Teeworlds - bridge

yes. it is due to buggy drivers. the helptext of that config var is something like "might fix some color bugs on some drivers"

00:01

Well, @redix, does your "fix" with setting all vars 0.3f work?

ah okay. some drivers have buggy implementations for deprecated opengl stuff =\ for me it did when i wanted to render a transparent tee. premultiplying by alpha is used at other places as well: https://github.com/teeworlds/teeworlds/pull/1676/files

thanks redix, pre calculating worked

Do you think we should get the new color code from ddnet into 0.7? It's been working pretty decently and there hasn't been reports of any bug in a while

12:06

Also I'd like to get 2479# in sooner then later

Learath2: could you explain?

Dune: the colors?

Yes

https://github.com/ddnet/ddnet/blob/master/src/base/color.h

ddnet/ddnet

DDraceNetwork, a cooperative racing mod of Teeworlds - ddnet/ddnet

12:08

It's a cleanup of the color conversion stuff that happened at the point of use in the past and it clears up what kind of color a function expects

Oh that'd have saved me days

12:10

Yeah we need proper abstractions for RGB/HSL/HSLA/RGBA

Not sure what your stance would be on the C++ style casts I introduced though. e.g. color_cast<ColorHSLA>(Rgb)

I prefer C stuff as much as possible, but casts seems very appropriate here, easier to read, and not adding complexity

Ah, there is also another concern there, we lighten tee colors so the colors saved in the config aren't actually the colors, there is some complexity introduced when reading those, hence the Lighten and Darken functions there. With ddnet I was trying to keep behaviour the same, but maybe we could come up with something better for vanilla

Do we have a way to store CNetChunks?

13:45

s/way/data structure/

yo perl hackers do you know what s/foo/bar/xg; does? I assume its like in sed being a replace and 'g' is replacing all matches in a line but what does 'x' do?

perl seems dead, but seemingly isn't as there are lots of perl jobs available. but I doubt any of the young'uns here do perl, grandpa chilli.

18:29

is there any proper explanation or documentation how the map format is constructed? or anyone's brain that is willing to explain it or a part of it?

Learath2, is diffie-hellman really mitm resistant?

Yep

the server could replace all the packets during the key agreement

Then you couldn't read the messages

19:25

As long as we both can read the messages, the exchange took place and no one else has the key

19:25

Oh, you could actually establish 2 keys and relay messages in between

19:26

okay then it's not safe :P

yeah

19:26

you need accounts

We need authentication somehow

so not possible with the current infrastructure :/

I have an implementation of accounts working with private keys, it's just not easy to make accessible to users

what's the accessibility issue?

19:27

you can't really auth users without a central point, but I guess you can say that Learath is the same as last time's Learath

Yep exactly what I had in mind

but then can't the server intercept that, unless you do p2p

The accessibility issue is that there is no recovering a pricate key

19:28

private*

19:28

the server can never authenticate as me as it lacks the private key

how do people know what your pubkey is then

19:29

registered at a trusted place?

that's another accessibilty issue I'm having, but yes that's the idea I had, there could be central places you could trust to get the pubkeys

I guess we can make teeworlds.com show a pubkey for everyone yeah