DDraceNetwork - developer

[ddnet/ddnet] Pull request opened: #8707 Initial work on 0.7 dummy seperated translation context

This is WIP, don't have the time right now to fully test & finish this. It doesn't work correctly yet, some skin information for example is still not applied correctly and probably other things. F-Client 0.7 dummy fix: https://github.com/fokkonaut/F-Client/commit/628ea8abe03e910e2f4a6b6f4cb106cdad584431 I think the 0.7 ddnet client should still send it's ddnet packet, so the server has information about the ddnet version and so on... F-Client currently does this, and therefore it can...

@fokkonaut you can mark PR as draft if it's WIP

Does anyone know how to generate DDNet.exe (as well as steam_api.dll) in /build instead of /build/Debug using cmake? (edited)

louis

good change

although i feel the slider buttons should move out <> as well as up down

00:58

animations on un-hover would be dope too but idk if there's support for that

louis

although i feel the slider buttons should move out <> as well as up down

I didn't want to do that because it hard to tell where the end of the slider is when the bar is overlapping it

01:08

but maybe it would be better idk

louis

animations on un-hover would be dope too but idk if there's support for that

wdym by animations on un-hover

Tater

I didn't want to do that because it hard to tell where the end of the slider is when the bar is overlapping it

oh good point

01:12

i mean fade back to normal color and fast but smooth scale down to original size

<ChillerDragon> thanks poggers2

@Tater

eh, animations are pretty pointless for ddnet ui

ah

01:13

idk it's not like css where I can add the animation in 3 lines

01:13

it would be sorta annoying

yeah

ChillerDragon

[ddnet/ddnet] Pull request opened: #8708 Send the new is ddnet message in 0.7 connections

We already sent the legacy is dddnet message and so far the ddnet protocol over 0.7 seems to work fine. But for completeness and correctness it is now also sending the new ddnet version net message. !image

Checklist

[x] Tested the change ingame
[x] Provided screenshots if it can be seen in wireshark
[ ] Tested in combination with possibly related configuration options
[ ] Written a u...

01:43

@Ryozuki can confirm?

ddnet-maps

[ddnet-maps:master] 1 new commit

644d915 A Amethyst - ddnet-maps

@Kaffeine: did you name the folder skins7 or did I? If so why? If there is no reason I agree with @Robyt3 to rename it to just skins so its shared with the vanilla teeworlds client.

Probably it was me: https://github.com/ChillerDragon/ddnet/pull/3/commits/f0aa9fc520bb64086d088b23a1baa935e1ec728e

So what do you think about renaming it to skins? @Kaffeine

You named the class (classes) skins7 but didn't add any data. We already have maps7, so the skins would be consistent with that. I thought it would be easier to manage it this way. The skins7 dataset seems to be completely different, and not intersecting with skins6.

07:37

On one hand it means that we won't have filename conflicts, and on the other I don't see 'share with vanilla' be a good reason: ddnet is always installed to a different directory, and the local files are not in ~/.teeworlds but in ~/.local/share/ddnet.

07:42

IOW, what are the real cases of possible data sharing with vanilla 0.7?

07:42

Did I understand correctly that we're about the client installations? In case of the source repository, vanilla stores the skins in datasrc/skins, and DDNet stores it at data/skins (and ATM 0.7 skins in DDNet stored at data/skins7).

In my humble opinion, the renaming and possible sharing with vanilla does not worths it. I don't see cases where it will give us significant (or any measurable) benefits. Having different (incompatible and non-intersecting) data sets in different directories seems to be a better idea, and aligns with maps.

In a different situation it could be beneficial to change layouts to automatically get user's 0.7 custom skins.

<ChillerDragon> my ddnet is in ~/.teeworlds

08:35

<ChillerDragon> maps7 exists because maps are potentially different in 0.6 and 0.7 so they need to folders to avoid conflicts with the filenames. The skin files are comptatible and not conflicting.

08:37

<ChillerDragon> I see two advantages of skins over skins7 one is that players can switch from vanilla to ddnet without losing their skins. And the other is that we do not bloat yet another 7 folder into the file tree.

08:38

<ChillerDragon> In the code I used the 7 suffix to mark 0.7 stuff because otherwise it would conflict with 0.6 and also heinrich asked for 7 suffixes.

it seems so unfun to implement

. Btw @Learath2 do you know a reason why in CallMe it uses 8 bytes for local variable when the struct fits in 4? (edited)

https://github.com/fzyzcjy/flutter_rust_bridge

GitHub - fzyzcjy/flutter_rust_bridge: Flutter/Dart <-> Rust binding...

Flutter/Dart <-> Rust binding generator, feature-rich, but seamless and simple. - fzyzcjy/flutter_rust_bridge

08:52

https://cjycode.com/posts/rust-ui-flutter/

Rust GUI library via Flutter, done simple

Background #Rust has been “the most desired programming language” for 8 years (by StackOverflow and GitHub 1 2), and many people want to write programs with a GUI in Rust.

08:54

https://chevyray.dev/blog/creating-175-fonts/

Chevy Ray | How I Created 175 Fonts Using Rust

How I used my own Rust tool to ship 175 quality pixel fonts.

MilkeeyCat

it seems so unfun to implement

. Btw @Learath2 do you know a reason why in CallMe it uses 8 bytes for local variable when the struct fits in 4? (edited)

sysv amd64 abi wants rsp to be aligned to 16 before call, after the call rsp will be 8 from 16 alignment, so the new function needs to allocate another 8 to get it back to alignment

09:21

What's more interesting is that even if you don't call Yep but use -mno-red-zone it'll still keep the stack in 16 alignment, guess it's being conservative

Learath2

sysv amd64 abi wants rsp to be aligned to 16 before call, after the call rsp will be 8 from 16 alignment, so the new function needs to allocate another 8 to get it back to alignment

i mean, would it not work if it used DWORD PTR instead of QWORD PTR? (edited)

Ewan

hey this is the first time ddnet has worked in fullscreen for me on native wayland

Ohm, may I ask you which Linux distribution are you using and which SDL2 version you're linking against?

MilkeeyCat

i mean, would it not work if it used DWORD PTR instead of QWORD PTR? (edited)

The abi doesn't split registers. I'm guessing it also isn't any faster to mov 32bits rather than 64 so no point bothering with special codegen for smaller than eightbyte arguments. But idk for certain

guys, I started writing my client, how can I update it correctly, if a new DDNet update comes out, should I make my changes to the updated DDNet source code? Sorry for such a stupid question.

DynamoFox

Ohm, may I ask you which Linux distribution are you using and which SDL2 version you're linking against?

he was using nix, so he might've used nixOS, depending on his channel (stable or unstable) he either linked against 2.30.2 or 2.30.4 (edited)

unkown user

guys, I started writing my client, how can I update it correctly, if a new DDNet update comes out, should I make my changes to the updated DDNet source code? Sorry for such a stupid question.

you should merge new changes from ddnet updates into your own client

10:50

are you using git?

ok, thanks

if it's a fork, you should be able to just update your fork via github.com if you cloned it and removed/changed the remote - you want to add ddnet as a new remote (e.g git remote add ddnet https://github.com/ddnet/ddnet) and pull it on top of your feature branch (edited)

1

louis

are you using git?

+

Then merge it with git as they suggest (edited)