DDraceNetwork - developer

[ddnet/ddnet] Issue opened: #7401 Follow Ghosts with /pause /spec or spectate mod

Many Speedrunners often wanna see their run quickly without need of opening demo, to make that possible u could make it possible to follow ghost.

Good chance to also enable Multiview Ghosts, to compare multiple ghost precisely or with ur mates,
or maybe in the future to compare with the WR ghost of the Map.

teini94

[ddnet/ddnet] Issue opened: #7402 src/game/voting.h VOTE_DESC_LENGTH = 64 to 128

Problem if 2 many mappers and/or long map names. "Arctic Festivity by louis, Im 'corneum, Silex, Voxel & Pipou + + + + | 2/5 ★" would be 2 long to show up so u cant see or vote the map we should increase this vote_desc_length to like 128? in src/game/voting.h enum { old VOTE_DESC_LENGTH = 64, VOTE_CMD_LENGTH = 512, VOTE_REASON_LENGTH = 16, MAX_VOTE_OPTIONS = 8192, }; old VOTE_DESC_LENGTH = 64, new VOTE_DESC_LENGTH = 128, change it to like 128...

good morning. good night

3:27AM

and tomorrow work will be done

I have a tag here

5:07AM

Anyone know why?

5:07AM

Can't find it

tag, you're it! (edited)

Blade

Anyone know why?

Because your Doppelganger was trolling us by showing his coding skills

typedef ivec4 CColor; in the map format pepeW

9:29AM

How many maps might have color component values outside the 0-255 range?

i dunno, but some have values outside of range because of env calculations

You mean calculated values? The values in the map should always be in 0-255 range, right?

Robyt3

How many maps might have color component values outside the 0-255 range?

I think none, twmap is strict there

9:41AM

well at least no ddnet maps

Robyt3

You mean calculated values? The values in the map should always be in 0-255 range, right?

i dunno if we need to generally force it

If it could cause rare issues with some drivers then it would be better to avoid I guess

9:43AM

It also breaks the editor if colors components are not in the 0-255 range though

9:44AM

Since all components are packed into one int at some point

and I mean, colors only really make sense up to 255, no? thr only thing I can think of is trickery with envelopes, but that should be possible through other means anyways

Patiga

and I mean, colors only really make sense up to 255, no? thr only thing I can think of is trickery with envelopes, but that should be possible through other means anyways

i mean u could say an env alpha of 2.0 allows to multiply all color values by 2

9:45AM

i dunno if this is nice

9:46AM

but if the color values are <= 0.5

9:46AM

it's not breaking anything

Yeah, it's possible that clamping color values would change some maps that used this weird behavior with old envelopes. Color envelopes cannot have components outside 0.0-1.0 either in the latest client.

9:47AM

I don't know if it was ever possible to input an out of range color value in the editor, since we probably packed the components into an int since forever (edited)

Jupstar ✪

i mean u could say an env alpha of 2.0 allows to multiply all color values by 2

I thought we were talking about tiles layer colors or quad colors

yeah but like, why enforce it for those, but not for env?

If the quad/layer color is 511 but an envelope limits it to 255 then it's fine, but it shouldn't be possible anymore for color envelopes to cause out of range values on their own, because they are already limited to the 0.0-1.0 range

but do you want to have it a validation check or what?

9:52AM

i'd say just be careful with rounding errors, if the editor uses floats or smth

9:52AM

that we dont use a u8 vec sucks here ofc

Jupstar ✪

yeah but like, why enforce it for those, but not for env?

the biggest upside I saw for envs is that high values allow you to create differen interpolation effects. was brightening with env values a thing? I don't remember

At some point it should be added as validation, but it's not on my list right now. I just found out we use an ivec4 for colors

I'm not sure if there is any map out there with an invalid value there

Patiga

the biggest upside I saw for envs is that high values allow you to create differen interpolation effects. was brightening with env values a thing? I don't remember

there is one city map that accidentially caused this brighten effect because we changed from clamping to "i dont care"

9:54AM

but i'd say it's not used yet

ah

9:55AM

I mean, with bezier curves, higher values should not be required anymore for different interpolation

Robyt3

[ddnet/ddnet] Pull request opened: #7403 Add more assertions to datafile writer, refactoring

See commit messages.

Checklist

[X] Tested the change ingame
[ ] Provided screenshots if it is a visual change
[ ] Tested in combination with possibly related configuration options
[ ] Written a unit test (especially base/) or added coverage to integration test
[X] Considered possible null pointers and out of bounds array indexing
[X] Changed no physics that affect existing maps
[ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet...

GitHub

Click to see attachment 🖼️

@Patiga Does it make sense to support map items with size 0 (i.e. only the type and ID meta data exist)? This would disallow items without any data.

sos how to memory manage

12:42PM

https://github.com/ddnet/ddnet/blob/03400f6ff384a9c8a64f3ebf26f1003858e362a1/src/game/client/components/chat.cpp#L76

ddnet/src/game/client/components/chat.cpp at 03400f6ff384a9c8a64f3e...

DDraceNetwork, a free cooperative platformer game. Contribute to ddnet/ddnet development by creating an account on GitHub.

12:42PM

CChat::RegisterCommand(const char *pName, const char *pParams, int flags, const char *pHelp)

12:43PM

this copies the string pointers in some vector. which works fine because RegisterCommand is called on fixed strings from ddracechat.h

Robyt3

@Patiga Does it make sense to support map items with size 0 (i.e. only the type and ID meta data exist)? This would disallow items without any data.

depends on the type I would say. I think env points can have 0 size, and I'll have to debug but I think there are also maps with ddnet bezier data with 0 size

i now want to call RegisterCommand later in the code with some dynamic input. And I need to store that strings memory somewhere.

12:44PM

any suggestions on how to do that? would it be bad if RegisterCommand does a strcpy and allocates memory?

12:44PM

@Patiga sorry for hijacking your convo

all good ^^

Patiga

depends on the type I would say. I think env points can have 0 size, and I'll have to debug but I think there are also maps with ddnet bezier data with 0 size

Ah, true, envelope points are a special case. That explains why I added checks to allow empty map items on vanilla, but I didn't remember why

12:46PM

Currently causes a malloc of size 0 in that case

@Patiga if u want u can edit my pr

12:50PM

ill probs forget otherwise xd

I think its a rather small change

12:52PM

but I'm not at home rn, will have to remember it later

@ghrn Create a Report: https://discord.com/channels/252358080522747904/1124657351442579486

hi

hai, moin

lol I'm just testing this irc thing

1:50PM

looks interesting

swaggy

oldschool hipster

Jupstar ✪

oldschool hipster

lol

Wait a minute

1:54PM

I just realized... doesn't the ddnet chat works kinda like this?

1:54PM

like an irc thing

1:54PM

I still didn't checked the ddnet chat's code so I'm just guessing

1:54PM

in first place cuz of how mentions works

1:54PM

bridge: hi

u could argue both are rather simplistic chats xd

Robyt3

Thanks for checking. I think it's still better to fix the maps themselves instead of trying to read strings best effort (i.e. fix broken UTF-8 etc.) (edited)

it might be nice to show a warning when launching the map first

Jupstar ✪: makes sense

heinrich5991

it might be nice to show a warning when launching the map first

Already done, #7400

https://github.com/ddnet/ddnet/pull/7400

Allow loading embedded images without valid image name, show warnin...

The image name is only required for external images. For embedded images it was only used in log messages. This fixes rendering with various maps that have embedded images with invalid names. Show ...

Robyt3

@Patiga Does it make sense to support map items with size 0 (i.e. only the type and ID meta data exist)? This would disallow items without any data.

maybe for envelopes? they're stored in a weird way

Robyt3

Currently causes a malloc of size 0 in that case

a malloc of size 0 isn't a problem AFAIK

Robyt3

Already done, #7400

but not for internal image names?

heinrich5991

but not for internal image names?

No, I'm not sure it's would be that useful for users to know that a basically irrelevant string failed to load

2:02PM

I'd prefer fixing the majority of maps first before adding a warning message that will lead to countless bug reports

heinrich5991

a malloc of size 0 isn't a problem AFAIK

It's a problem for clang-tidy if it detects it (Call to 'malloc' has an allocation size of 0 bytes [clang-analyzer-unix.API])

Robyt3

It's a problem for clang-tidy if it detects it (Call to 'malloc' has an allocation size of 0 bytes [clang-analyzer-unix.API])

it can return nullptr

2:02PM

that kinda sucks to work with

Robyt3

No, I'm not sure it's would be that useful for users to know that a basically irrelevant string failed to load

if it's no problem, then we should fix clang-tidy instead

Yeah, if it returns nullptr on some implementations and we accidentally access the memory then that's bad

2:03PM

On other implementation it allocates some dummy memory that must be freed

why do we access the memory if we allocate 0 bytes?

2:04PM

that's a bug even if it returns a non-null pointer

Assuming all functions also check the data size then it shouldn't cause issues, but I don't think that's the case yet

I don't see what this fixes

2:05PM

an access to memory of size 0 is a bug already

what if we have some logic that does if ptr != nullptr

2:05PM

independent of if we access it

yes, that's a problem I can see

heinrich5991

a malloc of size 0 isn't a problem AFAIK

yeah, it's allowed. IIRC it was used to retrieve heap pointers directly

i once tried to make a pr that added ui sounds

2:09PM

for hovering / clicking / contextual related things

Voxel

i once tried to make a pr that added ui sounds

make cool menu music xdd

i dont know how to make music

buy instrument (edited)

i mean i have fl 20

2:13PM

but idk music theory

lmao

2:14PM

in the modern times of electronic sounds, ppl just move some sounds into their music editor until it sounds cool

2:15PM

the question is: do you have good taste

i can try and replicate good taste

epic

2:15PM

i mean you are an artist, i bet you have a better feeling what could work out than many xd

not going to lie all the updates i wanna do completely go against the "making ddnet lightweight" thing.

2:16PM

i wanna so much crap but itd mostly be bloat

lightweight doesn't mean "not extendable"

several music tracks, more weapons, a better skin system

i gotta make more concept videos

Jupstar ✪

in the modern times of electronic sounds, ppl just move some sounds into their music editor until it sounds cool

what

Ewan

what

Hi

omfg

2:44PM

@Jupstar ✪ images with normal maps

2:45PM

if we ever set up a lighting system we gotta get images that pair with normal maps

2:45PM

like in spelunkey

Voxel

if we ever set up a lighting system we gotta get images that pair with normal maps

yeah interesting, u could try patigas godot ddnet renderer

2:55PM

and add lights to that to test around with

oh yea good idea

3:04PM

time to learn godot LOL

dont think music will ever work with current ddnet

Jupstar ✪

@Voxel here a present for you

not like a give a shit but this outputs images, right

Voxel

not like a give a shit but this outputs images, right

it outputs pixels and width & height

3:11PM

i also have it as outputting images tho

3:11PM

https://github.com/Jupeyy/dd-pg/tree/test5/src the -convert stuff (edited)

dd-pg/src at test5 · Jupeyy/dd-pg

don't use or read this code. Contribute to Jupeyy/dd-pg development by creating an account on GitHub.

3:12PM

you can do cargo build -p x-convert

3:12PM

where x is one of the items in that list

3:12PM

if u want i can also build them for u xd

3:12PM

on the other hand side.. you need to learn rust anyway

3:13PM

xDD

im only asking because i still wanna do the whole remodel of the images things

ew

3:16PM

ln 61

else if height % 4 != 0 {
        Err(anyhow!("width is not divisible by 4"))

3:16PM

lmao

xd

3:17PM

https://github.com/Jupeyy/dd-pg/blob/test5/game/client-extra/src/skin_split.rs#L62

dd-pg/game/client-extra/src/skin_split.rs at test5 · Jupeyy/dd-pg

don't use or read this code. Contribute to Jupeyy/dd-pg development by creating an account on GitHub.

3:17PM

fixed in current version actually

3:17PM

seems like i already noticed it when implementing the other splitters

oh i forgot to see the link

3:20PM

the X * segment_width/height looks so clunky LOL

3:22PM

why not have it times by segment_width/height in the actual single_img function

which converter are you reading?

3:23PM

note that i automated all other converters using regex xd

3:23PM

and the content.py from ddnet

game_splitter

Jupstar ✪

note that i automated all other converters using regex xd

that makes much more sense

I should do a ddnet remake too, that seems to be the hype

epic

3:27PM

Learath & ?

3:28PM

whom do u team up with

Do I need a duo?

currently we have Me + Robyte vs. Patiga + Zwelf vs Ryo + Voxel

Learath2

Do I need a duo?

1 vs 2

3:28PM

what a pro

Ryo and Voxel are making a ddnet remake?

yes

Learath go with zogtib ;)

in c xDDD

Learath2

Do I need a duo?

every team has to have someone who gives you a smack and does nothing

Jupstar ✪

in c xDDD

Portable C99

3:29PM

Or maybe I can get heinrich onboard

3:29PM

Though then it has to be in Rust

it's settled

3:30PM

Jupstar + Robyte

️ Patiga + Zwelf

️ Ryozuki + Voxel

️ Learath + heinrich (edited)

2v2v2v2

4 teams fng

3:30PM

xd

was about to say it

lmao

Hm, though with heinrich it'll also have to be fully backwards, forwards and multiverse compatible

xDDDDDDDDDD

btw, has there been any discussion/solutions about accounts since the last comment on github in July?

Nope

sounds like it's time for the bimonthly accounts debate

It's biannual not bimonthly

i have a plan for my implementation, but more than changing some fields and sql strings isn't really done

3:41PM

i got kinda distracted with other stuff xd

Robyt3

[ddnet/ddnet] Pull request opened: #7404 Add freeze/unfreeze testing commands

See #7370.

Checklist

[X] Tested the change ingame
[ ] Provided screenshots if it is a visual change
[ ] Tested in combination with possibly related configuration options
[ ] Written a unit test (especially base/) or added coverage to integration test
[ ] Considered possible null pointers and out of bounds array indexing
[X] Changed no physics that affect existing maps
[ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#us...

hey is there a easy way to completely isolate instances of CGameWorld in the clientside prediction?

5:20PM

i want them to suffer for what they've done to me

Learath2

Ryo and Voxel are making a ddnet remake?

"""making"""

6:11PM

wait wtf why am i paired with ryozuki?? LOL

Jupstar ✪

in c xDDD

nuh uh ryo's doing all the coding

6:13PM

im the ideaman

Voxel

wait wtf why am i paired with ryozuki?? LOL

https://discord.com/channels/252358080522747904/293493549758939136/1168495500954185758

shrug. i guess if it becomes a serious thing ill make proper concept art for how everything should function.

ideas man

6:23PM

lmfao

6:23PM

everyone needs an ideas man pepeW

Voxel

im the ideaman

gonna be real the only thing ddnet needs is sentry gun from tf2

oh man cant wait for ddnet to have classes

6:34PM

the medic tee

its just rifle but curved

@Ryozuki i saw you typing

xd

have u never played infclass

Ewan

have u never played infclass

cyber forced me to grind for him

did the behavior of these buttons change?

6:55PM

i click square.. nothing happens

6:55PM

ah lol nvm

Voxel

cyber forced me to grind for him

thats not infclass

do we somewhere print the zoom level in the client? xd

7:25PM

i have the feeling u can't get its current value

Yeah, doesn't look like we show it anywhere. Using zoom in console also doesn't seem to print the current value

Jupstar ✪

do we somewhere print the zoom level in the client? xd

No

Robyt3

Yeah, doesn't look like we show it anywhere. Using zoom in console also doesn't seem to print the current value

it sets it to the default zoom

7:47PM

bcs con zoom value is immediately converted and stored in ccamera so the original value is never kept

Ewan

the og is pretty much like m_Zoom = pow(ZoomStep, g_Config.m_ClDefaultZoom - 10) if ZoomStep = std::cos((30.0f * pi) / 180.0f) (about 0.866025, the constant the game uses) then u can reverse it w this m_ConsoleZoom = std::log(m_pClient->m_Camera.m_Zoom * std::pow(ZoomStep, 10)) / std::log(ZoomStep)

cyberFighter

thats not infclass

mb

jesus wavpack is fucking garbage

yeah

T datafile: loading data. index=23 size=8 uncompressed=0 nouis

9:26PM

At least only one official map has invalid map settings

Robyt3

[ddnet/ddnet] Pull request opened: #7405 Add gameclient warnings to queue instead of overriding others

Using CMenus::PopupWarning will immediately set the popup and override the current warning, which can hide warnings when multiple are shown at the same time. Now all warnings are added to the queue with IClient::AddWarning.

Checklist

[X] Tested the change ingame
[ ] Provided screenshots if it is a visual change
[ ] Tested in combination with possibly related configuration options
[ ] Written a unit test (especially base/) or added coverage to integration test
[ ] Co...

Robyt3

At least only one official map has invalid map settings

in what way invalid?

Not really invalid, but the string array data has size 0, which I consider an error now, as empty string arrays should not be packed into datafiles at all (edited)

10:03PM

In DeadCore 1

ah, I get it

10:03PM

but it doesn't sound too bad, no?

I checked for valid utf-8 in the settings too, but that's all good it seems (edited)

Patiga

but it doesn't sound too bad, no?

Yeah, I already fixed the map by resaving it, should be deployed by the bot to github at some point (edited)

remember that every map update requires every player to redownload it

10:06PM

I do think issues in maps should get fixed, but some things might just not be worth it

github-merge-queue[bot]

[ddnet:master] 10 new commits

f7722eb Remove separate CDataFileWriter::Init/OpenFile functions - Robyt3 284390c Add assertions to ensure correct map item/data sizes are written - Robyt3 6471607 Add assertion to ensure correct item IDs are written - Robyt3 7343ca2 Add CDatafileHeader::SizeOffset to replace magic 16 - Robyt3 9b8eb9d Use for- instead of while-loop, improve comments - Robyt3 75debad Add freeze/unfreeze testing commands - Robyt3 ebe2dde Add gameclient warnings to queue instead of overriding others - Robyt3 9b75f85 Merge pull request #7405 from Robyt3/Client-Warning-Queue - def- 3ec5211 Merge pull request #7404 from Robyt3/Server-Testing-Commands-Freeze-Unfreeze - def- 80c4940 Merge pull request #7403 from Robyt3/Datafile-Writer-Refactoring - def-

10:25PM