I am the proper guy to spread my word in that topic ( don't know teeworlds logic at all). But if I had to choose, I prefer to have everything 1-0. I mean, without randomness

if you have a github account, please comment there ^^

22:31

your input is appreciated

Yea i have, but im web dev

22:31

instead of software ;D

you're a dev, that's enough. but I'd have no problems with opinions of players either

22:32

anyway, so which of these options would you prefer?

22:32

only 1?

Don't know how spawn logic works

you get the spawn at the position that is the furthest away from any other player

oh, right

22:34

like on CTF/DM. I should recognize that.. I'm using spawn logic while playign dm1

^^

To predict where play will respawn

22:35

anyway..

Give me a second

Meh, i had idea

22:50

then i started writing and i forgot

22:50

:V

thanks for commenting on the issue!

No problem

22:55

btw. I have to say that I started learning about graphics processing (or more started reading). By using OpenGL, huh. Sounds pretty hard

interesting ^^

VBO, VAO

I don't have much experience with that so far

I never know what does it mean

22:55

now I at least, know abbreviation

https://en.wikipedia.org/wiki/VBO

VBO

VBO may refer to: Vainu Bappu Observatory Vertex Buffer Object, an OpenGL extension for faster rendering of triangles Verbond van Belgische Ondernemingen, the Federation of Belgian Enterprises

https://en.wikipedia.org/wiki/VAO

VAO

VAO may mean: Eastern Administrative Okrug of Moscow, Russia Vertex Arrays Object, an OpenGL feature similar to Vertex Buffer Object (VBO) Vienna Art Orchestra VAO (sports club), a Greek sports club

And, If I understand! You can correct me. VBO, saves positions of vertices

wikipedia helps with abbreviations

and it's calculate position once

22:57

in space

22:57

then it's used with GLSL shaders

22:57

to render graphics

22:58

In a nutshell

22:58

(hoho, i learnt new sentence !!

0

I always was like, wtf

22:58

"nutshell"?

22:58

what they want from them..

^^

About 3 option. It would be reset if no one take any output?

no

23:03

it would be resetted on map reload

It wouldn't be better, if it would be resetted if there's no people around output? (edited)

that sounds more complicated, so it should have an advantage

23:05

did you think of any?

Yes

23:05

For example, if you are speed racing, then you will have always

23:05

same output

23:06

Or maybe it could be resetted

23:06

for teams?

teams have their own world state

oh, like that. Didn't know

23:07

there's any big changes in performence and complexity between 2 and 3 ?

2 is probably less complex than 3

23:07

because we don't have to carry around state

Okay, then i would stay with second option.

Even without my idea described in issue

23:09

about computing first time of using teleport/anything else

23:09

In map or something, it was just idea

I don't quite understand that idea

I will try to describe it better

as I understand it, the furthest teleport output can only be computed at race time

23:11

because you don't know the other player's positions beforehand

Hmm, I started thinking more. About 2 and 3 option. Generally, the first idea in my mind about 2 option, was to make first output for a map static. For example, if one guy from team jump into teleport, he will be teleported once to the same place, every first time. (every reset). But after thinking a bit more, even if third option is more complex, in meaning of fair, it would be best option. Because, while playing you will get the same output (depends on world state)

23:15

so even prediction would works fine, im right ?

no

23:15

the client doesn't know how often the teleporter has been entered

Funny but some days ago I had a situation where the tele randomness helped my team to continue a race

but if he would know, then it would work yes?

23:16

jao, yea. Randomness can be a cure also a curse

what do you mean, it helped?

and while speed racing, you need to have "luck". And i bet, there are maps

23:17

which can be finished faster, with enough luck

There were 3 TO teles and we could only survive if 2 of 3 of us get tped to the same TO tele to tower above freeze

23:22

You probably didn't think about it, but options 1 and 2 can be deadly for solo maps that have several TO teles for the same tele. If you get always tped to the furthest away now, records could become impossible to beat

furthest away from player

from other players

23:23

since you're solo, there is no one

23:23

but having to rely on randomness for good ranks doesn't sound good either

23:24

in the best case, the maps would just get fixed

Always use the first output teleporter.

to only have one TO teleporter

this can be the furthest away

In my opinion, is houdl rely on world state ( which is reseted for teams ). So, it would be always the same (edited)

well, furthest away also relies on world state.

fixing all maps should work, just wanted to mention it

it would be computed by how many times certian teleport "teleported", or in different way ?

yes, just a counter for each teleporter

23:27

(for each teleporter number)

Okay, we have to think about it a bit more. Let's give time

good idea. good night

good night