unix

[upper prairie]

The requirements for WSLg are that you be on Windows 11

[olive jungle]

i wonder if i could use xrdp or something as a workaround

[upper prairie]

it used to be available to Windows 10 insiders, but not anymore

[somber ermine]

does anyone know how to do this?

[olive jungle]

you probably don't know how to fix this but for some reason bet works inside of a WSL 2 terminal but it wont' work inside of an xrdp session

[upper prairie]

what does where bet output from WSL 2?

[olive jungle]

😁 I don't have where

haha

[upper prairie]

Why don't you have where? O.o it's a bash built-in I thought

[olive jungle]

¯_(ツ)_/¯

[upper prairie]

You really don't have where?

[olive jungle]

@upper prairie

[upper prairie]

what's your shell?

echo $SHELL

[upper prairie]

how weird

but if you look at $PATH on your xrdp session, you probably won't see /usr/local/fsl/bin/ in there

also if you're using zsh, then editing .bashrc won't do much for you

I don't think so, anyways

you'd use .zshrc

[olive jungle]

I thought about amending .zshrc and I may still try that to get my prior problem solved - i mean attempt to

[upper prairie]

well

try using zsh in xrdp

[olive jungle]

but currently i'm more confused why the terminal emulator in the locally hosted xrdp session is different

[upper prairie]

see if bet works

[olive jungle]

bet works in zsh and bash inside of a wsl2 window inside of windows 10, however where it don't work is inside the xrdp session

[upper prairie]

no clue, then, honestly

[olive jungle]

what a funny little thing

just for the sake of trying I'll add those lines to my zshrc and see what happens 🙃

[olive jungle]

@upper prairie I got somewhere

https://tenor.com/view/anime-tea-drink-cute-bangs-gif-15778933

[upper prairie]

I've only ever had a problem like that when messing with NVIDIA drivers

[olive jungle]

new error is good error

time to investigate 🧐

so I see the gears

but they're frozen T-T

[olive jungle]

in order to get fsleyes to run (it still crashes) I have to run this export . . . command first.

[upper prairie]

Is it not in your zshrc?

[olive jungle]

even after adding it, the behavior is the same

[upper prairie]

Well the GL error is going to be more involved

if I had to guess, maybe it's something driver related

[olive jungle]

yeah that has to be correct because fsl starts the application but once the graphical load gets up there something wonky happens and it produces that GLXBadContext error or GL error. maybe in the process of learning what fsleyes depends on I can get the right nvidia driver...

@upper prairie great news, it wasn't driver related.

[upper prairie]

What was it?

[olive jungle]

@upper prairie I was starting xterm with a .bat script that was crap. if i did it manually it worked out okay.

[olive jungle]

tbh I doubt my troubles are over considering glxgears still won't show turning gears

[upper prairie]

Sounds painful

[lofty sorrel]

any ubuntu user can tell me why this alias doesnt work, but when i run the command directly it works? the other aliases in bashrc work fine

[lofty sorrel]

its because i forgot the quotes around the alias. duh

[sleek remnant]

It’s only needed because of the space. So the other alias for python3 is fine

Also, you can install a package called python-is-python3 to make that idea system wide

[opaque ginkgo]

why do you have an alias to sudo python

unless you're binding to a port <1000 you don't need root for a web serve

but idk why you would run the dev server if so

[lofty sorrel]

i only aliased python to python3, and i am binding to port 80 :P

[trail sapphire]

for development purposes (as you are using the flask development server anyways) you could just as well bind to 8080 or any other port that is > 1023, which is outside of the reserved range
it's also more secure to not run your server using superuser privileges

[lofty sorrel]

ah yea, i just didnt want to change my AWS ports because its a hassle for me as i cant do it myself

[silent hare]

Im using the recommended propriatary driver v515.48.07

[upper prairie]

They're using 3050 Ti Mobile

The latest version they can use is 515.57

[feral mirage]

#help-potato

[trail sapphire]

i was just responding to you in #help-potato when it timed out
it wasn't clear if you required it to work directly in a python script without calling for a shell command
but i just verified that xclip works with images, the key is to both set -t to the correct mime type and set -selection clipboard
at least all of these variations works independently of each other after setting the bash variable FILENAME to what ever filename you would like to operate on:

MIMETYPE=$(file -b --mime-type "$FILENAME")
xclip -t "$MIMETYPE" -selection clipboard -i "$FILENAME"

MIMETYPE=$(file -b --mime-type - < "$FILENAME")
xclip -t "$MIMETYPE" -selection clipboard -i < "$FILENAME"

MIMETYPE=$(cat "$FILENAME" | file -b --mime-type -)
cat "$FILENAME" | xclip -t "$MIMETYPE" -selection clipboard -i
```but that last one is not recommended as using pipes creates sub shells that redirection doesn't
`$()` used to find the mime type of the file with the `file` command and assigning it to `MIMETYPE` does create one sub shell for a short while
but at least the other two solutions doesn't require two more sub shells each for the two pipes that is present in that last one

i verified by pasting the image into a blank gimp canvas using either `ctrl-v` or the `edit` > `paste` menu option in gimp

[feral mirage]

I needed it to work directly yes

[feral mirage]

As a semi-newbie to Linux, the usages you gave were kinda hard to understand, I think I got it, but just to be sure, could you give an example?

Sorry in advance if I'm causing trouble and thanks for the help

[trail sapphire]

no worries, i just don't have a python solution for you yet
but a bash shell script that might be named add-to-clipboard.sh could look something like:

#!/usr/bin/bash

FILENAME="$1"
MIMETYPE=$(file -b --mime-type "$FILENAME")
xclip -t "$MIMETYPE" -selection clipboard -i "$FILENAME"
```and after running `chmod 0755 add-to-clipboard.sh` you could run it like `./add-to-clipboard.sh screenshot.png`

or manually without the bash shell script, if you don't know the mime type of the file you could just run:
```bash
file -b --mime-type screenshot.png
```and then take the output (which in this case is `image/png`) and use that for the `-t` option of the next `xclip` command, like:
```bash
xclip -t image/png -selection clipboard -i screenshot.png
```now i can paste the image in to gimp

[feral mirage]

Hmmm, alright, will take a look

My main problem tho, is that I've already tried solutions that say they work, but in the end they do nothing

[trail sapphire]

i've verified this on my linux system using the command line and gimp less then two hours ago

[feral mirage]

Did the python gtk extension work for you?

[trail sapphire]

i haven't gotten that far yet, the above was just a non python solution

[feral mirage]

Oh ok

[trail sapphire]

so, that was just on the command line in a bash shell

[trail sapphire]

somethings strange is going on
if i take an image from the clipboard
store it in my program in a pixbuf
clear the clipboard with my normal clipboard manager
verify in gimp that i can't paste anything
with my program put the pixbuf back on the clipboard
everything works and i can paste it again in gimp

if i instead load a png file from disk and try to put it on the clipboard, nothing happens 🤔

[feral mirage]

Huh

That's weird

[trail sapphire]

yeah, that's what i think as well, can't really explain it, i'm going to dig a bit deeper

[molten wedge]

Is there a way I can pipe the stdout from one program into stdin of another AND also to stdout. Usecase is like so:

I have an application thats logs to stdout and I want to collect the logs using vector (https://vector.dev/). But when I do node application.js | vector, I can not see the logs anymore (because they are being redirected to the vector instance). Is there a way I can redirect the logs to both stdout (so that i can see them) and stdin of vector (for log aggregation)?

[shrewd stratus]

use the tee command

[molten wedge]

can you give a small example?

[shrewd stratus]

node application.js | tee | vector

wait. I might have misremembered

[molten wedge]

nope its not working

[shrewd stratus]

node application.js | tee >(vector)

[molten wedge]

wow that works

what is >() though?

i noticed it is not working on the fish shell only bash

[shrewd stratus]

() creates a subshell

() redirects stdout to a subshell

dunno if/how fish does subshells

[molten wedge]

ehh its fine

i will write a bash script anyway

thanks again for your help! @shrewd stratus

[leaden venture]

hello everyone! i just wanted to see that what happens if i delete the partition i use for linux ?
cuz i need space for my windows. and will it harm my windows?

[upper prairie]

Close to nothing, but you'll wanna get rid of GRUB first if it's installed

I honestly don't remember how you're supposed to do that, though

[vital bay]

you can use a windows recovery disk to "fix" the windows install so that it uses the windows bootloader instead of grub

[vocal wasp]

are you booting via legacy BIOS or UEFI? I'd wager it's the latter... I don't think windows allows booting via legacy BIOS these days..

But if that's the case, all you have to do is set the windows bootloader in your UEFI firmware options as the primary boot method. Bootloaders these days are just specially named files located in FAT32 partitions (no more crazy stuff in the first sector of the disk). So most UEFI firmware is able to scan your disks for FAT32 partitions and find the bootloaders there. You'd just have to tell it to use the windows one

[wheat grove]

Hello I am trying to uninstall ubuntu from my dual boot but I am having this problem

[fickle granite]

well if you want to remove the "boot" folder, you'd type "rd /s /q boot"

I don't know if that's the right folder, or if that will work (you might not have the necessary permission) but "rd ubuntu" won't work if there's no folder named "ubuntu"

[wraith gate]

you have EFI

just go to the BIOS settings and remove the Ubuntu boot menu entry

if you don't have the option to delete entries

boot from a Linux USB stick

in EFI mode

open a terminal,

determine the number of the ubuntu entry using the efibootmgr command

and then

remove it using efibootmgr -b 1111 -B (place the number of the entry instead of the 1111)

[near ravine]

I have a weird idea of piping in my mind but never read the actual term. Anyone knkws it

[fickle granite]

are you asking "what does 'piping' mean"?

https://tenor.com/byC3L.gif

[near ravine]

Pretty much

Lmfao

[fickle granite]

depends on the context

[near ravine]

💀💀💀💀

[fickle granite]

in Unix, a "pipe" is a thing that lets one process send bytes to another

[near ravine]

Yeah thx

[fickle granite]

https://www.gnu.org/software/libc/manual/html_node/Pipes-and-FIFOs.html

[near ravine]

Thx

[main olive]

it's only me or this website is broken?
as soon it opens the arguments for like less than a second, it closes it and gets back to the index
https://developer.apple.com/library/archive/documentation/OpenSource/Conceptual/ShellScripting/ForMoreInformation/ForMoreInformation.html#//apple_ref/doc/uid/TP40004268-TP40003514-SW1

does anyone has any other source, so I can take a look at how does the mac terminal work?

[leaden venture]

can any1 recommend a good vpn for ubuntu?
i tried everything but nohting works

[trail sapphire]

if it's bash you want to learn about there is other sources then apple for that, not much should be apple specific other then file system locations of stuff
that said, apple has moved to using zsh as the main shell on mac a while ago, they are very similar but if you want your shells to be portable you probably should stick with writing them for bash

[leaden venture]

and how to download it?

[trail sapphire]

what kind of vpn are you looking for?
and what is your particular use case?

[leaden venture]

im in a country that a lot of websites and apps are banned so im looking for a vpn so i can visit those websites and launch those apps

i mean i downloaded rise vpn and it works but it takes mor than 10 minutes to connect

[main olive]

ok thanks

just one last question, is zsh the same in every their machines and macos versions?

[trail sapphire]

i was asking since nowadays vpns are mostly used for anonymity or being able to reach sites or content that is geo blocked or something similar as in your situation
but traditionally it has been about being able to reach the inside of a network from outside, like the home lan or remote work which has been really big ever since corona hit and the work from home ways of working

[opaque ginkgo]

almost all free VPNs are shit

except like protonvpn and some others I can't remember the name of rn

if you want paid, I hear mullavad is good

[trail sapphire]

have you tried https://1.1.1.1/?

[kind coral]

Do they offer a VPN?

[opaque ginkgo]

that only affects dns though
won't help if it's an ip block

[trail sapphire]

no, not only the dns service, they have "warp" as well which is their free vpn service

[opaque ginkgo]

oh is that an actual VPN? I'll have to try it out sometime

[trail sapphire]

yeah, they use the wireguard protocol in the warp app

[trail sapphire]

if you try it please report back if it works good for you or not

[trail sapphire]

yeah, see above

[kind coral]

Thought they were just a DNS

[fickle granite]

https://en.wikipedia.org/wiki/1.1.1.1

[trail sapphire]

that's how cloudflare started the service, but after a little while they expanded the service to include a wireguard based vpn

[tawdry pilot]

@soft dirge

[soft dirge]

yes, i though we concluded the talk, u can daily drive kali

[tawdry pilot]

u can daily drive kali

YOU CAN, you can but it will only make it worse

Too much bloat and you won't use all the tools

[soft dirge]

it did not for me

[soft dirge]

u use everything on your arch ?

[tawdry pilot]

Yes.

Xfce

Picom

Lightdm

Discord

VSCode and intellij

[soft dirge]

ok good for you ¯_(ツ)_/¯

[tawdry pilot]

I use almost everything i've installed because i know what i needed

[soft dirge]

i can say xfce is also bloat as i use dwm

[tawdry pilot]

Bloat ?

You probably mean the plugins

xfce4-goodies

[soft dirge]

naa i meant DE

[tawdry pilot]

Compared to KDE and GNOME i would say no

[soft dirge]

compared to dwm yes

[tawdry pilot]

And how is it "bloat"?

[soft dirge]

cuz its a DE

[tawdry pilot]

You know what discord is bloat because it has too many channels in it

It has only the essential stuff to run it

I mean adding new features that makes you like it more is not bloat

Bloat is something that you don't even need and no application uses it but the bloat is hogging up space

[soft dirge]

ok broo.. i was pulling your leg...

[tawdry pilot]

Makes me wonder if you have used manjaro instead of arch

[tawdry pilot]

Bruh

[soft dirge]

yes i did, i have used gentoo too 😄

[tawdry pilot]

Same had to get my laptop runnin for 3 days for that to compile

My first install took me 4 due to the errors

[soft dirge]

i am still building my LFS build, idk when that will finish

[tawdry pilot]

MAKE FLAGS did you use em?

[soft dirge]

u mean make.conf ?

[tawdry pilot]

I don't know about LFS

[soft dirge]

ahh in LFS

[tawdry pilot]

But -j4 stuff

[soft dirge]

i though u meant gentoo

[tawdry pilot]

ram / 2 = threads you can use

1 thread for every 2 GB ram

[soft dirge]

yh i use those

[tawdry pilot]

How many threads do ya use?

My guess would be atleast 16

[soft dirge]

12

[tawdry pilot]

How many total threads and how many ram ?

If my guess is right you have 24 GB ram ?? judging by the amount of threads you got

I use 3 ;-;

Got only 6 Gb ram

And 4 total threads

[soft dirge]

i have 32 but i think i only used 16 at that time, cant remember the reason

[tawdry pilot]

Well 32 / 2 = 16

I guess this is why

[deep talon]

If you want to create IMG files of a certain size try truncate instead.

truncate -s 4k ext4.img

With these sizes it doesn't really matter, but for bigger files it would make a big difference.

[wet folio]

Thanks for your reply. Actually each mkfs contains different config param, so I was thinking how can assign certain config for certain number of img files specifically, also what if I wanna modify the bs or count as well, so is there any way to put those variable at the begining of for loop to use and modoify accordingly

[deep talon]

Since the params for mkfs are so different and you only do 3 images, you should probably just not use a loop.

[wet folio]

For examples, suppose I keep dd one a single one, which will be same for those 3 differnet mkfs. how i can improve code structure of mkfs part to make it short or assign particular config param sperately

just trying to create a script to do it automatically rather doing it manually, as it take some time:)

[trail sapphire]

what tools or environment are we talking about here, python or just the shell, what's the constrains if any?

[wet folio]

Guys, my bash script does not work:( anyone would like to help me a bit to fix it, Thanks:)

[prime hawk]

hey what did you get so far?

[prime hawk]

no, they upgrade its version here and there with macOS versions

[shy yoke]

:incoming_envelope: :ok_hand: applied mute to @prime hawk until <t:1657618102:f> (9 minutes and 59 seconds) (reason: duplicates rule: sent 4 duplicated messages in 10s).

[thorn shale]

xd

[wet folio]

Thank you for your response. I did 2 small bash/shell script but none of them working when run it. Would you like to see if I wrote it wrongly or something need to add in proper way:)

[trail sapphire]

one more note about zsh on mac, it was set as the default shell for all new user accounts created with macos catalina (released during the autumn 2019) and onward
this is because apple will not update the bundled bash version to version 4.0 or above due to the change of licensing of bash from GPL 2.0 or above to GPL 3.0 or above, which is not a license apple seem to agree or want to comply with

[fickle granite]

I still use bash to write scripts for some reason

habit, I guess

[wet folio]

I am quite new. Trying to read online resources to get some idea before I wrote it. Manually it works but I was trying to make a small script so I can use it anytime whenever I need to do that. But it's throwing error:)

[fickle granite]

well if it's really small, post it as text (not a screenshot)

[wet folio]

Ok

[wet folio]

Gimme a moment please

[wet folio]

So, here I was trying to achieve: Use a DD command to create multiple images, and then use mkfs command to create ext4 file system.. So for 1 DD syntax will generate 4 ext4 files for this case, where on mkfs I used different configuration parameter with different properties.

But when I run it, it shows invalid block, does create ext4 images but its still throws error that mkfs invalid block. While I did it manually it created ext4 successfully, without such error message

[fickle granite]

lemme try it

@wet folio here's what I see https://paste.pythondiscord.com/manixidofe

is that what you're seeing?

[wet folio]

Yes

It Generate the .img file but it didn't do it properly, like the inavlid message. It usually print differently when I did it manually

[fickle granite]

this is your real problem ```

• mkfs -t ext4 -c -F extent,dir_nlink -b 1024 -O ext4-1.img
  mke2fs 1.45.5 (07-Jan-2020)
  The file extent,dir_nlink does not exist and no size was specified.

I think you've got the arguments to mke2fs wrong

[wet folio]

:) I see. Let me modify it a bit again then. But mkfs also create ext4, right

[fickle granite]

why are you running mkfs four times on the same file?

[wet folio]

I wanna keep same block size and count from DD command. But I wanna change it tho, but didn't able to do it:) lack of knowledge on it

[fickle granite]

no idea what you're talking about

[wet folio]

I wrote like 4 times DD command for mkfs one.

[fickle granite]

I'm not an expert but I'd imagine you'd run mkfs exactly once per file, not four times

yes, and I think that's a mistake.

[wet folio]

So I need to wrote as many of it based in mkfs. Right

[fickle granite]

I'm also having a terrible time understanding you

[wet folio]

:)

[wet folio]

Is it ok now, if I modify like this?

[fickle granite]

I dunno, try it and see. I expect it's wrong, since (as I've said three times before) I think it's wrong to run mke2fs more than once on the same file

but you do you

[wet folio]

:/

[fickle granite]

https://tenor.com/3juR.gif

[wet folio]

If I forget about DD command syntax, I actually wanna create multiple ext4 file system using mke2fs. Does this make sense if I able to explain now:)

[fickle granite]

do you want those four file systems to all live in the same file?

I'm not sure that's possible; I've certainly never heard of it being done

and I can't imagine how you'd mount any but the first

[wet folio]

Yes,

[wet folio]

Yes, mounting not possible. I didn't see anything that can help tho, I guess it need to do manually, mounting part.

[fickle granite]

er

why would you make a filesystem that you don't intend to mount?

I personally have never made a filesystem for any other reason

[wet folio]

I want to mount it, but the things I was trying to it's not working. Like mount multiple file in a single mounting point at same time.

[fickle granite]

I don't think that's possible.

if it is, I don't know how to do it

[wet folio]

I also, didn't see anything can help tho. I give up:)

[fickle granite]

are you trying to "layer" file systems, the way docker does it?

if so, maybe https://en.wikipedia.org/wiki/UnionFS will help

[wet folio]

I guess yes. looks pretty much same as what I want

[fickle granite]

🎉

disclaimer: I have no idea how that works; I just dimly recalled that it did something like what you want

[wet folio]

Sure, no problem

[prime hawk]

yea sure go ahead share them here <3

[wet folio]

Just a question, is it possible to mount multiple file system in a single mount point at same time?

[prime hawk]

I don't think so, no, even if it is possible it's not very advisable

[wet folio]

I see.

Another question about ext4 file system creation

For say you use DD command first to create a empty file. Then you use mke2fs to create ext4 file system.

So, if I use a single DD command syntax for creating multiple for say 3 or 4 ext4 file system using mke2fs, is it possible to do it with script

I wrote one small script with my understanding but it's not working actually.

[prime hawk]

@wet folio apparently you can use LVM to mount multiple fs on a single mount point

https://unix.stackexchange.com/a/32853

there's also UnionFS and mhddfs

but I've personally never tried to do this

[prime hawk]

from what you are describing it seems to me it should yea

[prime hawk]

can you share it?

[wet folio]

Sure

[trail sapphire]

didn't you ask this question to @fickle granite whom also answered that question?

[wet folio]

Yes,

Never mind, Leave it:)

[trail sapphire]

i mean, seeing the script was good, then it's easier to help you, but now it's gone again

[prime hawk]

yea I couldn't even check it

[wet folio]

sorry

let me post it again

[prime hawk]

@wet folio if you're worried someone may steal what you're working on, can I suggest sending versions that are just boilerplate 💗

[wet folio]

#!/bin/bash

mke2fs_cmd="mke2fs -t ext4 -c -F"
for i in $(seq 1 5)
do
    time dd if=/dev/zero of=ext4-${i}.img bs=4k count=2048

    ${mke2fs_cmd} "extent,dir_nlink" -b 1024 -O ext4-${i}.img
    ${mke2fs_cmd} "bigalloc,extent" -C 16364 ext4-${i}.img 
    ${mke2fs_cmd} "flex_bg" -G 16 -O ext4-${i}.img
    ${mke2fs_cmd} -C 16K -i 16 ext4-${i}.img 
    ${mke2fs_cmd} -I 128 -r 0 ext4-${i}.img
done

I was trying to use DD command to create image file. Then I tried create multiple ext4 using mke2fs command while I try to use different configuration state to see the information of a ext4 file

manually can do it, but I wanna write a script to do it automatically for me. But I am not able to do it so far:)

[prime hawk]

the manual ones work but not the script?

the script was giving some errs

[wet folio]

yes

[trail sapphire]

you are creating 5 different files with dd but you are trying to create 5 different filesystems that will just replace each other on every one of those files

isn't the intention to create the different filesystems one in each of the files you create with dd?

[wet folio]

yes

[wet folio]

thing is i am trying to use different configuration parameter for each file system, not keeping the same. So, I am not sure what better way I can write it except the one I write with my understanding:) although its not correct i guess

Keeping Same DD to create file, but changing different fs by adding different parameters for each.

[trail sapphire]

then something like:

#!/bin/bash

mke2fs_cmd="mke2fs -t ext4 -F"
for i in $(seq 1 5)

    time dd if=/dev/zero of=ext4-${i}.img bs=4k count=2048
done

${mke2fs_cmd} -O "extent,dir_nlink" -b 1024 ext4-1.img
${mke2fs_cmd} -O "bigalloc,extent" -C 16364 ext4-2.img
${mke2fs_cmd} -O "flex_bg" -G 16 ext4-3.img
${mke2fs_cmd} -C 16K -i 16 ext4-4.img
${mke2fs_cmd} -I 128 -r 0 ext4-5.img

```might work better

[wet folio]

I can not find any better way to represent it within my search online

[wet folio]

Thank you let me try it

[trail sapphire]

if you run the script by running it like bash -x script_name_here.sh you can see all the full commands as they are being executed by the script

[wet folio]

ok, let me run it like this way

[prime hawk]

yea -x will print each shell command that will be run by the shell

[wet folio]

it gives me such error, i give file name like filesys.sh

+ mke2fs_cmd='mke2fs -t ext4 -F'
filesys.sh: line 7: syntax error near unexpected token `time'
filesys.sh: line 7: `    time dd if=/dev/zero of=ext4-${i}.img bs=4k count=2048'

[prime hawk]

where did you write the time command

[wet folio]

$ bash -x filesys.sh

-rwxrwxr-x permission level

I typed chmod +x to make it executable

I guess i did wrong

[dapper musk]

do not use dd to create files with given sizes use truncate it will be way faster + it wont generate a lot of unnesecary writes to disk. dd is only needed for creating swap disks

[wet folio]

I got this kind of result

[wet folio]

you mean replace dd with truncate command?

[trail sapphire]

@wet folio i see what the errors are from, it was an extra -O in the wrong place just before the filename, i've edited the code i posted that i also link to here as a reply

should take care of those errors you saw

[dapper musk]

yes truncate -s filesize yourFile

[wet folio]

`mke2fs -t ext4 -F -C 16K -i 16 ext4-4.img
mke2fs: invalid inode ratio 16 (min 1024/max 67108864)

• mke2fs -t ext4 -F -I 128 -r 0 ext4-5.img
  mke2fs 1.44.1 (24-Mar-2018)`

one shows such things

[wet folio]

like on dd I use bs and count, can I use such parameter on it or only filesize, no need count param

[dapper musk]

bs is unessecary with turncate since it infacts does not copy anything

[wet folio]

For example truncate -s 2k ext4-1.img ext4-2.img ext4-3.img

[wet folio]

is it a correct way to write it

[dapper musk]

yes

also if you have program that does not accept multiple arguments use xargs tool

[wet folio]

never used it:) i can take a look but i was thinking to script to work

[trail sapphire]

yeah, saw that too now, but that's the parameters you gave, i don't know what you want instead

[wet folio]

may be i used those param wrong way

Based on this ext4 i was thinking to do the mounting of multiple files in a single mount point

[trail sapphire]

i even says that you try to set inode ratio (the -i option) to 16 but it says in the error that minimum is 1024 and maximum is 67108864

[wet folio]

at same time*

[wet folio]

I see, I need to set it correctly then

[trail sapphire]

do you need to mount all of the images at the same time or are you processing them one after another?

[wet folio]

I actually have a ext4 default image fs which contains many files and folders once I mount it, I can see those tree structure

[wet folio]

So the 5 new image I created, I want to store all the files and folders of that default image at once, and then wanna unmount it

Usually I use for single file, but for multiple file i can not do it

[trail sapphire]

or you can process them in a loop one after the other

[wet folio]

$sudo mount default.image ex

$sudo mount ext4-1.image ext

$sudo cp ex/foo ext -a

$sudo umount ext

[wet folio]

Like I did it for one file like this manually one by one.

On second line I wanna add multiple files that I create, like shared mount point for all of it

I wrote such way:

#!/bin/bash

set -x
mkdir ex
mkdir ext
sudo mount -t ext4 default.image ex

for f in ext4-{1, 2, 3, 4, 5}.img
    sudo mount -t ext4 $f ext
    sudo cp ex/foo ext
    sudo umount ext
done

default image contain all files and dir, and I just wanna copy its all files and folders to new multiple 5 images mount point, so it can help me to keep same structure

[trail sapphire]

or do:

#!/bin/bash

set -x
mkdir -p ex
mkdir -p ext
sudo mount -t ext4 default.image ex

for i in $(seq 1 5)
do
    sudo mount -t ext4 "ext4-${i}.img" ext
    sudo cp -a ex/foo ext
    sudo umount ext
done

sudo umount ex
```and you could put that at the end of the other script instead of making a new script if you just omit the shebang line

[wet folio]

I was making two separate one, silly me:)

#!/bin/bash

mke2fs_cmd="mke2fs -t ext4 -F"
for i in $(seq 1 5)
do
    time dd if=/dev/zero of=ext4-${i}.img bs=4k count=2048
done

${mke2fs_cmd} -O "extent,dir_nlink" -b 1024 ext4-1.img
${mke2fs_cmd} -O "bigalloc,extent" -C 16364 ext4-2.img
${mke2fs_cmd} -O "flex_bg" -G 16 ext4-3.img
${mke2fs_cmd} -C 16K -i 16 ext4-4.img
${mke2fs_cmd} -I 128 -r 0 ext4-5.img


set -x
mkdir -p ex
mkdir -p ext
sudo mount -t ext4 default.image ex

for i in $(seq 1 5)
do
    sudo mount -t ext4 "ext4-${i}.img" ext
    sudo cp -a ex/foo ext
    sudo umount ext
done

sudo umount ex

Like this, right

[trail sapphire]

yeah, but you have an error in the first for loop, you need a do after the for as well, just like in the last loop

[wet folio]

updated now, is that ok now

I will change i value as you suggested

[trail sapphire]

just don't know what you want there instead of 16

[wet folio]

Thank yo so much for helping me. Just some random parameter, but I did not follow their specific range size

Now I understand from your explanation

[trail sapphire]

you can either skip the set -x line or set it higher up in the script, it does the same thing as running the script like bash -x script_name.sh but then you can run it with just ./script_name.sh and still have the same effect

[wet folio]

sure, I will do that and test it.

[trail sapphire]

and the -p option for mkdir is for allowing the creation of several levels of directories at once like one/two/three without any of them existing before hand
but in your case i added it to suppress warnings about directories that already exist if the directories ex and ext already exist when you run the script the second time and so on

[wet folio]

Thank you so much for the explanation, it makes sense. I will give it a try now

[trail sapphire]

and you can make it into on command like mkdir -p ex ext to create them both with just one command

[wet folio]

then I can delete this line from script, pre created dir to mount it

[trail sapphire]

or you'll just have it in the script if you start it when the directories doesn't exist already
it doesn't hurt to have it here as insurance and as a guard against the script erroring out instead

as -p will make it just skip any directories that already exist and without complaining about it

[wet folio]

got it, it works now but but ext4-4.img shows error i guess mount: wrong fs type, bad option, bad superblock on /dev/loop27, missing codepage or helper program, or other error.

seems like syntax is not right, I will take a look at manpage and see what parameter needed, i might add wrong number

[trail sapphire]

yeah, you probably still have an error when you call mke2fs for that file

[trail sapphire]

-i 1024 works for that file, it's the smallest number you can use there

[wet folio]

sure, I guess I do not need it. but yeah I will change all i to 1024 if I use it otherwise I will leave it.

[trail sapphire]

or whatever number you want that is valid

[wet folio]

got it. kinda random pick within that range

[trail sapphire]

hmm, something else is wrong with that file

mke2fs: specifying a cluster size requires the bigalloc feature

[wet folio]

one thing, like I mount multiple filesys using the script at same time on a single mounting point, does that mean all new file system have same dir and folders that I copied from Default filesys image file

[trail sapphire]

okay, you need to set the -C to something else as well

[wet folio]

sure, I will again read man page and put the value accordingly

[trail sapphire]

i don't really understand why you would do that

[wet folio]

I am actually running those all image on a afl-fuzz tool, just keeping their original image file structure, so I copy the mounted files and folders of default image one to new file sys

[trail sapphire]

either you would mount them on different directories or do it sequentially like you do with the loop in the script where you mount and then unmount the file

the copying has already been done

do you want to run the fuzz tool on all the images in parallel or one after another?

[wet folio]

but the way I wrote works same, sequentially process may be lengthy one?

[wet folio]

one after another, last time you fixed the code a bit for me. subprocess code

[trail sapphire]

if you want to run it in parallel you should mount them on different dirs

[wet folio]

I see, what if i run one after another, do you think the process i was trying to automate wont work in such situation

[trail sapphire]

with serial work (sequentially) it would be fine to reuse the same dir for all of them

[wet folio]

got it. means all new fs mounted successfully at once rather replacing the previous mounting point, right

[trail sapphire]

yeah, like:

for i in $(seq 1 5)
do
    sudo mount -t ext4 "ext4-${i}.img" ext
    fuzzer ext
    sudo umount ext
done
```if fuzzer was the name of the fuzz tool

[wet folio]

fuzzer ext in this place we copied the file to default images mount point to our desired one which is ext

[trail sapphire]

either directly after you have copied the files in the same loop or you can make a new loop that looks like that after

[wet folio]

you mean create multiple loops?

[trail sapphire]

if you want to, or you can put it in the same loop by running the fuzzer directly after the line that copies the files

it will all just run one after another then

[wet folio]

Like what i was doing with afl-fuzz is that, once I create the new image as you did on script, and finish mount process. Next step, I was using that py code last time you fixed to take those images as input and run on fuzzer gui

[trail sapphire]

if you want to run them in parallel you would need different directories, one for each image that you want to mount and then run the fuzzer as a background job

[wet folio]

right, as I am not doing it parallely, then I guess for now it will work fine

[trail sapphire]

ah yes, i don't remember all the details of that python script

[wet folio]

if parallel, then require more complicated code may be, which is difficult for me to organize

[trail sapphire]

yeah, and it will probably just run slower if it's a io bound workload as it sound like it would be, then all the processes would be competing for the disk io all at once

if it's benchmarks it would only give you bad data that way

[wet folio]

i see, then its better one after another, easier for me

[trail sapphire]

as it would take longer to run due to the disk having to try to service all the instances at once

[wet folio]

right, I will follow the easiest step to do it rather go more complicated one then.

[near ravine]

Hello, does anyone know if folder creation leaves any hidden "fingerprints" in Unix based systems

[fickle granite]

heh

I wonder why you're asking

[near ravine]

We had a small excersise from school more like a ctf, and after we get access to the machine we need to "clear" it from proof. I am sure he doesnt man to go burn the pc itself

[fickle granite]

well it might..

depends how you do it.

[near ravine]

Maybe our teacher meant to clear the history or something

But yean if it does i want to know

[fickle granite]

e.g. if you simply log in and type "mkdir /tmp/whatever", that command will probably be saved in your shell's history file.

depends on the shell and how it's configured.

[near ravine]

Yeah. So that is the onl

Only fingerprint that i leaves

[fickle granite]

apart from that, it's possible that the system as a whole "audits" stuff, which means something might get written to some centralized audit log

[near ravine]

Yeah

[fickle granite]

er, no, I'm not saying that at all

I'm saying almost the opposite: there is an infinite number of ways the machine might be logging what you're doing; I'm only listing the ones I can think of

[near ravine]

Yeah

Alrigt thanks

I have a ssh session with the victims machine so i will just try and dig into and find what i can

[fickle granite]

basically, it's not your machine, so you don't know what software it's running.

so: that software could do absolutely anything in response to your typing.

it could send an email to the NSA.

[near ravine]

Yeah.

Lmfao

It could

[fickle granite]

and if it were doing something weird like that, it's very unlikely you'd know.

on the other hand, part of your assignment might be to assume that it's an ordinary (say) RedHat box, configured in the usual way; in that case, you can study that setup and see what it does

[near ravine]

If someone cared that much bout his machine would make sure not to output anything so yep i wouldnt know

The machine run Ubuntu

[fickle granite]

the Big Boys run "honeypots" precisely to sucker bad guys

[near ravine]

Yeah

[trail sapphire]

there will usually be leftover traces in the filesystem table even after it's deleted and modern filesystems also many times have a journal of some sort that can have old traces left

[near ravine]

Any resources about that

[trail sapphire]

lots and lots, and it's very dependent on OS and filesystem in use too, but look up documentation for specialized file deletion software (shred and scrub being just two very basic examples for gnu/linux) as well as file and filesystem recovery software (photorec/testdisk is just one such software suite that happens to be free/oss)
also remember that ssd drives and the like has wear leveling (which is totally separate from the filesystem and os in use) can also retain information after it should have been deleted from a device

you might read up on data recovery and digital forensics in general to learn more on these subjects

[fast viper]

is windows not a type of unix?

[shrewd stratus]

nope. it has a different lineage

[trail sapphire]

no, not even a little bit

[prime hawk]

sadly no

[digital remnant]

Can someone explain what the tilda and the single quote does?

I know cp means copy, so we are copying the css534/hadoop... file

But what is the use of ~, and the '

[fickle granite]

~ means "my home directory"

~css534 means "the home directory of the user whose name is css534

[digital remnant]

ohhhhh

[trail sapphire]

single quote there doesn't make sense

[digital remnant]

and then the ~ at the end means MY home directory?

[digital remnant]

hmmm

[fickle granite]

ya

[digital remnant]

should I remove it?

The single quote was given for us to "type"

like we have to copy and paste it

but imma try to understand what the symbols mean

should I try doing it without the single quote?

😭

[trail sapphire]

yeah

it probably shouldn't have been copied

[knotty igloo]

that's an actual single quote, too - the keyboard key produces ' , and the screenshot has a different version, an actual apostrophe. methinks that was prepared in ms word or such, as actual apostrophes are usually invalid in terminals

[digital remnant]

okay I deleted it, and now there is the file in my home directory, so that is good!

[digital remnant]

yahhhh it was prepared in ms word and then I copy pasted it

All right all right, I have another question

So these were the three instructions given:

So far, we have solved the first one... but the second one gives me an "error"

is it because it is -d, should it instead by ~d? (tilda d)

Because I noticed step 1 has tilda, but not step two, and I assume step two wants to gzip the hadoop file into the home directory?

and to do that, wouldn't you need a tilda?

[trail sapphire]

unnecessary steps

you could just do tar xzvf hadoop_0.20.2.tar.gz which merges 2 and 3

[digital remnant]

All right - could you explain to me what that is doing? I'm tryna understand, conceptually, what is happening

so what does the "tar xzvf" mean?

why are we able to combine it?

(also what does the (-d) mean)

sorry for the many questions!

[fickle granite]

tar zxvf is basically unzip

[trail sapphire]

tar is a archive format, short for "Tape ARchive", but it's general purpose
x is for eXtract, z is for gZip, v for Verbose and f for file and then you give it the filename

it's what is commonly used in linux, bsd and other unix and gnu systems instead of zip, even if you can handle zip as well on linux

[digital remnant]

ohhhh gotchya

thanks!

[digital remnant]

What does this mean? Does the "export" essentially say, Run:[this]?
and what about the dollar sign?

[fickle granite]

"export" means: here comes a variable assignment; please put that variable into "the environment" (as opposed to just this shell) so that child processes can see it

the dollar sign means "here comes a variable name; please give me the value of that variable"

[fast viper]

are unix systems better for programming?

[mortal knoll]

There’s a certain amount Linux can do that’s better than windows, and some things windows can do better than Linux.

Things like c# would be best to do on windows because of .net stuff and whatever other cool things there’s to do. Many libs are supported mainly on windows but there’s a chance you can throw something together and it still work running it through wine on Linux

[summer surge]

hi, is there way to exclude some modules that can be used by unix users in shared environment ?

[fickle granite]

If you aren't doing GUIs, then yes, Unix is far better. cloud computing is probably 98% unix

[fickle granite]

Is there some reason you don't want to use the built-in TimeMachine app?

Not certain that'll back up to a network drive, but still, portable SSD drives are pretty cheap

in any case, if you don't insist on rolling your own, https://en.wikipedia.org/wiki/Network-attached_storage might be a good place to start. I've heard good things about Synology's NAS product, but tbh the one version I saw gave me the creeps

I'd expect dd to fail on all the interesting files, but then I'm a pessimist

I'm reasonably sure you can run an ssh daemon on your mac

ugh

well you could try rsync

but like I say: I'd expect that to either fail on lots of files, or to capture files in an inconsistent state, or similar

of course there's a chance it won't work. These are computers we're talking about 🤣

it's free though and you either already have it, or can get it via homebrew

it copies files

It will not copy the file system, and you probably wouldn't want it to anyway

every byte of what?

almost certainly not

but then why would you want to? Many file systems have large areas of unused space, that are filled with essentially random bytes; would you really want to copy those?

some of your hard drive isn't even part of any file system

some of it is part of some file system you're not supposed to look at, and is likely encrypted, etc etc

if you have 300gb of data to back up, then yes, you'll need to shove 300gb of data over the wire. Compression will help, but only so far.

definitely worth a look

it wins on "not bothering to copy stuff that hasn't changed since you last backed up"

[digital remnant]

I have this error:

This is the core-site.xml file:

I am really confused what the error means, and how to resolve it... anyone have any idea?

[fallow tusk]

dd will create a byte-for-byte copy of whatever you want. You can use it to image the disc (or possibly partition), and transfer that, then mount it from the image…probably more trouble than it's worth, though.

[dapper musk]

basically any *nix tool will create byte-for-byte copy of data using dd instead of cp or even cat for this purpose is just waste of time

[trail sapphire]

what do you even mean with "modules" in this context?

[steep sequoia]

I think they mean unix only modules such as termios

[summer surge]

If I have a small hosting server for few people I want to give them possibility to run python apps but without modules like os or subprocess

[summer coral]

hi, instead of removing - essential - modules, just run python in a sandbox eg WebAssembly

[trail sapphire]

i'm 100% with @summer coral on this, run it in a container or other kind of sandbox

[summer surge]

this is something new to me, I will of course read about this option
but maybe close user in some chroot/jail/cage is safe too ?

[trail sapphire]

you could go that route, are you running linux or bsd or something else?

[summer surge]

linux machines with Debian

I can switch to Centos its not problem to me

[trail sapphire]

for linux it's chroot rather then jail, but then you would need to look into cgroups, selinux, namespaces and seccomp if you want to have make it secure and properly sandboxed
or just use docker or some other container engine, i think it will be less complexity and much easier for you

[summer surge]

ok I know cgroups more than seccomp, maybe this is the way

[trail sapphire]

i would say you need all of those combined to make it safe

[summer surge]

I know docker a little bit and it's not so secure as it should be for this project 🙂

oh all of this, ok I understand

[trail sapphire]

they are for different things, so you'll combine them to get all the way

[summer surge]

sure, this is good answer I must know better seccomp but now it is only thing I haven't use in prod environment

thx @trail sapphire

[trail sapphire]

docker should also be able to be just as secure, as you would use all of those technologies (selinux, seccomp, cgroups and namespaces; that is: all of those four, not just one or a few of them) together to secure that as well
it makes docker a chroot on steroids, and BSDs jails are something much better then chroot

[summer surge]

I had bad expirience with docker and firewall, there was problem in one project to secure incoming connections

it was few years ago when docker was new project, maybe it has no place now, no idea

[trail sapphire]

i think a lot has gotten better during the years, among other things they haven't always used all those technologies mentioned above to secure things properly
and you wouldn't give the end user control of the host or docker itself, only the container and some control panel kind of portal where they have some very limited control over how docker provisions the container for them

firewalling would be mainly done either on the docker host system or with an external firewall

[summer surge]

propably I will decide on something other than docker but good to know

[digital remnant]

what does fs mean in unix?

in the context of this command:

[fickle granite]

"file system"

[digital remnant]

I know that ls means list files

[fickle granite]

although in that context, I'm not sure

I have only a dim idea what hadoop is

[digital remnant]

yeah when I searched it up I got that it means file systems but for me that is like a word, I don't understand what the word means 😦

[digital remnant]

alr alr

[digital remnant]

and do uk why they do ls /user/css534/output? That lists all of the files in the output directory correct? So are they doing something to those files?

[fickle granite]

nope, dunno

like I say: I know essentially nothing about hadoop

[high sparrow]

I think that’s to access remote file servers

You can type in fs by itself and it’ll give an explanation

[serene arch]

Hey what is kali Net-hunter

Is it an os for mobile

??

[wet folio]

Hi guys I have one question to ask, for say I have to run a program for few hours to train on linux server. So, here I first have to connect with a vpn to ssh to that server. Then on server I have to run the code for couple of hours.

So thing is, is there any way to run that code for few hours in the background without being vpn get disconnected to sign out from that ssh server. Is there any way to keep my macbook screen awake?

Even screen shuts in few minutes* like that but code still run in the background without the interruption of beging vpn get disconnected.

[tulip dock]

You don't need to keep your laptop on to have the Linux server keep running

you can use GNU screen or nohup

[wet folio]

how about the vpn too keep the ssh server keep running

vpn got disconnected if screen closed or off, and the program disconnected from being running

[tulip dock]

you don't need the SSH connection to stay on for the server to keep doing its thing

use GNU screen or nohup

[wet folio]

vpn one?

[tulip dock]

you don't need any sort of connection at all

[wet folio]

because i always vpn to ssh to the server, otherwise I can not login

[fickle granite]

or tmux; tmux is like screen, but it's actually maintained 😐

[tulip dock]

I don't care how you connect, I'm telling you that you don't need to stay connected for a program to keep running

[wet folio]

I got it, I will check:)

thank you for your the suggestion:)

[fickle granite]

@wet folio example https://asciinema.org/a/DNGyG2rjgS8zKC5LVbtwrDkMF

[wet folio]

Difficult to understand:/

[trail sapphire]

if you don't need to reconnect and continue the session you can just use nohup as @tulip dock talked about, it's the easiest option for you and you only need to type the nohup command in front of the command you would otherwise run on the server
it will create a file named nohup.out in the current folder where you run the command if that folder is writable by your user, otherwise the nohup.out file will end up in your users home directory or if you redirect your output to a file it will end up in that file instead and you can put it where ever you need to with a better name as well

if you instead need to be able to continue the shell session as it was you would need to use either screen or tmux on the server before you run your fuzzer

for screen you start a new shell with screen -S fuzzer where fuzzer is the session name and could be something else and run your script in that shell
you can detach from the screen session when ever you like using ctrl-a and then type d
regardless if you disconnect from the session or if you get disconnected from the server the shell and commands it is executing inside the screen session will just continue to run
you can list all the screen session you have using screen -ls
reattach to a screen session with screen -dr fuzzer whenever you like, as long as you have not exited the shell inside screen

for tmux you can use tmux new -s fuzzer to start a new session named fuzzer use it as you would use screen, but to detach from the session you instead use ctrl-b and then type d
to get a list of all tmux sessions you run tmux ls, and to reattach to a tmux session you run tmux attach -t fuzzer

screen and tmux are very alike for basic usage, tmux just offers more features for advanced users and is more modern

[wet folio]

Thanks for the detailed explanation. That make sense. But I have one problem with the vpn, it last 12 hours to stay login to my ssh server. So once 12 hours done, ssh server got stuck or doesn't work. In such case, does this really helps. Or it's something like I can exit from the ssh server and then disconnect the vpn and the code may run without an interruption using screen or tmux>. Is it the case.

I understand tmux and screen using technique now.

[fickle granite]

yes, this really helps.

you should try it out.

this is exactly what screen and tmux are for.

[trail sapphire]

no matter if you use nohup, screen or tmux the command will continue to run on the server even if your ssh and/or vpn connection gets disconnected
but you have to start the screen or tmux session on the server and not your computer, or else it won't work when disconnected

[wet folio]

thanks a lot, not it makes sense to me. i will follow this

[azure plinth]

What is unix

[trail sapphire]

a grandfather of a operating system to most operating systems in use today which is "unix-like" (linux, macOS and other bsd variants among others)
except the Windows and DOS families which are of separate lineages

[fallow tusk]

Even Android is a unix variant :)

[trail sapphire]

as they run linux under the hood

[fallow tusk]

Is it actually based around the same kernel? Or is the kernel itself tweaked?

[trail sapphire]

same kernel, do you mean the unmodified stock linux kernel?
more or less all linux distributions patch the stock kernel one way or the other
android is a linux distribution that runs a patched linux kernel, just like most others

[fallow tusk]

I wasn't sure how heavily modified it was. I knew it was "based on" but had made an assumption it was less like a regular distro. I thought most distros used the same kernel, just with settings tweaked to favour some use case or other.

[trail sapphire]

most distros patch their kernels pretty heavily
to the point that redhat and some others commercial distros now merge all patchen into one giant patch that they release as open-source (baked) and apply to the kernel to obfuscate exactly which patchen and in what order they apply them to throw a wrench to competitors like oracle that has based their distro upon redhats and sells support for it

[fickle granite]

that's mostly it

some other smallish differences:

• licensing
• choice of kernel

support, &c

personally I use the most popular one, because it's the most popular (ubuntu)

[prime magnet]

there are also licenses preferences. Debian is more strict than Ubuntu e.g. I recalled years ago there was something Debian didn't like about Firefox logo license and they decided to fork it and start weasel something

not meaning that you couldn't install Firefox. Just it was not included in default repositories

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

[trail sapphire]

also some of the file structure, scripts and configuration file formats

have you compared a redhat based system and a debian based system?
check the network configuration files for example
before systemd was used in any of the two there was a lot of differences among files due to the different startup scripts in the system v style init scripts

[prime magnet]

isn't that related to systemv vs systemd?

[trail sapphire]

they both run systemd now but still have the old network configuration file formats as far as i have seen, i guess it's kind of a backward compatibility thing and to not alienate it's current user-base 🤷

but sure, if you use networkmanager on both systems that will be another story, then it will be the same on both systems 🙂

yay, standardization 🎉

[flat echo]

My pip seems to be associated with python 2.7. How can I change it to one of the more recent python versions I have installed (CentOS)?```
[adankert@ppdo-dev-app-1 ~]$ pip -V
pip 8.1.2 from /usr/lib/python2.7/site-packages (python 2.7)

[ember quiver]

that might be required by the OS, I would use pip3 instead

[trail sapphire]

what @ember quiver said, or use something like pyenv (as a user, not as root) and set the local (for a directory and all its sub directories) or the global (for that user account) python version that is going to be used by default

[flat echo]

Back with a new question... Thank you @ember quiver and @trail sapphire

My new issue is that when I run gunicorn it fails saying it cannot find flask: ```ModuleNotFoundError: No module named 'flask'

[summer wasp]

It's better if you start gunicorn from inside that venv

are you creating a service that calls gunicorn ?

[trail sapphire]

a few different options, maybe you can find the answer there: https://stackoverflow.com/questions/48567245/pipenv-deployment-workflow

[flat echo]

I guess I'm not sure how to start gunicorn within the venv when using pipenv. End result is flask app using nginx and gunicorn. I am planning to use supervisor to run gunicorn. Hope that helps.

[summer wasp]

What I do is that I create a local service sock and then from nginx I call that service on whatever domain/port I want.

For the service itself, you can specify where does it start (which would be from inside the venv)

I'll try to share the code for my service

[Unit]
Description=Gunicorn instance to serve SI-CT
After=network.target

[Service]
User=si-panel
Group=www-data
WorkingDirectory=%s
Environment=\"PATH=%s/bin\"
EnvironmentFile=%s/local.env
ExecStart=%s/bin/gunicorn --workers 1 --bind unix:si-ct.sock -m 007 wsgi:app --log-level debug --error-logfile - --access-logfile - --capture-output --reload 

[Install]
WantedBy=multi-user.target

The four %s are place holders for the following variables

$localFolder $localPipVenv $localFolder $localPipVenv

[flat echo]

I'm still struggling with this. I would prefer to avoid added complications of a sock (though I sure @summer wasp solution would work.) ...but would just like to get everything installed in the correct location with executables in the correct in directories.

my gunicorn executable is in /home/<user>/.local/bin

[lapis tundra]

Guys. I'm starting with python, I don't understand much about this language, and I'm doing my first project with it: https://github.com/pandasoli/cat-code, but he's giving an error on line 47, some error with regex that I don't understanding

[flat echo]

I removed pipenv as even with @trail sapphire suggestion, that was just additional complications and used venv to finally get gunicorn to run without claiming a library is missing

...but it just crashes? I don't know where my bash terminal went:

[summer wasp]

It has to be something inside the venv, can you access the logs?

[fickle granite]

that doesn't look like a crash to me. In fact, I'm not convinced there's anything wrong.

Maybe it's just sitting there, happily waiting for a request?

[it looks like your cat walked on the keyboard, but that's something else 🤣 ]

[waxen tangle]

Hi Svengoolie, you're not in an interactive shell, so exit / exit() and any arrow keys won't work, you need to run CTRL-C to stop gunicorn

That is, if pipenv forwards signals properly Just saw you removed that 🙂

[fallen rampart]

is there a way to turn MinGW into a suitable environment for running bash commands (even the most basic like man, sudo, apt, etc.)?

i'm on windows 10 home right now and found a way to use WSL, but it can't connect to the internet for some unknown reason (i've been researching for 3 days and have come up with no solution or reason why it doesn't work). anyway, i turned to MinGW, which connects to the internet just fine by using ping but it doesn't have the basic functionality of most shells.

what packages do i need to install from the manager to make it (almost) indistinguishable from bash?

[fickle granite]

there's a "git bash"

[trail sapphire]

"git bash" that comes with "git for windows" that @fickle granite suggested is probably the easiest to get a basic environment
or maybe cygwin which is pretty easy to get going with as well and is a bit more complete (this is my go-to solution if i'm stuck on windows)
or you have msys2 which is kind of similar but might not be as easy to get working right but has three different repositories to install software from

cygwin has one large repository with most software that you'll need precompiled and compilers for most other things
it's a decent posix emulation layer, but everything for linux won't run in that environment, but hopefully enough for your needs
WSL will be a much more complete experience, or a linux VM under something like VirtualBox

[molten wedge]

is it possible to get values of environment variables dynamically?

for project in ${affected_projects}
do
    echo "$project"
done

I have this script and I get this output.

Now in each loop I want the value of ${TOKEN_<git branch>_branch}

I have env variables stored like TOKEN_main_branch=kjhkj, TOKEN_dev_branch=kjhkhjbcb etc

How can I do so?

Got it: https://stackoverflow.com/a/18124325/11667450