Counterintuitive word splitting

I've recently already made a post about word splitting, however, this seems to be another unrelated issue that I again can't seem to find any answers. Consider this setup:

$ #!/bin/bash
$ # version 5.2.26
$ IFS=" :" # space (ifs-whitespace), colon (ifs-non-whitespace)
$ A=" ::word:: " # spaces, colon, "word", colon, spaces
$ printf "'%s'\n" $A
''
''
'word'
''

As you can see, printf got 4 arguments, as opposed to 3, what I would've expected. First, I though my previous post might be related, however, adding another instance of `$A` to the end makes it 8 arguments, exactly double, so it's not related to stripping trailing "null arguments".

Why does this happen? Is there a sentence in the man page that explains this behavior (I couldn't parse it from the section about word splitting :'D)

Edit: I tested the following bourne-like shells:

bash
bash -o posix
dash
ksh
mksh
yash
yash -o posix
posh (policy-compliant ordinary shell)
pbosh (schilytools)
mrsh (by Simon Ser)

ALL of them do it exactly the same, except mrsh (it's doing what I expected). However, mrsh is quite niche and rather a hobby project by someone, so I wouldn't take that as any authority.

https://redd.it/1dnqswy
@r_bash

Question about stream redirection / file denoscriptors

**TL;DR - In bash, what is the significance of the `-` character in the following expression?: `${@}"; echo "${?}" 1>&3-;`**

Problem denoscription:

While trying to find a way to capture stderr, stdout, and return code to separate variables, I came across a solution [on this stackoverflow post](https://stackoverflow.com/questions/11027679/capture-stdout-and-stderr-into-different-variables).. I am mostly looking at the section labeled "**6. Preserving the exit status with sanitization – unbreakable (rewritten)**" which has this:

{
IFS=$'\n' read -r -d '' CAPTURED_STDOUT;
IFS=$'\n' read -r -d '' CAPTURED_STDERR;
(IFS=$'\n' read -r -d '' _ERRNO_; exit ${_ERRNO_});
} < <((printf '\0%s\0%d\0' "$(((({ some_command; echo "${?}" 1>&3-; } | tr -d '\0' 1>&4-) 4>&2- 2>&1- | tr -d '\0' 1>&4-) 3>&1- | exit "$(cat)") 4>&1-)" "${?}" 1>&2) 2>&1)

It seems to work ok. although I am making my own alterations. I've read through the post a couple times and mostly understand what's going on (short version is some trickery using redirection to different denoscriptors and reformatting output with `NUL` / `\0` so that `read` can pull it into the appropriate variables).

I get that e.g. `1>&3-; ` is redirecting from file denoscriptor 1 to file denoscriptor 3, `1>&4-` is redirecting from file denoscriptor 1 to file denoscriptor 4, and so on. But I've never seen stream redirection examples with a trailing hyphen before and I don't really understand the significance of having a `-` following `1>&3` etc. I have been hitting ddg and searx for the last 30 minutes and still coming up empty-handed.

Any idea what am I missing? Is there any functional difference between using `1>&3-; ` vs `1>&3; ` or is it just a coding style thing?

https://redd.it/1dnxehx
@r_bash

\e38;?;<hexcode>m ?

\e[38;2;<r>;<g>;<b>m sets the fg color using rgb values. I wonder whether there is a mode that uses the hex valve of a color?

[https://redd.it/1do1sov
@r_bash

Differences between (MacOS) 3.2.57 and 5.x?

Hi, folks. I'm sure this has been asked before. I've been doing searches but keep bumping up against posts about ZSH or how to upgrade with Brew.

Unfortunately, I'm in a bit of a tight spot. I have not found an answer to what I need and am hoping someone can point me in the right direction.

I wrote a BASH noscript that is fairly sophisticated. Nothing too crazy though. Lots of functions, a few run-of-the-mill commands like find, sort, uniq, awk. Keywords like 'local' and 'read.'

It works on my laptop (Windows running BASH 5.2.21 under Cygwin - I'm not allowed to run WSL) and runs perfectly on a Linux host. Idk the BASH version on the Linux side (and logging into it is a PITA which is why I'm not checking) but it's a modern Linux so probably 5.x. I handed the noscript to a coworker who ran my noscript on his MacOS laptop and found it didn't work. 🤦

Sigh. So, now I need to try to figure out what BASH feature I'm using that's not compatible with 3.x. I can't tell all my coworkers to upgrade BASH just so my noscript will work. I don't have time to make my noscript compatible with ZSH. I'm probably the only one in the dept NOT running MacOS. I'm starting to remember why 🤣😬

If anybody has ideas of where I can look for guidance on what features to avoid when making a BASH noscript work on MacOS, I'd appreciate it. Maybe 4.0 and 5.0 release notes on what features were introduced?

Is variable expansion ${} incompatible or running a subprocess with $() instead of backticks?

I wish I could share the noscript but I would be violating rules doing that.

Thanks in advance

https://redd.it/1doc89p
@r_bash

RAG in bash for MongoDB Atlas

Hi everyone, I made a small bash noscript (based on a JS noscript) that allows you do turn data/insights into AI and then query it directly in the terminal.

https://github.com/farspeak/farspeak-cli

https://preview.redd.it/pojyoheqlr8d1.png?width=1600&format=png&auto=webp&s=961b13e6f43af6c23a5fe326590968b683b2b1d5

Please let me know what you think


https://redd.it/1dodphp
@r_bash

Does anyone know of a good way to read raw hexadecimal / uint data using only bash builtins?

Im trying to figure out a way to convert integers to/from their raw hex/uint form.

Bash stores integers as ascii, meaning that each byte provides 10 numbers and N bytes of data allows you to represent numbers up to of `10^N - 1`. With hex/uint, all possible bit combinations represent integers, meaning each byte provides 256 numbers and N bytes of data allows you to represent numbers up to `256^N - 1`.

In practice, this means that (on average) it takes ~60% less space to store a given integer (since they are being stored `log(256)/log(10) = ~2.4` times more efficiently).

Ive figured out a pure-bash way to convert integers (between 0 and `2^64 - 1` to their raw hex/uint values:

shopt -s extglob
shopt -s patsub_replacement

dec2uint () {
local a b nn;
for nn in "$@"; do
printf -v a '%x' "$nn";
printf -v b '\\x%s' ${a//@([0-9a-f])@([0-9a-f])/& };
printf "$b";
done
}

We can check that this does infact work by determining the number associated with some hex string, feeding that number to `dec2uint` and piping the output to xxd (or hexdump), which should show the hex we started with

# echo $(( 16#1234567890abcdef ))
1311768467294899695

# dec2uint 1311768467294899695 | xxd
00000000: 1234 5678 90ab cdef .4Vx....

In this case, the number that usually takes 19 bytes to represent instead takes only 8 bytes.

# printf 1311768467294899695 | wc -c
19

# dec2uint 1311768467294899695 | wc -c
8

***

At any rate, Im am trying to figure out how to do the reverse operation, speciffically the functionality that is provided by xxd (or by hexdump) in the above example, efficiently using only bash builtins...If I can figure this out then it is easy to convert back to the number using printf.

Anyone know of a way to get bash to read raw hex/uint data?

https://redd.it/1dor7ss
@r_bash

Is it possible to prevent debugfs printing it's version?

Is there any way to not have debugfs printing it's version before outputting the result of the command?

This noscript always outputs "debugfs 1.44.1 (24-Mar-2018)" on the first line:

#!/bin/bash

file="/var/packages/Python3/INFO"

getcreatetime(){
# Get crtime or otime
inode=$(ls -i "$1" | awk '{print $1}')
filesys=$(df "$1" | grep '/' | awk '{print $1}')

readarray -t dbugfs < <(debugfs -R "stat <${inode}>" "$filesys")

echo "array line count: ${#dbugfs@}" # debug

for d in "${dbugfs@}"; do
echo "$d" | grep -E 'ctime|atime|mtime|crtime|otime'
done
}

getcreatetime "$file"

The noscript output:

# /volume1/noscripts/getcreatetime.sh
debugfs 1.44.1 (24-Mar-2018)
array line count: 15
ctime: 0x66348478:bc1cbfa4 -- Fri May 3 16:30:16 2024
atime: 0x6608e06d:0d3cf508 -- Sun Mar 31 15:02:53 2024
mtime: 0x65beb80c:054935ac -- Sun Feb 4 09:02:52 2024
crtime: 0x6607eb8f:2e7278fb -- Tue Jul 20 16:02:55 2432


&#x200B;

https://redd.it/1dou64x
@r_bash

Command result in terminal

Hi
I'm tryimg to use fzf inside a directory and the result should be pasted onto the command-line( not as a stdout, but should be available in the terminal)

I have something like this

#!/bin/bash


test() {
FZF_DEFAULT_OPTS_FILE='' fzf "$@" |
while read -r item; do
printf '%q ' "$item" # escape special chars
done
}

bind -m emacs-standard '"\C-t": " \C-b\C-k \C-u`test`\e\C-e\er\C-a\C-y\C-h\C-e\e \C-y\ey\C-x\C-x\C-f"'


Which is working, but i don't want to use the bind.
I want just to run the noscript from command line.

So instead of the bind i want only the call to test function.

In this case the result is simply printed to the screen.

Thank you.

https://redd.it/1douatw
@r_bash

Jesus i need help with this 4 year journey

Disclaimer: I do not have admin password, i cant enter bios, and i cant use certain process that require admin on cmd.



Background info: Norton family is a child monotoring software that i want to bypass, i have tried over, and over again, for 4 years, nothing has worked on the internet. Norton family has a range of things you can do in it, like set time limits on a user, and restrict websites (i've bypassed restricted websites with a vpn), norton can also monitor data and send it to my dad. There is also an application called norton 360 that is like an antivirus thing, and its so annoying because it will just remove applications even though its not even a virus. Time restrictions work like this: every 5 seconds there will be a pop up window saying "you have reached your time limit" or something like that and you will be logged out and have to enter your password and it repeats, so last night i was playing with task manager, and decided to end task on one of the norton family processes, and it worked! the pop up didnt happen.... But i realised it would make a new process of the one i ended and i would have to keep ending the task.



*What i need done (if you want to help)*:

Can someone make a software that will kill a proccess in my task manager called "Norton Family", the problem is:, When i do it manually i have to keep doing it every 5 seconds, so its kinda annoying.There are two proccess that are called norton family, the one i want to end the process with is the one without the drop down box (actually it might have a drop down box but we'll get to that later on). There should be a drop down box in the Norton Family process, and inside it there should be a gear icon that says "Norton Family", thats the process i dont want to end, but in the other process also called "Norton family" there should be no drop down box, thats the one i want to end, BUT i've noticed that the first time i end the Norton family application without the drop down box it will end for 5 seconds, before making a new process of itself and then i have a few seconds to end it before the pop up window.But when that new process is made again after i end the old one there is now a drop down box, and inside it theres a thing called "Norton family" (Again) that has a window icon, that thing inside the drop down box is just a window on my pc that says norton family needs a restart, i always click no, lol, so pretty harmless.

By the way, the reason i want to end a specific task of Norton family is because if i try to end the task with the drop down box that has a gear icon in it, it will just say you need admin privileges where as the other one wont say that.



If you have any questions feel free to ask, and i also have norton 360 on my laptop

https://redd.it/1dp844f
@r_bash

how do you put human format in command identify for file size?

Hi, I use the comand identify (from IamgeMagic version6, the old version built-in at Lubuntu OS).

I'd like to retrieve in Vim the output of this command with file size in Kb or Mb, like using the flag -h in ls -lh ...

the command that I use in Vim is this:

r !identify -format "\%f [\%m \%xx\%hPixels \%[size]ytes] \n" path/to/*

this comand only shows %[size]ytes like this 444323bytes

I'd like to see 444.323Mbytes

The command work well fine and I understand the command, I only need what letter should and where put it in the command.

help man identify in C L I and https://www.imagemagick.org/noscript/identify.php

Thank you so much and Regards!

https://redd.it/1dpgw7j
@r_bash

Is bash good for the task I need?

Hi everyone,
I need to parse a directory structure and store each symlink in an array as a "source destination" pair, so each entry in the array will look like this:

source-a.png dest-a.png, source-b.png dest-b.png, source-c.png dest-c.png, ... and so on 1000 more entries

Once I have this array, I need to create a second array but now instead of parsing a directory tree, read out similar source-destination pair information from a file instead. After both arrays are filled, I need to compare them to identify which elements are missing or have been added.

The idea is that if I remove a source-destination pair (source-a.png dest-a.png) from the file, the next time the noscript runs, it removes the source-destination symlink from the filesystem as well. So, the point is to keep the symlinks on the filesystem in sync with the information from the file. That's why I need two arrays.

Given that there might be 1000 entries in both arrays, is it practical to do this with Bash, or should I consider implementing it in Go, which is a language I'm more familiar with than Python? I heard that Bash is slow when it comes to comparing two arrays.

Any suggestions appreciated.

https://redd.it/1dpqa09
@r_bash

Where to Implement noscripts and how to manage them?

I have a noscript I made (my first), but want to know

1. Where to store it (I've read this is the best location: /usr/local/bin )
2. How to manage them with Github and across multiple machines

I'm looking into Ansible for automating my environment setup (current machine is dying plus I anticipate a new job soon). And I just figured out GNU Stow for dotfiles. So in writing my first noscript (well it was actually my second time writing it), as well as the fact that I'll likely have 2 new machines to setup soon, I need to understand properly managing noscripts & between machines.

My problems:

1.) if I put noscript files on Github I believe they must be in a directory (for example: noscripts ). The problem is I've read that user noscripts should be stored at /usr/local/bin not /usr/local/bin/noscripts for example.

2.). There is already a lot of crap in /usr/local/bin and I am wary of adding it all to Github/source control for fear of fouling something up.

I've already figured out:

1. How to get rid of my noscript's extension (.sh) by making this the first line: #!/bin/bash plus runningchmod +x
2. how to make it so that you don't need to whole file address by putting it in a directory that is known to my PATH.

I am sorry I if this is a dumb question - honestly I'm far enough in my career I should already know this but I went through a bootcamp and have some knowledge gaps like this I'm working to fill.


I realize I'm probably over-thinking this. And should just add my personal noscripts to /usr/local/bin/noscripts , add it to my path, and make the "noscripts" directory my git repo.

Any help appreciated. Will post to a few relevant communities.

In summary:

1. Where to store personal noscripts
2. How to manage them with Github and across multiple machines
3. Any thoughts on managing noscripts with Ansible or similar?

https://redd.it/1dq0obx
@r_bash

Will it work or not
https://redd.it/1dq43hf
@r_bash

Make my code even better; which tools are you using?

Hello everyone,

I spend most of my time coding with PHP and, from time to time, I create Bash noscripts that can be several thousand lines long (I have a main noscript and "helpers" that I load as external files).

I use /bin/bash -s mynoscript.sh to identify syntax errors, ShellCheck ( https://github.com/PeterDaveHello/docker-shellcheck) to identify certain errors and shfmt (https://github.com/PeterDaveHello/docker-shfmt) to force formatting of noscripts and I'm hard pressed to find any other tools.

I attach the greatest importance to the quality of my code, its readability, etc. so I'd be happy to read any ideas you have for tools I could use to analyse the quality of my code and make suggestions for improvements.

Thank you very much.

https://redd.it/1dqiq20
@r_bash

Get first output of continous command

Hello, I'd like to only have the first output of a continous command, like pactl subsribe or hyprland-workspaces ALL

https://redd.it/1dqjmig
@r_bash

Ssh into servers and show custom ps1prompt

I have a .bashrc file. Which has alias colors and custom ps1 prompt. In my job we ssh into a passwordless server and from that server we ssh into multiple servers(in those server we have to enter password).

Is there any way to use my local .bashrc file in those ssh servers without modifying the .bashrc file in those servers?

https://redd.it/1dqn22x
@r_bash

what are these things? do they have a name? like the "file permissions letter grid"?
https://redd.it/1dr4q04
@r_bash

portmanager: A Bash Function

## Sourcing the Function
You can obtain the function [here](https://github.com/slyfox1186/noscript-repo/blob/main/Bash/Misc/Functions/portmanager.sh) on GitHub.

## How It Works

The function uses system commands like ss, iptables, ufw, and firewall-cmd to interact with the system's network configuration and firewall rules. It provides a unified interface to manage ports across different firewall systems, making it easier for system administrators to handle port management tasks.

## Features

1. Multi-firewall support: Works with iptables, UFW, and firewalld.
2. Comprehensive port listing: Shows both listening ports and firewall rules.
3. Port range support: Can open, close, or check ranges of ports.
4. Safety features: Includes confirmation prompts for potentially dangerous operations.
5. Logging: Keeps a log of all actions for auditing purposes.
6. Verbose mode: Provides detailed output for troubleshooting.

## Usage Examples

After sourcing the noscript or adding the function to your .bash_functions user noscript, you can use it as follows:

1. List all open ports and firewall rules:

   port_manager list
   

2. Check if a specific port is open:

   port_manager check 80
   

3. Open a port:

   port_manager open 8080
   

4. Close a port:

   port_manager close 8080
   

5. Check a range of ports:

   port_manager check 8000-8100
   

6. Open multiple ports:

   port_manager open 80,443,20000-20010
   

7. Use verbose mode:

   port_manager -v open 3000
   

8. Get help:

   port_manager --help
   

## Installation

1. Copy the entire port_manager function into your .bash_functions file.
2. If using a separate file like .bash_functions, source it in your .bashrc file like this:

   if [[ -f ~/.bash_functions ]]; then
       . ~/.bash_functions
   fi
   

3. Reload your .bashrc or restart your terminal.

https://redd.it/1drgck1
@r_bash

Does anyone know how to change dualsense led color

So I'm working on a noscript that allows you to change dualshock led colors I'm trying to implement a dualsense led changer so it's compatible with ps4/ps5 controllers but ran into multiple dead ends on Google if anyone has any idea on how to do this pls let me know

https://redd.it/1drj1s3
@r_bash

Beginner-friendly bash noscripting tutorial

There are plenty of excellent bash noscripting tutorial videos, so I thought one more is not going to hurt.

I've put together a beginner practical tutorial video, building a sample noscript and explaining the concepts along the way. https://www.youtube.com/watch?v=kFovBYgtEuI

The idea is to take you from 0 to 60 with creating your own noscripts. The video doesn't aim to explain all the concepts, but just enough of the important ones to get you started.

https://redd.it/1ds25n3
@r_bash

Share your $PS1 prompt config

Intrested how people use prompts to get most of it.

https://redd.it/1ds7dkv
@r_bash