DDraceNetwork - developer

hello. what is the format for map files and is there a way to load a map on startup?

GitHub

Click to see attachment 🖼️

@Skeith

why did he step down

Load map where on startup of what?

Replying to @david loves Jellycar Worlds hello. what is the format for map files and is there a way to load a map…

cyberfighter 2

just dont use speeders with a speed limit

wait does this bug not happen without max-speed? This would indeed be very helpful, can anybody confirm this?

this code is a mess and I see at least 2 divisions by zero.

ddnet/src/game/server/entities/character.cpp at 85a30b6c6dcda31e791...

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

Assa

this code is a mess and I see at least 2 divisions by zero.

Looks messy indeed. Looks they actually wanted to compare against < -0.00001 So it can't be null

agreed, because the else case asin(-1) would be the same as atan(Infitiy) here (in terms of angle)

10:04

same below for TempVel btw

10:07

going to make a PR

AssassinTee

[ddnet/ddnet] Pull request opened: #9658 Fix speedtile max-speed bug

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/#using-ad...

Ewan

why did he step down

Nothing interesting, wanted some time off

just leave it on by default

GitHub

Click to see attachment 🖼️

it's used in some maps, no?

@Discord Mod

goddamn bro its not even night

14:28

keep yourself safe (kys)

zhn

it's used in some maps, no?

mentioned maps: Pharaoh, Broken brain, Cosyris, Binary

GitHub

Click to see attachment 🖼️

if anyone knows any other map that relies on speedtile max speed bug please mention it in this pr!

Fly to the moon xdd

14:31

it doesnt rely on it, but if u want r1

zhn

mentioned maps: Pharaoh, Broken brain, Cosyris, Binary

Binary is a different mapbug, has nothing to do with this

14:36

@zhn Jelly is another map, but this uses yet another speedtile bug

zhn

keep yourself safe (kys)

what happened

Hecta

what happened

just typical discord spam

Jupstar ✪

it doesnt rely on it, but if u want r1

can you show me where? I can't find a single speedtile with max-speed set in that map

Assa

can you show me where? I can't find a single speedtile with max-speed set in that map

in the rocket

14:43

bottom of map i think

@Robyt3 can we talk about your suggestion?

yes

how is this not the same (logically) else if(Direction.x < -0.0000001f || (GameServer()->EmulateBug(BUG_SPEEDTILE_MAXSPEED) && Direction.x < 0.0000001f)) vs else if(Direction.x < (GameServer()->EmulateBug(BUG_SPEEDTILE_MAXSPEED) ? 0.0000001f : -0.0000001f))

Jupstar ✪

in the rocket

can you send a screenshot, I may be blind ;_;

@Robyt3

Assa

how is this not the same (logically) else if(Direction.x < -0.0000001f || (GameServer()->EmulateBug(BUG_SPEEDTILE_MAXSPEED) && Direction.x < 0.0000001f)) vs else if(Direction.x < (GameServer()->EmulateBug(BUG_SPEEDTILE_MAXSPEED) ? 0.0000001f : -0.0000001f))

If GameServer()->EmulateBug(BUG_SPEEDTILE_MAXSPEED) is true then the first is equivalent to else if(Direction.x < -0.0000001f || Direction.x < 0.0000001f) which is not the same as the original buggy version

it is

15:08

else if(Direction.x < -0.0000001f || Direction.x < 0.0000001f) is the same as else if(Direction.x < 0.0000001f) which is the original buggy version

ah, true, but I think my variant makes this a lot clearer

just make a new speedyile bug that actually works

15:10

speed tile*

15:10

just woke up justatest

Robyt3

ah, true, but I think my variant makes this a lot clearer

if GameServer()->EmulateBug is expensive, my variant is cheaper, that's why I wrote it like this, I am going to take a look at EmulteBug

15:14

@Robyt3 it's a bit more expensive with multiple function calls, but might be very well optimized out by branch prediction, what is your final verdict? I guess switching to yours

I prefer readability unless there's a measurable performance improvement. Not sure about the PR in general though, maybe other maintainers prefer not changing physics

Only problem is that it probably affects many maps

15:21

from a speedrun pov

i say make a new tile

15:21

speedup layer is effectively empty anyways

15:22

having multiple behaviors for a single feels like bad practice, unfortunately ddnet already has a lot of these cases

Jupstar ✪

Only problem is that it probably affects many maps

I should probably run a script checking all maps for speedtiles with max_speed set

louis

having multiple behaviors for a single feels like bad practice, unfortunately ddnet already has a lot of these cases

this already has an edge case with speed 255, which is also bugged. I also like the idea of adding a new speed tile instead

Assa

I should probably run a script checking all maps for speedtiles with max_speed set

yes, would be good, if we want to merge this, if we can be relatively sure that all maps that use this bug are marked properly

15:27

existing affected maps also have to be changed to add the mapbug server setting, otherwise the hardcoded mapbug hashes are brittle and will break if any of the maps are changed later

@Learath2 hi, can you help me xd. I was trying to make an interference graph using this vid (https://www.youtube.com/watch?v=eeXk_ec1n6g), at ~11:30 there's an algorithm how to calculate it, I implemented it and wrote a test program

+--------------+ defs {%1, %2}
|%0:           | uses {%1}
|%1 = 1 + 1    | in {%1, %4}
|%2 = 0 + %1   | out {%1, %2, %4}
|goto %3       |
+--------------+
+--------------+ defs {%4}
|%3:           | uses {%1, %2, %4}
|%4 = %1 + %2  | in {%1, %2, %4}
|ret %4        | out ∅
+--------------+

but it gives a weird in variables for block %0 block, is it my skill issues or it's the correct answer? UPD. uses were calculated wrongly (edited)

Assa

this already has an edge case with speed 255, which is also bugged. I also like the idea of adding a new speed tile instead

id much rather have a new speed tile

16:51

maybe a tile that can forcefully set a tees speed to something specific

16:52

as well

16:52

instead of needing some stinky speeder setup

what was i on last night

cyberfighter 2

id much rather have a new speed tile

I am going to explore this idea

would be funny to have nade speedup tiles

17:31

but i dont think it would lead to good maps

I wonder, would it be possible to have tiles applying tunes to a player instead of having the tunes as we know them (applied to the server or area)?

Pipou

I wonder, would it be possible to have tiles applying tunes to a player instead of having the tunes as we know them (applied to the server or area)?

a pr was merged with this behavior recently

18:27

it's mainly for mod support but it can be extended to support tune lock (which is a pending pr)

Oh I wasn't aware

18:29

It could be fun to have different settings for different players

18:29

Like one has a short hook, or a stronger grenade

@Pipou we could make arctic festivity v2 with that much power

1

louis

it's mainly for mod support but it can be extended to support tune lock (which is a pending pr)

pending since 2022 :/ (edited)

but im MAD because tune lock is at the back burner!

Everything is

when tune lock gets merged im gonna start mapping

kebs

pending since 2022 :/ (edited)

its incomplete iirc

louis

its incomplete iirc

someone should take over the pr i guess

18:51

same for 128p support

^^^^^

https://x.com/melqtx/status/1888140533215662231

mel (@melqtx) on X

just realized open source is just one person who actually does the whole thing and the rest of the 1000+ contributor is there to just edit the readme

19:42

kebs

https://x.com/melqtx/status/1888140533215662231

looks at robyt3

Hecta

what was i on last night

were you affected by the clean code propaganda?

this doesnt look clean

20:28

at least the whole std::any part

It doesn't even work correctly AFAICT because std::find compares the addresses of the strings and not the strings' contents

ohhh

20:30

so how would i fix it, in tangent of the new CheckIndexFlags function?

20:30

maybe return -1 if none match and an input bool is true?

I'm not really familiar with the automapper code. Either use a plain old loop instead of std::find, or use std::find_if and pass a predicate lambda that uses str_comp to find the matching entry. By default std::find will compare objects with ==, so it will compare the strings' addresses.

Hecta

maybe return -1 if none match and an input bool is true?

Prefer using std::optional if you want to represent the absence of a value instead using a magic value like -1

hell yeah options mentioned

if u use string but not const char*, i think u can use std::find

I did have a silly idea that I want to try, but I'd need to be home in order to try it.

Pioooooo

if u use string but not const char*, i think u can use std::find

We only use std::string sparingly though

Somthing like:

NewIndexInfo.m_Flag = CheckIndexFlag(NewIndexInfo.m_Flag, aOrientation1, true);
if(NewIndexInfo.m_Flag == 0 && str_comp(aOrientation1, "NONE"))
    NewIndexInfo.m_TestFlag = false;

Robyt3

We only use std::string sparingly though

we could change that greenthing

dunno what use itd have that we cant already do

We already use it were it makes sense, e.g. when we want to handle any length and the alternative is just doing malloc manually, but we don't want to use std::strings where performance matters

I admit having written java before

Robyt3

I prefer readability unless there's a measurable performance improvement. Not sure about the PR in general though, maybe other maintainers prefer not changing physics

yes please, don't change physics...

Robyt3

[ddnet/ddnet] Pull request opened: #9659 Refactor keys enum, key names and scripts/gen_keys.py

Enable pylint for the scripts/gen_keys.py script and fix various issues. Avoid duplicate code and make script more extensible by using loops and functions. Avoid clang-tidy NOLINT(misc-definitions-in-headers) by moving the definition of the key names from the keynames.h header to the keynames.cpp source file. Reorder the joystick hat keys so the order is consistent with the mouse wheel directions. Improve various related comments. Also check style of keys.h, keynames.h...