DDraceNetwork - developer

fc5cfb6 Remove CFlag dead code (fixes #5677) - def- b33cbdd Remove unused backend_opengles.h - def- 1e47c28 Automatically detect unused header files - def- 34fb372 Merge #5680 - bors[bot]

Robyt3

[ddnet/ddnet] Pull request opened: #5683 Fix accesses to vector with index -1 by adding static button IDs

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
[ ] Changed no physics that affect existing maps
[ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresss...

Robyt3

[ddnet/ddnet] Issue opened: #5684 Output stdout/stderr of integration test when server/client process exits abnormally

For example if the server crashes due to an assertion, the workflow will only log

./integration_test.sh: line 169:  6420 Illegal instruction     (core dumped) $tool ../DDNet-Server "sv_input_fifo server.fifo;
	sv_rcon_password rcon;
	sv_map coverage;
	sv_sqlite_file ddnet-server.sqlite;
	logfile server.log;
	sv_register 0;
	sv_port $port" > stdout_server.txt 2> stderr_server.txt
[-] server exited with code 132

which does not help to diagnose the problem.

Chairn

do you agree that it copies array pEntry->m_Info.m_aAddresses into aAddresses and then from aAddresses into pEntry->m_Info.m_aAddresses ?

AFAICT that code is very much useful, wanna figure it out yourself why that is the case?

13:18

@Tater @Ryozuki ^

bubliman

64 pixels, im not sure what is used for these calculations

no, tiles aren't bound to be 64 pixels, they just are in the default tilesets

hmm

13:21

you declare a array on the stack, copy the addresses into it, then update the entry, without ever modifying the addresses, and then copy them back

13:21

idk

bubliman

So hypothetical question, Would I get banned if you proved that I was using this ?

unclear, we'd have to talk about it

fokkonaut

not really, there is even a config setting to show everyone's hookline always

yes, really. the client respects the server settings for this IIRC

thats not my point. the point is that the server can not detect whether a client is using this, because we talked about these other prediction lines before

13:22

and that they are not detectable

13:22

and therefore, also the server can in theory not disable hooklines

yes. that doesn't make them not-cheats

thats not the point

okay, that point is enough for me

??? but its completely out of context in that sense

fokkonaut

Thing is, you cant prove that someone uses this. There are other clients which pre calculate grenade and laser shots already.

.

trml

the server can choose not to allow hookcoll though

And this was nuborns msg

13:23

Which I basically responded to

we will ban anyone who gets detected using one of these clients

Cuz the server can not really control it, if the client just renders it, like the prediction lines

heinrich5991

we will ban anyone who gets detected using one of these clients

Yeah, thats the thing, its not detecable xD

if you use one of these clients, you're not safe from being banned, should we invest some time detecting them

Same for hookline if it was forbidden

usually, this comes from a fixed number of cheat clients

13:24

sometimes, they behave differently in small ways

i know, i was just talking theoretically

great, I'm talking pratcially

nice, missed my point

Ryozuki

you declare a array on the stack, copy the addresses into it, then update the entry, without ever modifying the addresses, and then copy them back

but they are part of the entry

I basically want to inform people that your point isn't everything

yeah, thats obvious

it's nice that it's obvious to you

13:25

I wanted to make that clear for everyone

ah

ohh

13:26

pEntry->m_Info = Info;

13:26

i missed this xd

13:26

ye the code makes sense

fokkonaut

thats not my point. the point is that the server can not detect whether a client is using this, because we talked about these other prediction lines before

and the first thing I said was actually just that the hookline can be disabled, which seemed like a false statement on your part

yea, cuz thought a bit further it cant be disabled, if the client decides to render it not depending on the flag, as my example with cl_show_hook_coll_other 2 said (edited)

that's okay, but not what you said

13:28

hence my comment

fokkonaut

not really, there is even a config setting to show everyone's hookline always

.

yes, that config setting is disabled when the server requests it

that means, in the end the server can not control it

it can control it in the ddnet client

heinrich5991

yes, that config setting is disabled when the server requests it

yes, but in the end the client can render it anyways.........

yes

13:29

but that's not what you wrote ^^

this setting shows, that the client can render it, not depending on the flag

you gave an example that doesn't work

it does work

13:29

as long as the server allows it, which can be changed in the client

13:29

was just an example

13:29

idk why you take it so hard xD

fokkonaut

as long as the server allows it, which can be changed in the client

this is the important extra point

13:29

same for you ^^

you know exactly what i mean

it always takes two stubborn people to land in this situation ^^

and i didnt make it too unclear

I know I tend to be one, but you apparently also are

of course

13:30

:D

fokkonaut

idk why you take it so hard xD

if you undestand that, why do you talk like I'm solely at fault here?

13:31

that's a bit annoying ^^

i didnt lol

why are you taking this so seriously?

why are you trying to make me look weird all the time?

13:31

its so useless

heinrich5991

why are you taking this so seriously?

see? this reply from me was doing what you did above

heinrich5991

why are you taking this so seriously?

wrong assumption

I wanted to demonstrate how you say that it's my fault

13:32

and it seems you agree that a statement like this assigns fault

13:32

because you went on to "defend" yourself

i clearly explained my points, and you just talk against it with things that are not involved

13:33

thats the difference

13:33

and you knew exactly how i meant it

I acknowledged your points, you didn't seem to acknowledge mine

oaky whatever dude :D

13:33

have a good one

Another heated developer moment

4

14:52

@Learath2

bors[bot]

[ddnet/ddnet] Pull request closed: #5683 Fix accesses to vector with index -1 by adding static button IDs

17:29

bors[bot]

[ddnet:master] 3 new commits

344e50f Remove dead code - Robyt3 1d6dca8 Fix accesses to vector with index -1 by adding static button IDs - Robyt3 9df5a74 Merge #5683 - bors[bot]

@c0d3d3v i dont know where you got that info from, regarding the spec inputs xd

17:50

its EXACTLY the other way around

def-

[ddnet-web:master] 1 new commit

632deaf 3548 - def-

Vy0x2

[ddnet/ddnet] Issue opened: #5685 Freezebar behaves weird after using spec

The freeze bar behaves very weird when you do following steps: 1. Jump into freeze and land on a block 2. /spec 3. Wait some seconds 4. /spec -> The remaining freeze bar is very small but you are still frozen for ~2 seconds !de1e459165 1 In this picture i was still frozen for 2 seconds

Jupeyy

[ddnet/ddnet] Pull request opened: #5686 Fix editor image delete

fixes #5675

Checklist

[ ] 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
[ ] Changed no physics that affect existing maps
[ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#u...

heinrich5991

AFAICT that code is very much useful, wanna figure it out yourself why that is the case?

i didn't see the info assignment in the middle

^^

Although, you could use an assignemnt to TmpInfo

19:52

to have less code

how to? if it improves the clarity, you could make a PR

TmpInfo = pEntry->m_Info;

19:54

pEntry->m_Info = Info; // restore other data from TmpInfo

bors[bot]

[ddnet/ddnet] Issue closed: #5673 Editor: Often crashes when recoloring many quads, after switching between quad layers

20:23

bors[bot]

[ddnet/ddnet] Pull request closed: #5678 Reset selected quads when switching layers (fixes #5673)

20:23

bors[bot]

[ddnet:master] 2 new commits

76e3db9 Reset selected quads when switching layers (fixes #5673) - def- 730c54a Merge #5678 - bors[bot]

Chairn

[ddnet/ddnet] Pull request opened: #5687 Use copy assignment rather than copy variable one by one

Less code and i find it more clear :)

Checklist

[ ] 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
[ ] Changed no physics that affect existing maps
[ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https:/...

bors[bot]

[ddnet/ddnet] Pull request closed: #5687 Use copy assignment rather than copy variable one by one

21:04

bors[bot]

[ddnet:master] 2 new commits

c651936 Use copy assignment rather than copy variable one by one - Chairn e5e2609 Merge #5687 - bors[bot]

21:11

def-

[ddnet/ddnet] Issue closed: #5674 Editor: Random quad has random colors after moving around other quads + Crash

bors[bot]

[ddnet/ddnet] Issue closed: #5675 Editor: Crashes after removing specific mapres.

21:32

bors[bot]

[ddnet/ddnet] Pull request closed: #5686 Fix editor image delete

21:32

bors[bot]

[ddnet:master] 2 new commits

9821399 Fix editor image delete - Jupeyy 1110d0e Merge #5686 - bors[bot]

why does it take so freeaking long to link with mingw

23:02

45 seconds for the client...

LTO?

$ time C:/MinGW64/bin/ar.exe cr CMakeFiles/DDNet.dir/objects.a @CMakeFiles/DDNet.dir/objects1.rsp

real    0m12,465s
user    0m0,000s
sys     0m0,031s

$ time C:/MinGW64/bin/g++.exe -fdiagnostics-color=always -Werror=class-memaccess -g  -mwindows -Wl,--whole-archive CMakeFiles/DDNet.dir/objects.a -Wl,--no-whole-archive  -o DDNet.exe -Wl,--out-implib,libDDNet.dll.a -Wl,--major-image-version,0,--minor-image-version,0 @CMakeFiles/DDNet.dir/linklibs.rsp

real    0m32,080s
user    0m0,000s
sys     0m0,031s

23:27

i don't see any lto flag in there, just slow linking

23:28

meanwhile, the server takes only 3 seconds:

$ time C:/MinGW64/bin/ar.exe cr CMakeFiles/DDNet-Server.dir/objects.a @CMakeFiles/DDNet-Server.dir/objects1.rsp

real    0m0,723s
user    0m0,000s
sys     0m0,015s

$ time C:/MinGW64/bin/g++.exe -fdiagnostics-color=always -Werror=class-memaccess -g   -Wl,--whole-archive CMakeFiles/DDNet-Server.dir/objects.a -Wl,--no-whole-archive  -o DDNet-Server.exe -Wl,--out-implib,libDDNet-Server.dll.a -Wl,--major-image-version,0,--minor-image-version,0 @CMakeFiles/DDNet-Server.dir/linklibs.rsp

real    0m2,804s
user    0m0,000s
sys     0m0,015s

(edited)

23:31

hmm, one more flag in the client compile line: -mwindows

23:33

doesn't change anything with or without it