unix

[narrow fern]

Is there any way to access these functions without using C?

[summer trail]

You mean, other than calling time.process_time()? Let's back up, what are you trying to accomplish?

[narrow fern]

I need this function in golang

https://github.com/python/cpython/blob/2cd1c87d2a23ffd00730b5d1648304593530326c/Modules/timemodule.c#L1344 there it just adds utime and stime

[shy yoke]

Modules/timemodule.c line 1344

_PyTime_t total = utime + stime;```

[narrow fern]

So that would be correct, right?

[summer trail]

Well, it is utime+stime, but it's getting it from getrusage, not from procfs

But sure, I suppose getting it from procfs works too, at the expense of portability

[narrow fern]

Portability is not a problem, only needs to run on Linux

Are the values from getrusage and procfs the same?

I believe there is also a wrapper for getrusage in Go

[summer trail]

They should be, though getting them from procfs will be (slightly) slower

[narrow fern]

Should be fine

[craggy kite]

Does anyone use Python CLI tools which are annoyingly slow? E.g. I find the AWS CLI annoying, it never responds in less than about half a second.

[prime magnet]

yes, poetry

[main olive]

unrelated but I just did speed testing on curl vs wget vs certutil vs other system commands and never realized how slow curl is relatively

[craggy kite]

Do you mean Poetry is slow in things like resolving dependencies, or just slow to respond in general?

[radiant kraken]

for my raspberry pi it takes a couple minutes for poetry to resolve dependencies, which is my main gripe with it

but looking at other commands they all seem to have about a 1s delay minimum

the main bottleneck for a recent CLI i made was the imports so i had to work around it by moving imports inside functions and not importing certain submodules in __init__.py

[prime magnet]

well, modifying list of installed packages is basically most of its job. I can understand it relies on connection for that. Other like "env list" is good. Still annoyingly slow when updating package list 😅

[keen barn]

one way to quickly profile how long imports take: python3 -v 2>&1 | ts -i "%H:%M:%.S"| grep '.*import'

[misty meteor]

Anyone knows if there is a way to usee the 'ping' command through python?

[knotty sentinel]

@misty meteor https://pypi.org/project/ping3/

[misty meteor]

Thank u so much!

[main olive]

woa neet

[craggy kite]

Why not subprocess.run(["ping", ...])?

[warped nimbus]

Curious why you ask. Are you investigating something or just trying to find some comfort in knowing others deal with the same problem?

[spark mulch]

Because of jumpthegun, I assume :)

https://github.com/taleinat/jumpthegun/

[craggy kite]

Indeed, for my latest project JumpTheGun, I'm trying to gauge interest and find areas where it can be most useful.

[craggy kite]

Ah, that seems like exactly what JumpTheGun could solve for you! Next time you're working on a CLI and have such issues, I'd be happy if you considered it.
github.com/taleinat/jumpthegun

[wise forge]

in my opinion AWS CLI will not be improved by any kind of CLI interface
https://gist.github.com/jboner/2841832
Because network latency is unavoidable

network communicating CLI is a network communicating CLI

it is not really for python to solve it

plus, there is already perfect tool for CLI tools, named Golang.
Any tool is made as fast as possible, lightweight as possible, with minimum dependencies, and maximum possible multihtreading if necessary and at the same time as maintainable as possible.
kind of no point for smth else in terms of CLI, if u already learned golang in my opinion

i choose for CLI not golang, only if projects are written in another language and i am forced to choose another language

[craggy kite]

@wise forge

in my opinion AWS CLI will not be improved by any kind of CLI interface [...] because network latency is unavoidable
Yes, JumpTheGun isn't going to resolve network latency, but it can avoid long startup times, which are often an issue even without network latency.
plus, there is already perfect tool for CLI tools, named Golang.
I have two main responses to this:

1. The effort to rewrite many such tools in Go (or Rust etc.) would be immense. In some cases it is worth it, see for example Ruff, but those are the exception to the rule. JumpTheGun makes existing tools faster with zero development or maintenance effort.
2. The first time I ran into such an issue was actually with a CLI tool written in Go, which did not suffer from network latency: go-license-detector. It simply does a lot of setup every time it is run. I ended up forking it and creating a version which worked as a simple HTTP server; with something like JumpTheGun I could have trivially done something similar with zero development.
   https://github.com/go-enry/go-license-detector

[wise forge]

have u measured time for startup of AWS CLI in comparison to network delays? have you run profiling?

comparing network delays with startup times is supposed to be many times different values

i would expect, network delays being way higher to notice any startup delays

[craggy kite]

I have measured the startup of the AWS CLI with an without JumpTheGun, and on my machine it went from ~450ms to under 200ms, nearly 3x faster.

[wise forge]

eh, and the tool works without rewriting code base. yeah, sure. great thing

although to be honest, for my human eye, 250ms is not very noticable

[craggy kite]

That depends on the specifics of your networking, where you are in the world relative to the AWS region, etc. However, it's not unreasonable for a simple query to require a single back-and-forth, and that can take well under 100ms. Compared to that, an extra 450ms of startup time is massive.

[wise forge]

i am more horrified by Java CLI tools delays... their delays are measured in seconds...

[craggy kite]

Maybe Java should be my next target language then 😎

[wise forge]

https://media.tenor.com/jNURRppBLewAAAAC/dew-it-galactic-republic.gif

[wise forge]

if u will make Gradle working faster, it will be amazing

Java ecosystem has no choices. Maven and Gradle

slow as f stuff for any, any command launch

launching unit tests, is many many seconds adventure for hello world program

[craggy kite]

Ah, yes, launching tests is a great use case! I'm aiming to do something about that for pytest to making running tests as near-instant as possible.

(It already shaves off ~100ms of every run of pytest, but there's likely more that can be achieved.)

[wise forge]

have u tried running it with integration to vscode? checking visual debug is working during test launch?

[craggy kite]

Not yet, but I don't use VSCode. Still, whether it could be used with a debugger attached is an excellent question. At the moment I don't think that will work, unfortunately.

[wise forge]

naa@naa-MS-7C89:~/repos/pet_projects/darklab_avorion_api$ grad test

BUILD SUCCESSFUL in 2s
19 actionable tasks: 9 executed, 10 up-to-date
naa@naa-MS-7C89:~/repos/pet_projects/darklab_avorion_api$ grad test

BUILD SUCCESSFUL in 672ms
19 actionable tasks: 19 up-to-date

launching tests for hello world

first launch 2s, second launch 672s

for hello world, Carl!

[main olive]

How to make a command? Have different permissions everytime?

Like when running a command how to specify for only that time it is ran it has only read permissions?

[ember quiver]

There's probably a better way to solve whatever you think you're trying to solve here... You would probably need to create a read only user and su with that, unless the specific command has a read only option

[humble falcon]

To answer your question directly, you would use chmod to change permissions of a file

[heavy scroll]

So i was just messing around slightly and this issue kept popping up

error: externally-managed-environment

× This environment is externally managed
╰─> To install Python packages system-wide, try apt install
    python3-xyz, where xyz is the package you are trying to
    install.
    
    If you wish to install a non-Debian-packaged Python package,
    create a virtual environment using python3 -m venv path/to/venv.
    Then use path/to/venv/bin/python and path/to/venv/bin/pip. Make
    sure you have python3-full installed.
    
    If you wish to install a non-Debian packaged Python application,
    it may be easiest to use pipx install xyz, which will manage a
    virtual environment for you. Make sure you have pipx installed.
    
    See /usr/share/doc/python3.11/README.venv for more information.

note: If you believe this is a mistake, please contact your Python installation or OS distribution provider. You can override this, at the risk of breaking your Python installation or OS, by passing --break-system-packages.
hint: See PEP 668 for the detailed specification.

Which made me use a venv instead

That was fine till I had to use a jupyter notebook for a project, for some reason the venv never really worked for vs-codes jupyter notebook extension

----> 1 import pandas as pd

ModuleNotFoundError: No module named 'pandas'

I kept getting this so I just ended up

sudo apt remove python3

Than reinstalled it, hoping that would fix it but same issue

Now kinda out of ideas in fixing it

The same issue also pops up in jupyter-notebook

[main olive]

Ik what chmod is but like give certain permissions to a command when executing like different permissions per execution

[humble falcon]

Why would you want that?

[main olive]

I do not want it was just curious if there was a way

[heavy scroll]

sudo apt install python3-packageName

This solution works for now but is there a better way to fix the train-wreck that's known as the external environment

[ember quiver]

You should definitely be using a virtual environment of some sort instead of messing up your system. See if this helps? https://stackoverflow.com/questions/58119823/jupyter-notebooks-in-visual-studio-code-does-not-use-the-active-virtual-environm

[heavy scroll]

This helps a lot thanks

[tulip dock]

Is there a way to pass a specific file descriptor to a process using subprocess? I see the pass_fds argument but you can't remap file descriptor numbers with it, only pass them through

I have a script that expect something as fd 7, this is done currently from another shell script (using 7>), I want to rewrite that script from bash to Python

I can call dup2 but there might already be something as fd 7, really I want to to dup2 after fork and before exec, but preexec_fn seems deprecated, I'm not sure how to do this

I just want to translate cmd 7> output.log to Python Popen(['cmd'], ???)

[crystal granite]

Hello guys and girls.
any ideas how to fix this (whatever the reason) ?

im not familiar with python (just average 3d designer) but i ran in such issue when tried to launch UI for image generator.
in other words i would appreciate any help with installing missing dependencies - tried allready on my own but i failed

PS. Running on latest fedora linux

[ember quiver]

If you're not a developer, this shouldn't be your problem. Report it to whoever maintains the install script.

[odd sage]

What is Unix?

[fallow tusk]

It's an operating system. Or rather, a whole swathe of operating systems that conform to a similar set of ideas and ideals. Linux is a Unix operating system (or at least Unix-like), BSD is another, and there are many realtime OSes which are Unixen or Unix-like

[rugged sentinel]

https://en.wikipedia.org/wiki/Unix

[main olive]

i use endeavourOS btw

[vague ginkgo]

Got any shell experts here? Is there a more concise way, that doesn't use the temp variable to do this sort of variable expansion?

export foo1=2
export num=1

echo $(tmp=foo$num; echo ${!tmp})

It does what I want but I'm trying to use this in qsub command being submitted through drmaa python interface

[summer trail]

what shell are you using? POSIX sh doesn't have ${!tmp}, so you must be using bash or ksh or something?

[vague ginkgo]

Bash my good sir

[fickle granite]

can you explain in simple English what it does? 😕

[also, if at all possible, do that computation in Python, not shell]

[fallow tusk]

You can have arrays in bash...

https://stackoverflow.com/a/11856262 has an example using a variable as an array index, too.

[vague ginkgo]

Cannot export arrays or associated arrays and this needs to be evaluated by the shell before it's passed to qsub to be ran on a SGE cluster

[fallow tusk]

The bane of all programmers - incomplete or unclear or just downright misleading requirements.

[summer trail]

can you elaborate? I don't think I understand, still. What's using the environment variables? The shell itself? What's setting them?

maybe you're looking for ```bash
export foo1=2
export num=1

declare -n the_foo="foo$num"
echo "${the_foo}"

[modern fog]

hello, im using ubuntu wsl and i am tryna uninstall python3.10.6 and install python 3.9 instead, how do i do this?

[summer trail]

https://tooling.bennuttall.com/deadsnakes/

[modern fog]

i get this error: add-apt-repository: command not found

[fickle granite]

huh, I have that binary

❯ whence add-apt-repository
/usr/bin/add-apt-repository
❯ apt-file search add-apt-repository
software-properties-common: /usr/bin/add-apt-repository

this is Ubuntu 22.04 "Jammy"

[modern fog]

i dont have apt-file either

lmao

[fickle granite]

well, "apt-file" itself is not installed by default. But "software-properties-common" sure sounds like it ought to be installed by default

[modern fog]

i installed it

i think it should work now

[fickle granite]

huh, good

[vague ginkgo]

So everything I define, needs to be exportable because qsub will export the current shell environmental variables when submitting the executable (in this case a python script) to the Sun Grid Engine cluster.

So associative arrays and arrays are out of the picture

I'm gonna be making a full blown stack overflow post with more info because I cannot access discord at work

[summer trail]

well, does the declare -n option I suggested work?

[modern fog]

ok now im getting module distutils.cmd not found when i try to pip install

i have a feeling my ubuntu corrupted cause now apt-get install isnt working

its giving some lock error

[summer trail]

what error?

[summer trail]

how did you run pip? Do you have the appropriate pythonX.Y-pip package installed?

[modern fog]

Pythonx.y -m pip install

[summer trail]

ok, and what error are you getting here?

did you install the pythonX.Y-pip package via apt-get?

[modern fog]

I tried using the cmd on the website for deadsnake

The curl one

[modern fog]

I fixed it with a restart

[summer trail]

ok, so try doing sudo apt install pythonX.Y-pip and see if that helps anything

[vague ginkgo]

I've seen that as a solution but I think I have a bigger issue and this is a red harring. https://unix.stackexchange.com/questions/747692/shell-variable-expansion-in-qsub-command-through-drmaa

[vague ginkgo]

Is the above, better defined?

[summer trail]

you switched from bash to zsh?

[vague ginkgo]

Yes, but the shell scripts are functionality equivalent

My shell at work is zsh, home is bash

[summer trail]

what do you see if you do: py task_id_env_vars = {"foo": "bar"} and ```py
template.outputPath = f":{INPUT_BASE_DIR}/output/$foo$TASK_ID.o"
template.errorPath = f":{INPUT_BASE_DIR}/error/$foo$TASK_ID.e"

I have a guess that this might result in files being created named $foo1.o and $foo1.e - and if so, that would strongly suggest to me that the $SGE_TASK_ID isn't being expanded by the shell at all, but by something else

I've never used (or even heard of) SGE before, but my educated guess is that it's expanding the $SGE_TASK_ID in the output filename itself, and then using the rest as a literal string, without the shell having any ability to further expand it

https://docs.oracle.com/cd/E19957-01/820-0699/i999036/index.html says:

To build custom but unique redirection file paths, use dummy environment variables together with the qsub -e and -o options. A list of these variables follows.

HOME – Home directory on execution machine
USER – User ID of job owner
JOB_ID – Current job ID
JOB_NAME – Current job name; see the -N option
HOSTNAME – Name of the execution host
TASK_ID – Array job task index number

The use of the word "dummy" there strongly suggests to me that you can't use arbitrary variables, and instead only these fake variables can be used - and if they're "dummy", the shell wouldn't be able to expand them, so SGE must do so itself

[vague ginkgo]

Ah

[vague ginkgo]

This is horribly disappointing

[summer trail]

well, it's just an educated guess at this point - but it's the only way I can think of that you could have wound up with $(tmp=SUFFIX_TASK_ID1; echo ${(P)tmp}).e as a file name - I can't think of any way that the shell could have expanded one of the $ and not the other 2.

[vague ginkgo]

I'm not at work so I cannot answer your question.

Looking over those docs I see these too things.

SGE_O_SHELL – The content of the SHELL environment variable in the context of the job submission command.

SHELL – The user's login shell as taken from the passwd file.

Note –
SHELL is not necessarily the shell that is used for the job.

I don't recall what SHELL was but I recorded SGE_O_SHELL in the SE post. Maybe it's executing it as a different shell...I can only hope

I'm not specifying the -S option

[dapper gust]

I am looking to automate the creation of ubunutu machine images using Packer, however while executing commands like "sudo apt-get upgrade" I'm both asked to enter Y/n and to restart services. Are these things going to interfere with the commands sent in Packer? If so, how could I set the default behavior/bypass them?

[ember quiver]

You've tried and encountered this problem or are you just speculating that this could be an issue?

I haven't used Packer but if it can't handle such a basic thing I can't imagine what the point of it would be.

[wise forge]

https://stackoverflow.com/questions/33370297/apt-get-update-non-interactive
U need to Google
"Debian not interactive apt upgrade"

[wise forge]

Not a packer fault, it was made to be CLI

[dapper gust]

Just something I noticed while trying a dry run manually executing everything myself. I also tried the export command and it didn't seem to actually do anything

[wise forge]

https://askubuntu.com/questions/1364742/how-to-do-apt-upgrade-with-noninteractive
More recent version applicable

[dapper gust]

I'll give that a shot, thanks!

[crystal venture]

That's what my bash prompt gets set to

[crystal venture]

Nice!! That's real simple, I like it

[main olive]

ty

[crystal venture]

I do mine with two colours because it looks cyute

[fickle granite]

I use zsh fanciness

https://asciinema.org/a/CrN6mcBq9IB7ABYFymKTyewvX

[crystal venture]

Fancy shmancy

[fickle granite]

Fency Shmency

[fallow tusk]

So....is fencepost your alter ego? The thing that causes offby1?

[fickle granite]

It's my "someone already got the name offby1" ego

probably this guy https://twitter.com/offby1

bastard

[naive oasis]

How would I go about deactivating my venv from a new shell?
I am running my code on an Ubuntu VPS and when I deploy to my github I have the server automatically update.
This is the command I am currently using to start my code in a venv and also not require the terminal to stay open at all times.
source discord_bot/bin/activate && nohup python bot/main.py & echo $! > bot.pid

This is how I am stopping the bot before I update it
kill -2 bot.pid This will leave the venv still running though.

[fickle granite]

it will?

[naive oasis]

Nevermind. I am like 85% sure this did not work before now it almost does. It is storing the pid of the venv and not the python code

is it okay to kill the venv? I know kill -2 allows for a clean way to kill python with KeyboardInterupt

[fickle granite]

do you mean "is it okay to kill the shell"?

I think that's what you're currently doing anyway

[naive oasis]

Still storing the wrong pid though.

414010 pts/0    00:00:00 bash
 420290 pts/0    00:00:00 bash
 420291 pts/0    00:00:00 python
 420304 pts/0    00:00:00 ps

bot.pid: 420290

When I kill the python the second bash dies with it.

[wise forge]

don't use nohup for deployment, be a man and use systemd
https://linuxhandbook.com/create-systemd-services/
put config /etc/systemd/system/service_name.service with configurations

and run it as sudo systemctl start SERVICE-NAME.service (optionally enable its autolaunch)
on stop it will stop

that is vanilla linux solution

if u want to be especially cool and modern, use docker to run your services
docker run --name my_service -d image_name, that will be enough to launch it in background
docker stop my_service to stop, and docker rm my_service to erase

[naive oasis]

Why would I want Docker ontop of my VPS?

[wise forge]

source discord_bot/bin/activate && nohup python bot/main.py & echo $! > bot.pid
also u can run it as
/path_to_python/bin/python bot/main.py
that will make not required venv activations despite using venv

[wise forge]

capturing your application system level dependencies into it

+having frozen its python dependencies in addition (kind of alternative to lack of compiling)

ability to build and test once, and deploying already confident being valid application

[prime magnet]

easier env spin up for local dev and reduce "but works on my machine" cases

[naive oasis]

systemctl --user enable surveywolfbot.service
Failed to connect to bus: No such file or directory

[wise forge]

start with systemctl start surveywolfbot

.service is automatically ommited from name

[naive oasis]

still same error
systemctl --user start surveywolfbot

[wise forge]

what yields ls /etc/systemd/system/surveywolfbot.service

also could u remove --user from your command

[naive oasis]

I put it in ~/.config/systemd/user/ as per the tutorial you sent

[naive oasis]

That requires root access

[wise forge]

do it with root access then

when u will grasp how to use it with root

feel free to modify to work without

[naive oasis]

Failed to start surveywolfbot.service: Unit surveywolfbot.service not found.

[wise forge]

ls /etc/systemd/system/surveywolfbot.service ?

[naive oasis]

Not found

[wise forge]

make file with configs available at /etc/systemd/system/surveywolfbot.service

[naive oasis]

now it does not give any errors but it does not appear to be running

[wise forge]

systemctl status surveywolfbot for short logs

https://www.digitalocean.com/community/tutorials/how-to-use-journalctl-to-view-and-manipulate-systemd-logs
journalctl -u surveywolfbot.service for full logs
see variations of command in a guide

[digital cairn]

Xcode updates for Apple M1 broke my Django projects.  (mach-o file, but is an incompatible architecture (have 'x86_64', need 'arm64')) I have arm64 installed. “ arch command returns ‘arm64’  on the terminal ”. How can I re-build my projects using ‘arm64’?

[naive oasis]

Ok, it is now running via root. How do I go about running it without root? Do i need to give permissions to a specific folder/file/command to my user?

[naive oasis]

Also it does not seem to be running in my venv with this command
/home/statwolfbot/discord_bot/bin/python3.10 /home/statwolfbot/bot/main.py
It is not properly accessing the environment variables.

[wise forge]

define env vars in systemd config too
https://www.flatcar.org/docs/latest/setup/systemd/environment-variables/

[Service]
# Client Env Vars
Environment=ETCD_CA_FILE=/path/to/CA.pem
Environment=ETCD_CERT_FILE=/path/to/server.crt
Environment=ETCD_KEY_FILE=/path/to/server.key

[naive oasis]

Is there somthing similar I have to do to set up the DNS server? I cannot connect to outside websites
aiohttp.client_exceptions.ClientConnectorError: Cannot connect to host discord.com:443 ssl:default [Temporary failure in name resolution]

[rotund girder]

Wow, do we have this channel? 😮 hello!

[woeful flint]

where do i download the apt-pkg for python 3.9

[rotund girder]

What does apt show python3 show you?

[woeful flint]

shows version 3.10

[rotund girder]

What about apt show python3.9?

[woeful flint]

where do i download the apt-pkg for python 3.9

[rotund girder]

Wow, do we have this channel? 😮 hello!

great, apt install python3.9, you might need sudo

[woeful flint]

i did but still get ModuleNotFoundError: No module named 'apt_pkg'

[rotund girder]

oh, I misunderstood your question

you want to interact with apt in your python code?

[woeful flint]

i wanted to download heroku and got this error

i wanted to download heroku and got this error

tried installing apt-pkg specifically but got this

[wise forge]

answer depends on which cloud provider u a using

[naive oasis]

As in hosting provider?

The thing is ping -c 4 discord.com works just fine

[Unit]
Description=Survey Wolf Bot
After=network-online.target

[Service]
ExecStart=/home/statwolfbot/discord_bot/bin/python3.10 /home/statwolfbot/bot/main.py
Restart=no
Environment=bot_token=
Environment=db_name=
Environment=db_user=
Environment=db_host=
Environment=db_password=

[Install]

This is my updated file (with passwords removed)

[wise forge]

what yields curl https://discord.com?

[naive oasis]

curl not found :/

[rotund girder]

I dont know Heroku, did you follow this guide? https://devcenter.heroku.com/articles/getting-started-with-python

I dont know Heroku, did you follow this guide? https://devcenter.heroku.com/articles/getting-started-with-python

I dont know Heroku, did you follow this guide? https://devcenter.heroku.com/articles/getting-started-with-python

[woeful flint]

yeah

is there any other way to externally download the apt-pkg of python3.9

[naive oasis]

I did not reload the service file properly or something. IT seems to be fully working now.
Now i just need to find out how to make it run without root

[wise forge]

https://www.cyberciti.biz/faq/how-to-install-curl-command-on-a-ubuntu-linux/
sudo apt update && sudo apt install curl

[rotund girder]

@woeful flint What are you running when you get ModuleNotFoundError: No module named 'apt_pkg' ?

[naive oasis]

I will do this in a bit but it does not seem 100% needed now that I reloaded the service file

Now i just need to find out how to make it run without root

[woeful flint]

the command to download heroku

[wise forge]

Now i just need to find out how to make it run without root
good luck ^_^

[wise forge]

actually it is very easy

https://askubuntu.com/questions/676007/how-do-i-make-my-systemd-service-run-via-specific-user-and-start-on-boot

You can specify the directives User= and Group= in the [Service] section of the unit file.

https://www.golinuxcloud.com/run-systemd-service-specific-user-group-linux/
this one better instruction

[rotund girder]

Did you follow the heroku "Getting Started" guide I linked?

[wise forge]

User=deepak
Group=admin

just a matter of adding to [Service] section, few lines

You can specify the directives User= and Group= in the [Service] section of the unit file.

[woeful flint]

yeah i did

[woeful flint]

yeah i did

[rotund girder]

so what command was that, brew install heroku?

[woeful flint]

no its curl https://cli-assets.heroku.com/install-ubuntu.sh | sh since i'm on ubuntu

[rotund girder]

oh right, the guide only showed mac and windows

[rotund girder]

I would investigate which part of the script https://cli-assets.heroku.com/install-ubuntu.sh that comes from. It looks super strange

why would it be running python o.O

oh right, the guide only showed mac and windows

oh right, the guide only showed mac and windows

[finite crater]

I miss when heroku had free tier 😢

[rotund girder]

it's some cloud thingy?

Is it Japanese or just trying to play the part?

[fickle granite]

the latter iirc 🙂

yes, it's a cloud thingy -- someone once described it as "like AWS, but comprehensible"

[light haven]

The term is merger of "Hero" and "Haiku". The Japanese theme is a nod to Matz for creating Ruby.

[naive oasis]

Thank you for helping me everything seems to be working as it should :)

[rotund girder]

Anyone noticed that git diff hunk headers for python are kinda broken? Even when turning on the builtin python diff - that one is just bad in different ways 😢

[wise forge]

show examples

bug reports without evidence are useless

and preferably with version of OS, software and etc and steps to reproduce it

[rotund girder]

It's just a git thing, so OS is uninteresting. Yeah I might do a writeup in the coming weeks. It's a bit annoying to collect examples. Was hoping that someone else had already done an inventory.
Perhaps anyone has gotten tips about how to effectively collect hunk header examples. My idea approach would be:

1. Produce example python code with classes/methods/multiline strings, commit
2. Make changes to the code git diff
   repeat 2 about 20 times

[wise forge]

https://static.wikia.nocookie.net/star-wars-memes/images/6/6d/Dew-it-galactic-republic.gif

do it, at the moment we have even no idea how to recognize it / what to look for.

[rotund girder]

Sorry, here is some examples and discussion around it: https://stackoverflow.com/questions/2783086/how-does-git-diff-generate-hunk-descriptions

The problem I have seen is with methods inside classes, when there are also multiline strings in the mix. Perhaps decorators can cause problems too.
But a systematic inventory is probably needed to understand it 🙂

[main olive]

what distro do you recommend

[ember quiver]

If you have to ask, I say it's a toss up between Mint or Ubuntu. But it really doesn't matter and it's a matter of personal opinion, just try anything you like

[main olive]

i am compiling gentoo atm

[fickle granite]

"whichever distro your favorite cloud provider supports"

which in my case means Ubuntu and AWS Linux

[humble falcon]

Depends on if they're asking for desktop or server use. Also some cloud providers have "bring your own iso" to install whatever you want

Personally for servers, I like debian. I'm also exited for debian 12 in a few days, there's some nice upgrades and improvements

[fickle granite]

true; I was thinking for server use

[vagrant fern]

Late to the party.
Debian Stable for desktop and server.

[main olive]

depends on your level, your needs etc etc

i use endeavourOS

which is basically arch linux with gui installer and several other stuff

and im not touching ubuntu since it stinks

[dense kernel]

what is the difference between a path that starts with backslash and one that dosent in Linux?
i.e. /some_directory/file.txt vs some_directory/file.txt

[spark mulch]

That's a forward slash. If it starts with one, it's an absolute path (similar to C:\ on Windows), otherwise it's relative.

The difference is that on unix, all drives are mounted somewhere on /; as if D:\ was instead at C:\drives\D\ or something like that.

[dense kernel]

so you can have multiple paths that have a different root directory on the same machine? how does that work?

oh so / is basically the root node, and you cal always create a new sub-tree under it by adding a new child, is that correct?

[spark mulch]

No, everything is under the root directory /.

[spark mulch]

You can "mount" a device under a path inside this structure. Typically, the main disk is mounted at /, and other ones are often under /mnt/some_other_drive

There's a bit of magic here as well: What's at / is not only the main drive, but the kernel also injects other magical directories into / that don't correspond to data on disk, e.g.:

• /proc, containing information about running processes
• /dev containing "device files" (those are the things you then mount to a different location

[dense kernel]

I'm not talking about mounting different devices, I'm just trying to understand how this works, cause its kind of like you are adding the path to PATH and you can access it from everywhere, but now I understand that you are placing it under the root node of all paths so it becomes an absolute path that you can access from anywhere.

but how do you create it? because when I do Path('/code').mkdir() it gives me an error (PermissionError: [Errno 13] Permission denied: '/code')

[spark mulch]

Since / is so important, you can only do that as root

[dense kernel]

so with sudo ig

ok thanks

[wise forge]

Minor anti pattern in using Path btw

(Path("/") / "code").mkdir()

The point is, that u can insert different root folder to begin path, which can be even windows one. And same code will get into code folder in OS agnostic way

[spark mulch]

But... with (Path("/") / "code").mkdir() you still hardcode that it's a unix path. If you want it to be injectable, you'll have to change this line regardless. So I disagree about this being an anti-pattern.

[wise forge]

😛 but when making it dehardcoded from unix only path, u will need changing only "/" variable to smth else, instead of changing much more code of nested folders and files, since u have already OS agnostic folder moving

less code riples.

[shrewd stratus]

AFAIK, either way pathlib will resolve it appropriately

[spark mulch]

YAGNI, do it when needed

Especially here, where they were explicitly experimenting with a Linux system and how one adds folders to the root.

[wise forge]

YAGNI is not applied to everything. if u capture errors handling like

try:
  # code
except Exception:
  pass

it will be already not a YAGNI to do it better from a start.

[south fossil]

silent supress of Exception is not good

[dense kernel]

it drives me nuts when I see ppl doing:

try:
  # some untested code
except Exception as e:
  print(e)

[main olive]

does someone have written python script to execute a command in the bash before?

if so, can u share me a code? im trying to make it so i can play my music directly from clicking a file on desktop. i partially succeeded, but its still isnt something i wanted

[neat jasper]

You should use subprocess.run. there is an example in the doc

https://docs.python.org/3/library/subprocess.html#module-subprocess

[main olive]

yeah, i was using that, but i didnt achieved what i wanted. also, thx!

[neat jasper]

Can you explain what the problem is?

[main olive]

the script i wrote could only be executed, by writing a command in bash. also,pretty funny fail - i forgot to write a way for me to quit xd but thats okay, i will try using loop

[ember quiver]

Anything you do in bash should be callable with subprocess. Try subprocess.Popen if you need more control

[fallow tusk]

Subprocess on linux tends to prefer the list form of a command than the string form, too.

[muted fossil]

Hi, I'm looking for help with a couple of linux libraries at #1118810078422777946 message
If anyone knows how I could get a pyudev observer to make threads to watch devices with evdev, (or if anyone can suggest a better approach) help would be appreciated.

[vagrant fern]

Morning folks, what's the clean way to run CGI scripts these days?
I have a static website and want to add a tiny bit of dynamic content to it. cgi from the stdlib deprecated and I'm on NGINX so only FastCGI is an option anyways.
Would be happy for good alternatives 🙂

[vagrant fern]

Some things I found:

• Twisted's .rpy functionality is very similar
• wsgiref still includes a CGI handler that could be used with fcgiwrap

[main olive]

i will try that out, ty!

[keen wraith]

fr

[north briar]

Hi all, Please refer me to a better channel if my question is better placed there. I have a program that I want to record the peak memory usage. I want to create a line plot with the peak memory. Any recommended libraries to get the peak memory. Ideally, I would be able to reset the peak memory counter during the execution of the program since I have the input sizes programmed as a for loop.

[main olive]

what DE are u using?

[north briar]

emacs

[main olive]

in xfce i think there is already program for that

[main olive]

DE = desktop environment, u dummy

[north briar]

please behave respectfully, I don't appreciate that

[main olive]

💀💀

since when "u dummy" is smth derogatory? but well, im sorry if u didnt liked that

[main olive]

now, if u could answer that question

[main olive]

oh, im sorry. there is actually a de like that. i thought u were referring to text editor

[main olive]

have u tried tracemalloc?

hmm, wrong library

[river raptor]

Should i Put Swap before or after /

[fickle granite]

after

are you talking about the order of entries in /etc/mtab ?

[south fossil]

On a regular rotating HDD it would be faster in the beginning of the disk. Becasue linear read/write is faster at the beginning. For the SSD doesn't matter.

[daring seal]

Ok, so I have an odd question revolving around X11, matplotlib qt windows and ipython, I hope this is the right place. I need the x11 window class name (to set a floating window rule in bspwm), if I create a plot from a python terminal this is "matplotlib", but if I do it from ipython its a single space " " and I'm not sure how to set a bspc rule otherwise (I tried on the name, but don't think it supports wildcards "Figure*"). Any ideas?
The problem seems to be specifically with the plt.ion() option, but again, only in ipython...
This ipython -i -c "from matplotlib import pyplot as plt; plt.ion(); plt.plot(0,0)" works and so do subsequent interactive calls to plot, while
this ipython -i -c "from matplotlib import pyplot as plt; plt.ion()" with subsequent plots does not.

I guess for now, as a hack, I can start ipython -i -c "from matplotlib import pyplot as plt; plt.ion(); plt.plot(); plt.close(plt.gcf())" instead of ipython and maybe create an alias.

[wispy ingot]

I've been running sudo mount /dev/sda1 /media/t5 to mount an external drive to an Rpi. The SSD is exFAT. Writing to it from the dan user is impossible, I just get permission denied, e.g. mkdir /media/t5/foo yet prepending sudo works.
Not sure what to do to resolve this, as sudo chown -R dan:dan /media/t5 just spams Operation not permitted.

[fallow tusk]

You can specify the user in the mount options -o user=dan,group=dan (or something similar - I always have to check the mount docs to get the right option format)

[fickle granite]

I'd also try changing ownership of the mount point before mounting the file system on it

don't honestly know if that'd work, but it should be easy to try

[fallow tusk]

I don't think it does. I recall it being something I used to do, but I think that's for being able to mount, it doesn't affect the perms of the mounted partition

(One of those - got it working, now, leave well enough alone; now forgotten how I got it working - things!)

[wispy ingot]

aye -o uid=1000,gid=1000 worked, tyvm!

[fallow tusk]

You might also want/need to set exec and/or other options

[tulip vortex]

hahahahahahah

no Mac channels here

lol

does anybody here has access to a MacOS machine ??
I want to test something

[lavish storm]

This channel does include MacOS

Tooling, setup and configuration related to Python development on unix or unix-like systems such as Linux or MacOS.

[tulip vortex]

wow

I thought Unix only refers to linux-ish things

do you guys think this is a good code to determine which OS my app is being executed?

https://gist.github.com/yolowex/e130ac0c451776717b7b5624817443c4

I don't have access to a mac device so I won't know if it's exectuing correctly

[lavish storm]

Traceback (most recent call last):
  File ".../some/dir/test.py", line 1, in <module>
    from common.names import *
ModuleNotFoundError: No module named 'common'

@tulip vortex

[tulip vortex]

oh wait

import os
import sys
import time
import random
import subprocess
import traceback

put these instead of the common.name one

[lavish storm]

Current platform:  Darwin
macOS Version:
Build Number:

[tulip vortex]

dam

[lavish storm]

I think it's because version_info is defined only for windows

[tulip vortex]

do you know the correct way?

[lavish storm]

sure

[tulip vortex]

funny fact:
in my language, Seaman means cement

[lavish storm]

under IS_MAC: I'd do:
version_info = platform.mac_ver()

but it won't show the build number

[tulip vortex]

does it show the os version?

[lavish storm]

yes

❯ python3 test.py
Current platform:  Darwin
macOS Version: 12.6.6
Build Number:

[tulip vortex]

it seems that there are three outputs for the mac version

def mac_ver(release: str = ...,
versioninfo: tuple[str, str, str] = ...,
machine: str = ...) -> tuple[str, tuple[str, str, str], str]

can you print them all?

[lavish storm]

In [1]: import platform

In [2]: v = platform.mac_ver()

In [3]: v
Out[3]: ('12.6.6', ('', '', ''), 'arm64'

here's my output from ipython

[tulip vortex]

MACOS_VERSION
MACOS_BUILD_NUMBER
these are the two I'm using right now

what is the name of the third one

[lavish storm]

architecture

[tulip vortex]

so I should say MACOS_ARCHITECTURE ?

[lavish storm]

sure

[tulip vortex]

are they in correct order?

mac_version = platform.mac_ver()

MACOS_VERSION = mac_version[0]
MACOS_BUILD_NUMBER = mac_version[1]
MACOS_ARCHITECTURE = mac_version[2]
        
print("macOS Version:", MACOS_VERSION)
print("Build Number:", MACOS_BUILD_NUMBER)
print("Architecture:", MACOS_ARCHITECTURE)

[lavish storm]

yes

though, as you can see above, the tuple for build number was empty

might be different for intel-based macs

[tulip vortex]

maybe it's the case only with your machine

could it be a permission issue-ish thing?

try running the code as sudo

[lavish storm]

No difference 😄

[tulip vortex]

dam

Thanks for the help!

[lavish storm]

np!

[fallow tusk]

Does the mac have an /etc/os_release file? Most of that info should be in there, and should be parsed by platform (although the interface in platform has changed over various Python versions)

[spark mulch]

It does not

[fallow tusk]

:/

I guess that is lsb, not usb :D

[spark mulch]

There is the sw_vers command though

[covert bramble]

I've got terminalexia. I've got a terminal in vscode, one in docker app, one in terminal app, screen terminals, python env terminals. Hard to find the right one.

[sonic pendant]

maybe try https://zellij.dev to see if it suits you

[wise forge]

Try terminator. Terminal multiplexor

I use default one Konsole from Kubuntu ^_^

[shrewd stratus]

tmux 👍🏼

[covert bramble]

it's hard to get away from from the integrated terminal in VScode when using Platformio

[river raptor]

How do i remove Partition 2 and move Partition 3 and make It füll size

[fickle granite]

there's an old tool https://www.gnu.org/software/parted/ that I think can do that

[main olive]

im trying to run snekbox on my rpi, ive installed amd64 support or well i tried with this command: sudo docker run --privileged --rm tonistiigi/binfmt --install amd64 and then i installed snekbox with this command as per the docs sudo docker run -d --rm --platform linux/amd64 --ipc=none --privileged -p 8060:8060 ghcr.io/python-discord/snekbox, however, i get an error ```
2023-06-26 12:45:43,649 | 13 | snekbox.nsjail | ERROR | Couldn't launch the child process

2023-06-26 12:45:43,671 | 13 | snekbox.nsjail | INFO | nsjail return code: 255``` which leads me to believe theres some shenanigans going on. I would like to know if i should try to mess around with it more or if its just useless to try stuff out because it wont work i also tried to build it locally and run it that way but still doesnt work
i get this error: https://paste.pythondiscord.com/ovadovetet

[south fossil]

fdisk can only create partitions and delete them (which will destroy data as well). In case you need to preserve data, use software like GParted, EaseUS, or SystemRescueCd which has GParted included. Also make some backups of the most important files in case partition operations goes wrong.

[soft dawn]

guys i need help with my shellscript

someone can help me

dm me

[lavish storm]

Could you provide code samples or error messages

[river raptor]

Can anyone Help with OpenBSD? I somehow overwrote my Linux Partition and only have a USB with OpenBSD, If i Plug in a USB WiFi Thing, it Shows a blue Message saying it isnt configured

Please mention If Response

[river raptor]

Okay nvm

I'll use my Sisters Laptop to Install alpine in my usb

[fallow tusk]

That was going to be my suggestion, too ;) Use another machine (or even the same machine if you have two USB ports) to write an ISO liveboot of your distro to a USB stick

[empty barn]

[root@archi system]# systemctl status http-drop
× http-drop.service - Dropbox-http
     Loaded: loaded (/etc/systemd/system/http-drop.service; enabled; preset: disabled)
     Active: failed (Result: exit-code) since Tue 2023-07-04 22:28:06 CEST; 3min 17s ago
   Duration: 194ms
    Process: 3319 ExecStart=/usr/bin/python3 /home/user/code/filenet.py (code=exited, status=1/FAILURE)
   Main PID: 3319 (code=exited, status=1/FAILURE)
        CPU: 189ms

Jul 04 22:28:06 archi systemd[1]: http-drop.service: Scheduled restart job, restart counter is at 5.
Jul 04 22:28:06 archi systemd[1]: Stopped Dropbox-http.
Jul 04 22:28:06 archi systemd[1]: http-drop.service: Start request repeated too quickly.
Jul 04 22:28:06 archi systemd[1]: http-drop.service: Failed with result 'exit-code'.
Jul 04 22:28:06 archi systemd[1]: Failed to start Dropbox-http.

Trying to run an Tornado web server

this is on arch linux

[slate crypt]

check journalctl -u http-drop

that should show the full log of http-drop

[empty barn]

[root@archi system]# journalctl -u http-drop
Jul 04 22:28:01 archi systemd[1]: Started Dropbox-http.
Jul 04 22:28:03 archi python3[3313]: Traceback (most recent call last):
Jul 04 22:28:03 archi python3[3313]:   File "/home/user/code/filenet.py", line 61, in <module>
Jul 04 22:28:04 archi python3[3313]:     app.listen(100)
Jul 04 22:28:04 archi python3[3313]:   File "/usr/lib/python3.11/site-packages/tornado/web.py", line 2134, in listen
Jul 04 22:28:04 archi python3[3313]:     server.listen(
Jul 04 22:28:04 archi python3[3313]:   File "/usr/lib/python3.11/site-packages/tornado/tcpserver.py", line 183, in listen
Jul 04 22:28:04 archi python3[3313]:     sockets = bind_sockets(
Jul 04 22:28:04 archi python3[3313]:               ^^^^^^^^^^^^^
Jul 04 22:28:04 archi python3[3313]:   File "/usr/lib/python3.11/site-packages/tornado/netutil.py", line 162, in bind_sockets
Jul 04 22:28:04 archi python3[3313]:     sock.bind(sockaddr)
Jul 04 22:28:04 archi python3[3313]: PermissionError: [Errno 13] Permission denied
Jul 04 22:28:03 archi systemd[1]: http-drop.service: Main process exited, code=exited, status=1/FAILURE
Jul 04 22:28:03 archi systemd[1]: http-drop.service: Failed with result 'exit-code'.

eh

thanks for helping ig

ill try to figure out the rest on my own

thanks

it gave me a lecture

💀

Jul 04 22:42:54 archi sudo[3919]: pam_systemd_home(sudo:auth): systemd-homed is not available: Unit dbus-org.freedesktop.home1.service not found.
Jul 04 22:42:54 archi sudo[3919]: pam_unix(sudo:auth): conversation failed
Jul 04 22:42:54 archi sudo[3919]: We trust you have received the usual lecture from the local System
Jul 04 22:42:54 archi sudo[3919]: Administrator. It usually boils down to these three things:
Jul 04 22:42:54 archi sudo[3919]:     #1) Respect the privacy of others.
Jul 04 22:42:54 archi sudo[3919]:     #2) Think before you type.
Jul 04 22:42:54 archi sudo[3919]:     #3) With great power comes great responsibility.
Jul 04 22:42:54 archi sudo[3919]: For security reasons, the password you type will not be visible.
Jul 04 22:42:54 archi sudo[3919]: sudo: a terminal is required to read the password; either use the -S option to read from standard input or configure an askpass helper
Jul 04 22:42:54 archi sudo[3919]: pam_unix(sudo:auth): auth could not identify password for [user]

i dont have recieved lecture from system admin

[slate crypt]

you are the system admin

[empty barn]

yeah

eh

any ideas on

eh

[slate crypt]

Is that binding to port 80 or something?

[empty barn]

100

[slate crypt]

use a higher port

[empty barn]

but i have been using

85 86 87 89 90 100 45

without issue

except 100 one

[slate crypt]

did you change the config to allow users to bind lower ports?

[empty barn]

how

[slate crypt]

users can bind to ports 1024-65535

only root can bind to 1-1023

and technically 49152-65523 is or dynamic port assignment.

8080 is a common port to use

[empty barn]

eh

that would need to

can i do it in systemctl?

[slate crypt]

https://superuser.com/questions/710253/allow-non-root-process-to-bind-to-port-80-and-443

[empty barn]

https://superuser.com/a/1733073

oh

uh

it fix

thank u

[valid stream]

arch btw

[lavish storm]

I see, thank you

[main olive]

I think I'm going to use Redhat as my main OS, and use windows for a little gaming on a vm

I tried ubuntu for a little bit (a few projects) and I've been really liking it

[humble falcon]

Isn't redhat mostly for enterprise and servers?

[main olive]

mint is just like ubuntu but better. if you want somth similar

[main olive]

Maybe? I’ve heard it can be good for desktop

[main olive]

I’ll check it out

🙏

[main olive]

It comes pre installed with blender??

https://linuxmint.com/

[main olive]

nah but you can easily install it with the package manager afaik

apt install blender

[main olive]

Oh alright, it said on their website the apps it contains

[main olive]

noice

[valid ibex]

preventing any think to be done

i am on ssh server, what to do?

[fallow tusk]

Defunct processes should (iirc) eventually time out.

I've not had much luck in getting them cleaned up beyond waiting or rebooting, unfortunately. However, if they're defunct, they won't be using up processor time

If they're owned by you, you may be able to just log out from all seats and get them cleaned up that way (I believe…I could be wrong!)

[valid ibex]

i am unable to use GPUs

it hangs on doing python ... or nvidia-smi etc

[spark mulch]

I know about synchronize-panes, but is there also a way to synchronize tmux sessions? Currently I'm using a weird setup with nested sessions :/

[summer trail]

I wouldn't think that's true... If you don't have a subscription, you won't get driver updates, security patches, etc...

you can't even submit bug reports without a subscription, AFAIK

[main olive]

Yeah, I’ve seen that with a subscription it’s a really nice OS, but I’m thinking Mint looks good

https://linuxmint.com/

[summer trail]

that's definitely way more targeted at personal use than Redhat Enterprise Linux is.

[main olive]

RH is in it for the money

[summer trail]

well, yeah - selling support is literally their business model

[main olive]

But I believe at enterprises, they have a large Linux server, and everyone just SSH’s into it

I have an uncle who’s an IT Director for a hedge fund and he told me this is how they operate

[summer trail]

that's one possibility, yeah. Nowadays it's a bit more common for developers to have their own dedicated VMs they can ssh into (still shared hardware under the hood, but they have the illusion of having a private machine)

[main olive]

Mint looks good, 3 weeks until I build the pc 👍

[reef token]

This is what we do at my company

Although yeah we are looking to migrate to individual VMs at some point

[main olive]

one thing. get an amd card if you can. nvidia's linux drivers can be problematic

[main olive]

In what sense?

https://pcpartpicker.com/list/sZyYpH

[languid vortex]

Im a student and am learning programming.
I wonder would u install an Ubuntu dual boot system for getting used to linux/learning linux commands?

[warped nimbus]

Yeah it's a good skill to have but may not be a priority for you at your current stage. Instead of dual boot, I'd recommend using a VM, or WSL if you're on Windows

[languid vortex]

thank you yeah I think so
especially others told me I can run VSCode with WSL

[wise forge]

Wonderful idea to go dual boot right away.

And switching to Linux as primary OS 😊

That is how I started to be more and more familiar with it every month for last years

[languid vortex]

didnt u feel any inconvenience?

as I cant find many softwares on ubuntu

[wise forge]

The only inconvenience I feel after some time... I fell in love so hard with how developer friendly environment in it, that I can't force myself opening windows back more than for very tiny amount of time

[wise forge]

It has its own software available, that is not available for Windows. Just a matter of getting to know stuff

[languid vortex]

I fell in love so hard with how developer friendly environment in it
I havent known about this yet
maybe very interesting to discover

[wise forge]

In fact, Linux flares best if u a doing stuff related to web development and cloud server infrastructure.
It has programs that are not available to windows or work very very very slowly in limited fashion at Windows
At linux they work natively with lighting speed and full functionality

[languid vortex]

thank you!

do you have plans to uninstall windows someday?

I dont like dual boot personally tbh

[wise forge]

I keep windows at dual boot as secondary OS for gaming only, and for that super legacy project in C++ visual studio that I did not bother to migrate to make Linux friendly

[languid vortex]

I see, thanks

[wise forge]

It is nightmare how badly 😔 this legacy project setup. Not automatable at current state to build and unit test.

All my projects at Linux usually do it all on their own in CI tools, since they all run Linux too

[languid vortex]

it's mainly the interface which is called GRUB2 I think
I had problems with that

[wise forge]

I just installed to different SSD drives OSes

Windows at one SSD without awareness Linux exists (plugged out Linux drive during installation)

Linux at another, with grub and discovered both oses
Linux drive is main to boot

Makes simple to reinstall stuff to me

In case I delete windows, repairing grub is easy to remove no longer present windows

Doing same in windows would be unreliably hard to me ^_^

[languid vortex]

thank you, I see

this seems to be a pretty good way

[ember quiver]

Yeah, dual boot is hell for so many reasons and getting Windows to behave with GRUB2 is one of them. I have a dedicated Linux box and I use XRDP to connect to it from my various Windows devices. You quickly forget you're on Windows with no need to worry about certain hassles of Linux (like the wifi drivers that always seem to break). If you ever do need Windows, just close RDP and no need to reboot.

I also use WSL on my company laptop and it's pretty great too

[languid vortex]

I agree with it
Im too newbie to fix any grub problems

another reason is most laptop I bought only has one SSD I think

[main olive]

sorry i forgot to reply.
i dont use nvidia so idk much but theres lotsa arguments about it online. for example see the comments under this:
https://www.reddit.com/r/linuxhardware/comments/p63ka0/how_bad_is_nvidia_gpu_support_on_linux/

[fallow tusk]

Nvidia drivers sometimes get "blips" in their update cycle. It doesn't seem to have happened much recently, but it has in the past. Usually, though, it's just a matter of waiting a day for someone to fix the repository

[south fossil]

You can, but you will really used to it if you use as a primary OS

[runic goblet]

any tmux user in the house?
is there a way to move only the cursor to the top of the page without moving the page content itself in copy mode? (this is H in vim in case it's not clear)
the closest i can find is ctrl+b in copy mode (i have vi mode on if that matters at all) but that changes the content of the page which is disorienting to me

[spark mulch]

I've never tried it, but in the docs it looks like M-R is what you are looking for.

The corresponding command is called top-line, you might want to bind that to something more reachable. At least for me the whole meta key stuff never works consistently.

[runic goblet]

interesting, looks like it's an issue with my config then.

Command                                      vi              emacs
top-line                                     H               M-R

implies H should have worked for me

thanks!

[spark mulch]

Oh, that's why it didn't work for me :D

[river raptor]

You could do that

When i started with Linux i Just overwrote the Disk with Windows

Entirely

[river raptor]

@ivory moat lets talk here

[ivory moat]

sure

Have you tried ubuntu previously?

I'm trying to look for a comparative experience to decide

[river raptor]

my first distro was ubuntu with kde, kubuntu

then switched to arch and now i use opensuse

[ivory moat]

but does opensuse have better support for proprietary softwares?

like ubuntu

[river raptor]

why would you even want proprietary software

[south fossil]

straight downgrade

[river raptor]

why?

i don't want to risk breaking my installation just by not doing pacman -Syu every day

and i dont really like the AUR, too much malware

[spiral yoke]

hi o/

i've made a package manager in python

https://git.iohio.xyz/FPKGProject/fpkg (i know it looks ugly)

currently, i'm working on 0.1.0 fractal

[ivory moat]

might be needed at times

[kindred tartan]

yo uh

getting this issue

any way to downgrade the driver so it supports chromium v113?

running raspbian on Raspberry pi 4b with box86 emulation for linux64-based packages

which in my instance is chromium webdriver

Doing this in python btw*

[upper notch]

yeah there is

@kindred tartan just download the correctly driver

https://chromedriver.chromium.org/downloads

In your code or application just specify this executable. Alternatively export it to PATH 🙂

[kindred tartan]

I found the fix

yeah i downgraded the driver

[upper notch]

Oh cool

[kindred tartan]

all good

many thanks

[upper notch]

sweet

[kindred tartan]

just had an issue doing it through the manager

instead did it manually

[upper notch]

Ah, makes sense

what distro?

Debian, Arch, Gentoo?

[kindred tartan]

oh uh

Raspbian

apparently it had some form of x86 emulation

but it didn't so i had to install box86

worked like a charm

I needed it for some python scripts i had

@upper notch

[upper notch]

I see

[patent mural]

Why is pip not installing anything... im using arch btw

[shrewd stratus]

could you elaborate?
are you getting any errors?

[patent mural]

Im really sorry for not giving the error message to begin with

$ python3 -m pip install -U discord
error: externally-managed-environment

× This environment is externally managed
╰─> To install Python packages system-wide, try 'pacman -S
    python-xyz', where xyz is the package you are trying to
    install.
    
    If you wish to install a non-Arch-packaged Python package,
    create a virtual environment using 'python -m venv path/to/venv'.
    Then use path/to/venv/bin/python and path/to/venv/bin/pip.
    
    If you wish to install a non-Arch packaged Python application,
    it may be easiest to use 'pipx install xyz', which will manage a
    virtual environment for you. Make sure you have python-pipx
    installed via pacman.

note: If you believe this is a mistake, please contact your Python installation or OS distribution provider. You can override this, at the risk of breaking your Python installation or OS, by passing --break-system-packages.
hint: See PEP 668 for the detailed specification.

[shrewd stratus]

as mentioned in the error output, you should use a virtual environment

[patent mural]

What if i want something for system config

[shrewd stratus]

do you have an example?

[patent mural]

Like for a window manager named Qtile

It need some optional dependencies

I dont remember the specifics

[shrewd stratus]

you would install qtile using pacman

[patent mural]

Is there any fix to this pip issue tho?

Like why is it happening all of a sudden

[shrewd stratus]

^

[shrewd stratus]

recent versions of pip include this check because installing packages directly into the system environment can sometimes break your system

because a lot of Linux distros make use of Python for their core functionality

[patent mural]

Ic

[patent mural]

Thanks for the explanation

[vernal kayak]

anyone here use docker-compose on linux and have experience with setting up a MYSQL DB through it?
Mine keeps wiping the DB and creating a new file structure once I use docker-compose down

mysql:
    container_name: **
    environment:
      - MYSQL_ROOT_PASSWORD=**
      - MYSQL_DATABASE=**
      - MYSQL_USER=**
      - MYSQL_PASSWORD=**
    command: 
      --character-set-server=utf8
    volumes:
      - my_volume:/var/lib/mysql  # path in the container where MySQL stores its data
    ports: 
      - 3306:3306
    image: "docker.io/mysql:latest"
    restart: always
    networks:
      - mynetwork
volumes:
  my_volume:
networks:
  mynetwork:

Example structure of my file

[wise forge]

try

    volumes:
-      - my_volume:/var/lib/mysql  # path in the container where MySQL stores its data
+      - ./my_volume:/var/lib/mysql  # path in the container where MySQL stores its data
...
-volumes:
-  my_volume:

[vernal kayak]

will give it a try, thanks

[shrewd stratus]

just to check, are you using docker compose down --volumes?

[vernal kayak]

just docker compose down is that bad?

[shrewd stratus]

no, just checking, because with --volumes the volumes get deleted as well.

[vernal kayak]

I'm just a bit worried about the stability of things, i'm importing a lot of backdated API data etc now and don't want my data team to randomly message me saying "where did all the data go" one day 😅

I've been thinking about using AWS or Google Cloud Platform to host a DB and to just update it remotely instead

definitely open to any recommendations

[wise forge]

that's very good and most recommended option btw (for real prod)

using stuff like AWS RDS MySQL (or usually better postgresql) for example

it allows to go fully immutable regrading deployments

deploying straight containers into AWS elastic beanstalk / or AWS ECS or AWS EKS and etc

or just wiping servers without care and deploying from zero each time (some people deploy as immutable built AMI images)

allows managed RDS db, easy to update in terms of engine and any params

easy to backup

easy to replicate

everything easy

[vagrant fern]

yeah, except your bill.

or, you know, the fact that cloud providers (azure at least) actively recommend you add retries to every single one of your queries, because "due to the cloud, sometimes requests fail". super easy

just add a retry to each of your queries.

[vernal kayak]

I work for a big e-commerce agency so the bill is fine

As long as I can justify it

I don’t like the sound of retries though

Amazon is a pain in the ass when it comes to large amounts of data as it is

I suppose I could store it in my DB on the server I’m renting, then retry to there

Then have a daily push to AWS

[vagrant fern]

the retry pain bit me badly in my previous place. I ended up quitting

The problem is that devs don't expect their world of databases and caches to suddenly vaporize themselves

[vernal kayak]

I understand it all too well

I woke up one day right before a meeting with my commercial director and data team to find my database just vanished

That’s what I want to avoid moving forward

I want to be able to change everything, constantly

But for the DB to remain there

Like a roman toilet

[fickle granite]

🤔

[fickle granite]

Every program that does cloud stuff should do retries.

unless you don't want it to actually be reliable.

[wise forge]

🤣

[vernal kayak]

Honestly I haven’t been doing this, and I need to

Even when it comes to calls with one variable, I just trust that it does every one of them correctly

Need to find a better way to checking that everything is done and repeating x amount of times until everything is done

[fickle granite]

https://pypi.org/project/retrying/

[vernal kayak]

anyone know why I would be getting the following when using docker, i'm trying to not use root and create a user with specific perms:

flask-celery-worker-1       | PermissionError: [Errno 13] Permission denied: '/usr/src/app/project/logs/celery.log'

the command in docker-compose file:

celery --app project.server.tasks.celery worker --uid=celeryworker --loglevel=info --logfile=project/logs/celery.log

my DockerFile:

WORKDIR /usr/src/app
RUN adduser --disabled-password --gecos '' celeryworker
RUN mkdir -p /project/logs/
RUN chown celeryworker:celeryworker /project/logs/

[south fossil]

/project/logs/ and project/logs/ are different directories

[vernal kayak]

damn.. I need more sleep or coffee 😅

thank you

0 errors and im no longer using root ⭐

[formal schooner]

i greatly dislike that docker volumes don't allow you to mount relative to WORKDIR

it's one of the only places where you must use absolue paths

[wise forge]

-v $(pwd)/relative_folder:/inside_folder

[formal schooner]

i mean on the right hand side, you can't specify that relative to the working directory inside the container

the left hand side needing to be an absolute path is a separate annoyance

[wise forge]

If I am unsure regarding inner workdir, I just overwrite workdir too
-v $(pwd)/relative_folder:/inside_folder -w /inside_folder

[formal schooner]

but that's even worse!

[fallen jasper]

How do I get the gid (group id) of a group given it's name? For example, on a university computer I have three groups 'student', 'faculty', 'staff' and want the gid for each of them. One option is to import grp and use grp.getgrname('staff').gr_gid, but the getgrname call also tries to populate the gr_mem field of the struct and if i use it with 'student', it tries to pull the name of all students which takes quite a long time

[fickle granite]

oops! didn't read closely enough

damn, looks like there's no portable way. I'd just parse the damned file by hand 😐

def get_gid(name: str) -> str:
    with open("/etc/group") as inf:
        for line in inf:
            fields = line.split(":", maxsplit=3)
            if fields[0] != name:
                continue
            if len(fields) >= 3:
                return fields[2]


get_gid("certusers")
``` wfm

if your group database is remote, like on a Windows server or something, this won't work 😕

[fallen jasper]

hmm. thanks, then! looks like i'll have to stick with the slow solution

[minor cypress]

I want install python on Windos and Linux with Commands

[fickle granite]

I don't know what "with Commands" means, but on Linux: use your system's package manager (dnf, apt, whatever); on Windows, download and run https://www.python.org/downloads/release/python-3114/

[prime magnet]

I think MSI packages can be automated on Windows

[minor cypress]

Because I want to write malware and I don't know how to install python on my victim system for my malware!!!

[fickle granite]

...

[spark mulch]

!rule 5 You won't receive help for that here.

[shy yoke]

Rules

5. Do not provide or request help on projects that may violate terms of service, or that may be deemed inappropriate, malicious, or illegal.

[wise forge]

https://youtu.be/69Hitt9Dw4Y

[minor cypress]

this just for fun

[opaque ginkgo]

and how can we prove that that's the case

[south fossil]

script kiddie

[tiny terrace]

Ever heard of a compiler

U sound middle school ngl

[spiral yoke]

as i mentioned a week to two ago, fpkg is my project and it's developed into quite something

https://git.iohio.xyz/Felis/fpkg

(it does somewhat break pep8, with the line limits and stuff)

but the code in of itself is pretty much functional

[river raptor]

!rule 6

[shy yoke]

Rules

6. Do not post unapproved advertising.

[normal dagger]

what files does the scanner generate exactly?(wayland)

[spiral yoke]

Oh, sorry

[river raptor]

What scannee

[normal dagger]

arry poeh

[languid vortex]

What (graphical) text editors do you use on linux? I like notepad++ on Windows.
Im currently using notepadqq on Ubuntu

[shrewd stratus]

VS Code

[river raptor]

https://pulsar-edit.dev

[languid vortex]

thank you
oh I think I mean a pure text editor
I use VS Code for codes and syntax highlight

[wise forge]

Kate. Really good one.

Simple, yet with syntax highlight for all langs

Default in KDE plasma GUI (like Kubuntu)

Not an ide at all

[languid vortex]

thank you!

[flint rover]

This is a dumb question, but does linux/*nix actually stop you from removing a file which is being used by another program

[wise forge]

why should it? no.
everything is solved by user permission levels

database for example often created by user postgres

so only postgres user or root can change it

[flint rover]

fuckkkkkkkkkkkkkkkkkkk

Ngl was really hoping it didn't do that.

why should it? no.
Something something, no one likes a sigbus

[spark mulch]

It won't cause an error. If process A opens a file, and afterwards process B removes the file, process A can continue using the file handle, reading from and writing to the file. The inode still exists, just no link to it. Only when the file is closed by A is it really lost.

(And before it is closed you can also undelete the file by finding it in /proc/pid_of_process_a/fd)

(On macOS there's no /proc, but I believe the same can be achieved through other means)

In a sense, unix files work a lot like Python's names. rm a_file does something similar to del a_variable — the GC will only delete the variable if there's no references left, and same on the file system (with hard links and file descriptors both being counted as references)

[frigid barn]

shouldnt it be preinstalled sometimes? (sorry for screenshot and breaking conversation)

[spark mulch]

What OS are you on? If it's something debian-y, then you'd need to install python-tk

[frigid barn]

solydx. yep, debian-based

(didnt even changed home menu logo xD)

[river raptor]

why do you prefer it over debian?

[frigid barn]

network problems 🙂

[lusty dust]

#media-processing message

[silk herald]

maybe winget for windows?

[silk herald]

hi opensuse

i might try leap as sles might be good for my software

[river raptor]

Nice

[main olive]

NICE

[vapid notch]

hi

ok to ask about access rights here of linux or must i create a help channel. its more about development of a linux-like FS rather than the actual OS

[spark mulch]

Sure, you can ask here.

[glad fiber]

program without icon: https://paste.pythondiscord.com/WOFQ

python, tkinter

I am trying to make a simple App with tkinter and I am having a problem with looping, I have two buttons in my simple app.. and I want it to work when another loop is working in a background...

when I press "start" Button, It runs a function "main_prog" with loop, so until the the loop is not complete the window says "not responding" and I can see it printing in the terminal so it's working in background... and this is a problem because I want it to respond and when I press "Stop"
I want it to take that response and stop...

for this to happen I also tried using "threading" module but it felt really slow and unresponsive most of the time it was not responding. also it was starting the function before I wanted it to... so I think threading is not a viable option..

Do you have any idea that I might be able to fix it?

thank you

[fallow tusk]

This feels like you'll get a much better response as a help thread (see #❓｜how-to-get-help ) #unix tends to be a lot slower, and your question is not unix-related, so people are going to be less inclined to respond.

[normal dagger]

how can I do mouse interactions on Wayland similar to Xlib?

[sterile eagle]

realized I probably needed to take this question here.
dumb question because I can't find a good answer. Is it possible to have anaconda installed just in a venv? I've checked the PKG-INFO and it states that I can't install with just pip, but not sure how to keep it just in the venv, since I'd have to use yay to install it from the aur?
trying to get jupyter notebooks to work with vs code having issues with kernel because I need anaconda (to my understanding anyway)

[west musk]

one of my drives just stopped mounting in linux

$MFTMirr does not match $MFT (record 3).
Failed to mount '/dev/nvme1n1p2': Input/output error
NTFS is inconsistent. Run chkdsk /f on Windows then reboot it TWICE!
The usage of the /f parameter is very IMPORTANT! No modification was
made to NTFS by this software.

Failed to open '/dev/nvme1n1p2'.

$MFTMirr does not match $MFT (record 3).
Failed to mount '/dev/nvme1n1p2': Input/output error
NTFS is inconsistent. Run chkdsk /f on Windows then reboot it TWICE!
The usage of the /f parameter is very IMPORTANT! No modification was
made to NTFS by this software.

Unable to read the contents of this file system!
Because of this some operations may be unavailable.
The cause might be a missing software package.
The following list of software packages is required for ntfs file system support:  ntfs-3g / ntfsprogs.

or rather, one partition, the partition after that mounts just fine, has some small error about the name but i guess thats related to this

do i need to install windows to fix this or what?

really hope its not dying or anything, quite new ssd, but i think its mounted under my gpu...

ah ntfsfix seemed to have solved it

time to migrate to a different filesystem i guess

[fickle granite]

I certainly wouldn't use NTFS unless I was sharing the partition with Windows

the Linux driver code is almost certainly incomplete and buggy

[glacial valley]

https://towardsdatascience.com/virtual-environments-in-anaconda-jupyter-notebook-f92cda7184fa

[river raptor]

how to check which window manager i am running (both wayland and X)? should i check for all processes to check for the window manager?

in python

[light haven]

check for the WAYLAND_DISPLAY environment variable

if it exists you are on Wayland, if not fall back to checking if it's X or a tty

[river raptor]

I wanna Know the WM im using

Not either wayland or X or tty

I wanna Know which WM i am using on both wayland and X

[light haven]

right

I think you can check who's using the display with lsof but if it requires root I think you just need to basically list all processes and have a hardcoded list of the options

[river raptor]

Alright

And then Check to which Display it belongs to to Check for the current display

[cunning nexus]

what distro youre on?

[frigid barn]

debian-based

[cunning nexus]

sudo apt install tktinter should be sufficent

[south fossil]

neofetch detects it somehow, and it seems to use xprop command:
https://github.com/dylanaraps/neofetch/blob/master/neofetch#L1957-L1960
https://askubuntu.com/questions/72549/how-to-determine-which-window-manager-and-desktop-environment-is-running

[river raptor]

olkay thanks

[rustic crater]

hello

Is unix has only terminal in ?

[fickle granite]

try asking again, I didn't understand that

https://translate.google.com/

[south fossil]

Modern UNIX OS does not exist. Someone says MacOS is true UNIX. Okay fine, but in this case it's obvious it is not only terminal

So in modern world UNIX is just a specification / standard. However, we have lots of unix-like OS'es, and they definitely include not only terminal

[wise forge]

it is common for Server linuxes to run only terminal though. no point to have gui there

it is always an option to install missing GUI if desired though and connect over RDP

there is just no point in it

it is enough doing actionvs over ssh

or even if desired u could connect IDE over ssh/sftp and etc

there is even Linux GUI with RDP offered to be run as container somewhere 😅

that makes pressence of GUI temporal at server if necessary

but there is no point in it

the most used option is usually just SSH tunneling through server, in order to access local/private network resources via your dev machine
(Sometimes people set VPN as an option too)
that makes forwarded resources accability to local dev machine

there are always an option to use programmatic tools to configure server, like Docker/Ansible and etc

people are too used to linux terminal of doing things over console, and not really needing GUI at servers. which would be just consuming extra resources for nothing

[fallow tusk]

That's a lot of words to answer a questioo that has yet to be clarified…

[south fossil]

yeeeeee

[mint sage]

To summarize, Unix is not an OS, and it doesn't only carry out only a terminal either.

[tribal finch]

Eh, this isn't really true

AIX, Solaris, HP-UX, they are all still in existence and updated, and are proper Unix Operating Systems rather than just "Unix-like"

but they are primarily used in large (and usually old!) enterprises, and not very much in smaller business, clouds, consumer environments, etc.

Solaris is a proprietary Unix operating system
...
Latest Release: 11.4 SRU57[1] / May 25, 2023; 2 months ago
https://en.wikipedia.org/wiki/Oracle_Solaris

AIX is based on UNIX System V with 4.3BSD-compatible extensions. It is certified to the UNIX 03 and UNIX V7 marks of the Single UNIX Specification, beginning with AIX versions 5.3 and 7.2 TL5 respectively
...
Latest release: 7.3 TL1 / December 2, 2022; 7 months ago
https://en.wikipedia.org/wiki/IBM_AIX

HP-UX (from "Hewlett Packard Unix") is Hewlett Packard Enterprise's proprietary implementation of the Unix operating system, based on Unix System V
...
Latest release: HP-UX 11i v3 2023[1] / May 2023; 2 months ago
https://en.wikipedia.org/wiki/HP-UX

[west musk]

anyone here familiar with grub the bootloader? guy in ot0 needs some help, maybe grub can help

[frigid barn]

i have problems with installing requires for some script.
my OS is solydx 10 (debian). python yet 3.7
here error log:

      x86_64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -DFFI_BUILDING=1 -DUSE__THREAD -DHAVE_SYNC_SYNCHRONIZE -I/usr/include/ffi -I/usr/include/libffi -I/usr/include/python3.7m -c c/_cffi_backend.c -o build/temp.linux-x86_64-3.7/c/_cffi_backend.o
      c/_cffi_backend.c:15:10: fatal error: ffi.h: No such file or directory
       #include <ffi.h>
                ^~~~~~~
      compilation terminated.
      error: command 'x86_64-linux-gnu-gcc' failed with exit status 1
  
      ----------------------------------------
  Command "/usr/bin/python3 -u -c "import setuptools, tokenize;__file__='/tmp/pip-install-kw4hzj5w/cffi/setup.py';f=getattr(tokenize, 'open', open)(__file__);code=f.read().replace('\r\n', '\n');f.close();exec(compile(code, __file__, 'exec'))" install --record /tmp/pip-record-25eidkg7/install-record.txt --single-version-externally-managed --prefix /tmp/pip-build-env-71uy8s1h --compile" failed with error code 1 in /tmp/pip-install-kw4hzj5w/cffi/
  
  ----------------------------------------
Command "/usr/bin/python3 -m pip install --ignore-installed --no-user --prefix /tmp/pip-build-env-71uy8s1h --no-warn-script-location --no-binary :none: --only-binary :none: -i https://pypi.org/simple -- setuptools>=61.0.0 wheel "cffi>=1.12; platform_python_implementation != 'PyPy'" setuptools-rust>=0.11.4" failed with error code 1 in None

command: pip3 install impacket==0.9.23

[spark mulch]

Install libffi-dev

[frigid barn]

k

[fickle granite]

Might be safest to install everything listed at https://github.com/pyenv/pyenv/wiki#suggested-build-environment under your particular distro

[lean blaze]

What's a lightweight linux distro with a graphical environment which runs python well (&tkinter)?

[ember quiver]

If you want extremely light, maybe Alpine (it does include tkinter in it's main repo apparently). If you just mean something that doesn't need 8gb of ram to be useable, I'd try any normal distro with a lighter window manager (Lubuntu or Xububtu for example)

[main olive]

Artix,Arch, Alpine, Parrot, Arco, many others. Just use Xfce, Lxde or lxqt desktops

[formal schooner]

yeah the distro usually matters less than choosing the right desktop

debian with xfce is an old staple

lately i've been enjoying endeavour, it's very nicely set up. i believe xfce is one of its default options

[sleek swan]

I'm trying to output a string like this echo "string" > file.txt but when I cat the string it's not the same as what I put inbetween the quotes. Why is this?

[formal schooner]

can you give an example that actually shows the problem?

[sleek swan]

This was the string "$y$j9T$dDRO50LNqijRmgC4eO4TA/$MbRSqDrgELP1JjxbW/BslB4Qtumg.nipkMsGg3ffhM4"

[formal schooner]

without quoting, those $ will get processed as parameter expansion by the shell. so $y will expand to whatever y is set to, or the empty string if unset. and so on

echo '$y$j9T$dDRO50LNqijRmgC4eO4TA/$MbRSqDrgELP1JjxbW/BslB4Qtumg.nipkMsGg3ffhM4' > file.txt
cat file.txt

i also hope you didn't just share an actual API key or something

[sleek swan]

nah just a hash of the word test lmao

[sleek swan]

wdym by without quoting?

[formal schooner]

putting quotes around the string

[sleek swan]

so using double quotes doesn't "count" as quoting?

[formal schooner]

it counts, but parameter expansion still occurs inside double quotes

double quotes mostly protect against whitespace being interpreted as an argument separator

single quotes stops all expansion, including parameter expansion and escape sequences

[sleek swan]

aha! thank you so much!

[south fossil]

it is probably a salted hash

[fickle granite]

mmm corned beef hash

[hexed rose]

Anyone here understand swap partitions on Linux? I was wondering if I even needed it and wanted to free up the space and use for other partitions.

[fickle granite]

you probably don't need it

you only need it if there's a real risk that you'll run out of RAM, and if you're willing to accept a 1000x slowdown to prevent that

(ok, ok, maybe it's only 100x)

[hexed rose]

I have 32 GB ram on this machine. I looked at my system and found that it is using something called a swapfile. I am confused. Here are the outputs:
`:~$ swapon
NAME TYPE SIZE USED PRIO
/swapfile file 2G 0B -2

$ sudo swapon -show
Filename Type Size Used Priority
/swapfile file 2097148 0 -2`

[fickle granite]

you have a swapfile, but it's not being used.

unless you're low on disk space, don't worry about it.

[main olive]

parts are arriving in 3 days, still debating on mint or ubuntu

my only issue with ubuntu is the desktop and i can change that

idk about mint though

[hexed rose]

I'm only worried about the part where 20 GB of my disk size has been allocated to swap partition during first install of the OS. If the understanding here is that I am not even using that partition, I'd love to allocate that 20GB to the linux partition.

[fickle granite]

looks like it's 2GB, not 20GB

I mean, sure, get rid of it if you want.

I think you need to run sudo swapoff /swapfile or something, then sudo rm /swapfile

now, "using it for other partitions" might be tricky.

[ember quiver]

You can use Ubuntu with pretty much any desktop environment you want. There are official spin offs like Kubuntu, or you can install the server version which comes with no desktop and add whatever you want. Then there are unofficial spinoffs like Regolith (which is Ubuntu with i3 window manager and no real desktop).

If you're a Windows user and just want something intuitive and easy by default Linux Mint is a good bet

[formal schooner]

it's really useful if you want to hibernate, that's often where your memory state gets dumped for hibernation. in general it's a place where the OS can dump stuff that it doesn't want to hold in memory but doesn't want to completely free either. that's possibly because you're out of memory, but also can be for things like filesystem caching

[formal schooner]

you will need to remove the corresponding entry from fstab. otherwise your system will go looking for the swap file on next boot and fail to find it. you will likely end up in the grub console in that case.

[formal schooner]

fedora kde is a great experience if you can get past the installer, which is very much not a great experience

[main olive]

I’ve never used windows, only macOS

[formal schooner]

endeavour kde has also been excellent but i don't recommend rolling release to someone who isn't prepared to deal with rolling release

[main olive]

I have some experience with ubuntu

[west musk]

is drag an drop worse on linux in general? or is it just kde/plasma?

i was an avid drag and dropper on windows, but theres a lot more friction on linux

could just be that im not used to it ofc

[fickle granite]

I personally gave up on Linux guis ages ago.

[west musk]

ijust tried to drop a file into vscode, and every time i mouse over vscode in the taskbar it darted away to the other side of the task bar

actually all non-terminal items on the taskbar does that, how strange

gotta investigate if its a stupid feature or a bug to report

[ember quiver]

I'm not a KDE/Plasma user so I'll blame that ...

To me half the point of using Linux is to need the mouse as little as possible, but I've not noticed any major problems with drag and drop when I do it

[west musk]

i probably just use it 100 times more than the average person

dont really use my left click normally when browsing, most links i open i just drag and drop into the tab bar

[hexed rose]

2GB is the swapfile
There is 20 GB I allocated to swap partition during install. I was referring to that.

[fickle granite]

oh, well those are entirely separate

they solve the same problem, but they're otherwise unrelated

[hexed rose]

I am asking about the orange one.

[fickle granite]

sure

same answer: if you're not worried about running out of RAM, then sure, use it for something else

[hexed rose]

I don't use hybernation. Also, please check the screenshot I posted. I am not talking about the 2GB swapfile I found on the system but the 20 GB swap partition I allocated during original instal.

[fickle granite]

yes, I understand that

the swap file and the swap partition solve the same problem. If you don't have that particular problem (running out of RAM) then you don't need either.

[hexed rose]

Since it looks like the 20 GB partition isn't even being used, I will probably have to use GParted to allocate this 20 GB to my linux partition. I thought I could do it from inside the OS but it may require a flash drive.

[fickle granite]

it might; I haven't repartitioned in so many years that I assume the technology is now different

[fallow tusk]

If you're moving the system partition, yes, you'll need a live system to do that from.

Installs will often automatically give you a swap partition, unless you actively ask them not to

Although I'm surprised that you got both a swap partition and a swapfile from an installer (I'm assuming that's where they both came from)

I'm also a tad surprised it bothered with an extended partition for the ext4 (system?) partition. You only have three partitions. You don't need one in an extended partition.

(Not that I know of any downsides to that, beyond having the extra partition entry)

[hexed rose]

during the install, it let me allocate some memory for swap. I gave it 20 GB. I have no idea why it ignored that and just made a 2gb swapfile

[fallow tusk]

Well, it didn't ignore that. You got both

Oh - memory, not drive space?

[hexed rose]

drive

[hexed rose]

IDK what that extended partition even is. I just know that it is there. lol

[fallow tusk]

An extended partition allowed you to have more than four partitions on a drive (I don't think that's a limitation now, either)

[hexed rose]

Well! I'll have to find a way to safely backup the linux partition first and then try to GParted resizing thing.

[fallow tusk]

Do you have 100 Gigs of storage somewhere else?

The easiest is to use dd to create a raw image of the partition

[hexed rose]

yes

[hexed rose]

But wouldn't that be a problem for resizing?

[west musk]

is it a hdd?

[hexed rose]

ssd

[west musk]

resizing isnt an issue on ssds is it?

[hexed rose]

no idea

[west musk]

im not 100%

[fallow tusk]

If it's running the current system? I think it still is

[west musk]

but theres just no way it would take forever like an hdd

even if data had to be moved, its so much faster

[hexed rose]

in my head, for an SSD it should only be updating some list of pointers or something

[west musk]

it probably doesnt have to be moved either

yeah i mean its shuffling data around for wear levelling anyways

[hexed rose]

The bit I didn't understand is if my dd backup is image of a 100gb partition, wouldn't restoring from it force the new partition to be 100?

[west musk]

depends on what type of image

but not necessarily

[fallow tusk]

If you found you needed the backup, it's either that your partition is borked, in which case you just write the data to a new 100G partition and start over. Or it's that you need some of the data, in which case you mount the image and grab what you need

[hexed rose]

oh sorry! Yeah! My bad. For some reason I was thinking I'd make a 120 GB partition after dd and try to slap on this 100 gb image on to that.

I hope resizing the linux partition isn't going to mess up the data partition

[fallow tusk]

The NTFS partition? No, it shouldn't

If it does, something more major is going wrong!

[hexed rose]

From all the gparted tutorials I've seen, it doesn't look like people are usually afraid of resizing.

[fallow tusk]

I'm always tentative - I've had things go wrong enough times in the past that I am just extra cautious, however much things have improved over the years ;)

[hexed rose]

same here. Dealt with horrible nightmares. My system is Ubuntu 18.04. Not sure if it has the improvements your disasters should have had.

[finite crater]

I've definitely observed people having disaster situations after messing with the partitons to where I would say no matter how much more feasible it is than it use to be, it's not something you should do without backing up unless you are prepared to lose the partition.

I think it often has to do with, they downloaded some piece of junk free trial crapware to do it, or...they just didn't understand what they were doing enough and did the wrong thing

But it happens

[fickle granite]

'"The servers in today's data center are like puppies -- they've got names and when they get sick, everything grinds to a halt while you nurse them back to health," Joshua McKenty, co-founder of Piston Cloud, is quoted as saying in a recent company press release. "Piston Enterprise OpenStack is a system for managing your servers like cattle -- you number them, and when they get sick and you have to shoot them in the head, the herd can keep moving. It takes a family of three to care for a single puppy, but a few cowboys can drive tens of thousands of cows over great distances, all while drinking whiskey."

sounds like y'all treating your ubuntu box like a pet

or else you're not drinking enough whiskey, or saying "yee-haw" loudly enough 🙂

[fallow tusk]

Yeah. I treat my data with tenderness because I never get around to backing it up until it's possibly too late.

[finite crater]

my nginx containre uses FROM nginx:latest
my python container uses FROM python:3.11.3-alpine

if I shell into the nginx, I will not have features like tab completion or ANSI colored output. However if I shll into the python one, I will.

Specifically for the tab completion; what is missing that makes it not available on the nginx container?

[fallow tusk]

Assuming it's using bash as its shell, then https://stackoverflow.com/a/5574402 shows how that works - but it may just be that you need the (equivalent of the) bash-completions package (which is mentioned towards the bottom of the post). [Note: I don't see a bash-completions or similar in the Arch repositories, so ymmv.]

My /etc/bash.bashrc references /usr/share/bash-completion/bash-completion, but while I have /usr/share/bash-completion/ with content underneath it, I don't appear to have /usr/share/bash-completion/bash-completion itself, which seems a bit odd.

(I'm running xonsh, though, so things may not all be tied up for bash)

[finite crater]

ah its probably as simple as just installing that in the Dockerfile.

[harsh willow]

do anyone has idea abount knn?

@finite crater

[finite crater]

nah but someone here might #data-science-and-ml

[west musk]

i looked it up earlier

but i dont know much about it

didnt seem that interesting

i cant open links from discord, this happened before too, but the solution last time doesnt seem to work now

last time i had to uninstall wslu im pretty sure

googling now says something about setting some dbus variable to autolaunch

anyone had the issue? how do i fix it?

[long burrow]

When I do rsync -avP, I get a copy of the files I rsynced in a directory that's named as the ssh user@ip. Is there a way for me to modify the files in that file so it automatically links to the file in the VM I sshed into?, e.g changes I make to the file in my local directory is reflected to those files in the VM?

[fickle granite]

no

that's not what rsync does

you might be able to get something like that with sshfs though

[tribal finch]

alternative, Clear Linux is pretty neat! I love Alpine, but it's missing a lot of key things for use in things other than containers IMO

[scenic fjord]

is unix especially great for programming?

[south fossil]

Nope. Linux is just a tool, same as Windows

[scenic fjord]

cool

[fallow tusk]

It is a bit more oriented towards having tooling for programming easily available

You are expected to want to tinker with Linux, less so with Windows

[fickle granite]

I'll just come out and say "yes".

To a first approximation, the only programming that works well on WIndows is: writing Windows apps.

[scenic fjord]

thats right to an extent, but what about cross-platform languages?

or tools

[fickle granite]

what about them?

[scenic fjord]

they can work outside windows

[fickle granite]

sure. I don't see how that relates to your original question, though.

[scenic fjord]

so if you use unix, is it more useful?

[fickle granite]

depends what for

[scenic fjord]

ok then thanks

[finite crater]

Never stepping outside of windows is similar to never leaving your home country. You can live your whole life that way... but youll be missing perspective of the world.

[main olive]

ubuntu on da floor

[main olive]

i’m a masochist so im going to game on ubuntu

it’s not as bad as i remember it

[main olive]

it's not LTS lmao

[south fossil]

because of LTS

[neat jasper]

you are running that off the garbage can ?

[latent topaz]

vim moment

[neat jasper]

I have fedora but have not been able to get Steam working on it. Do you have to use Ubuntu for gaming ?

ive tried converting the deb package to redhat package but just its not easy

[humble falcon]

Have you tried flatpak?

And no, you don't need ubuntu

[main olive]

yes

[neat jasper]

How many Banana core is it ?

or are you using apple core

[neat jasper]

i will have to look that up, thanks !

[neat jasper]

Just messing looks great

[main olive]

lmao

pc specs are going to be good for programming and a little gaming

gpu hasnt come yet but it still performs well with everything i threw at it

[shy elm]

regex moment

[lapis cloud]

Hit some weirdness. According to https://en.wikipedia.org/wiki/Comparison_of_file_systems#Limits, the filename length limit on ext4 is 255 chars. Yet I'm able to create (in my folder, haven't tested elsewhere) files with ASCII filename of lengths only up to 143. Any idea what that may be caused by?

[fickle granite]

is this something weird like wsl?

I imagine the file system's limit is only one of many; the kernel has a say in the matter, too

[lapis cloud]

Nope, just an ordinary linux mint system.

ah-ha, got it maybe. If I write to /tmp/filename, I can get all 255 chars of length.

So it's only for my home directory. And it's encrypted, IIRC my decisions from like a year ago.

so maybe it's an encryptfs weirdness.

[fickle granite]

likely

betcha they encrypt the plaintext filename, then tell the kernel to use that encrypted wozzit as the file's name; and the encrypted wozzit probably has some header

[normal dagger]

My cibuildwheel runs on Alma_Linux but this building process fails. what could be wrong?

[tool.cibuildwheel.linux]
before-all = [
    "git clone https://gitlab.freedesktop.org/libinput/libei.git",
    "cd libei",
    "pip install Jinja2",
    "pip install attrs",
    "yum install -y pkgconfig systemd-devel libxkbcommon-devel libxml2 doxygen python3-pytest",
    "yum install -y meson",
    "meson --prefix=/usr builddir/",
    "ninja -C builddir/ ",
    "ninja -C builddir/ install"
]
ninja: build stopped: subcommand failed.

I only can do 2 things, find a way to install it from the source by using AlmaLinux or go with an image which supports the package.

yeah....

[fickle granite]

perhaps "ninja -C builddir/ " should have a verb? Disclaimer: I have no idea what any of this stuff is 🙂

I'd spin up an Alma VM and try those commands by hand

[normal dagger]

What do you mean?

What's wrong with the ninja command?

[normal dagger]

Btw this is a cibuildwheel.

[fickle granite]

Well, you've got two invocations of the "ninja" command. The second of them has a "verb" -- "install" -- but the first one doesn't. I was wondering if that first one might be wrong.

[normal dagger]

there's been going on some discussion about the issue here: (if you are interested)https://discord.com/channels/267624335836053506/1138844639449858130

[silk herald]

i think its almost comparable with windows barring games with ( anti gnu/ linux)->( anti cheat)

[main olive]

yeah

it’s nice that amd drivers were engraved in the kernel

[silk herald]

ah cool, i hear nvk may be a suitable open source alternative to the nvidia propriatary drivers when the project is more complete

[main olive]

cool

this was first build so it was nice not to have to worry abt drivers

[silk herald]

ah

[silk herald]

i was interested in a non c++ but performant programming system and was therefor interested in rust, do you prefer one such as go?

[main olive]

why not c++?

[silk herald]

because i find it wordy and without already knowing the system, i put more faith in a newer system

[main olive]

as someone who has been learning c++ for two months

it is as bad as everyone says

i did python for 2 years

and c++ was a HUGE transition for me

everyday it is something new that makes my brain mushier

[silk herald]

lol mood

[normal dagger]

Anyone knows how can I get this folder-sharing work in a vm(Fedora)?

[main olive]

how do I activate the venv on macos?

I created the virtual environment

now I have to activate it

[fallow tusk]

Should just be django_test/bin/activate, I think. Might want to check whether the activate script in bin needs any suffix…probably not

doesn't Mac :D

[lapis cloud]

might need to do something like source django_test/bin/activate but the path should just be that

[main olive]

I could finished it, thanks

[main olive]

it worked

now I am at this part, but I want to use VSCode

should I've done all these steps on VSCode tho?

[fallow tusk]

No, VSCode will pick up that venv

[main olive]

or if I just open the project it will work just fine?

that's the project I've created, like this?

[fallow tusk]

Yes, if you start a project in that directory, it should find the venv Python. If it doesn't, you just select it from the menu on the bottom-right (or Select Python Interpreter in the…shift-ctr-p menu (whatever that might be on Mac! shift-cmd-p?)

[main olive]

you use linux right?

[fallow tusk]

Yep

[ebon dove]

some_command > /dev/null 2>&1 & is the standard way to run something in the background without being bothered by stderr and stdout right?

[fallow tusk]

Yep.

[ebon dove]

thanks!

[ebon dove]

is there a unix-interpretable character that has no unicode representation?