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 2021-01-13 00:00:00Z and 2021-01-14 00:00:00Z

this is ok?

Image attachment

the brutal and solo rows continue there, but yeah looks buggy. Needs horizontal scrollbar

the scroll bar is at the bottom, a bit awkward

Oh well

Ryozuki: do you use yay with sudo or without? For me it was passwordless the first time i installed it that was nice

yay w/o sudo

https://xkcd.com/1200/

xkcd: Authorization

13:44

Doesn't yay use sudo to get root?

yes

13:45

but the build process shouldn't run as root

13:45

so you start it without sudo and let it escalate privileges itself

yes, you should always use it without sudo like in the docs

I mean i think the last time i installed it it had a SUID set or something it never needed the admin password to do things

13:54

oh btw yay i managed it to run memtest its running since a few mins im not sure how to read the output but there is alot red stuff and it says 2501 errors so far

do you have passwordless sudo?

does that mean my ram is broken?

2501 errors means something is broken

no i did not had passwordless sudo

there should be 0 errors

/usr/bin/yay doesn't have suid at least on my machine

i run windows for like 2 hours and it run super smooth

i do have passwordless sudo

so buying new ram should fix my problems?

either a ram chip is broken, or it's not properly in the socket

i did not touch my pc

idk, do oyu have a different ram stick to test whether that makes your problem go away?

you gotta figure out which ram stick is broken

yikes

13:55

can i simply unplug one?

you can try, some sticks?/motherboards? require ram sticks to be in pairs, I think

It might be that you gotta have it in the 1st and 3rd slot when you only have 2 or something like that

13:56

depends on the mainboard

oof

nah 3 is fine

13:56

you just gotta figure out which you can remove

and run memtest again to figure out which ramstick is bad

so there is no more value in running memtest anylonger with 11k errors i assume

nope, you need 0 errors

ok

13:58

i should have done that before trying to wipe my system :/

my dad had this problem on my parents home server and it only occurred after the system was running for days

14:02

i just asked him and he said there's no way he knows of you can see which ram stick is broken from the test. so i'm afraid you need to take them out separately and rerun the test (or maybe binary search ^^ ) (edited)

ChillerDragon: If memtest shows that always the same part of memory is broken, you can disable parts of the RAM stick on Linux

14:03

Cheaper than buying new RAM

oh wow, the things linux can do

14:04

but kinda makes sense, if it can be fixed in software, no need to replace hardware

memmap=1$0x0007cec2d74 can be passed to the kernel to disable a single byte for example

14:05

at address 0x0007cec2d74

14:05

Actually I should have noticed a few days earlier that ChillerDragon's problems sounded familiar: https://hookrace.net/blog/broken-hardware-fixes-hacks-8-years/#random-access-memory

did you have similar symptoms?

14:08

i wonder why windows run so well tho

for my dad it just crashed linux after a few days

perhaps windows happens to map less stuff to the bad range

my hardware is like 8 years old i am fine with investing some money again tho im a bit scared to touch hardware

don't do it like a friend of mine and buy incompatible ram sticks 2 times

(edited)

haha classic

14:10

happend to me 3 times to me with razer blades xd

oh shit

oh not the gaming brand razer blade

14:10

but like actual

14:11

u know

14:11

xd

14:11

blades for a razer

razor?

ah ye

haha

oh man

mostly nothing can go wrong with touching hardware

14:11

just don't use excessive force

14:12

and make sure to apply thermal paste!

i can cause segfaults by just looking at software trust me

14:12

i break all the things

Embedded image

oh boi ram is expensive af i did not know that

https://www.youtube.com/watch?v=VwR3VALnycQ

Thermalpaste: Part II

thats the wrong place i assume ._.

14:18

dont show me such stuff i might follow the steps lul

don't ^^

14:19

https://www.youtube.com/watch?v=3GsRWZcuM0M

Installing memory

on it :)

argh dont' follow it

xxd

14:29

i removed two slices yikes

14:30

the 4th one is unreachable :/

14:30

yikes 10 errors in 1 second

14:31

ok so the ram that is under the cpu thingy is broken i guess i wont touch that :/

u mean the fan? xd

14:47

you can just take it off... more often than not theres no screws involved

heyho, friends and me developed a register for a homepage. was wondering if that could be a thing for teeworlds as well?

i guess people here will appreciate the efforts, but i dont think that idea is too innovative

socket could be broken aswell btw

ye the fan is big fat

15:01

what is a "register"?

I've got the Alpenföhn Brocken, love that name

xd

15:05

i tried violence :)

15:05

0 errors so far

15:05

ah nvm

15:05

fok

:(

can i see which ram is afftected in memtest? I mean u see the regions tested but are those from left to right the physical memory?

15:07

wat?

Original message was deleted or could not be loaded.

i have literally no idea what you are talking about... but now im curious to see the code

what code what are you all talking about?

15:18

oh that kind of register

Original message was deleted or could not be loaded.

ok fam gotcha

https://github.com/ddnet/ddnet/issues/3411

Accounts · Issue #3411 · ddnet/ddnet

If anyone wants to summarize the discussion, feel free to. I just wanted to create the issue now.

[quakenet] ChillerDragon

https://github.com/ddnet/ddnet/issues/3411

alright, i will stop here and leave it how it is

i like "Zoozti" argument. never really thought about it consciously, but thats a huge reason why i love this game

15:25

ty boys

since i cant reach the ram i guess i have to use memmap to turn it off

15:41

where do i provide this kernel arg never did something like this

what is your bootloader?

15:42

there

Grub cfg

if grub is your boot loader

so in /boot/grub/grub.cfg i put GRUB_CMDLINE_LINUX_DEFAULT="memmap=xyz"

15:43

?

15:43

ye i use grub

does the file say something about being automatically generated?

im in memtest but i think not that it is auto generated since i generated it manually

for me, it says # DO NOT EDIT THIS FILE

oh

15:44

then /etc/default/grub regarding to archwiki

Ye

does memtest also use this then?

Probs

nice

15:46

ok ima try

Remember to generat the config again

ye :)

15:46

learned it the hard way during arch and memtest install xd

not sure about memtest tbh

15:48

these are linux command line options

15:48

I doubt memtest runs linux

on the ddmax map "NUT_hardcore_bestof" the (completely unused) envelopes are broken in a way that makes the editor crash when viewing the third envelope and hovering over the broken looking stuff

15:50

the fix (removing all envelopes) can be replicated like usual with https://gitlab.com/Patiga/twmap/-/blob/master/src/bin/fix_map.rs

15:50

NUT_hardcore_bestof-fix.map

270.13 KB

https://gitlab.com/Patiga/twmap/-/blob/1357369ad0998ead9233377446516e5f61589866/src/bin/fix_map.rs permalink

src/bin/fix_map.rs · 1357369ad0998ead9233377446516e5f61589866 · Pat...

Map parsing library

fix_map --ddnet --remove-all-envelopes NUT_hardcore_bestof.map NUT_hardcore_bestof-fix.map

(press y on github/gitlab to get the permalink)

yikes htop says i have 3gb ram

16:08

hmm so memmap=5000M$00210e6e270 shows 3gb in htop and memmap=4000M$00210e6e270 shows 31gb i am confused

16:09

well ill try the 31 and see if my system is stable

[ddnet/ddnet] Pull request opened: #3506 Send the NETMSG_ITIS/NETMSG_IDONTKNOW from the correct connection

Noticed after #3505, thanks @fokkonaut.

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 memcheck](https://github...

[ddnet/ddnet] Issue opened: #3507 Pylint failure centrally but green locally

``` datasrc/network.py:1:0: E0611: No name 'NetMessageEx' in module 'datatypes' (no-name-in-module) datasrc/network.py:1:0: E0611: No name 'NetObjectEx' in module 'datatypes' (no-name-in-module) datasrc/network.py:1:0: E0611: No name 'NetStringHalfStrict' in module 'datatypes' (no-name-in-module) datasrc/network.py:360:1: E1123: Unexpected keyword argument 'teehistorian' in constructor call (unexpected-keyword-arg) datasrc/network.py:399:1: E1123: Unexpected keyword argument 'teehistoria...

[ddnet:master] 2 new commits

dccbf26 Differentiate hook symbol - def- 8c392ee Merge pull request #3503 from def-/pr-hook-symbol - def-

17:05

[ddnet/ddnet] Pull request closed: #3503 Differentiate hook number by showing symbol

[ddnet/ddnet] Pull request closed: #3501 RFC: Introduce CGameContextDDrace and take some very first benefits from it

What are freenode/quakenet? Is the ddnet irc hosted on a ddnet vps or on one of those providers? are passwords on irc something specific to freenode/quakenet?

hosted on both of them, connected with bridge

19:12

usually we used quakenet, the most popular irc server for games

19:12

freenode is the most popular one for open source software

19:12

only Zwelf is on freenode I think

19:13

each server has its own pw implementation

is it unusual to host your own irc?

yes

that's all i wanted to know, thanks :)

similar to mail isnt it

hmm weird since i reinstalled arch my 3rd screen does not work and if i install the package nvidia ONLY my third screen works lol

20:29

why is life so hard

yeah sometimes linux feels like a constant fight against entropy ^^ (edited)

i cant remember any problem when i installed arch the first time on this machine tho its so weird

20:54

i should have listend to deen and backup my whole drive to then restore it :D

ChillerDragon thats why u get amd gpu if u use linux

21:09

i mean, u wont use the rtx shiny stuff on linux anyways

21:09

i think amd drivers are better in 2021

21:09

on linux

21:10

ChillerDragon whats ur xrandr command?

21:10

maybe it fixes it

Ryozuki: you know arch packaging, why is here pkgver defined and then a function called pkgver as well? which one is use? https://aur.archlinux.org/cgit/aur.git/tree/PKGBUILD?h=passff-host-git

i use xrandr --output DVI-D-0 --primary --left-of HDMI-A-1 --mode 1920x1080 --rate 144.0

21:11

@timakro its cuz its a -git package

21:11

they use dynamic version

arandr is a nice GUI for xrandr

21:11

but it can't do everything xrandr can

21:11

I still use it for setting up projectors

@timakro pkgver=1.0.1.r7.1feab77 is a version set when the pkgbuild was released but it doesnt mean anything for a git pkg

21:12

so they implement a function to get the true version

ok

21:12

thanks

@timakro https://wiki.archlinux.org/index.php/VCS_package_guidelines

VCS package guidelines

21:12

look this

21:13

The pkgver autobump is now achieved via a dedicated pkgver() function. This allows for better control over the pkgver, and maintainers should favor a pkgver that makes sense. To use pkgver(), you still need to declare the pkgver variable with the most recent value. makepkg will invoke function pkgver(), and update variable pkgver accordingly.

@Ryozuki Is this sed some kind of standard?

21:13

git describe --tags --long | sed 's/\([^-]*-\)g/r\1/;s/-/./g'

i guess its standard if u follow semver or something

It changes 1.2.1.r1.2bceedd < 1.2.1-1-g2bceedd (edited)

21:14

no

ah

21:14

i see what u mean

other way around xD

1.2.1.r1.2bceedd

21:14

this is how it ulooks

21:14

usually

21:14

from my experience

why is this better

21:15

the thing is, this breaks the package haha

i think the r means something

21:15

what do u mean

breaks this ^^ https://github.com/passff/passff/blob/c3456219d204c7df4b9230ba267da4ece6e8810d/src/modules/util.js#L217-L222

The revision number delimiter ("r" right before REVISION) is important. This delimiter allows to avoid problems in case if upstream decides to make its first release or uses versions with different number of components. E.g. if at revision "455" upstream decides to release version 0.1 then the revision delimiter preserves version monotonicity - 0.1.r456 > r454. Without the delimiter monotonicity fails - 0.1.456 < 454.

21:15

this the reason

ah, it helps ascii sorting

21:16

thats nice

21:17

Now where should I file the bug report xD

21:17

I guess not on arch

what does that extension do

It's for the pass password manager https://www.passwordstore.org/

ah

21:18

the extension wants to read the pass pkg version?

well actually the pkgbuild writes the version into a placeholder in the python script it then installs

Image attachment

21:19

whats wrong with this one?

21:19

it doesnt have r stuff

the browser extension then talks to the python script and checks the version and then just stops working without an error message -.-

21:19

Oh this is the firefox extension for pass

21:20

Comes in two parts, the firefox extension and a python script on the host which talks to firefox

u need this

21:20

https://github.com/passff/passff-host

passff/passff-host

Host app for the WebExtension PassFF. Contribute to passff/passff-host development by creating an account on GitHub.

That's the aur package

21:21

I'm making a pr, gotta understand js regex first though

Image attachment

21:21

use the non git version

21:21

xd

Oh awesome

do u use a aur helper?

I use yay

maybe u should improve ur search or smth xd

21:21

Image attachment

21:21

monkalaugh

I looked on aur.archlinux.org xD

xd

21:22

u can search using yay <regex>

Ok never used that

21:22

Awesome, works ^^

poggers

hi@deen, for https://ddnet.tw/mapstats/ personally I would sort in reverse order, top of the page being most recent. list would keep growing and users will keep scrolling.

could u also show the number of finishes for each map in the month too if possible?

hover the image

ah nice

multeasymap

25k finishes

Even crazier, in september 2020 sunny side up had 30k+ in a month, which is 1027 each day

both august and september

Exported 243 message(s)