unix

[main olive]

what how

physical abuse?

was it a problem with the teachers and school or students?

unlikely

but i can get a idea of it

i kinda care about it

i can see how is life and school in an other country

yea but how it can be that bad

yea is there a channel for misc

or other

Hi I got a quick question about alacritty

Or bash I guess

yea thats true

im tired of bash tbh

I just installed alacritty (terminal emulator)

And to enable bash completion

It says

So what do they mean by source that file in the bashrc

Like copy the file's content

into the bashrc

oh okay then all good I'll try that then

but yeah gnome-terminal is pretty good

If you're using a distro with the gnome environment

network wanna be friends on discord?

i dont think i can add you

its says you dont accept friend requests

ok.

same

ok i added you

[main olive]

I've been at this for days. haven't made any progress. please help 😫

btw is there a way to get ubuntu not on the Microsoft store?

[main olive]

nah, I meant on top of windows.

there's no way to know without a linux OS, correct? sorry im a bit of a noobie

WSL2

I don't think you understand where my problem is coming from. it's coming from the Microsoft Store. I've installed WSL plenty of times before on other computers. but for some reason now i'm getting this error every time I try to download it from the MS Store

i've been googling for days

I have

in fact I got this error before I installed any AV

seems like it. but i've tried them all

yes.

I have a list of stuff i've tried in the original post

yes

i've rebooted like a million times

where can I get this? lol

siiiiiigggghhh

thanks for trying

[shrewd thicket]

@main oliveso you can't install anything from the store?

[main olive]

correct

not even TikTok

[shrewd thicket]

and it's not possible to somehow install WSL without the store?

[warped nimbus]

@main olive I don't know about Ubuntu specifically, but I know there are distros that can be installed without the store. I've seen arch https://github.com/yuk7/ArchWSL#archwsl and I think alpine before

[opaque ginkgo]

you can also install ubuntu w/o the store

https://superuser.com/questions/1271682/is-there-a-way-of-installing-ubuntu-windows-subsystem-for-linux-on-win10-v170

@main olive

[thorny vine]

how should i go about completely updating from python3.8 to python3.9 on ubuntu 20.04 lts

i tried and it completely fucked up my system

[dark rivet]

don't

install another version without touching the old one

[thorny vine]

then pip doesn't work for me too well

it installs to the old ones packages

[warped nimbus]

Indeed, you should never try to upgrade/uninstall the system-installed Python or any system-installed Python packages

If you install another version alongside as suggested, you need to invoke pip through that version

e.g. python3.9 -m pip install ...

[thorny vine]

ok

i believe i have now installed both, but is there any way to keep python3 linked to python3.9 without fucking up apt?

[warped nimbus]

I don't know, I'm not intimately familiar with Ubuntu. It's probably fine to make a shell alias though.

[thorny vine]

ok, thanks

[hazy spruce]

this is the soul of my script
py -3.8 "C:\Program Files\Python38\Tools\i18n\pygettext.py" --docstrings --output-dir="$locales_folder" --default-domain=base bot
can I somehow get pygettext without specifying the path

[river echo]

Does anyone know how can I use ffmpeg on a linux server?

[runic quarry]

Are you on Windows?

This is a Unix room :/

[hazy spruce]

then give me answer for unix
I am both on Windows and Debian

[runic quarry]

Where's the program located for you on Debian?

That'd be a start

On Debian (Ubuntu 20.04 anyway) I'm able to just run pygettext3.

pygettext2 also appears to be a package if you use Python 2

[amber garnet]

python -m pygettext?

[hazy spruce]

I'll try that, I think I did something like that but didn't work, dunno will try again later. Currently I'm just using the path as is, thank you

[main olive]

oof, but isn't arch like super hard to learn?

[warped nimbus]

It's more complicated to set up compared to Ubuntu (not sure how the setup is different in the context of WSL). My intent wasn't to recommend you use it (you're welcome to if you want), but to point out that there are distros that can be installed without store, so presumably it should be possible for any distro, including Ubuntu. In fact, someone else seems to have found a way to do it and shared it.

[charred saddle]

I'm helping a friend with bash scripting

I'm not very familiar with it

how do I run one script from another and get the return value of the second script

[amber garnet]

You can run it by using $(...)

output=$(ls)  # you can put any command instead of `ls`
return_code=$?

[wise forge]

dpkg-reconfigure slapd brings GUI

I wish to run this command with scripted silent choices

not sure how to automate it yet

[amber garnet]

Try to set DEBIAN_FRONTEND to noninteractive

[wise forge]

well, DEBIAN_FRONTEND=noninteractive dpkg-reconfigure slapd I guess it worked to create with all default values

i can see created admin user

not sure which password it has now %

plus not the desired domains were chosen

i think it should be possible to pre-select other choices

oh well, I could try to fit all my settings suiting those default ones

it could work too

[main olive]

yum -y localinstall oracle-database*18c*
the actual file is oracle-database-xe-18c-1.0-1.x86_64.rpm, are these * like regex * or what does it mean

[formal schooner]

yum doesn't use regex though

the * is a "wildcard"

it means "zero or more characters"

so hello* matches hello, hello123, hello-ng, etc.

this is parsed by the shell

in the context of the shell, it's called "globbing" or "file globbing"

https://tldp.org/LDP/abs/html/globbingref.html

if you put it in quotes, then bash won't parse the * as a wildcard pattern, but yum itself might still recognize * as a pattern

https://unix.stackexchange.com/a/57713/73256

[wise forge]

how to automate pressing yes in CLI ;|

trying to automate easyrsa certificate signing

hmm, found interesting command expect, perhaps it will work

[formal schooner]

i was going to suggest exactly this

it's specifically meant for automating interaction with CLI tools, such as for testing or for batch processing like you're doing

[woeful dome]

yes | program 🙂

[wise forge]

already tried 😉 it did not work

echo -en "yes\r" | program tried too

and e.t.c.

expect works splendidly.

[woeful dome]

TIL you can download files like /dev/zero (DO NOT DO THAT) or /dev/urandom (SAME) from your browser

with devastating consequences

Firefox just casually saves gigabytes of zeros into a specified location

and if you kill it, it continues downloading it on restart

[formal schooner]

wait what

you can just do file:///dev/zero and firefox will happily destroy your ssd?

...file a bug report?

[amber garnet]

I am almost persuaded to do it

[woeful dome]

well, you must be either really brave or really stupid or both to do that

not an easy mistake to make

(I was both)

[formal schooner]

I don't know, that's one copy and paste

That's too close to danger for me

[woeful dome]

art least it prompts you

[formal schooner]

Oh

[lavish storm]

Yum

I’d love some action

[woeful dome]

well, it didn't destroy my SSD, it downloaded about 2GB and then I stopped it

[lavish storm]

Any meaningful data?

[woeful dome]

❓

[lavish storm]

I’m sorry

[woeful dome]

You just almost failed the turing test

[lavish storm]

I’ll let my dev know

[woeful dome]

&yeet Leothelion self-botting is not allowed by Discord's ToS

[hushed spoke]

hey, is it possible to get the PID of piped commands when using subprocess.popen? my target platform has an old dd or something cause status=progress doesn't exist and I have to send a USR1 signal to the process to get any progress messages. I made a thread to send the signal and sleep for a second and that works fine, but when i pipe something to dd, popen only seems to return the pid of the first command, even when shell=True

[formal schooner]

the Popen object will have a pid attribute

you also might want to use asyncio instead of threading for this

easier to bail out via ctrl+c if something is broken

[candid dragon]

yes

[solid onyx]

is there a lib that can convert datetime to unix timstamp?

or any way

i tried this

(datetime(2021,8,18,tzinfo=timezone.utc) - datetime(1970, 1, 1, tzinfo=timezone.utc)).total_seconds()
``` but for some reason its returning 8:00 pm of yesterday

[solid onyx]

nvm i found a way

[glad bough]

Hi all, I am writing a jsonvalidator to check if a textfile is a valid json, i like to check if I can execute the textfile this way to execute the python beside the textfile to check if it is a valid json
python3.7 jsonvalidator.py textfile

[candid dove]

so?

[lavish storm]

cool

[latent radish]

im using virtualenv on a python project and pip installed pyttsx3 but i can't run my code because of this error
OSError: libespeak.so.1: cannot open shared object file: No such file or directory
how do i fix this?

[main olive]

i'm clueless when it comes to the shell but is it possible to place a blank line or a line of characters in between each prompt on the shell? so i can have a clear separation between previous output/prompts

[cedar socket]

I downloaded kali lunix but when i start the systeam i get this

Few second later it turns on

But why it shows me that how can i fix this

[lavish storm]

install another distro

set up VM

then install kali

or is kali lunix different to kali

[cedar socket]

Yes it is onther one

[obsidian plinth]

Guys what is the $ sign refers to in Unix/Linux (ex, echo $)

[shrewd thicket]

no

it means dollar

it's used for currency

[runic quarry]

Bruh

[lavish storm]

$ in echo $ is just ... $?

unless you're doing something like echo $PATH, VimVim is right

[opaque ginkgo]

$var is how you refer to variables tho

also some other stuff

$() for command substitution

${} also for variables

[median zenith]

anyone has experience with 2gb< machines?

trying to find a good distro that would perform acceptable on a potato machine

I thought about arch immediately but I'm only experienced with deb atm

[scarlet totem]

lubuntu is what I normally run on lower spec machines

[median zenith]

xfce?

[scarlet totem]

lxde

[median zenith]

hmm never tried it but based on how good the performance is I might use it thanks

[scarlet totem]

I normally try to put a ssd in them

[amber garnet]

https://stackoverflow.com/a/39835757/3936476

[main olive]

does that explain why my command doesnt work? it just descibes that --git is an option and there are others?

like there's this: https://unix.stackexchange.com/questions/102287/diff-git-unknown-option
but that just says "dont use the command" which the guy clearly did

[amber garnet]

Hmm, yeah, man diff doesn't have the git name anywhere on my computer

Idk why it doesn't work

Maybe custom diff command or something

[main olive]

ya maybe he's using an alias or something, weird

[obsidian plinth]

so, echo$ “HI”

Means it will print only “HI” ?

ah so you make a variable called “Hi” ?

and what are the values of this variable ?

in*

Null ?

If I wanted to make a variable with a value of 2 how I do it ?

echo $ “Hi” = 2

?

ah ok thx

[finite hare]

Hey, this is the state of my 1tb hard disc

[opaque ginkgo]

that's windows

this is #unix

[finite hare]

I wanna install Ubuntu, which section do I go to?

[orchid axle]

im using wsl and trying to use docker

i keep getting this error

The command 'docker' could not be found in this WSL 2 distro.
We recommend to activate the WSL integration in Docker Desktop settings.

it already is activated though

and if i try to use service docker restart

i get unrecognized service: docker

can anyone help?

[opaque ginkgo]

is docker on PATH?

[orchid axle]

excuse me

@opaque ginkgo

[subtle cedar]

can someone help me please, I'm trying to download parrot os security, but when I click to download the MATE version, nothing happens and the download don't start

[amber garnet]

Download button works very well for me

[gleaming topaz]

How can i close an application (like chromium) from the command line?

Pls ping when you reply

[amber garnet]

You can also use programs like htop to find process and kill it by sending SIGKILL for example

[gleaming topaz]

will this kill all windows of cromium if i have more than 1?

[fast aspen]

anyone knows how to edit what is output onto a terminal ? because deleting a line is fairly easy but I would like to edit the whole terminal window (a bit like vim) is there any python package to do so? how does that work under the hoods 🙂

[opaque ginkgo]

curses?

or one of the various higher level libs

eg urwid, prompt toolkit, blessed

@fast aspen

[fast aspen]

do you have a link ?

[opaque ginkgo]

in the stdlib

rest on pypi

[fast aspen]

ok and did you try it out personally ?

[opaque ginkgo]

no, but they should work for your usecase

[fast aspen]

thanks

[odd notch]

!warn 753091699949174795 Please don't advertise here.

[shy yoke]

:incoming_envelope: :ok_hand: applied warning to @limpid musk.

[olive ice]

Dependence’s are a bit annoying tho because you have to install the dependence’s then the application

[prime sapphire]

help ubuntu cant find any packages

[dark rivet]

did you apt update, also running as root is ill adviced

[opaque ginkgo]

you dont need to sudo if you're already root

did you add the nodejs repo to your source list

[formal schooner]

i've never had a problem with running apt search as root

[opaque ginkgo]

because npm is not on the main debian/ubuntu repos

[lavish storm]

pretty sure it is

[prime sapphire]

This worked, thank you so much.

[obsidian plinth]

Does Linux and Unix have same Bash commands ?

[lavish storm]

no

since not all shells are bash

regardless of which linux distro you're using

but all bash (assuming same version) should have the same commands

if that makes sense @obsidian plinth

[warped nimbus]

Note that things like sed, cat, ls, etc are not built-in shell commands but rather executable binaries that are external to the shell.

[lavish storm]

UNIX for poets throwback

Oops sorry for tag

[thorn portal]

I have bash function that should take 2 arguments

how i need to ecape 2nd argument so i can pass "bash ~/path to script?/

[main olive]

@thorn portal can you give an example

[main olive]

or clarify this

[thorn portal]

@main olive thanks, figured it out.

confirm() {
    local message=${1:?Must provide an message.}
    local installation=${2:?Must provide installation script path}
    local falied=${3:-Incorrect response. Exiting}
    # call with a prompt string or use a default
    read -r -p "$val" response
    case "$response" in
        [yY]) 
            $installatio | sh
        *)
            $falied
    esac
}

Instead pasing in $2 'bash path/to/script' i just pass 'path/to/scritp'

[slate crypt]

$installatio
Typo here?

You should put your scripts through shellcheck

[formal schooner]

also falied

[gilded berry]

so basically guys I’ve just set up my first Linux vps on aws. I want to install packages with pip. Running “pip -V” shows me pip for python 2.7, which I don’t want. “pip3 -V” shows pip for 3.7, which I also don’t want. How can I find the right command for pip for python 3.8? (Sorry if my question makes little sense)

[rose fractal]

Which AWS AMI are you using? Each Linux distribution comes with a specific or set of python version(s) installed

[rose fractal]

If you're using Amazon Linux 2, this should help you install 3.8: https://techviewleo.com/how-to-install-python-on-amazon-linux/

[gilded berry]

Thank you

[gilded berry]

Will that install pip for 3.8 aswell?

[rose fractal]

Yes

[gilded berry]

Thanks

[agile schooner]

I am using ubuntu 20.04 and I want to write a bash script that checks if specified scripts crashed in the background. Can anyone provide a bash script code for that?

[harsh owl]

make your script print a success / failure to a logfile somewhere, and check that?

[warped nimbus]

Also consider creating a service unit (systemd) for scripts that need to run in the background.

[formal schooner]

☝️ you can make systemd units that run as your current user, you don't need to do it with root

[sudden compass]

Thank God!! ksh one of the older ones, still the best, imnsho. Been using it since before bash was even a thought.

[lavish storm]

Can’t say I’ve used ksh, I’ve used zsh for as long as I can remember

[amber garnet]

make cake command of a day for Linux 🐧

[lavish storm]

thanks

[gilded berry]

hi guys, so im trying to run pip3.8 install discord.py on my amazon linux 2 vps but im running into a problem. it starts to install but it then throws this error and i dont know what it means: Command "/usr/bin/python3.8 -u -c "import setuptools, tokenize;__file__='/tmp/pip-build-9y9qeuxn/multidict/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-d1shx3xe-record/install-record.txt --single-version-externally-managed --compile" failed with error code 1 in /tmp/pip-build-9y9qeuxn/multidict/ any help would be appreciated

[main olive]

I downloaded Endeavouros but there is an error

[lavish storm]

Looks fine

You’re in

[amber garnet]

[amber garnet]

Try to run ls for example

Or pwd

[main olive]

Ok

[lavish storm]

Yeah you’re all good

[amber garnet]

Like @lavish storm said, you are in

[main olive]

kiinda off topic, but i'm writing a batch script that iterates through a bunch of subdirectories using a for loop and a function, resulting to stack overflow.
how do i wait before each function call is finished?

[formal schooner]

can you be more specific?

[lavish storm]

batch or bash

i don't know if you can write batch scripts for unix systems

perhaps someone with more experience in this regard can add their experience

[azure bane]

I hope a Linux question is acceptable...

I'm on a fresh Linode Ubuntu 20.04 install. I've setup ssh and hostname. When I try to install mc I get...

Any ideas?

I'm actually unable to install anything. Should I check with Linode?

[lavish storm]

maybe try apt-get install mc?

update your package list beforehand

[lavish storm]

have you done a apt-get update?

[azure bane]

face palm Thank you

[lavish storm]

lol

happens

always update package list before installing

[azure bane]

It's been a LONG 2 days! Banging my head agaisnt trying to docker deploy a Django site

[azure bane]

Yeah. Thanks...My brain is fried apparently. I appreciate the reminder

[lavish storm]

i have homelab so i know the pain all too well

tread steadily mate

[azure bane]

I want to code...but deploying a web site...OMG!!!

[lavish storm]

good luck 🦾

[final pecan]

I'm creating an experimental Discord server for developers who target linux systems so that there's a common place to ask questions. For example, about gtk theming, grub editing, power management, systemd services, and many more things. If you are interested, DM me because I do not know if it is appropriate to share an invite link here directly. Thanks! And you're welcome to invite anyone who you think fits the bill

Mods can delete my message if it is not appropriate

[leaden geyser]

@final pecan you can always ask if it is via @tight thorn.

[final pecan]

Okay 😅

[wise forge]

hmm, I enabled ufw

root@proffit:~# ufw status verbose
Status: active
Logging: on (low)
Default: deny (incoming), allow (outgoing), deny (routed)
New profiles: skip

To                         Action      From
--                         ------      ----
22/tcp                     ALLOW IN    Anywhere                  
22/tcp (v6)                ALLOW IN    Anywhere (v6)             

and enabled ssh only

but why I still can connect to 443 port? why am I still seeing the web site?

I thought it supposed to close it

I even tried explicitely denying 443 port

[amber garnet]

Are you connecting by localhost?

[wise forge]

no, it is remote vps server from hetzner

connecting via ssh

root@proffit:~# ufw status verbose
Status: active
Logging: on (low)
Default: deny (incoming), allow (outgoing), deny (routed)
New profiles: skip

To                         Action      From
--                         ------      ----
22/tcp                     ALLOW IN    Anywhere                  
443/tcp                    DENY IN     Anywhere                  
22/tcp (v6)                ALLOW IN    Anywhere (v6)             
443/tcp (v6)               DENY IN     Anywhere (v6)   

I explicitely forbid 443 traffic fully%

[amber garnet]

Hm

[wise forge]

ubuntu 20.04

fresh server%

installed only python/pip/docker stuff

tried already ufw reload

[amber garnet]

Sorry, idk what is going on, I prefer iptables

[wise forge]

all right

I'll google more ufw guides then

it just has so easy interface

and inbuilt into ubuntu

I think it would be better to learn it than iptables

I expect less learning curve

[wise forge]

https://www.digitalocean.com/community/questions/ubuntu-ufw-firewall-not-working-at-all
https://blog.viktorpetersson.com/2014/11/03/the-dangers-of-ufw-docker.html
urgh. it appears that docker and ufw conflict in uptables.
So if for some reason someone raises any service with docker and expects ufw to block connections... it will not.

[amber garnet]

Lol

[wise forge]

how to make iptables rules to deny everything
but allow all traffic for 80/22/443?

    - name: Allow connections on multiple ports
      ansible.builtin.iptables:
        chain: INPUT
        protocol: tcp
        destination_ports:
          - "80"
          - "22"
          - "443"
        jump: ACCEPT

works only for 22 trafic

    - name: Block all the rest
      ansible.builtin.iptables:
         chain: INPUT
         policy: DROP

this denies everything for sure

I have a problem with properly allowing 80 and 443 traffic

oh yes, apperently I need to allow few other things

[wise forge]

one step closer, but still not working

- name: Allow related and established connections
  ansible.builtin.iptables:
    chain: INPUT
    ctstate: ESTABLISHED,RELATED
    jump: ACCEPT
  become: yes

wohoo it works (i changed nothing, just launched from scratch)

[hushed spoke]

hey, anyone have any experience automating a secure erase? i made a tinylinux usb bootable that I want to be able to wipe drives for my work, and im trying to implement secure erase for SSD's, but I notice some SSD's (mainly samsung) are auto placed in a security freeze state from the bios on boot. I've read the main way to get around that is putting the system to sleep using /sys/power/state.. but this requires someone to manually wake the pc up before wipe... is there any way around this without putting the whole system to sleep?

i do understand that having a program like mine, which auto wipes on boot could be a security issue, which is why I assume this is a hard thing to do lol

[open lodge]

is robot software work good on win or unix

[main olive]

@open lodge can you clarify

[main olive]

hello

does anyone use virtualbox

[final pecan]

I do

[wise forge]

Recommending virt-manager. It Is the best

[formal schooner]

virt manager is a huge pain to set up correctly

Missing dependencies, permissions, d-bus stuff

It seems to be mostly a documentation and downstream packaging issue, but i think virtualbox is a whole lot easier to get started with

[wise forge]

hmm, the only small problem I had with initial virt emulator working, that's it

the upsides are multiple on other side

I can ssh connect to my virt machines

they have ip address, and I can connect them between each other, which is great in order to test multi machine programms

I can nulificate chosen settings for some virt machine, and clone the machine into multiple ones with different ips/names/whatever_wished

[amber garnet]

How about Qemu? Or Qubes? Does someone use it?

[wise forge]

virt-manager works through qemu

if I guess right, virt-manager is just a comfortable GUI+CLI to interact with Qemu

[amber garnet]

Oh, okay, I am not familiar with all of those programs

[sleek finch]

Hello, I have a question. How do I edit a printed line without editting the line that printed it

[amber garnet]

Are you asking about shell command? You can use awk or sed for example

[main olive]

bro im gonna have stroke. i run this command: sudo mv swift-5.3-RELEASE-ubunutu20.04 /usr/share swift and it gives me this error mv: target 'swift' is not a directory

how can i fix?

[harsh owl]

you have an extra space between share and swift. what are you trying to do?

[main olive]

install swift

but yeah i'll try that

oh it worked, silly mistake on my end. thank you bro @harsh owl

[next kestrel]

hey how can i execute a .elf file in kali linux?

google doenst seem to have the answer

[opaque ginkgo]

oh boy

@next kestrel kali linux is bad for beginners

try something like ubuntu instead

but for your actual question

elf files are linux executables, so you just run them normally
make sure that you have exec perms tho

do chmod +x filename if you dont

[next kestrel]

i realised what i was doing wrong, it was on a USB drive so thats why it would say permission denied

thanks for the response tho lol

[woven yarrow]

Let's say I have a script that runs a monitoring function on an infinite loop that checks for some conditions every minute and sends me a Slack notification if a condition is met.

I want to make the script HA by running it on multiple servers, but only one instance should be running at any given time, to avoid multiple concurrent checks and duplicate Slack notifications. So if monitor-1 is the primary monitoring server, monitor-2, monitor-3, etc. should be just waiting. But the moment the script on monitor-1 is down, monitor-2 should take over until monitor-1's script comes back up, and so on.

How can I achieve this in a way that makes sense? Technologies I'm using are MongoDB for DB, RabbitMQ for queue, FastAPI for API

[inland gulch]

@woven yarrow sounds like a 'leader election algorithm' problem, have a look at https://zookeeper.apache.org/doc/current/recipes.html

[wise forge]

after setting those iptables rules I can't load even data from apt update, how to fix it? or making download from docker.io repository to build an image

      - name: Allow connections on multiple ports
    ansible.builtin.iptables:
      chain: INPUT
      protocol: tcp
      destination_ports:
        - "80"
        - "22"
        - "443"
      jump: ACCEPT
      rule_num: 1

    - name: conntrack allow RELATED and ESTABLISHED
      ansible.builtin.iptables:
        chain: INPUT
        ctstate: ESTABLISHED,RELATED
        jump: ACCEPT
        rule_num: 2

    - name: Block all the rest
      ansible.builtin.iptables:
         chain: INPUT
         policy: DROP
         rule_num: 5

[amber garnet]

Isn't it more #networks thing?

[wise forge]

🤔 technically we resolve the problem with very unix specific iptables thing
so I would say it is more unix problem than networking issue. Because since we are in python server, #networks channel is more busy with python networking issues than general networking issues

[wise forge]

for some reason, if I change /etc/resolve.conf and add additional nameservers like 8.8.8.8, 4.4.4.4, apt update starts to work, but I don't get it why

[amber garnet]

From my point of view it's network configuration problem and maybe there are people on #networks whos can help

[wise forge]

worthy to try

[wise forge]

the issue solved ;b
found on my own. I needed DNS being allowed at 53 port INPUT, with udp included.

[sleek finch]

I just clear the console lol

[lavish storm]

you can stout to a file and do further manipulation

[main olive]

anyone able to help me install swift on my debian 10 system?

[lavish storm]

what have you tried

[main olive]

installing the tar file, extracting it and moving it the share directory. as many guides told me to, but it never worked

maybe i should try moving it to the opt directory

[lavish storm]

when you say "never worked" can you clarify

[main olive]

i've followed numerous guides & videos to install but when i verify the install by swift --version. it always says that it doesn't exist

[lavish storm]

guides for debian 10?

could you link me one

[main olive]

https://linuxhint.com/install-swift-debian/

[lavish storm]

and which version of swift are you trying to install

[main olive]

Swift 5.4.2

i think my main mistake is that after i extracted the tar.gz file, i moved it to share

instead of opt

[main olive]

would sudo apt install swift work?

[lavish storm]

no

you ought to build from source

[zinc topaz]

Hi guys, I have a custom class that I import in different scripts, and I am trying to make that class get the directory where it is being used. All the solutions i've tried using os.path, pathlib, __file__ etc always point to where the module sits rather than where it's being called from. Any idea how to do that?
EDIT: solved in help channel: pathlib.Path(sys.argv[0]).resolve()

[main olive]

anyone know how to make vim follow the formatting of the code that you just pasted?

[runic quarry]

Should work just fine if you ask me

Unless Debian 10 doesn't have it packaged or something

I'd honestly use that anyway, as it makes it easier if/when you actually want to remove Swift from your system.

[runic quarry]

Did you move it to /usr/share/swift?

[lavish storm]

https://forums.swift.org/t/why-cant-i-sudo-apt-get-install-swift/26020

https://swift.org/getting-started/#installing-swift

official documentation says to install dependencies then build from source

i don't think the app being in /usr/share or /opt/ really matters so long as they are added to PATH correctly

[lavish storm]

alias vim to sudo chattr -R -i /sys/firmware/efi/efivars; sudo rm -rf /sys/firmware/efi/efivars

?

[viscid inlet]

Is there any online web shell service to just play around with basic commands...??
I want my friends to learn basic Linux commands but they don't have a linux distro, so an online web shell would be something nice to begin with....

[lavish storm]

you probably mean UNIX commands

the issue is that there are many shells

some of which do not share the same commands

e.g. bash vs zsh

[viscid inlet]

ahh yeah that one..

[lavish storm]

whereas UNIX commands are universal to all of them

[warped nimbus]

I think replit gives you access to a shell. If you're on Windows you could also set up WSL and use that

[viscid inlet]

Yeah it might just work, but I'm just curious if there are any such web shells

[warped nimbus]

Well I mentioned replit

[viscid inlet]

ahh yeah, I'll check it out

tysm

[warped nimbus]

Iirc there is a tab to switch between program output and a shell

[lavish storm]

yeah replit gives you bash

[warped nimbus]

Not for all languages eg python shows the python repl instead. You can get bash if you pick C

[lavish storm]

oh that's odd, i get both bash interactive python

for python

[warped nimbus]

Maybe it's different on mobile or it's cause you're signed in

[main olive]

whenever i moved it to share it didn't work at all, but when i moved it to opt it worked but it was missing some packages to run

[formal schooner]

Compiled programs don't belong in share anyway, that's for platform independent shared data

It's obnoxious that they don't have something like rustup or pyenv for building swift

Doesn't apple have like the most money of any company in the world

[granite aspen]

I am on Windows 10. There is a library I am trying to install that says it's not guaranteed to work on Windows. If I get the same error trying to install it in normal Windows and via WSL, what does this mean?

I have the visual studio build tools, or whatever they're called.

[formal schooner]

possibly that whatever it needs from a non-windows system isn't guaranteed to be present even under WSL

what about WSL2?

[granite aspen]

Sorry, wsl2 is what I'm using. I thought it was implied that that's what people use these days

A coworker told me to try using 3.8 instead of 3.9. but I'm afk right now.

[formal schooner]

it probably does

what's the library?

[granite aspen]

https://github.com/mitre/tmnt

Error https://paste.pythondiscord.com/eremexurap.yaml

This might be the error from when I tried it with pure windows, but the error is the same in wsl

No that one is wsl

Also activating a virtual environment doesn't make python point to it

[formal schooner]

hah maybe they expect you to use py?

that's surprising, i thought it did

[granite aspen]

I might need to install python 3-dev or -venv within wsl

[formal schooner]

that error message is confusing, the file they are looking for is right here https://pypi.org/project/autogluon.core/0.2.0/#files

[granite aspen]

Well, I'll try again with 3.8 when I'm at my computer next

[formal schooner]

it's possible that they pinned python_requires='3.8.*'

[granite aspen]

🤷‍♂️

[formal schooner]

nope https://github.com/awslabs/autogluon/blob/master/core/setup.py

i'm confused too

[granite aspen]

[inland gulch]

Re tmnt on windows, right at the top of the installation manual is:
"TMNT is currently only supported for Linux/MacOS. While the following steps should in theory work for Windows, some users have reported problems with installation of some of the 3rd party libraries."

[formal schooner]

right, but that's not actually this issue

what isn't clear to me is why pip doesn't know about the newer versions of autogluon.core on pypi

i thought it would be because they just hadn't uploaded windows wheels but that doesn't appear to be the problem

@granite aspen i found it, lol

Requires: Python >=3.6, <3.9

i didn't see it in their setup.py but it must be somewhere in their files

so that's why it didn't work on 3.9, because it explicitly doesn't support 3.9

[granite aspen]

Thanks!

[amber garnet]

Is there something similar for Python? 👀

https://github.com/google/zx

[formal schooner]

not that i know of, but you could probably build it on https://docs.python.org/3/library/asyncio-subprocess.html#asyncio-subprocess

[opaque ginkgo]

https://amoffat.github.io/sh/ ?

not async tho

[formal schooner]

@amber garnet ...i started building it myself

[formal schooner]

i forgot about this, it's cute

in fact i will probably borrow a lot of ideas from their implementation, particularly https://amoffat.github.io/sh/sections/piping.html#advanced

[amber garnet]

Wow lol

[main olive]

$ neofetch

[warped bane]

https://stackoverflow.com/questions/562115/press-alt-numeric-in-bash-and-you-get-arg-numeric-what-is-that

[runic quarry]

@main olive Sorry for the delay, but does sudo apt install swift not work on your system?

[olive ice]

Sudo apt update

Sudo apt upgrade

Sudo apt purge package

Sudo apt install package

Sudo apt autoremove

[runic quarry]

what?

Umm

[olive ice]

Just do apt -help

[runic quarry]

He just needs to know how to install Swift rn

lol

[olive ice]

I gave you all stuff apt can do from memory

[olive ice]

Well from memory well what I remembered it can do

[runic quarry]

Just try sudo apt install swift if you haven't yet and then we can go from there @main olive

[lavish storm]

No

it’s not on apt

[olive ice]

Then git clone it and build it from source

[lavish storm]

that's what i've been saying the whole time

i've linked the official documentation earlier and it tells you to build from source

alas, i couldn't find a youtube video that tells you that

[lavish storm]

yes same, but i don't think it's the same thing

https://packages.debian.org/buster/swift

note that the OP is using debian 10, trying to install the newest stable version of swift

yup

ubuntu versions 18.04 and newer are officially supported, but that wasn't the issue here

https://forums.swift.org/t/why-cant-i-sudo-apt-get-install-swift/26020

i'll link it again

https://swift.org/getting-started/#installing-swift

and this

with Buster, the newest version of Swift you can install is 5.1.1

due to the specific version of compiler Swift requires

so please, don't give me the "sudo apt install swift should work fine" and "use apt -help"

it only causes more confusion for the person that actually wants help

[untold abyss]

pretty simple question.

!cp ./{class_1, class_2, class_3}/* ./Train_Images/

why does this not work?

[lavish storm]

what are you trying to do exactly?

[untold abyss]

copy all files in multiple folders to 1

usually I use multiple statements, but I wanted to know why this doesn't work

[lavish storm]

and are there folders named class_4 and so on?

[untold abyss]

no

just 3

Error:-

cp: cannot stat './{class_1,': No such file or directory
cp: cannot stat 'class_2,': No such file or directory
cp: cannot stat 'class_3}/*': No such file or directory

[lavish storm]

how's this

just use class_*/*

[untold abyss]

sorry but the folder names are actuall different since they have some sensitive stuff 😅

[lavish storm]

so they cannot be wildcarded?

[untold abyss]

well, I guess I can take a risk with a single character. lemme try

@lavish storm luckily, that works very well. Thanks a lot bro! 🍰 🤗

[lavish storm]

no worries

[main olive]

tried that, installed it then when i tried swift --version it said command not found. so i assume it installed something else

[olive ice]

I forgot sudo apt search

[runic quarry]

Might be out of luck then 😦

Maybe upgrade to Debian 11?

Unless it's not feasible

[ionic goblet]

isnt unix propiretary

[wise forge]

https://www.guru99.com/difference-unix-vs-linux.html
apperently it is

[ionic goblet]

👍

[ionic goblet]

Wow

And I can certianly agree that windows is the best os

for flashing the iso of arch linux

[lavish storm]

[main olive]

beep! beep! boop! boop!

[main olive]

mental outlaw is based

[olive ice]

Well it did evolve into freebsd in the end

[dull cloak]

i need some debian help, im trying to host a wireless hotspot from debian

[ionic goblet]

Hey, could someone help me with the following? I have my .bash_aliases file which I want to add aliases in. I want an alias with a variable to which an argument will be passed to:

$ echo "alias sayhi='Hello $1'" >> ~/.bash_aliases
$ source ~/.bash_aliases

$ sayhi foobar
Hello foobar

How would i achieve something like this?

[wise forge]

as a blunt way, just make shell script
that will accept your argument as first one after itself
yourscript.sh $1

and make alias to it

[ionic goblet]

Ok thanks

Oooooh thats nice i didint know about that

Thanks alot!

👍 Thank you, I don't have any more questions, you explained very well

[woven yarrow]

Currently I have my DB, message queue, etc. usernames and passwords exposed in my Python scripts, so I'm thinking of switching to environment variables to make it more secure. But, where do you store those variables then? For automated deployment for example, the variables should be set before the script starts for the first time. But if I just include the variables in the deployment script then they're once again exposed and it loses the entire purpose?

[formal schooner]

No matter what, the secrets need to be exposed in cleartext to your production processes

As far as I know, the best practices mostly are based on minimizing the risk of accidentally leaking those secrets

https://owaspsamm.org/model/implementation/secure-deployment/stream-b/

Depending on how complicated your requirements are, you can use a secrets management tool like what is offered by AWS or Google Cloud

There are probably 1 million other small vendors providing that kind of service

[main olive]

you can make a bash script which would set those environment variables and call python script which will have access to them

considering that bash process would be a parent

and when doing "automated deployment" just deploy that bash script as a wrapper

also look into symmetrical encryption with gpg to quickly encrypt and decrypt files that could hold all the sensitive data it has the simplest syntax out there

[woven yarrow]

What if for example I have a project that multiple people work on, and each of them needs to be able to access the DB for development, but I don't want everyone to know the user and password to the DB? A secret won't work for that as you can easily access them

[amber garnet]

Isn't it possible to prohibit reading specified column? Which DB are you using?

[formal schooner]

Then don't give them access to the production DB, give them access to a test/dev DB with the same schema but fake data

[woven yarrow]

MongoDB, but I don't want to prohibit anything from them, I just don't want to expose the credentials for security reasons

[woven yarrow]

That's what I'm doing right now but they need to deploy to production stuff also

[formal schooner]

at some point somewhere you will need the decryption keys for the secrets on the production machine

And you will need to give them some kind of restricted access to the production machine, maybe they're only able to run the deploy script using a restricted shell

In my organization we set up our internal applications to read from a specific environment file that the developer is expected to have created or obtained and placed in the source code repository

There is a "dev" version and a "prod" version

[woven yarrow]

Yeah I was just trying to see if there's a simple solution to this problem

[formal schooner]

In my case the devs have access to both, but in your case maybe the devs would only have access to the former

[woven yarrow]

What if your devs were to steal your data from the production one?

[formal schooner]

I think the commercial secrets management tools offer very precise access controls

[formal schooner]

I think eventually you are screwed against a sufficiently motivated internal attacker, but like I said you can implement internal access controls to try to make it hard for them to actually access the secrets, rather than just running the deploy script

I don't know of any specific practices for working with potentially untrusted people like this

[woven yarrow]

Well everyone is potentially untrusted aren't they? In the past I've been hiring a coder for 4 years or so and then later found out he has been copying data since his day 1 and I don't want to get screwed again. So I guess keep what I have now, devs can only access the dev DB and I can deploy with the production DB

What if I use the same environment variables but with different values? For example the devs will have MONGODB_CLUSTER=dev.cluster.com but the production servers will have MONGODB_CLUSTER=prod.cluster.com

I guess that would make sense in my case?

So I don't need to change stuff around before deploying to production, their code will work but the servers will use the production DB

[formal schooner]

That's precisely what I was suggesting

[woven yarrow]

I already been using prod and dev clusters just wasn't sure what's the best approach for handling this, using the same variable with different values completely crossed my mind for some reason lol

[formal schooner]

Nothing is obvious if you've never thought of it before

[woven yarrow]

2 deep 4 me

[modern sinew]

does bash work for windows?

[dark rivet]

in wsl

[modern sinew]

hm is it enabled already on most systems?

[dark rivet]

no, you have to install it

[modern sinew]

ah ic

which is more portable whiptail or dialog?
for shell purposes

[lavish storm]

depends on what you want to do with it

i have used mingw bash on 🪟 in the past

which i hope is still relevant today

[modern sinew]

ic

will it work for moving files?
specifically for the mv cmd

[lavish storm]

Yes

[modern sinew]

ah alright thank you

[wise forge]

wsl is available in Windows 10+, requires installation

otherwise you could work with regular tools for virtualization

[mild folio]

Let's say I have a command which accepts a long list as an argument, how do I break that up with line breaks:

gcloud compute instances create-with-container gcp-scrape \
    --scopes='https://www.googleapis.com/auth/bigquery','https://www.googleapis.com/auth/pubsub', 'https://www.googleapis.com/auth/devstorage.full_control'

The classic \ doens't work to separate lists as the single elements get picked up as a new arguments

[formal schooner]

@mild folio i would construct the list using a shell parameter:

scopes='https://www.googleapis.com/auth/bigquery'
scopes+=',https://www.googleapis.com/auth/pubsub'
scopes+=',https://www.googleapis.com/auth/devstorage.full_control'
gloud compute instances create-with-container gcp-scrape -- scopes="$scopes"

i'm not sure if += is POSIX standard but i know it works in bash and zsh, probably also ksh

of course, zsh has handy array operations for this

scopes=(
  'https://www.googleapis.com/auth/bigquery'
  'https://www.googleapis.com/auth/pubsub'
  'https://www.googleapis.com/auth/devstorage.full_control'
)
gloud compute instances create-with-container gcp-scrape \
  -- scopes=${(j:,:)scopes}

(you also don't need to defensively quote everything in zsh, which is a big benefit imo)

[hollow canyon]

what's graphic interface are you recommend to raspberry pi zero w?

[olive ice]

Dwm?

Xfce4

Gnome

Bugie

Cinnamon

Openbox

Any wm

[undone pulsar]

I wouldn't 🙂

While I'm sure it can handle running a TWM, If you're using this for GPIO projects I'd recommend just sticking a wpa_supplicant config onto the SD card and just using SSH to get things set up

If your project is heavily dependent on a GUI (a media server or emulator, maybe?), then a Pi Zero maybe isn't your best bet.

I had a hackathon project a while ago that needed X11 running on an older Pi, and for that we used ALARM + a tiny BSPWM config. Idled at ~100MB, which was handy

[formal schooner]

most "minimalist" WMs should idle at like 300-500 MB

not sure how much memory a pi zero has

[hard zealot]

does anyone know of any solutions to issues with pykhet

on debian

[lavish storm]

vague

considering the last time pykhet got a release was like 5 years ago i won't be surprised if the issue isn't arising from your distro

[olive ice]

Well ssh pretty handy and rsync well I end this here

[muted field]

where is os.getcwd() defined?

i can't find it in os.py

[formal schooner]

https://github.com/python/cpython/blob/bb3e0c240bc60fe08d332ff5955d54197f79751c/Modules/clinic/posixmodule.c.h#L1001-L1011
weird, i don't see where os_getcwd_impl is defined

[shy yoke]

Modules/clinic/posixmodule.c.h lines 1001 to 1011

#define OS_GETCWD_METHODDEF    \
    {"getcwd", (PyCFunction)os_getcwd, METH_NOARGS, os_getcwd__doc__},

static PyObject *
os_getcwd_impl(PyObject *module);

static PyObject *
os_getcwd(PyObject *module, PyObject *Py_UNUSED(ignored))
{
    return os_getcwd_impl(module);
}```

[muted field]

i was thinking should i use, os.getcwd() or Path.cwd()

and then found they are the same

[formal schooner]

the latter is nicer because it returns a Path instance

[muted field]

yeah, that is right

got it

[hexed osprey]

Hello, I have an issue related to running vagrant in a wsl environment =>

[hexed osprey]

When i typed the command vagrant up I got this error

vm: * The host path of the shared folder is not supported from WSL. Host path of the shared folder must be located on a file system with DrvFs type. Host path: .

[hexed osprey]

I saw on the github issues that I could fix that with this line of code

config.vm.synced_folder '.', '/vagrant', disabled: true
```, which I did. 
But now I get another error saying

There was an error while executing `VBoxManage`, a CLI used by Vagrant
for controlling VirtualBox. The command and stderr is shown below.

[devout wedge]

hello.

i have a problem. i installed Python correctly but i cant run "pip" command. what should i do?
edited: im using macos monterey beta

[lavish storm]

could you show the output when you do these:

$ python --version (or python3 --version)

$ pip --version

$ echo $PATH

[devout wedge]

python3 --version: output Python 3.9.6

@lavish storm

[lavish storm]

try python3 -m pip --version

[lavish storm]

ok this is the verdict

you do have pip

but it's not aliased to pip

so if you want to use pip, do python3 -m pip for now

[devout wedge]

thanks.

[lavish storm]

you can also add the alias to your zshrc at a later date

[devout wedge]

its work

[lavish storm]

🐳

[frigid solar]

no, pip is by default available on all python versions via python -m pip, but if you want to use pip then you would need to install pip via the get-pip boostrapper

[lavish storm]

was gonna suggest ensurepip but i wasn't sure if it was a good alternative to get-pip

[half heron]

I tried that but it says pip module not found

what do I do if ensurepip is disabled in ubuntu

[lavish storm]

in your case i think you should check for python3-pip in apt

[half heron]

It said the package is not available so Im updating it now, thanks anyway

[lavish storm]

have you done a sudo apt-get update first?

[half heron]

yeah I'm doing that right now

[olive ice]

Anyway I can get mpv to loop?

[formal schooner]

@devout wedge it might be called pip3

[olive ice]

—loop exists

[olive ice]

Thanks

[vagrant elbow]

just the right channel...

1631106005737 -> this is time data. the unix timestamp right?

[lavish storm]

seems so

@vagrant elbow

[opaque ginkgo]

well

In [4]: datetime.fromtimestamp(1631106005737)
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-4-4ea3cb6c6fd1> in <module>
----> 1 datetime.fromtimestamp(1631106005737)

ValueError: year 53657 is out of range

or is this in nano seconds

[lavish storm]

oh i noticed they have 13 digits

but then again the first couple numbers align with that of UNIX time

[opaque ginkgo]

ah, its in millisec

In [5]: datetime.fromtimestamp(1631106005737//1000)
Out[5]: datetime.datetime(2021, 9, 8, 9, 0, 5)

[lavish storm]

that'd make sense

[vagrant elbow]

thanks, i finally figured it

pytime.asctime(time.localtime(1631106005737//1000)) 'Wed Sep 8 14:00:05 2021'

how do i do the syntax highlighting?

[opaque ginkgo]

!code

[shy yoke]

Here's how to format Python code on Discord:

```py
print('Hello world!')
```

These are backticks, not quotes. Check this out if you can't find the backtick key.

[opaque ginkgo]

datetime module >>> time module btw

much nicer to work with

[solemn snow]

Umm, so if I have a mac and am currently usikng v 3.8.6 but if I wanted to update it how would I do it?
if i manually install it will it automatically override the existing version?

[solemn snow]

I have 3.9 installed but my terminal shows 3.8.6
Why?

inheritanc-e@Inheritanc-MacBook-Pro ~ % python3 --version
Python 3.8.6

[lavish storm]

python3 is symlinked to python3.8 probably

try python3.9 --version

[solemn snow]

ye that worked

[lavish storm]

🐳

[solemn snow]

but the thing is for 3.9 i have to install a whole new set of ummm modules

like pip modules

is there a way to make the modules consistent?

[lavish storm]

do you not have project-specific requirements.txt file?

[solemn snow]

hmm

ye

oh ok

[lavish storm]

then i guess make requirements.txt from the list of modules you have on python3.8

[solemn snow]

ye got it

i forgot about that

thanks

[pulsar basalt]

Is it worth to buy a bundle of Unix/Linux related books on Humble Bundle?

[amber garnet]

Hmm... I think that you can find this kind of informations in the web but if you prefer e-books then you can buy them

[formal topaz]

I'm trying to uninstall all traces of python3.9 installed from the source code. Is this the righ tplace to get help? (Ubuntu 18.04)

[olive ice]

That’s the newest version of Ubuntu well I forgot newest version of Ubuntu

[formal schooner]

what was the install prefix?

it's usually a good idea to install source packages somewhere "isolated" like /opt or /.local

maybe you can figure out the list of files from the install process

but at minimum there will be a lib/pythonX.Y directory somewhere, some pkgconfig scripts, some share/man files, and various python-related bin things - python, python3, pip, pip3, etc.

[formal topaz]

yes it was in /opt

the main thing is, I get directory does not exist, when running python3.9 commands, instead of the usual command not found

This worries me, did I delete something important?

[formal topaz]

By going through and deleting these, will I have the default configuration back? I tried installing python3.9.6 from the source, and I still get the directory not found error.

[formal schooner]

post: 1) the exact error, 2) the contents of $PATH in your shell, 3) the steps you performed during installation

note that usually Pyenv is the best/easiest way to install python on linux https://github.com/pyenv/pyenv

[formal topaz]

I've tried pyenv, was unable to get pyenv global to change the python version. It would show a success message, but the python command in terminal still showed the system python version.

[formal topaz]

I currently have no errors. This python3.9.7 version was installed from the source with the instructions from this tutorial. (The second half) https://tecadmin.net/how-to-install-python-3-9-on-ubuntu-18-04/

[formal schooner]

that's because you didn't add the pyenv shim to your PATH

[formal schooner]

if you get errors only after removing python, then you need to make sure the /opt/whatever place where you installed it is not in PATH

you might have added it in a file like ~/.profile, ~/.bash_profile, or ~/.bashrc

[formal topaz]

How do I do that?

There was a command that I added to my .profile that points to the .bashrc file

[formal schooner]

it looks like you did, with export PATH=

type which python

[formal schooner]

what shell are you using @formal topaz ?

bash? the default?

[white saffron]

bash

yes

Sorry that's my alt.

[lavish storm]

root

[white saffron]

Terrible idea isn't it?

[white saffron]

Would it help if I show the .bashrc file as well?

[white saffron]

Ahh I just found the issue regarding the problem I had with pyenv

export PATH="$PYENV_ROOT/bin:$PATH" should have been export PATH="$PYENV_ROOT/shims:$PATH"

[formal schooner]

good catch, that is indeed the right value

i also tend to get a lot of use out of the where builtin in Zsh, which shows all possibilities for a command, not just the first

it can help a lot when diagnosing path issues

/Users/srl/.local/opt/pyenv/shims/python
/Users/srl/.local/opt/miniconda3/bin/python
/usr/bin/python
/opt/local/bin/python
/opt/pkg/bin/python

i currently have five python's in my path and have zero issues

but that's because i've set things up somewhat carefully

in fact i'm missing one that should probably be there

/Users/srl/.local/opt/pyenv/shims/python3
/Users/srl/.local/opt/miniconda3/bin/python3
/usr/local/bin/python3
/opt/local/bin/python3
/opt/pkg/bin/python3

here, 5 non-"system" python 3 installations

[white saffron]

Oh, very interesting. About deleting the installations of python directly from compiling the source, should I be worried about leaving a mess there?

[formal schooner]

it would bother the hell out of me, yes

it probably won't cause harm if it's somewhere out-of-the-way like /opt, but it's probably safer not to have it

[white saffron]

Yeah, I got pyenv to work well and probably won't have issues. But anyways, I installed both 3.9.6 and 3.9.7 directly from the source using the commands from this tutorial. However I tend to make a bunch of fatal errors when deleting stuff like this, so I really want to make sure I delete them the right way. https://tecadmin.net/how-to-install-python-3-9-on-ubuntu-18-04/

[lavish storm]

is there a reason why you're choosing to build from source rather than through apt with PPA?

[formal schooner]

sounds like they wanted/needed a specific version and thought that pyenv didn't work for them, even though it appears to work fine now

it's hard to tell where make altinstall will install things to

@white saffron you said it was somewhere in /opt? try find /opt -iname '*python*' to at least get a list of files that could be python related

then we can start working backwards to surgically remove the things we don't want

[lavish storm]

ah i see 👍

[formal schooner]

(this is why installing things to non-isolated directories is dangerous)

[white saffron]

/opt only has two folders. Both from the altinstall. (Python-3.9.7 and Python-3.9.8)

[formal schooner]

then you can probably rm -rf them both

but double check that they aren't referenced in PATH or any installed scripts

[white saffron]

so just check the ~/.bashrc?

I know /opt didn't have any directories before I installed those two so I'm just going to go ahead and delete them.

What else do I have to remove though?

Oh this is odd.

[formal schooner]

first of all, open a new shell

who knows what's in this one

second, where and how did you install pyenv?

[white saffron]

This is a new shell.

I installed pyenv by cloning

git clone https://github.com/pyenv/pyenv.git ~/.pyenv

This is the end of my ~/.bashrc file

[formal schooner]

ok, good

what is $PATH right now?

[white saffron]

Also, I really appreciate all the time you've taken to help me out. Even if things don't end up working out.

Ok this is very odd.

[white saffron]

Pyenv automatically starts when I use ```sh
export PATH="$PYENV_ROOT/bin:$PATH"

in the `~/.bashrc`
However, `pyenv global` doesn't make any changes to the system python when using "bin" instead of "shims" in the bashrc.
If I change the path from "bin" to "shims", then pyenv does not start on reboot.

[formal schooner]

change shims back to bin, you're probably missing the pyenv script itself

[white saffron]

I changed it back. and rebooted the shell

[formal schooner]

export PATH="$PYENV_ROOT/bin:$PATH"
if command -v pyenv >/dev/null 2>&1; then
  eval "$(pyenv init -)"
fi

[white saffron]

Yeah that's what I have now

[formal schooner]

that should work. what does pyenv versions say?

[white saffron]

This is the same config/issues I was having just before I gave up on pyenv to install from the source

[formal schooner]

perl \
  -e 'print "PATH:\n";' \
  -e 'foreach (split(/:/, $ENV{"PATH"})) { print "\t$_\n"; }'

what does this print?

[formal schooner]

and what does pyenv which python show?

[lavish storm]

root

[formal schooner]

wait you're not currently root are you?

[lavish storm]

they are

[formal schooner]

don't do that

holy

no

[lavish storm]

every screenshot has been in root

[white saffron]

I am in root

I never even configured any other users, except the postgres user

[lavish storm]

https://tenor.com/view/do-you-know-how-much-power-it-is-the-encore-do-you-know-how-powerful-it-is-do-you-know-the-impact-it-has-you-know-how-mighty-it-is-gif-22365591

[white saffron]

Yes, my entire vps host is a complete mess

[formal schooner]

why do you have /root/.pyenv in there twice?

[white saffron]

oh my! I didn't even realize. I suppose I must've installed it awhile ago and reinstalled.

Oh I messed up the shell script or something

I ran exec "$SHELL" again, and your shell script, and now I have three /root/.pyenv/bin

[formal schooner]

stop execing, they all inherit env vars

[lavish storm]

have you quit being root

[white saffron]

No I have no other users

[white saffron]

Oh I didn't know. I restarted the shell, and here's the new output.

[lavish storm]

https://unix.stackexchange.com/a/52269

[formal schooner]

that looks good. but what happened to your bashrc setting up pyenv?

[white saffron]

Same as before, pyenv seems to function normally, but pyenv global does not alter the system python version.

[formal schooner]

it shouldn't

pyenv global changes what pyenv uses by default

you have to make sure that pyenv's shim is in PATH before the system python

so in that output with the perl incantation i sent you, the pyenv shim needs to be higher in the list

[white saffron]

I see. How would I do that? I need to keep the bin path it seems. I suppose the whole issue is a result of me updating alternatives a few times before.

[olive ice]

Why I talk to myself

[formal schooner]

Yes, i said to keep it

Look at your bashrc script

You need pyenv itself

When you eval that pyenv init thing, pyenv will set up the shim path for you

[formal schooner]

You might want to re-ask your question in its entirety, and maybe also be more specific/detailed when asking. If nobody is responding, it's because nobody knows how to help you

[white saffron]

Sorry I don't follow you. It runs the pyenv init command though.. This is what's currently at the end of my .bashrc file #unix message

[formal schooner]

but command -v pyenv will return non-zero status if pyenv itself isn't visible in PATH, so the stuff inside the if will never run

and if you don't have$PYENV_ROOT/bin in PATH, then indeed pyenv won't be visible in PATH

[white saffron]

Oh I see! So should I remove the if statement?

[white saffron]

This is what I don't understand. I cleared everything related to pyenv from the .bashrc, and wrote

export PYENV_ROOT="$HOME/.pyenv"
export PATH="$PYENV_ROOT/bin:$PATH"
if command -v pyenv 1>/dev/null 2>&1; then
 eval "$(pyenv init -)"
fi
``` If I run `exec "$SHELL"` after editing, pyenv command work. (except pyenv global)
But on a new shell, pyenv raises `command not found`

[white saffron]

That's my issue. I removed all the stuff from the .bashrc and put it in .profile. I figured that was better because it goes to .profile before the .bashrc file. Now, the only thing that doesn't work is pyenv global. Even though pyenv init - is getting run

[formal schooner]

you also have a .bash_profile right? but it's just set to source .profile?

[white saffron]

No I have no .bash_profile

[formal schooner]

ok

.profile is only run in a login shell

.bashrc is only run on a non-login interactive shell

[white saffron]

Oh.. so it should be in both?

[formal schooner]

pyenv is meant to work more or less inside an interactive shell session, so you should put it in .bashrc

what you can do is source .bashrc from within .bash_profile, but put a check at the top of .bashrc to not run if the shell isn't interactive

basically circumventing bash's strange behavior

[white saffron]

create a bash_profile? my current .profile already sources .bashrc by default

[formal schooner]

ok good

that's fine

.profile will be used if .bash_profile doesn't exist

to make things simple, do all of the pyenv stuff in ~/.bashrc

[white saffron]

Ok let me change that. Give me one second.

[formal schooner]

that will mean pyenv is only "active" in interactive shells, but that's okay because pyenv is mostly meant for interactive shell work

~/.profile:

if [ -r ~/.bashrc ]
then
  . ~/.bashrc
fi

~/.bashrc:

# Don't run in non-interactive shells (in case it was invoked from a non-interactive shell)
case $- in
  *i*) ;;
  *) return;;
esac

export PYENV_ROOT="$HOME/.pyenv"
export PATH="$PYENV_ROOT/bin:$PATH"
eval "$(pyenv init -)"
eval "$(pyenv virtualenv-init -)"

[white saffron]

Yup. That was there by default

This is my ~/.bashrc now

And this is ~/.profile

[formal schooner]

in the future it's a lot easier to help if you post text and not screenshots

!code

[shy yoke]

Here's how to format Python code on Discord:

```py
print('Hello world!')
```

These are backticks, not quotes. Check this out if you can't find the backtick key.

[formal schooner]

```bash
echo 123
```

[white saffron]

👍

[formal schooner]

yes, that check for BASH is good too

those files look correct as-is, but i recommend using $PYENV_ROOT and not hard-coding $HOME/.pyenv

[white saffron]

Ok. just add that variable?

[formal schooner]

like i did in my example

[white saffron]

Got it.

[formal schooner]

for example i use a totally different location

export LOCAL_OPT="$HOME/.local/opt"

export PYENV_ROOT="$LOCAL_OPT/pyenv"
export PATH="$PYENV_ROOT/bin"

(my actual scripts are more complicated, and i use zsh instead of bash because it's a lot saner)

[white saffron]

I see. I suppose I should install zsh eventually.

[formal schooner]

and i suppose i should write a bash->zsh migration guide eventually

i think a lot of people open zsh for the first time and see the scary "new user" prompt and give up

[white saffron]

I use zsh on my local computer, decided to not install bash on my host because I thought it wasn't necessary

About the pyenv global command, am I missing some pyenv file? That runs when pyenv init - is called? It seems like the shims path is never set up.

[formal schooner]

show me the full contents if your bashrc file (but please redact any API keys and such)

and show me that perl script output again, from clean interactive shell

(i.e. not with exec, either make a new window in tmux or log out and log back in over ssh)

[white saffron]

Here it is. https://gist.github.com/Snowbot946/0c775678161a4a7d073866b100cd4171

[formal schooner]

[ -z "$PS1" ] && return

why not do the case $- idiom? this seems sketchy e.g. if you set PS1=

(most people wouldn't do that, but still)

[white saffron]

I have no idea, these are just there by default.

[formal schooner]

this is ubuntu?

interesting decision

[white saffron]

yes. 18.04

[formal schooner]

did you actually install pyenv to ~/.pyenv?

or somewhere else?

[white saffron]

~/.pyenv

PATH:
    /root/.pyenv/plugins/pyenv-virtualenv/shims
    /root/.pyenv/bin
    /usr/local/sbin
    /usr/local/bin
    /usr/sbin
    /usr/bin
    /sbin
    /bin
    /usr/games
    /usr/local/games
    /snap/bin

this is the result of your perl script after opening a new shell.

[formal schooner]

that looks correct to me

what does pyenv global say now?

and pyenv versions?

[white saffron]

pyenv versions:
  system
* 3.9.6 (set by /root/.pyenv/version)
  3.9.6/envs/.venv
  .venv
pyenv global;
  3.9.6
python -V:
   Python 2.7.17

[formal schooner]

interesting

it looks like you're missing the pyenv shims in PATH

i wonder why that is

what happens if you run eval "$(pyenv init -)" in the shell now? does the path output change?

[white saffron]

No. same as before

[formal schooner]

that's weird

this is in a brand new shell? logged out over ssh and logged back in?

[white saffron]

yes

[formal schooner]

do you get any weird output when starting a shell?

nothing that looks like a warning or error from pyenv?

[white saffron]

no

[formal schooner]

i find it strange that pyenv-virtualenv is working but not pyenv

oh

they changed pyenv recently

[white saffron]

Pyenv seems to work, it just doesn't alter the path

[formal schooner]

you need to manually add the shims to path

ugh

i forgot about this

so you need to eval both pyenv init --path and pyenv init -

[white saffron]

after pyenv init? or before?

[formal schooner]

~/.profile

export PYENV_ROOT="$HOME/.pyenv"
export PATH="$PYENV_ROOT/bin:$PATH"

if [ -n "$BASH" ]
then
  if [ -r ~/.bashrc ]
  then
    . ~/.bashrc
  fi
fi

eval "$(pyenv init --path)"

that's what i gathered from their docs https://github.com/pyenv/pyenv/

they don't give a complete example, annoyingly

[white saffron]

interesting. Should I just add the init --path to the bashrc where the rest of the script is?

[formal schooner]

and ~/.bashrc has

eval "$(pyenv init -)"
eval "$(pyenv virtualenv-init -)"

[formal schooner]

yeah, if you're doing it all in the same place that should be OK

but i would do it in the order they recommend, maybe it matters

my own setup has diverged from this slightly

[white saffron]

Ok so I'll add

eval "$(pyenv init --path)"
``` before the other two

[formal schooner]

it seems like you're supposed to add it after the other two

[white saffron]

Ohh i see it runs bashrc before

[formal schooner]

see the ~/.profile example?

[white saffron]

I get it

yes

[formal schooner]

i don't know if it matters, it probably doesn't

but that is what they recommend so it's what i'll recommend to you

i will have to propose a PR with complete examples of init files as per their recommendations

[white saffron]

Oh my goodness. It worked.

Thank you so much. I can't believe you wasted all this time on me!

[formal schooner]

i guess this is a complete example:

~/.profile:

export PYENV_ROOT="$HOME/.pyenv"
export PATH="$PYENV_ROOT/bin:$PATH"

if [ -n "$BASH" ]
then
  if [ -r ~/.bashrc ]
  then
    . ~/.bashrc
  fi
fi

if command -v pyenv 1>/dev/null 2>&1; then
  eval "$(pyenv init --path)"
fi

~/.bashrc:

case $- in
  *i*) ;;
  *) return ;;
esac

if command -v pyenv 1>/dev/null 2>&1; then
  eval "$(pyenv init -)"
  if command -v pyenv-virtualenv 1>/dev/null 2>&1; then
    eval "$(pyenv virtualenv-init -)"
  fi
fi

[white saffron]

Yes. I have everything in .bashrc though. Like you said, pyenv should be only for interactive prompts correct?

[formal schooner]

meh, if you're only logging in with SSH, then you should always be starting with a login shell, so you can do this setup

that way you get pyenv properly set up in your PATH

[white saffron]

Ok. So I'll most the path export and the pyenv init --path command back to .profile

[formal schooner]

the pyenv init - part is just to enable the pyenv shell and some venv-related things that need access to the current shell

yes, but make sure to do things in the order they specify

[white saffron]

👍

[formal schooner]

it probably doesn't matter, but do it anyway

[white saffron]

I will. Oh lastly, do you know any resources I can use to learn how to go about undoing a python installation from compiling the source code?

[formal schooner]

there probably won't be any resources on it

like i said, go look at the makefile to see what got installed where

try make -Bn altinstall to see what it would run, without executing anything

[lavish crest]

I am trying to setup this ssg https://rgz.ee/ssg.html , but I can't figure out this sentence:

Make sure ssg6 and lowdown or Markdown.pl are in your $PATH:

I am on mac so i tried adding "/bin/ssg6", "/bin/lowdown" in etc/paths.. Can anyone help?

[formal schooner]

it's telling you to add the directory containing those executables to PATH

so if ssg6 is in $HOME/bin, then you want to add $HOME/bin to PATH

[nimble coyote]

Hello i have unfortunetly deleted all python file on my pc in fedora, and now i got this when i try to run a command on the terminal : Could not find platform independent libraries <prefix> Could not find platform dependent libraries <exec_prefix> Consider setting $PYTHONHOME to <prefix>[:<exec_prefix>] Fatal Python error: initfsencoding: Unable to get the locale encoding ModuleNotFoundError: No module named 'encodings'

i'v tried to install python again but no sucess
when i'v run the 'make test' to install python i'v go all test done but the last test_socket have been running since 1hour and still not finished:/

and i got this when i did the make install
i cannot work since i got this probleme, someone know how to help me please?

[olive ice]

Test stuff in vms don’t go breakjaro your own computer

[lavish storm]

first off you are in root

jesus

not a good habit because if you mess something up there's no going back

[formal schooner]

Reinstall python with your package manager, and next time don't mess with system files + don't do things as root

[nimble coyote]

I cannot use it because i need python for my file manager

[formal schooner]

Use the shell for now

Log in as single-user on the linux console if you need to

[nimble coyote]

How Can i install python with shell ?

[formal schooner]

@nimble coyote use the package manager, i think on fedora it's dnf

[nimble coyote]

Yes but i need python to use it :/

It dont work anymore without python

[dark rivet]

that's why you don't touch the system python

reinstall linux?

[nimble coyote]

Yes but how Can i do it without loosing my files ?

[dark rivet]

no idea how much fedora uses python, but back them up to some external drive. if you can't do that, oh well

restore from your previous backups

[nimble coyote]

Can't i copy the /home of myy linux from my Windows ?

Because i have a dual boot

[formal schooner]

@nimble coyote log into windows, make a live USB, boot linux from the live USB, reinstall the base OS

in the future, use Pyenv to install python

[nimble coyote]

A live USB ?

[solemn cipher]

E: dpkg was interrupted, you must manually run 'sudo dpkg --configure -a' to correct the problem.
my apt package manager broke while trying to upgrade to apt 2.3.9, im getting this error "

and upon runnig the dpkg command, it gets stuck at setting up apt

can anyone fix this?

[woven burrow]

chessprogramming

[solemn cipher]

I did it

@main olive I tried kill all command too but I did a temporary fix

I downloaded the apt .deb package from debian package repository and it works fine now@main olive

@main olive thanks for assistance

Btw I ran dpkg

Yea

[formal schooner]

i usually do apt install on .deb files

[wise forge]

when I launch from same folder expect script

#!/usr/bin/expect
set timeout 5
set name [lindex $argv 0]
spawn ./easyrsa/easyrsa sign-req client $name
expect "Confirm request details:"
send "yes\r"
interact

it works, but... when I launch the same with ansible

  - name: Create cert if not exists
    ansible.builtin.shell: ./easyrsa/create_crt.sh {{ inventory_hostname }}
    args:
      chdir: files/
      creates: "pki/issued/{{ inventory_hostname }}.crt"
    delegate_to: 127.0.0.1
    become: no

it is stopped right before interact. it is not finishing itself

hmm, some sort of difference between how I launch it by hand and how ansible launches it

interact is not working from ansible i think

according to the logs

[formal schooner]

maybe you need a pty allocated?

not sure if there's a way to tell ansible to use ssh -t

[wise forge]

{
"stdout_lines": [
            "spawn ./easyrsa/easyrsa sign-req client 65.21.146.6",
            "",
            "Note: using Easy-RSA configuration from: ./easyrsa/vars",
            "",
            "",
            "You are about to sign the following certificate.",
            "Please check over the details shown below for accuracy. Note that this request",
            "has not been cryptographically verified. Please be sure it came from a trusted",
            "source or that you have verified the request checksum with the sender.",
            "",
            "Request subject, to be signed as a client certificate for 3650 days:",
            "",
            "subject=",
            "    commonName                = 65.21.146.6",
            "",
            "",
            "Type the word 'yes' to continue, or any other input to abort.",
            "  Confirm request details: yes"
        ]
}

right after yes, it supposed to finish it. but it is not

[wise forge]

pty allocated?

hmm, all right. so I will see arguments to the module

I think there it should be achieved easer

or I need to replace interact command with smth else

[formal schooner]

basically telling ssh that it has an interactive terminal available

[wise forge]

oh, all right. there should be argument for that

[wise forge]

solved 😉
replaced interact with expect eof

then it did not require any additional privileges

[formal schooner]

good to know

[plush vigil]

can someone help me with vscode and python interpreter? I'm trying to locate python 3.9.5 interpreter so I can use it in vscode. i know i have it because whenever I run python 3 in my terminal, that's the version I get. In vscode though, i can only find 3.8.8

I'm rather new to Linux, so this is proving more difficult than I thought

[opaque ginkgo]

which python3

will print the path to py 3.9

which you can then use in vscode