DDraceNetwork - developer

@deen wtf why do u delete atlas skin?????

atlas by whis o.O

yo @Soreu or anyone else any idea how to better organize graphics? Currently working on mapres and skins collections and its so hard to find duplicates.

09:03

People use different names for same graphics

09:03

and even sha1sums dont match becuase they are resaved and exported differently etc

09:03

especially for skin names its hard to decide which skin name is the "real" one or who is the author

I kept some under categories of style, if it could be named that way. Like "mountains", "Asian style", "jungle", etc

All these things also gameskins fly around on discord get edited and everyone changes names and format there is no real source of trust

for skins I think best would be by author, though this one is hard to find out after longer time

ye

09:05

well mapres names are not too important but still annoying to find duplicates

same with gameskins, it's rather uncategorisable

yea

especially after someone changes just color of hook

I have my mapres somewhat organized https://github.com/TeeworldsDB/mapres

TeeworldsDB/mapres

A sorted collection of teeworlds mapres. Contribute to TeeworldsDB/mapres development by creating an account on GitHub.

09:06

oh also @Soreu feel free to search dupes add credits rename mapres etc if you feel motivated contributing to this database

09:06

since you are also the author of many graphics there

ye, just found my spikes_colored though I remember that one under GSP-something name xd

yes you used a lot of GSP names idk what GSP stands for and i just ignored the mapres i already had under a diff name

09:09

i have no clue what names are best

tbh imo it would be better if all those graphics could be in more than one category, but for that there would be need of website, so you wouldn't have them duplicated in different directories

yes

09:10

there is https://teeskins.de/ currently

Teeworlds Skins, Mapres, Gameskins, Emoticons, Particles, Cursors, ...

Download & Share: Teeworlds Skins 0.6 & 0.7, Assets, Tools ▷ +2000 Teeworlds Graphics ✓ Biggest Teeworlds Skins Collection ✓ Join Now ✓

GSP stands for Get Some Popcorn (it will take a while) and was a 2-ppl clan xD weird and lovely times

09:11

ye, but kinda nobody checks whatever is uploaded to teeskins afaik, there have been a lot of crap & duplicates last time I browse thru it

09:11

maybe it changed since then

Yea my system is kinda bad because its hard to decide where to put it. I try to bypass that issue by having a metadata txt containing tags.

@ChillerDragon you could try to do some similarity comparison between the tilesets to find out what is a duplicate and what is not

Can we add a config that you can see who spec you xD? (edited)

who would enable this config? the person speccing or the person being specced?

I want to know who spec me

@heinrich5991 yea actually planned that i didnt look into it but do you have a tool already that you can recommend for that?

no, don't know any

k

@Trafalgar Law but maybe the people speccing you don't want to announce it

just update to 0.7

1

1

14:33

it has this feature

but there are a lot of results for similarity hashing if you search for it

yea sure

14:58

Im also not sure what to do about automapper rules and working files

14:59

like i have all the mapres.pngs but sometimes there are also .svg and .xfc and .psd

14:59

feels like chaos what i do :D

attach them to the image I guess?

I have idea for new feature

15:35

Cinematic mode

got a license for almost 3/4 of ddnet skins already

15:35

@TortiLeq what is that?

@heinrich5991 we don't use the l word anymore, please be mindful of our softer, more delicate tees

Something like when you spectating person in ath. But it can be option in hud settings. Cinematic mode can be used in recoding speedruns etc.

is it a smoother camera?

15:37

or what does it do?

Oh, now that we have bezier curves, a smooth camera shouldn't be too hard to implement

Wait i can show you but first gotta download ath to record it

@heinrich5991 you only implemented a 1d bezier curve though, right?

you'd have one for x and y separately

Though I guess we only need it to drive progress, the fact that tees move in small increments should make it look smooth enough even if we just go towards the tee in a line

15:44

Hm, but we do have teleports and tees can go very fast

15:45

not really sure how one would avoid the sharp changes in direction

except for teleports you don't really have sharp changes in directions, right?

15:45

ah, and ninja

some way to limit how much the direction vector can change per tick?

for the smoothing or for the tees?

for the smoothing

15:47

the direction vector of the cameras movement

Something like that smooth movement: https://drive.google.com/file/d/1lJzAB-zjt59dkIRP6DOuChpb1DzasFZu/view?usp=sharing

AllTheHaxx 22-07-2020 17-44-19.mp4

1

15:48

For new option in hud settings

yea, looks like simple smoothing

But for default its turned off

huh, that doesn't look very smooth even

we could add that

Oki

We should do better, ease in ease out looks better than linear interpolation

is that linear?

It looks linear to me

15:50

Though the duration of the switch seems to be constant, so far away switches happen realquick

15:51

The path looks linear though

15:51

Well both the path and the interpolation look linear

15:54

I guess the path being linear isn't too much of an issue except for the corner cases of ninja and teleport

one would probably want to center the tee somehow

?

I don't have an immediate on how to do that

15:59

idea*

The final position of the movement will be when the target tee is centered

but that would mean that it's always behind

not if we make sure the camera is moving faster than the tee

16:04

The first derivative of the curve should be target_v + default

16:05

hm I wonder how hard that would be to solve for though

@Learath2 got an idea how I can go from 0-255 to -128-127

16:59

context python

16:59

0 to 0, 127 to 127, 128 to -128, 255 to -1

eh?

I guess I could make a function, just a question if there's a nice formula

just subtract 128?

that maps 0 to -128, not 0 to 0

Oh, hm

17:03

255 to -1 sounds like a weird transform

it's like reinterpreting an unsigned byte as a signed char

Oh, a twos complement signed char?

yes

Hm, I don't see a nice clean formula