DDraceNetwork - developer

I'm downloading so many teehistorian files rsyncs --append-verify is taking a couple minutes to finish verifying on resume

accidentally unplugged the drive, now I have to wait half an hour for fsck to complete

exFAT is absolutely not made for storing files like this

07:16

teehistorian is really good at exposing the shortcomings of using a filesystem as a db

Some research reveals all cross platform filesystems are shite

sunl888

[ddnet/ddnet] Issue opened: #3968 It is recommended to add the automatic elimination logic after hanging up for more than 30 minutes

It is recommended to add the automatic elimination logic after hanging up for more than 30 minutes, which always prompts that the server is full. In fact, there are a lot of players hanging up

quicc tips by /me: If any of you have to mess around with exfat, you can be much faster than coretools if you examine the raw file system

07:56

e.g. I needed to count files, so I hexdumped the directory entries, and grepped that instead

just format ur drive a ext4

No good FUSE driver for ext

08:50

No thus ext support on mac or windows

08:50

Thus no*

who cares about mac or windows

08:56

yes yes linux superior and all but 90% of world desktop users don't use it, so until that changes a lot of people care

not me

08:59

@Learath2 what about btrfs

08:59

there is a open source driver for it on windows

08:59

it works well i tested it

08:59

https://github.com/maharmstone/btrfs

GitHub - maharmstone/btrfs: WinBtrfs - an open-source btrfs driver ...

WinBtrfs - an open-source btrfs driver for Windows - GitHub - maharmstone/btrfs: WinBtrfs - an open-source btrfs driver for Windows

I don't think there is a macOS driver for it :/

does rly macos not support it

08:59

shouldnt it be ez

08:59

since its unix

you'd think so, but porting filesystems is apparently not easy, for some reason

@Learath2 ntfs?

09:01

i saw talks about improving the driver for it on linux

09:02

maybe soon it lands

Sadly I think the only good choice for a cross platform fs is ntfs

https://www.phoronix.com/scan.php?page=news_item&px=Torvalds-PR-For-NTFS3

Linus Torvalds Calls On Paragon To Send In The New NTFS Driver

I'm surprised microsoft hasn't submitted a ntfs driver yet btw. They were doing so much on linux lately

bors[bot]

[ddnet:master] 4 new commits

8451775 Make components member of CGameClient - Jupeyy 26c5332 Remove components pointers - Jupeyy 7e55b64 Remove global CGameClient instance - Jupeyy b98263b Merge #3952 - bors[bot]

09:57

bors[bot]

[ddnet/ddnet] Pull request closed: #3952 Remove global components and global gameclient instance

@Ryozuki is your python good?

ye, he actually feeds it with mice

3

Learath2

@Ryozuki is your python good?

i know some python

mh, nvm I found a bug in pythons mmap module

10:59

or atleast in how it interacts with re

uh oh

oh uh

okay it's not mmap I have no idea what it is..

https://thehistoryoftheweb.com/a-fun-list-of-browsers-youve-never-heard-of/

LOL a single tempfile in the directory broke my assumption that the directory entries are properly aligned...

11:28

wasted an entire hour

that‘s why you use a library instead of looking at the memory yourself

@deen I doubt many exfat dirent parsing libraries exist

13:47

13:52

hm, FATtools does exist apparently, I wonder if it exposes it's internal data structures though

13:53

Actually if walking the list was fast ls would probably be fast too

@Learath2 https://github.com/rust-lang/rust/pull/86761 (edited)

Update Rust Float-Parsing Algorithms to use the Eisel-Lemire algori...

Summary Rust, although it implements a correct float parser, has major performance issues in float parsing. Even for common floats, the performance can be 3-10x slower than external libraries such ...

14:00

14:00

imagine c++

14:00

i wonder how fast stof is

@Learath2 it would be really helpful if it was possible for me to somehow debug the bot in this case for example handle_submission_approve fails when Cøke tries to approve the map of the bot, on my test server that seems to work all the time though, I can't get it to not work there

any nodejs express devs here?

Ryozuki

i wonder how fast stof is

C++ charconv on msvc already uses eisel lemire iirc

14:32

That's C++20 only tho

14:34

@Patiga well you can let me know what kind of access you need to debug it and I can help you. I can also run a debug build. I also have a version of discord.py that goes through mitmproxy if you have an issue on the api side

Learath2

@Patiga well you can let me know what kind of access you need to debug it and I can help you. I can also run a debug build. I also have a version of discord.py that goes through mitmproxy if you have an issue on the api side

yea the problem is that I don't know myself how to debug python nicely. I want to know where exactly the command returns early. On my test server I usually just do printf debugging, since the logs do not report when a function returns early

def-

[ddnet-web:master] 1 new commit

1d088aa Add speedrun.com news entry - def-

poggers

@deen Quentincube aka coradax

14:47

ok

def-

[ddnet-web:master] 1 new commit

076ab11 Quentincube -> coradax - def-

big dox coradax real life name is quentincube

@Learath2 access to logs would already help a lot

GitHub

Click to see attachment 🖼️

time to use "record best run" option

15:12

@Patiga I think murpi gave you related log lines. We can also do printf debugging when I get home

15:14

The way the bot is coded my guess would be a failed api call that doesnt result in a rollback of previous actions

Learath2

@Patiga I think murpi gave you related log lines. We can also do printf debugging when I get home

yeah murpi gave me super useful lines, I kinda assumed that everything fails silently somewhere

Boti.

time to use "record best run" option

nah fuck

@Learath2 did you also symlink the binary 'check_ddnet'? (there are 3 used binaries in total, 'debug_load', 'check_ddnet' and 'edit_map')

I think I did but you can get murpi to check

assuming that the logs of the bot are in logs/bot.log, could you check if the file data/map-testing/check_ddnet exists @murpi ?

Patiga

[ddnet/ddnet-discordbot] Pull request opened: #23 Pr minor changes

didn't intend to have the first commit in this pull request but I think it's fine. the first commit changes $optimize to not touch game layers anymore the second commit puts the positional file paths at the start of the command. I think it was safe before, but this way it is just better the last commit fixes the error in the log

GitHub

Click to see attachment 🖼️

@Learath2 would be nice if you could get those commits in when you have the time

any leet hax0rs upgraded to python2 yet?py2