DDraceNetwork - developer

I'll give it a try

def-

[ddnet/ddnet] Pull request opened: #5260 Pr thread safety negative

WorkerThread is hard because REQUIRES(!(CJobPool *)pUser->m_Lock) is not possible

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 if it works standalone, system.c especially
[ ] Considered possible null pointers and out of bounds array indexing
[ ] Changed no physics that affect existing maps
[ ] Tested the change with [ASan+UBSa...

00:21

Chairn

[ddnet/ddnet] Pull request opened: #5261 Fix possible misuse of comma

Found with -Wcomma using clang.

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 if it works standalone, system.c especially
[ ] 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/d...

bors[bot]

[ddnet:master] 3 new commits

4cba284 Fix possible misuse of comma - Chairn 39e3f52 clang-format - Chairn 73a0998 Merge #5261 - bors[bot]

00:44

bors[bot]

[ddnet/ddnet] Pull request closed: #5261 Fix possible misuse of comma

Jupeyy

[ddnet/ddnet] Issue closed: #1532 /showall 1 and go in spec, hear people double jumping from every distance

06:32

Jupeyy

[ddnet/ddnet] Issue closed: #1755 Some players hear jumping sounds from all players

06:37

Jupeyy

[ddnet/ddnet] Issue closed: #1451 Crash when accessing m_TeleOuts

06:41

Jupeyy

[ddnet/ddnet] Issue closed: #3111 laser bounce star thingie doesnt render in beta for some reason

06:42

Jupeyy

[ddnet/ddnet] Issue closed: #3491 Video recorder is not perfectly fluid

Jupeyy

[ddnet/ddnet] Issue closed: #1921 screen tearing even with vsync enabled

Jupeyy

[ddnet/ddnet] Pull request opened: #5262 Remove alpha threshold from map_optmize

Since it was removed from dilate's copy values, and also makes more sense anyway. Images should always be identifiable by all visible pixels

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 if it works standalone, system.c especially
[ ] Considered possible null pointers and out of bounds array indexing
[ ] Changed no physics that aff...

Such a hard read to wake up hangover to.So it seems like I gotta refactor my nice goto code to please the CI ._. (@deen)

deen

ChillerDragon: it allows gotos that jump strictly forward. In case your questino is serious: https://homepages.cwi.nl/~storm/teaching/reader/Dijkstra68.pdf

you should maybe use a function instead if you have a case where you use goto

chillerdragon

Such a hard read to wake up hangover to.So it seems like I gotta refactor my nice goto code to please the CI ._. (@deen)

08:15

show me your example code

I think I use it to cleanup stuff idk

08:15

Lemme search

use RAII

https://github.com/DDNetPP/DDNetPP/blob/master/src/base/system_ddpp.cpp

DDNetPP/system_ddpp.cpp at master · DDNetPP/DDNetPP

DDraceNetwork + city + block! Based on ddnet (www.ddnet.tw) which is based on teeworlds (www.teeworlds.com). - DDNetPP/system_ddpp.cpp at master · DDNetPP/DDNetPP

put above in a function and use return

08:17

same effect, no goto

Oke -./

if u use inline it compiles to pretty similar code, if you care about the assembly output ^^

xd

gimme regex support tho xd

chillerdragon

https://github.com/DDNetPP/DDNetPP/blob/master/src/base/system_ddpp.cpp

08:23

btw RAII still the "easy" c++ way, then u dont have to care about cleanup

08:23

the structure that wraps the regex has to garantuee it :D

What is even RAII

if u have a stack object

struct a {
 some_regex m_b;
 a () {
  b = some_regex_init(...);
}
~a() {
clean_some_regex(b);
}
};


int function () {
a MyVar;
...
if(...)
return;


[..]
if(...)
return;
else
return;

}

all returns garantuee that destructor of MyVar is called

08:32

so basically u have automated cleanup by the compiler

08:33

thats also what std::unique_lock etc. does under the hood

08:33

that's also why its considered more useful than manually locking, unlocking bcs u cannot forget to unlock

08:33

considered more safe too

C0D3D3V

[ddnet/ddnet] Issue reopened: #3560 Releasing a mouse button when entering the UI immediately results in a mouse click.

def-

[ddnet/ddnet] Issue closed: #5227 Initial map not found -> Server can't be killed

bors[bot]

[ddnet:master] 2 new commits

23683a7 Remove alpha threshold from map_optmize - Jupeyy fe8275b Merge #5262 - bors[bot]

10:56

bors[bot]

[ddnet/ddnet] Pull request closed: #5262 Remove alpha threshold from map_optmize

@c0d3d3v does freeze not always reset the freeze time?

11:08

or better, whats the reason for this^^

bors[bot]

[ddnet:master] 16 new commits

8010388 Use std::vector<char *> instead of array in config_store - Robyt3 0f097a0 Use std::vector<char *> instead of array in gamecontext - Robyt3 d3eb9d6 Use std::vector<std::string> instead of array - Robyt3 24ad383 Use std::vector<CSourceQueueEntry> instead of array - Robyt3 ce54a0d Use std::vector<CMenuImage> instead of array - Robyt3 4a02334 Use std::vector<CName> instead of array, adjust variable name - Robyt3 c6e195a Use std::vector<CNameBan> instead of array - Robyt3 3f8fb73 Use std::vector<CKey> instead of array - Robyt3 0c69495 Use std::vector instead of array in editor - Robyt3 ca309d6 Remove base/tl/array.h - Robyt3 47098da Remove base/tl/algorithm.h - Robyt3 f9d49bd Remove base/tl/allocator.h - Robyt3 a246560 Remove base/tl/range.h - Robyt3 6a72c5c Fix clang-analyzer-cplusplus.NewDelete (Use of memory after it is freed) - Robyt3 c609314 Use std::any_of instead of for-each loops - Robyt3 ccf0622 Merge #5252 - bors[bot]

11:09

bors[bot]

[ddnet/ddnet] Pull request closed: #5252 Use std::vector instead of array, remove base/tl/array.h, algorithm.h, allocator.h, range.h

def-

[ddnet/ddnet] Issue opened: #5263 Draggers are disabled too quickly

Reported by bencie on Fall into the Future. Bisect revealed: https://github.com/ddnet/ddnet/pull/5245 ``` 62c99515d564181e295a1528dc3dc23fbe3e2887 is the first bad commit commit 62c99515d564181e295a1528dc3dc23fbe3e2887 Author: c0d3d3v Date: Thu May 26 19:33:01 2022 +0200 disable dragger beams soon, not only every 150ms src/game/server/entities/dragger.cpp | 2 +- src/game/server/entities/dragger_beam.cpp | 15 +++++++++++++-- src/game/server/entities/dragger_beam.h...

def-

[ddnet/ddnet] Pull request opened: #5264 Prevent users from setting port 1-1023

Following user report that they set cl_port 1 and it didn't work

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 if it works standalone, system.c especially
[ ] 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...

That’s cool

Jupstar ✪

if u have a stack object

struct a {
 some_regex m_b;
 a () {
  b = some_regex_init(...);
}
~a() {
clean_some_regex(b);
}
};


int function () {
a MyVar;
...
if(...)
return;


[..]
if(...)
return;
else
return;

}

all returns garantuee that destructor of MyVar is called

Ideally some_regex also uses RAII; so you don't even have to worry about cleaning it up, your destructor will automatically call its destructor and clean it up

must asan be compiled while being in the root git directory?

why do u want to compile asan

I mean compile ddnet with asan

no

11:50

can be done from any build dir

should I copy ubsan.supp then, or is it not needed?

mkdir san && cd san && CC=clang CXX=clang++ CXXFLAGS="-fsanitize=address,undefined -fsanitize-recover=address,undefined -fno-omit-frame-pointer" CFLAGS="-fsanitize=address,undefined -fsanitize-recover=address,undefined -fno-omit-frame-pointer" cmake -DCMAKE_BUILD_TYPE=Debug ..

11:51

You can also just set UBSAN_OPTIONS=suppressions=../ubsan.supp

oh, I see, thanks

def-

[ddnet-scripts:master] 1 new commit

e07fadf port - def-

Jupstar ✪

@c0d3d3v does freeze not always reset the freeze time?

for what? freeze only sets freeze time every second

12:19

@deen do release tags match the last commit of a version? So If I checkout to tag 15.9.1 I get all commits in that version?

12:21

Mh even if not, should be close enough, for my investigation

yes, they should match (modulo mistakes by me)

c0d3d3v

for what? freeze only sets freeze time every second

i just saw weird flickering of the freezebar when dummy falls through 1 bar of freeze

Jupstar ✪

i just saw weird flickering of the freezebar when dummy falls through 1 bar of freeze

on latest master?

12:28

I thought I have patched it

yeah relativly latest master

12:28

master from like today morning xd

can you reproduce it? (edited)

Thats expected behaviour

the 2nd to be precise

trml

[ddnet/ddnet] Pull request opened: #5265 Cleanup prediction entity handling a little

I looked over the remaining code for creating/removing entities in the prediction, and only ended up making some smaller changes:

More consistent use of InsertEntity/RemoveEntity/Destroy
Fixed one instance in gameclient.cpp where an entity was removed but not deleted (could potentially have been a memory leak)
Removed some old code for retaining information about the characters after they left the snapshot (the client should only rely on the snapshot for prediction of players)

## ...

mh ok looks really weird imo

12:32

bcs the bar continues where it stopped pretty much

12:32

but flickers

Jupstar ✪

the 2nd to be precise

The freeze bar is turend off if you are in a freeze block. We could vote to let it stay on or we add a option. The freeze bar resets to full if you get contact with freeze after 1 second (edited)

yeah thats weird ddrace logic, not ur fault ^^

12:33

but stars are also always visible arent they

12:33

not even sure rn xd

yes

12:33

you can also see the stars in the video, as soon as the bar refilled 3 new stars poped up

yeah

actually it does not even refill in the video xD I did not see the tee unfreeze at the floor... It just refreeze again

well if we replace stars anyway probs should always show this bar

you probably only mean the freeze bar disapear in the block

yes

I would like a option for it, because I do not like it if a lot of tees sit next to each other in freeze and nothing happens except a lot of freeze bars refilling every second

well idc, can u turn that thing off?

12:39

then u can also just add a third option there

currently nothing happens except a lot of stars getting animated

12:39

yes

Jupstar ✪

well idc, can u turn that thing off?

there is already a option to turn it off, because some asked for it

12:40

Also have to think if we somehow can get the server inform, that freezebar is off, so he sends stars instead (edited)

maybe hackish, but maybe you could generate the stars in the client in that case?

trml

maybe hackish, but maybe you could generate the stars in the client in that case?

yes I also thought about that, but probably requiers prediction?

it would be a bit like airjump I think (which is also not sent), so it could be done without it. the code for it wouldnt be as clean as for just showing the freezebar though

def-

[ddnet/ddnet] Pull request opened: #5266 Use std::make_unique (fixes #5166)

No idea why clang-tidy's modernize-... didn't work

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 if it works standalone, system.c especially
[ ] 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](htt...

bors[bot]

[ddnet:master] 3 new commits

85e8677 Don't keep track of characters in the predicted gameworld that are not in the snapshot and require pred characters to be alive - trml c559520 Fix and slightly cleanup creation/destruction/insertion of predicted entities - trml ab28267 Merge #5265 - bors[bot]

13:02

bors[bot]

[ddnet/ddnet] Pull request closed: #5265 Cleanup prediction entity handling a little

what was the reason again that we overwrite system defaults linker? I dont think this is good style, also it probs wont help for packaging, bcs their CI always downloads packages they need, doubt they download smth like lld for no reason

People complaining that their local build is slow with IPO

14:10

since we enabled IPO by default

14:11

but now IPO is off by default, so probably doesn't matter much

i already said it in the pr, CMAKE has a parameter for that too

14:11

we can just enable it for our full release builds

C0D3D3V

[ddnet/ddnet] Issue opened: #5267 Server does no longer tell that the port is already in use

15.9.1: ``` ❯ ./DDNet-Server "sv_port 8305" [2022-05-28 16:24:07][engine]: running on unix-linux-amd64 [2022-05-28 16:24:07][engine]: arch is little endian [2022-05-28 16:24:07][storage]: couldn't open storage.cfg [2022-05-28 16:24:07][storage]: using standard paths [2022-05-28 16:24:07][storage]: added path '$USERDIR' ('/home/daniel/.teeworlds') [2022-05-28 16:24:07][storage]: added path '$DATADIR' ('data') [2022-05-28 16:24:07][storage]: added path '$CURRENTDIR' ('/home/daniel/Des...

deen

but now IPO is off by default, so probably doesn't matter much

The issue is that the option is a cache variable. You've turned it off, but for those who updated to "prev version with it being ON by default", it is still ON after update to the current master. (edited)

C0D3D3V

[ddnet/ddnet] Pull request opened: #5268 Execute the dragger_beam tick one tick earlier. And stop it one tick earlier

fixes #5263 @bencie could you please confirm it fixes the problem?

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 if it works standalone, system.c especially
[x] Considered possible null pointers and out of bounds array indexing
[x] Changed no physics that affect existing maps
[x] Tested the change with [ASan+UBSan or valgrind...

@bencie could you test this PR if you have time https://github.com/ddnet/ddnet/actions/runs/2401463925 (as soon as it has compiled)

yeah ^^

with a local server and the map Fall into the Future

Mh, how did an dragger rework even get merged without a side by side comparison to check the draggers behaviour?

I did do a side by side comparison

15:41

Its was one tick off... that was hard to see

15:42

one tick to slow activated

15:42

one tick to slow deactivated

15:43

and https://github.com/ddnet/ddnet/pull/5245 made it worse xD ... but that was not my initial rework (edited)

Learath2

Mh, how did an dragger rework even get merged without a side by side comparison to check the draggers behaviour?

we should do unit tests xD to prevent physics changes

15:46

but I do not know if tick related stuff is testable using unit tests :/ probably yes

Teehistorian based ones are possible if we ever get replay capabilities in the server

that would also be an option, but I mean just normal unit tests, without running a complete server, only a dummy, that executes a view ticks with specific inital data

Even without replay we could probably hand parse th files, it’s easy to check stuff like position at a certain tick

yes

The spaghetti we have isn’t really good for traditional mock tests

bencie

yeah ^^

build is ready

Learath2

The spaghetti we have isn’t really good for traditional mock tests

Not easy but with some work, probably doable (edited)

c0d3d3v

@bencie could you test this PR if you have time https://github.com/ddnet/ddnet/actions/runs/2401463925 (as soon as it has compiled)

am i supposted to try this on a local server? on official servers its still the same

with a local server and the map Fall into the Future

ah yeah

1

15:57

works

bencie evolved from cat to anime girl

1

bors[bot]

[ddnet:master] 3 new commits

a3c1a4c Prevent users from setting port 1-1023 - def- 19dcd44 Use std::make_unique (fixes #5166) - def- 719e3e4 Merge #5264 #5266 - bors[bot]

16:17

bors[bot]

[ddnet/ddnet] Issue closed: #5166 Update to `std::make_unique` now that we're above C++11

16:17

bors[bot]

[ddnet/ddnet] Pull request closed: #5266 Use std::make_unique (fixes #5166)

16:17

bors[bot]

[ddnet/ddnet] Pull request closed: #5264 Prevent users from setting port 1-1023

@Tezcan

Jupeyy

[ddnet/ddnet] Issue closed: #2698 Switch to our own config directory?

16:39

Jupeyy

[ddnet/ddnet] Issue closed: #2431 Follow XDG basedir specification

where is the demo render code?

video.cpp in client

ty

Kaffeine

[ddnet/ddnet] Pull request opened: #5269 Introduce and use constexpr CCharacterCore::PhysicalSize

Add Collision() getter to server CEntity to make it easier to sync the server and prediction implementations (e.g. IsGrounded() imps are now identical).

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 if it works standalone, system.c especially
[ ] Considered possible null pointers and out of bounds array indexing
[ ] Changed no p...

GiuCcc

[ddnet/ddnet] Issue opened: #5270 Add possibility to save/retrieve client settings from Ddnet.tw

I would find useful to save whole client configuration on ddnet.tw New client usually asks for player name, but it could also ask to retrieve everything (name, skin, bind, video config, etc...) from ddnet.tw, given login credential of course. This feature does not need much disk space (let's limit to 1KB per user or less) and it would be available only to registered user. It should be easy to implement as long as client settings are exportable/importable serialized. Bind export could...

bors[bot]

[ddnet/ddnet] Issue closed: #5263 Draggers are disabled too quickly

21:35

bors[bot]

[ddnet:master] 2 new commits

556b322 Execute the dragger_beam tick one tick earlier. And stop it one tick - C0D3D3V 5fdb94e Merge #5268 - bors[bot]

21:35

bors[bot]

[ddnet/ddnet] Pull request closed: #5268 Execute the dragger_beam tick one tick earlier. And stop it one tick earlier

def-

[ddnet/ddnet] Issue closed: #5198 MySQL connection stayed stuck for hours

I wonder why GitHub doesn't ping me, if a PR from me in vanilla teeworlds gets mentioned in ddnet

(edited)

hey is there a way to send commands to a running server from a script? I would like to send /reload when the map file changes on disk

def-

[ddnet/ddnet] Issue reopened: #5198 HTTP master registration post hangs on server shutdown

22:14

def-

[ddnet/ddnet] Pull request opened: #5271 Time out for POST requests too (hopefully fixes #5198)

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 if it works standalone, system.c especially
[ ] 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-addresssanitizer--u...

k2d222

hey is there a way to send commands to a running server from a script? I would like to send /reload when the map file changes on disk

you can use fifo file

k2d222

hey is there a way to send commands to a running server from a script? I would like to send /reload when the map file changes on disk

sv_input_fifo server.fifo in your server config, then you can write rcon commands into it: echo broadcast hello > server.fifo

k2d222

hey is there a way to send commands to a running server from a script? I would like to send /reload when the map file changes on disk

It should exist an headless client, ready to use

22:17

Unix fifo is nice! Didn't know it exist

def-

[ddnet/ddnet] Issue closed: #5270 Add possibility to save/retrieve client settings from Ddnet.tw

oho really nice way of doing it! thanks

(doesn't work on windows)

who cares

mkfifo server.fifo
nohup ./DDNet-Server "sv_input_fifo server.fifo" &
echo "command" > server.fifo

ddnet server creates the fifo itself too, no need to manually create

actually, does DDNet-Server accepts commands in command line?

yes

deen

ddnet server creates the fifo itself too, no need to manually create

ah nice

and deletes after shutdown, so you don't end up writing into an unconnected fifo

22:22

the fifos were probably the second thing I added after fixing the obvious cheats, so that we can run DDNet official servers at all

Chairn

actually, does DDNet-Server accepts commands in command line?

what you mean I can write commands in the running server terminal?

22:23

oh no ok

nah, you can give command when starting the server

k2d222

what you mean I can write commands in the running server terminal?

could be nice tough, I know e.g. minecraft servers can

not sure we want to reinvent the wheel

1

bors[bot]

[ddnet:master] 7 new commits

a5f4c66 vmath: Make vector2_base() constructors constexpr - Kaffeine a023671 prediction/entity: Sync constructor signature, add GetProximityRadius() - Kaffeine f934882 server/entity: Add CCollision() getter - Kaffeine f057dd3 server/character: Use CEntity::Collision() instead of GameServer() one - Kaffeine deb2e87 Sync prediction impl with server CCharacter::IsGrounded() - Kaffeine a39bbc7 Introduce and use constexpr CCharacterCore::PhysicalSize - Kaffeine 9e6ee21 Merge #5269 - bors[bot]

22:41

bors[bot]

[ddnet/ddnet] Pull request closed: #5269 Introduce and use constexpr CCharacterCore::PhysicalSize

22:42

def-

[ddnet/ddnet] Pull request opened: #5272 Handle failed socket creation (fixes #5267)

by returning nullpointer, as the calling functions expect Broken in https://github.com/ddnet/ddnet/commit/471bb441a14af9502558cfc7203de96f72547e87

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 if it works standalone, system.c especially
[ ] Considered possible null pointers and out of bounds array indexing
[ ] Changed no phys...

yay super cool !

can anyone answer me collision questions?

default hook length is 380.0 i read in the wiki. from where to where is that? middle of the hooking tee? or on the outline?
what is a unit? how does it translate to pixel?
what is the hook hitbox of tee diameter in units?
is the hook hitbox of tees a round circle? (ddfat skins are not round)?

(edited)

k2d222

yay super cool !

cool, a new map editor?

22:54

ddfat is just a fun way to make the tees look, it does not illustrate anything

22:54

(no idea about the other questions)

k2d222

Setup the repo for the web map editor at https://github.com/k2d222/twwe and a demo server is (sometimes) running at http://tw.thissma.fr (without tls, not setup yet) (...and it even works on mobile

)

.

really?, because i read that on the wiki

To see how big the actual hit box of tees are, you can turn on "fat skins" in Settings -> Tee -> Fat skins (DDFat).

I chose the size for fat tees based on what looks funny, so if it's close to the hit box, that's coincidence

1

k2d222

hey is there a way to send commands to a running server from a script? I would like to send /reload when the map file changes on disk

Another option is ECON. Long story short:

COMMAND=$@
printf "$EC_PASSWORD\n$COMMAND\n" | nc $EC_CONNADDR $EC_PORT -q 1

Very cool project @k2d222 . Collaborative mapping

1

so the hookcollline, does it start from middel of tee, or from the outline? in other words, is the hookrange radius of some tee box.. + hook length, well i'll just assume from middle

center

couldn't we use steam cloud to store client settings to fix this issue? you don't need ingame accounts for that https://github.com/ddnet/ddnet/issues/5270

Add possibility to save/retrieve client settings from Ddnet.tw · Is...

I would find useful to save whole client configuration on ddnet.tw New client usually asks for player name, but it could also ask to retrieve everything (name, skin, bind, video config, etc...) fro...