DDraceNetwork - developer

DDraceNetwork

Development / developer

Development discussion. Logged to https://ddnet.tw/irclogs/ Connected with DDNet's IRC channel, Matrix room and GitHub repositories — IRC: #ddnet on Quakenet | Matrix: #ddnet-developer:matrix.org GitHub: https://github.com/ddnet

Between 2020-10-14 00:00:00Z and 2020-10-15 00:00:00Z

[ddnet/ddnet] Issue opened: #3096 RFC: Smooth Dynamic Camera

This is my attempt of making a dynamic camera option for people with motion sickness or high mouse sensitivity. The camera provides two new settings:

smoothness (cl_camera_smoothness): determines the speed of the camera
- 0 being instant with no smoothness
- 100 being the most smooth but slower camera movement.
stabilizing factor (cl_camera_stabilizing): determines how fast your cursor need to move before the stabilization kicks in
- 0 being no stabilization, camera w...

@Deleted User How to enable or disable golden spikes in fng2?

01:17

Also: What is vanilla Antispoof?

01:20

Anybody can answer?

its exactly what it sounds like

Hola

01:36

alguien habla español

when vanilla teeworlds connection packets are sent from spoofed ips

01:36

that feature helps to mitigate it

?

vanilla antispoof should be turned on

01:37

this is english channel go to #off-topic

01:39

and @Люди = быдло. Discord = провал u cant "disable" golden spikes but u can set sv_player_score_gold and sv_team_score_gold to other values (to match the team spike points for example)

Ok, answer then how i can turn on random_map command

01:43

Server answered so: Random map not supported in file based servers

01:43

But where is supported? (edited)

are u sure fng2 has that command

haha, i talking about ddrace server huh ^^

when a command says "not supported in file based servers" it means u need the sql support to be enabled when compiling

01:46

not sure why that would be necessary for random_map tho

oh ye, thanks, that i wanted to know

Already told ya. If you don't want golden spikes don't put them on the map

^ just use openfng5_Beat map instead of AliveFNG map

so nice when the linter finds typos in variable names https://zillyhuhn.com/cs/.1602662856.png

Embedded image

08:08

#shellcheckmasterrace

[ddnet/ddnet] Issue opened: #3097 make gfx_quad_as_triangle default 0 again?

As far as i can see it was added for the android support, bcs Android does not support GL_QUADS. The config says "(fixes quad coloring on some GPUs)", but the question is, is it worth the performance loose, and is the info text only meant for android?? (only affects GL versions prior to 3.3, bcs in 3.3+ it was removed anyway)

@deen I suppose def- is your github accout? I made a simple video demonstrating overall new chat idea, in case you're gonna make a vote

@heinrich5991 are you available?

available

can you give me some hints on how to debug the snapshots

13:24

is a type with 32k value a extended snapshot

yes

13:24

can't be an ormal one

in NewItem: bool Extended = false; if(Type >= OFFSET_UUID) { Extended = true; Type = GetTypeFromIndex(GetExtendedItemTypeIndex(Type)); } this stays false

then there's something wrong with the code, either server or client-side

13:25

OFFSET_UUID is 64k

13:25

65k*

wait, is this server or client code?

client

13:27

CSnapshotBuilder::NewItem(

13:27

Image attachment

ah yea, extended not being true there is fine

alright

13:30

well maybe i should check if the linux behaviour changed and not the windows one xD

just for explanation @Deleted User

13:32

UUIDS have IDs > 64k

13:32

in teeworlds code

13:32

those IDs are never sent over the network

13:33

in snapshots, they are mapped to IDs <64k

yeah, i just know as much as i said in the issue, it tries to search uuids every tick for all 64 plüayers

the line you showed does this translation

ok seems like the linux behaviour changed.. thats good xd

14:10

the bad thing is, it calls the uuid search all the time for the nameplates

caching hte lookup might be a good idea

yeah, it litterally halves my fps xd

[ddnet/ddnet] Issue closed: #3071 Go through clang scan-build

14:15

[ddnet:master] 14 new commits

baa8b93 Add clang-tidy to CI and enable clang-analyzer checks - edg-l 7d476c3 Fix clang-analyzer-security.FloatLoopCounter - def- 476ac69 Fix memory leak in lock_create - def- 911cb5c CAutoMapper: Delete pUpdateLayer - def- 469888e array: Don't allocate size 0 - def- a4f13ed Fix clang-analyzer-optin.portability.UnixAPI - def- 7c86dd8 Fix external directory in .clang-tidy - def- 5c1b8cd Fix clang-analyzer-deadcode.DeadStores - def- 98d4baa Fix clang-analyzer-core.NonNullParamChecker - def- 0867424 Fix clang-analyzer-security.insecureAPI.strcpy - def- 6443e70 Fix clang-analyzer-core.CallAndMessage - def- 87b06d3 Fix clang-analyzer-core.UndefinedBinaryOperatorResult - def- 7f38bc0 Fix clang-analyzer-core.uninitialized.Assign - def- 812c5c1 Merge #3076 - bors[bot]

14:15

[ddnet/ddnet] Pull request closed: #3076 Add clang-tidy to CI and enable clang-analyzer checks (supersedes #3075, fixes #3071)

[ddnet/ddnet] Issue opened: #3098 ddmax_freeze has pixelated curvatures

left current version, right redrawn !screenshot_2020-10-14_16-24-17 Here is the newly drawn image: !ddmax_freeze_fixed

Why is a uuid search done for the nameplates?

14:38

I thought as soon as we receive an extended object, we "translate" it's ID

no, we translate it on the fly

14:44

not sure why I didn't implement immediate translation, perhaps there was no good place to store the reuslt, perhaps I was just lazy

@Deleted User is the uuid lookup in nameplates from the SnapFindItem?

yes

That should be using m_pClient->m_Snap.m_paPlayerInfos instead

ok

16:01

i'd prefer if someone of u patches it tho, dont know how safe these variables are

16:01

when they get destructed or anything

Who do I safely get a tee render info and render it with wanted size?

the getting isnt the problem, the refresh button just has to refind the skins for chatlines then too

Why RefindSkins()?

16:23

And when

its called when the refresh button is clicked for skins

16:24

just add a call in the gameclient.cpp

Yes, I see from code it's just loading skins from your game folder (edited)

16:25

So I ask why do I even need that

bcs it deallocates memory

16:25

so pointers of renderinfo would be invalid

dont do it

16:25

its a rootkit

no doubt

16:26

need curl to download virus btw

16:26

and git and wget xd

ChillerDragon: paste the command directly instead

16:27

or give us the url where we can get the command from

https://paste.zillyhuhn.com/c

I removed your curl … | bash message

wdym

16:28

i tried to make it short and convienient to get a diff

16:28

oh wait you mean because i control the paste?

16:28

oh its actually sketchy isnt it?

16:28

```

yea, and I made it so that the convenient solution is not the one that gives ChillerDragon root on your machine

@Deleted User I can't understand, can we talk more about it?

 cd /tmp && wget https://raw.githubusercontent.com/DDNetPP/DDNetPP/master/CMakeLists.txt && curl https://raw.githubusercontent.com/ddnet/ddnet/master/CMakeLists.txt | git diff --no-index CMakeLists.txt -

better

didnt think of that sowwy

well skins refresh deallocates all current skins

afk food

so all references to it are invalid

so pointers become invalid and, as I see, there is a garbage collector to deal with it

16:30

?

what garbage collector? XD

k, there is no

the refind function just updates these renference

But the references are used for drawing tees, isn't it? (edited)

16:31

I mean in-game players

yes, and thats why they need an update

16:31

they get updated

So when should I call for it

16:31

Not the every frame I suppose

just add a function UpdateLineTees() and refinds all renderinfo

16:32

and add that function call to the gameclient's RefindSkins()

Ok, I see it now. BUT

16:32

I still have a problem and I need a hint for that 'cause I suck in cpp

16:33

For now I use references from m_pClient to render info

16:33

which becomes invalid when player leaves the server

16:33

which causes bugs if there is still his message in our memory that wants to be drawn

yes, that doesnt sound good

16:34

the best is you copy the information you need and dont rely on it

Yes, I thought about it, but I don't know how to make a copy of renderinfo in c++ xD

16:35

help me pls)

16:36

also I'm not sure if it means that I should handle memory dealloc myself when there is no more line for such renderinfo

just list what information u want exactly

I need CTeeRenderInfo because RenderTools()->RenderTee() asks for that

btw @Дядя Женя


        CTextCursor AppendCursor = Cursor;
        AppendCursor.m_StartX = AppendCursor.m_X;

        if(m_aLines[r].m_TextContainerIndex == -1)
            m_aLines[r].m_TextContainerIndex = TextRender()->CreateTextContainer(&Cursor, m_aLines[r].m_aText);
        else
            TextRender()->AppendTextContainer(&AppendCursor, m_aLines[r].m_TextContainerIndex, m_aLines[r].m_aText);

except the ": " which is still added, it will align the text msg

16:47

Image attachment

spam

but the ": " could just be removed from the text(added seperatelly) (edited)

16:48

then it would fit perfectly

16:51

can you view a pr squashed into one commit in github

16:52

is "files changed" always all commits?

16:52

found it

16:53

ok @Дядя Женя , u only want it for the tee render info?

Image attachment

16:54

its not

as said, remove the ": "

16:54

then it is

16:54

anyway

void CGameClient::RefindSkins()
{
    for(int i = 0; i < MAX_CLIENTS; ++i)
    {
        if(m_aClients[i].m_aSkinName[0] != '\0')
        {
            const CSkin *pSkin = m_pSkins->Get(m_pSkins->Find(m_aClients[i].m_aSkinName));
            m_aClients[i].m_SkinInfo.m_OriginalRenderSkin = pSkin->m_OriginalSkin;
            m_aClients[i].m_SkinInfo.m_ColorableRenderSkin = pSkin->m_ColorableSkin;
        }
    }
    m_pGhost->RefindSkin();
}

16:55

here u need to add ur code (edited)

16:55

just do smth client m_pChat->RefindSkins();

16:56

and then implement that function like you already did with the m_AuthorRenderInfo

I did already

16:57

But I don't understand the part implement that function like you already did with the m_AuthorRenderInfo

but good that i read the code

16:58

its missing a updaterenderinfo() call

16:59

have to patch that, gimmi a minute

ok, should I commit my last version?

16:59

ok, nvm, there should be only 1 small fix

[ddnet/ddnet] Pull request opened: #3099 Add missing UpdateRenderInfo call

ok, well idc, if you implemented it its fine

17:05

tbh alot of stuff confuses my, you can retreive most text stuff from the cursor. the chat text height can just be stored in the chat line instead of the text container

17:05

the text container shouldnt care about such information

feels like it was fixed somehow

Image attachment

17:08

magic, Idk

i think we should remove the : anyway, else its colored

17:10

even tho its not part of the message

17:10

the ": "*

will see what I can do

17:15

there are some problems in old code with width calculations and margin (edited)

17:15

that I still did not rework

Image attachment

17:17

would that be good enough? XD

17:19

@Developers: The repository 'https://dl.bintray.com/sbt/debian Release' no longer has a Release file.

17:19

@Дядя Женя https://github.com/Jupeyy/ddnet/commit/4d8acf8fb53af63865e75972cb5c44d06fd3958a

this feature bugs out really hard in the new version

Image attachment

@Learath2 did u change this ^ ?

17:22

i just remember alot of ppl worked in the editor files xD

it looks weird

Image attachment

17:36

I think everybody used to see white ":" as not the part of player's name

but also not part of the chat

17:40

u can also split it completly

oh this looks much better

17:44

i can only suggest to delete conrners in central messages

17:45

like this mb

Image attachment

nice sketch

17:46

totally can see the corner

that's not so ez

there is a DrawUIRect that can pass corners.

17:47

but i guess determine the top chat might be difficult

it's ok, but that will look like shit

why is the perfectly good chat getting redesigned? xd

so you should make a corners like you do on maps

only left side with up and down corners?

17:50

Image attachment

Don't you think it looks bad?

dk, i can decide it only after look-up on finished chat (edited)

@Дядя Женя https://github.com/Jupeyy/ddnet/commit/76d870346c88da0b69cf5ac072d9a7d9446fb569 splitted, can be used as you want, and colored as you want

Image attachment

18:59

Just generated some more bugs, cool

bcs u have to revert the text stuff (edited)

19:39

Image attachment

[ddnet/ddnet] Pull request opened: #3100 Fix offline client that is not in menu

E.g. popups like "ddnet-info.json" failed will cause weird NaN glitches, bcs the impl relys on ChangePosition to clean movetime and animation start

[ddnet/ddnet] Pull request opened: #3101 Fix hookchains

Its already increased by the for loop. Might have made the hook shorter(for very very long hooks) the past 3 years xd

@noby @Learath2 crashed again, luckily heinrich got us debug info in time: https://github.com/ddnet/ddnet/issues/3028

Crash in AntibotOnCharacterTick() · Issue #3028 · ddnet/ddnet

Just happened on GER2 on full server on Ultranova map: Core was generated by `./DDRace64-Server_sql -f servers/8298.cfg'. Program terminated with signal SIGSEGV, Segmentation fault. #0 0x00...

ohh shit

21:13

do we have backtrace / server paused in debugger to get other stack variable values

21:13

im abit preoccupied for the next hour or so but i wanna look into it as soon as i can

21:14

the thing id look at first tho

21:14

is dist < 0 by chance?

[ddnet/ddnet] Pull request opened: #3102 Update ddmax_freeze (by saltyElefant, fixes #3098)

[ddnet:master] 4 new commits

6ff5335 Remove code for CRC in most places, except where needed for net compat - heinrich5991 070504b Update ghost format to v6: Replace CRC with SHA256 - heinrich5991 a604d91 Move ghost code to client - heinrich5991 3fdfe7e Merge #3054 - bors[bot]

[ddnet/ddnet] Pull request closed: #3054 Remove CRC in all possible places

maybe we should create an issue for the config( what we want to keep etc. ) instead of spamming that pr

[ddnet:master] 2 new commits

fd08482 Add missing UpdateRenderInfo call - Jupeyy 6ed79ff Merge #3099 - bors[bot]

[ddnet/ddnet] Pull request closed: #3099 Add missing UpdateRenderInfo call

[ddnet:master] 2 new commits

34c4166 Update ddmax_freeze (by saltyElefant, fixes #3098) - def- 66248ef Merge #3102 - bors[bot]

[ddnet/ddnet] Issue closed: #3098 ddmax_freeze has pixelated curvatures

21:42

[ddnet/ddnet] Pull request closed: #3102 Update ddmax_freeze (by saltyElefant, fixes #3098)

@noby might not have debug info because I only deployed the binary earlier this day

but the header lines looked plausible didn't they?

well okay, let's check the coredump

21:48

but let's take it with a grain of salt

21:51

@noby do you have any variable whose value you know?

21:51

in that function?

21:53

@noby here rn?

i will be here in like 15min sry

kk

[ddnet:master] 2 new commits

ca645bf Fix hookchains - Jupeyy 09549eb Merge #3101 - bors[bot]

[ddnet/ddnet] Pull request closed: #3101 Fix hookchains

[ddnet:master] 2 new commits

6eb6fad Fix offline client that is not in menu - Jupeyy 9e060d7 Merge #3100 - bors[bot]

[ddnet/ddnet] Pull request closed: #3100 Fix offline client that is not in menu

@heinrich5991 what kind of variable would be useful

anything in that function or class would be nice

rlh_record_in should be 3 at most and 0 at least

22:33

is that useful atall

22:33

can we look at the value of dist

22:33

or did it not reach that point

yes, it's a large negative value

which

22:33

dist?

dist

O_O

22:33

well

22:33

i guess thats the problem ..

22:33

lets add check for that

but I don't know if we can trust the debug info

        int dist = (hx * hx) + (hy * hy);
        dist = dist ? (int)sqrt(dist) : dist;

22:34

i figured

22:34

that this shouldnt be negative

22:34

wait why

this is the debug info of another binary

22:35

if it also matches this one, it's pure luck

f

22:35

4:51 PM] heinrich5991: @noby do you have any variable whose value you know? the problem is that pretty much every variable in that function is dependant on the position of the tee and the mouse

22:36

so if we cant tell those things then im not sure :x

need to wait for the next crash I guess

D:

22:37

why is the dbg info on this one messed up?

22:37

was it using the older binary

yes, it was using the one I hadn't updated yet

ohh

22:37

aw

22:38

lets pretend that the update fixes it

22:38

Is there any config for repeating chat counter

22:49

cl_friend_messages cl_show_ids Is there any other commands that can affect chat now?

Is there any way friend will not pass that exp? if (m_aLines[r].m_ClientID >= 0 && m_aLines[r].m_aName[0] != '\0') (edited)

mhhh, maybe if he has no name

is it the valid reason

23:06

to break the chat

you can also change that line, the problem was just then the chat message window itself was r == 0

23:06

so it did add ": " for me

23:06

maybe u can check it somehow else

23:07

like when u are typing stuff

fuck it, it will just not show the heart if you're friend with player named null

xD

Does it look like an old chat?

Image attachment

23:08

If so, I'm ready. The only thing left is to make a toggle in game setting hud

well the : is obv different but i already knew that

I will not fallback the ":" thing

so yes, looks same

don't even ask me

idc about it btw

the most thing I was unsure about is heart

i just dont want code in the text.cpp xD

23:09

bcs the cursor can solve all your problems

I've removed my "NewLineOffset" related code from text.cpp (edited)

23:10

Even improved old code a bit

23:10

for chat.cpp > OnPrepareLines

in chat.cpp change whatever u want

Im doing exactly that

23:11

Didn't even want to change anything else

23:11

but unfortunately I need width/height info of text containers

23:12

for my future updates as well. And someone else's I think

@heinrich5991 @Learath2 pls check repo

@Дядя Женя ok, but i'd prefer to have a proper reason to why u want it, bcs the textcursor can already return width and height

23:13

so its just bloating the already bloating text.cpp even more

23:13

makes a rewrite just harder and harder

@Deleted User what members exactly

m_X returns the current X value, so you can check it against the X value before the function call, same for Y aligned size gives you the exact font size

23:16

and so on

23:16

we can discuss this tomorrow or smth, just look around a bit

but you didnt have any width/height of text container

well x advances the whole lines

23:18

if i am wrong, i'd still prefer to put it in the cursor not in the text container

@Deleted User But if you're using multiple cursors to edit the same container?

23:27

which I do ofc

yes u can just add floats to each other if u want the full size

I've made last commit, you can look at it if you want

23:29

and give some commentary ofc

yeah but not today, tomorrow

k

23:30

I'll just add toggle in setting menu and I'm done with chat, so that will be your problem now

23:30

is it ok that my PR fails nearly all tests?

Made a checkbox for old chat style in settings

23:46

Chat is done now

Exported 309 message(s)