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 2020-12-07 00:00:00Z and 2020-12-08 00:00:00Z

@Ryozuki same mistake for me for part 2

[ddnet-maps:master] 1 new commit

0c03ed2 A Lavender Forest - ddnet-maps

[ddnet-web:pr-funding-2021] 1 new commit

e37cf31 Update funding for 2021 - def-

sadSnail

my regex aint workin sadSnail

sadSnail

Image attachment

@Ryozuki try +?

09:20

matches the least amount of chars

09:21

but at least in the py regex engine, you can only extract the last match of something, so this wouldn't work

yeah im doing a different aproach

09:21

i use 2 regexes

09:21

Image attachment

also what I did

POGGIES

rust regex is okay with \w in []?

why wouldnt it?

09:22

isnt \w a standard

idk, let me test python's. I don't know if these \ classes are okay in []

i always used em

09:23

and they worked

k, seems to work in py as well

09:23

TIL

Image attachment

09:24

:O

09:24

i knew smth heinrich didnt?

09:24

monkaSteerS

POG

this doesn't say anything about \w in []

true

09:25

i guess u thought it didnt work cuz \w translates to [a-zA-Z0-9_] and it would mean a [] inside a []

09:25

was that it?

yes, kinda

09:25

. also loses its special meaning in []

09:25

. usually matches an arbitrary character, but in [] it only matches a dot

oh did not know that

09:26

so what is [.] then?

just matching "."

09:26

one of the ways to escape a dot in regex

so [.] and [.] is the same

huh? did you write the same string twice?

xd

09:27

did the bridge eat my backslash?

yes

Embedded image

09:27

rude bridge

might be discord

sadSnail

Image attachment

Image attachment

sadSnail

Image attachment

09:43

aah

09:49

my program is running for more than 1 minute pepe_straight

pepe_straight

09:50

2 minutes more or less

09:50

but i got the answer lmao

Does anyone know why my CI is failing?

10:14

I set the minimum CMake version to 2.8.12 already, but it still fails

10:14

https://github.com/fokkonaut/F-DDrace/runs/1509892132

yeah, all fail. I'm waiting for a good idea from @heinrich5991 about how to fix it

10:32

it's also because of gmock

sorry

10:39

let me promise to fix it today

10:39

it's been on my todo list the whole weekend

using btreemap and btreesets i got it down 0.13s

10:50

aPES4_Steering

10:53

https://github.com/edg-l/adventofcode-2020/blob/master/day7a.rs

day7?

10:59

there must be an algorithmic improvement lying around ^^

https://github.com/edg-l/adventofcode-2020/blob/master/day7.rs with 2 parts

10:59

yeah i used a mess before

10:59

with hashmaps and recursion

10:59

xdd

my python solution runs in 0.04s, and I didn't optimize it for speed, but for speed of writing

it literally took 2 mins to get the answer

11:00

ah

11:00

i "time" ed it using cargo run

11:01

which adds a significant delay

ah

11:01

you traverse the data structure for each query

this does both a and b problem

Image attachment

yea, mine also does a and b

how long is ur input?=

11:02

it also depends on that

11:02

i guess

594 lines

11:02

the problem is that you do a search for each bag

45.25 KB

11:02

try with this input

11:02

yeah i guess its not optimized still

 time python day7.py ryozuki7
229
6683

________________________________________________________
Executed in   71.08 millis    fish           external

0.07?

yes

:o

11:03

i win then

11:03

sadSnail

I guess? ^^

thats a weird time output

fish shell

11:04

yours is bash's

yea

11:05

the problem is that you do a search for each bag

11:05

so how can i improve it

do it backwards

11:05

see which bag contains a shiny gold bag

11:06

and then check which bags contain those bags, etc.

ah

11:06

so i dont go that deep

11:06

everytime

yep

11:07

you'll only have to check exactly as many bags as the output is

i guess my second solution is good as it is right now right?

yes

problem b

hm

11:09

it could be better, but I didn't do it better either

11:09

if I may comment furtherly: don't pass &String as parameter, pass &str. it allows for strictly more types to be passed

11:09

same for &Vec<Type> → &[Type] (edited)

11:10

(but that doesn't occur in this code)

yeah but i didnt want to type lifetimes everywhere

11:10

xd

&String has the same amount of lifetime parameters as &str, none

11:10

try it

oh

just replace &String with &str and the code should still work. then you can remove .to_owned() from the function call

it doesnt complain on the .get

no

nice

11:11

Image attachment

I don't see what you're trying to show me ^^ explain?

that it works

11:12

xD

kk ^^

my vim can see errors

11:12

with rust-analyzer

fancy

see, it complains q is unused there

11:12

Image attachment

and now do println!("count {}", count(&bags, &"shiny gold".to_owned()) - 1); → println!("count {}", count(&bags, "shiny gold") - 1); for one less allocation

11:13

&"shiny gold".to_owned() → "shiny gold"

ye works

nice

11:14

all I had to complain about, now I'll be quiet

11:14

if you want to fix the warning, you can replace the identifier by _ or _q

ye did it

nice that you're doing it in rust ^^

fancy neovim POGGIES

POGGIES

Image attachment

11:15

ofc

11:15

i love rust

meh demos are broken again

14:53

"Invalid demo"

14:54

for like 8 of them

14:54

using save_replay

Am I overthinking day7?

Image attachment

probably?

yeah figures, kinda wanted to put it all in a graph and do a topological sort

oh, that worked

maybe u should spred more the nodes

15:20

its a bit messy in the center

15:20

xd

15:20

spread*

15:21

@Learath2 r u using python still?

I really don't like Perl: https://github.com/def-/adventofcode-2020/blob/master/day07a.pl

@Ryozuki yeah using networkx

15:27

The second part is a little tricky though, still thinking about it

second part was easier for me

15:27

xd

I know what I want to do, not really sure how to put it in graph algorithms terms

prolog and perl have the same file extension?

15:27

TIL

I need a sort of depth first search

15:29

but I need to know where I'm coming from too and the billion functions in networkx don't make it easy to find what I'm looking for

hmm

15:29

idk i just used btree map and set and a bit of brain

15:29

pepe_straight

15:29

sadSnail

A btree? Huh, that graph is most definitely not binary, I'll take a look at your solution when I'm done, I'm intrigued

https://doc.rust-lang.org/std/collections/struct.BTreeMap.html

std::collections::BTreeMap - Rust

API documentation for the Rust BTreeMap struct in crate std.

15:31

pepeH

[ddnet/ddnet] Issue opened: #3385 [question] sharing dotdir with teeworlds

Hi, I'm currently working on official DDNet package for FreeBSD, and I've noticed that it uses the same dotdir as teeworlds (package of which I also maintain). The question is whether such configuration is safe, e.g. can ddnet modify (and possibly break) teeworlds configs or can teeworlds modify (and possibly break) ddnet configs, taking into account that either teeworlds or ddnet configs format may change in the future. Wouldn't it be safer to use dedicated dotdir for ddnet?

I can't believe how rusty I've gotten with algorithms...

today problem

Image attachment

Pf, I give up. I must be misunderstanding something

Heh, on Windows you can use Win+. to enter emojis, kind of works in DDNet

16:30

But with the mouse you can't select the emoji, only with keyboard. @TsFreddie could that be imrpoved?

16:31

and of course our font doesn't support emojis well. Should we add more, and if so, in color? I noticed colored fonts don't work currently

probably not tbh

16:31

I mean improve the experience with external program.

How about switching to not capturing input? then it works with external program. As is currently done in F1

We probably couldn't know when to switch.

16:33

I don't think win+. works like ime either

@deen IIRC we backed out of emojis last time

yes, because they were totally broken

[ddnet/ddnet] Issue closed: #3385 [question] sharing dotdir with teeworlds

uhm, I think because people abused them

but using a proper emoji font would also be better

16:34

Oh well

Our font texture is currently a 8bit channel. So basically only alpha iirc

16:35

We will need a separate texture for colored emojis

can't find anything on the issue tracker

16:37

https://forum.ddnet.tw/viewtopic.php?f=3&t=6058

I do want a better selection of emoticons in teeworlds in general. Some of them works similarly like the angry ones (looks like we have four of them) while not having sad / crying ones.

TsFreddie

We probably couldn't know when to switch.

Also probably will break in exclusive fullscreen mode like ime does.

I'm still way overcounting :/

[ddnet/ddnet] Pull request opened: #3386 Add FreeBSD instructions

https://www.freshports.org/games/DDNet Thanks to @AMDmi3. Not sure if "pkg install DDNet" can also be used.

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
[ ] Te...

How am I supposed to clear out a folder in the tw source?

17:41

i wanna delete all files in a folder (or delete whole folder and re-create the empty folder :D)

17:41

how?

17:41

I cant use ListDirectory, as the file is still opened then

ListDirectory shouldn't open the individual files in the directory, then you can call RemoveFile on them

Does not seem to work

17:58

i will try it once again and make sure i am doing it correct

18:00

Image attachment

18:00

this is correct, is it?

18:00

within the callback of the listdirectory

[ddnet-web:master] 1 new commit

7cd633a Update servers - def-

You should check the return value of RemoveFile. pName might be missing the prefix of the directory, and just be the filename

18:02

so you might have to put that together yourself

but that should be correct already

18:03

before calling ListDirectory i passed the right path that also worked before when directly calling

18:03

let me try it again

18:04

it found both files, but they are not deleted

Embedded image

18:04

xd

oh @deen

18:05

thanks

18:05

it really cut off my path

18:05

in the front

https://github.com/Learath2/aoc/blob/master/2020/7/solve.py

Advent of Code. Contribute to Learath2/aoc development by creating an account on GitHub.

18:15

Absolutely shameful how long this took me :/ Need to brush up on graphs again

Though it does print a cute graph if you use the dbg function

I was also thinking about solving this with a graph db

[ddnet-web:master] 1 new commit

cfac1c0 ger3 - def-

@deen what assumption allowed you to reduce the problem so much?

tomorrow is the day of cmake fixes. I'm going to bed. sorry

oops

@Learath2 which problem? AoC? I don't think I reduced anything

AoC, I had to use like an entire graph

I just got really annoyed by the way types work in Perl, it's horrible

21:43

I have a hashmap too

21:44

but who knows that $m{$a} = ... makes $m a hashmap...

You seem to have reduced it to a hashmap somehow

I feel like there is no more development on DDNet since AoC

tho I guess hashmaps can be interpreted as graphs somehow?

Exactly, I just recurse on subroutine (function would be too boring) f

I'm really lacking in theory

I just store the graph as key = from and value = to in the hashmap

Exported 242 message(s)