Pull variables, with specific prefix, from env and create request

I have a bunch of noscripts like this one with a bunch of variables defined.

I'm looking for a way to not define all of those and perhaps pull them from the env if there's a variable prefixed with IM_.

Any suggestions on how to...

1. Pull all the variables?
2. Create the request body string?

https://redd.it/12n0jrl
@r_bash

Bash

What should i do to be bash & shell god?

https://redd.it/12n05wc
@r_bash

how to make a noscript write to a permission dienied file?

I want to make a noscript to change the brightness of my screen, but i cant seem to get permision to write to that file. ive tryed all sorts of sudoing, even tee , but notthing works. help. Also is it possible to make this write to the file without asking for sudo password everytime i want to use it, without haveing to change the permissions for the brightness file itself. ther is a program called brightnessctl that basicly does this and i wanted to try to write my own version.

\#!/bin/bash

echo "How bright do you want me? (0-255) : "

read brightnessAmount

echo $brightnessAmount | sudo tee /sys/class/backlight/brightness

https://redd.it/12o4tr9
@r_bash

Curious, how long did it take you to learn bash?

Also, did you have any programming knowledge prior?

https://redd.it/12p1oit
@r_bash

behaviour of GLOBIGNORE changed?

So far I have used GLOBIGNORE to prevent ssh-add from trying to load my public keys. My .bashrc contains (contained ...)

GLOBIGNORE="*.pub"
ssh-add ~/.ssh/id*

For a few days now I notice that ssh-add complains about permissions 0644 for my *.pub files so obviously it's trying to load them.

I changed the command to GLOBIGNORE="/home/*/.ssh/*.pub" and this helps, but I'm wondering if, when and why this changed, or am I totally on the wrong track?

I also tried the more generic /*/.ssh/*.pub (which would also work for /root) and this does not match for user home directories (only /root).

FWIW: our systems were upgraded from RHEL 7 to 8 recently, is there a connection?

https://redd.it/12p7218
@r_bash

A new paradigm file-manager

By accident I might stumble upon a new way of doing command line – in the project: [https://github.com/psprint/n-commodore](https://github.com/psprint/n-commodore). it is written in Zsh but can be run from any shell.

Basically it’s about 3 factors:

* panelize everything,
* grep everything,
* save everything.

Panelization is known from Midnight Commander - it means to capture command output into a list that can be browsed. Grepping is known from fzf. Screen saving is a new paradigm

In short, you have a new screen (a greppable panel) for each new command, which is saved to the disk (GDBM), and which can be fetched/navigated to, having also PWD dir and position in panel restored.

Asciicasts:

[https://asciinema.org/a/578076](https://asciinema.org/a/578076)

[https://asciinema.org/a/578088](https://asciinema.org/a/578088) (new features: 2 column view, man page viewing and ag/ripgrep/ack interface).

​

https://preview.redd.it/71toi7je0mua1.png?width=767&format=png&auto=webp&v=enabled&s=e0f08dcb1fa9245a5b4d38e4061fb21ab4f83490

https://redd.it/12qh07j
@r_bash

Interesting grep behaviour after power outage

Flagged as help but just to draw an expert opinion if you care to comment. I am can sort this out with another pass of grep.

I have many folders called say “monthA, monthB, monthC… monthZ”

Each contains numerous text files either do or do not contain the pairs of letters AA, BB,CC. I only want to keep the ones that do have them.

I started a bit of grep and pipe with this:

 -v -E -l -r ‘ AA| BB| CC’ month*/ | while read line; do ram “$line”; done;

It takes an hour to do each monthX file. We had a power cut. It stopped grepoing halfway through say monthD.

Power came back on. When I booted up and started the grep and pipe again it left monthD partially deleted and started on monthD.

I guess there’s a flag or algo decision somewhere and that is expected. I would have expected it to either start again at monthA or halfway through monthD. Anyone shed any light? I suppose it marks monthD as done when it closes it abruptly or something.

https://redd.it/12r1uj4
@r_bash

Is it okay to include bash noscripting in an embedded systems end of studies project?

I would love to see if there is anyone here who included it in his/her end of studies project. I tried to only use python with Raspberry Pi, since it seems more known in my university, but I keep having problems with the Speech Recognition module.

Now I'm considering using a bash noscript (and a microphone) to record audio and turn it into text, and using python within the noscript to read (via speakers) the final output based on the text.

TLDR: Do you think the professors within the jury will approve of using bash in an embedded systems end of studies project?

https://redd.it/12rskkq
@r_bash

Split full path and filename into path and filename.

I am currently doing this using a for loop. There must be an easier way.


fullpathfile="/path/to/where/file/is/stored.txt"

I want path="/path/to/where/file/is/" and file="stored.txt"

https://redd.it/12rt67y
@r_bash

Why is it that a command run from the terminal command line produces line break but the same command run within bash noscript has no line break/

If I run the following commands

lsblk > junk2.txt

cat junk2.txt

I see output with line breaks in the output like

NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS

loop0 7:0 0 285M 1 loop /snap/atom/282

loop1 7:1 0 284.9M 1 loop /snap/atom/286

loop2 7:2 0 4K 1 loop /snap/bare/5

​

But if I put assign the output of that command to a noscript variable and then echo that variable

temp1=$(lsblk)
echo $temp1
My output has no line breaks

NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS loop0 7:0 0 285M 1 loop /snap/atom/282 loop1 7:1 0 284.9M 1 loop /snap/atom/286 loop2 7:2 0 4K 1 loop /snap/bare/5

So why do I get line breaks with the first approach but not the second? Is the difference explained by different behaviors between cat vs. echo?

Thanks very much.

https://redd.it/12rzx3b
@r_bash

Storing a directory path as a variable, then passing it on to the 'ln -s' (symbolic link) command?

I stored directory in a variable:
var=$"'/media/disk2/video/001.mp4'"

I confirmed the directory was stored with the echo command:
echo $var
The directory is displayed with the quotation marks as desired.

I then tried to pass it on to ln -s:
ln -s $var 001_link.mp4

I get the error message:
ln: target '001_link.mp4': No such file or directory

https://redd.it/12s6aze
@r_bash

Intro to Bash noscripting for scraping and automation
https://www.trickster.dev/post/intro-to-bash-noscripting-for-scraping-and-automation/

https://redd.it/12sutav
@r_bash

DISCUSSION What if BASH could be used as a web server interpreter?

I've been BASHing for about 2 years now... it's changed my life.


The other day I was writing some code in PHP and it dawned on me, can I not find a way to run BASH as a web server interpreter?


Already I am managing SQL databases with BASH? I have used it to generate static HTML pages at regular time intervals. I have even used it to generate SVG and convert them into PDF files. It would be so nice to be able to make an index.sh rather than index.php... and I haven't yet found a problem I couldn't solve with BASH, except this?


Would like to get a discussion going and see what other people's views are, solely as a thought experiment. Would such an interpreter be practical and useful, or would it just be another dev tool in an ocean of dev tools?


No flaming please, I just want to have a righteous discussion/debate about if this sort of thing did exist now and how it would affect us as BASHers and/or the industry as a whole.

https://redd.it/12t6nnr
@r_bash

Aggregate multiple lines

I have a file that looks like this -

a 6509568
a 8011776
a 8011776
a 8011776
a 8011776
b 1112800
b 1369600
b 1369600

How do I make it look like -

a 38556672
b 3852000

https://redd.it/12t9df7
@r_bash

kube-dialog update

Hi, I've just updated kube-dialog with a filter command, you can filter namespaces and objects(pods,jobs...) with ease now, enjoy)

https://preview.redd.it/d8zv5c0jp3va1.png?width=742&format=png&auto=webp&v=enabled&s=7da1fc340ca5be8cfd2bf0ac3a709b435c3b0bc9

https://redd.it/12teich
@r_bash

Pure bash password generator

I wrote something like this but *waay* too complicated a few months ago. I fully realize it's not up to CISA standards, but I needed a way to auto generate dev Mysql passwords inside containers on startup.

genpw(){
declare -i limit="$1"
declare pw=''
_instr(){
declare -r str="$1"
declare -i sl=${#str}
declare -r chr="${2:0:1}"
declare -i ndx=$((RANDOM % ${sl}))
echo "${str:0:ndx}${chr}${str:$((ndx+1)):$((sl-ndx))}"
}
while [[ ${#pw} -le $limit ]]
do
if [[ $((RANDOM % 2)) -eq 0 ]]; then
# lc ascii
pw="$pw"$(printf "\x$(printf %x $((97+(RANDOM % 26))))")
elif [[ $((RANDOM % 2)) -eq 0 ]]; then
# uc ascii
pw="$pw"$(printf "\x$(printf %x $((65+(RANDOM % 26))))")
elif [[ $((RANDOM % 2)) -eq 0 ]]; then
# digits
pw="$pw"$((RANDOM % 10))
fi
done
declare c
# often mandatory non alpha characters
for c in '#' '@' '%'
do
pw="$(_instr ${pw:0:$limit} $c)"
done
echo "$pw"
unset -f _instr
} # genpw()

ahare@box:~/wk$ genpw 30
lQki068h@kUqa#8ewbtqw%t8eZkGcN
ahare@box:~/wk$ genpw 40
zwxNg7xl5Fz%EJBzajXdAn2dPTh#kdcJp@gk3fgM

https://redd.it/12trtg2
@r_bash

Can anyone PLEASE help with this noscript!

When run it says : versions found then SELECT not found and Syntax error: "do" unexpected.

Can you help solve this (probable really simple error)?

​

\#!/bin/bash

\# Get the name of the song from the user

read -p "Enter the name of the song: " song_name

\# Use find to search for all versions of the song in the Music folder

song_paths=$(find /home/Music -name "$song_name*")

​

\# Print out a numbered list of the song versions and ask the user which one they want to play

echo "Versions of '$song_name' found:"

select song_path in $song_paths;

do

if [ ! -n "$song_path" \]; then

break

fi

done

\# Play the selected song using the default music player on the system

xdg-open "$song_path"

exit

https://redd.it/12txs4b
@r_bash

Help/Direction creating a commandline AI?

I understand that their is a tool out there currently by the name of CLAI, however I'm wondering if it would be possible (as an independent project) to build/design a better command line artifical intelligence similar to the CLAI bot that is currently available. If so what reading materials should I start with? Is it possible to leverage "zsh"/"bash" and other cli tools with python AI systems.

https://redd.it/12uhaak
@r_bash

getting used and free disk space before and after rsync/trash noscript

Hi. Iconsider the following noscript:

if mountpoint -q /media/external; then
printf "Disk space before:\n"
# <usedGB>/<totalGB> Free: <freeGB>
rsync -avzP --ignore-existing --update --delete --modify-window=2 --stats --human-readable --delete --exclude={'*.db','*.tmp','recycle/'} /sambashare/ /media/external

find /sambashare/recycle -mtime +$30 -delete +
printf "Disk space after:\n"
# <usedGB>/<totalGB> Free: <freeGB>
else
exit 1
fi

the challenging part is, getting the disk space right (for me). tried.

df -ht ext4

but it outputs something like this

ilesystem Size Used Avail Use% Mounted on
/dev/nvmexxxx 21G 6.8G 13G 35% /
/dev/nvmexxxx 8.4G 2.3G 5.7G 28% /home

Can you point me to the right path ?

also, can it be done more efficiently what I am trying to achieve ? Thanks

https://redd.it/12upmle
@r_bash

Parse XML from a larger non-XML string in BASH

Hello people, I'm using a Calibre utility to fetch e-book metadata (via a BASH noscript):

$ fetch-ebook-metadata --isbn="9780791085455" --opf

The utility works by trying several resources, it returns errors as well as the metadata in XML format,

see output:

Failed to get cached URL from google for URL: https://www.amazon.com/Peyote-Mescaline-Drugs-Straight-Facts/dp/0791085457 with error: HTTP Error 429: Too Many Requests
Failed to get cached URL from google for URL: https://www.amazon.com/Nanny-Lana-Ferguson-ebook/dp/B0B6B41KFP with error: HTTP Error 429: Too Many Requests
Failed to get cached URL from google for URL: https://www.amazon.com/Basic-Geriatric-Nursing-Patricia-Williams/dp/0323554555 with error: HTTP Error 429: Too Many Requests
Response from wayback machine: {'archivedsnapshots': {},
'url': 'https://www.amazon.com/Basic-Geriatric-Nursing-Patricia-Williams/dp/0323554555'}
Failed to get cached URL from google for URL: https://www.amazon.com/-/es/M-Foster-Olive/dp/0791085457 with error: HTTP Error 429: Too Many Requests
Response from wayback machine: {'archivedsnapshots': {},
'url': 'https://www.amazon.com/-/es/M-Foster-Olive/dp/0791085457'}
Failed to get cached URL from google for URL: https://www.amazon.com/Responding-Emergency-American-Red-Cross/dp/1584805544 with error: HTTP Error 429: Too Many Requests
<?xml version='1.0' encoding='utf-8'?>
<package xmlns="http://www.idpf.org/2007/opf" unique-identifier="uuidid" version="2.0">
<metadata xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:opf="http://www.idpf.org/2007/opf">
<dc:language>eng</dc:language>
<dc:subject>Juvenile Nonfiction</dc:subject>
<dc:subject>Health &amp; Daily Living</dc:subject>
<dc:subject>Substance Abuse</dc:subject>
</metadata>
<guide/>
</package>

I only need this part (obviously):

<?xml version='1.0' encoding='utf-8'?>
<package xmlns="http://www.idpf.org/2007/opf" unique-identifier="uuidid" version="2.0">
<metadata xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:opf="http://www.idpf.org/2007/opf">
<dc:language>eng</dc:language>
<dc:subject>Juvenile Nonfiction</dc:subject>
<dc:subject>Health &amp; Daily Living</dc:subject>
<dc:subject>Substance Abuse</dc:subject>
</metadata>
<guide/>
</package>

Running the BASH noscript via Cygwin, but I can move to a Linux box.

I Googled it, but the Stack Overflow threads I got don't mention this scenario, at least not that I saw. My Shell capabilities aren't that strong.

https://redd.it/12v1lo6
@r_bash

sleep in while clause

Hello,

i found a noscript where there is sleep () inside done clause, i am not sure... but from my point of view its pure mess?

&#x200B;

while [[ something \]\]; do

.....

sleep X;

done << (sleep Y);

&#x200B;

does it make any sense to use 2 times sleep and especially inside done clause...

&#x200B;

&#x200B;

thanks!

https://redd.it/12v4cu7
@r_bash