r_bash – Telegram
Parsing byte counts

A few noscripts I wrote have "byte count" as an optional input. Id like these to accept using prefixes (e.g., 64 kb or 128 MiB). But, there are 2 competing systems at play here.

kilobyte is 1000, megabyte is 1000^2, etc.
kibibyte is 1024, mebibyte is 1024^2, etc.

Is there some universally agreed upon syntax for which prefic abbreviations map to 1000^n vs which map to 1024^N?

NOTE: for my use cases it doesnt make sense to specify bit count, so wshether or not there is a trailing b or B it will always refer to bytes.

My intuition here is that

1000^N:

k, kb, kB --> 1000
m, mb, mB --> 1000^2
etc.

1024^N:
K, Ki, ki, Kb, Kib, kib, KB, KiB, kiB --> 1024
M, Mi, mi, Mb, Mib, mib, MB, MiB, miB --> 1024^2
etc.

Are there any commonly used programs that would conflict with this mapping?



As far as the actual implementation, I use something like

getBytes() {

local +i nn
local -A byteMap

byteMap=([k]=1 [m]=2 [g]=3 [t]=4 [p]=5 [e]=6)

for nn in "${@}"; do
nn="${nn//[bB ]/}"
case "${nn}" in
kmgtpe)
echo "$(( ${nn//^0-9/} ( 1000 ${byteMap[${nn//[0-9]/}]} ) ))"
;;

KMGTPEIi)
nn="${nn,,}"
nn="${nn%i}"
echo "$(( ${nn//^0-9/} ( 1024 ${byteMap[${nn//[0-9]/}]} ) ))"
;;
)
echo "${nn//^0-9/}"
;;
esac
done

}

but if anyone has a better implementation please do suggest it!

https://redd.it/1h230p8
@r_bash
Understanding heredoc variable substitution

Hello, I'm confused about the output of this noscript:

Foo="bar"
cat << EOF
a $Foo
$Foo
EOF

This outputs:

a bar
Foo

It looks like variables at the start of a line don't get substituted. Can I work around that?

https://redd.it/1h2kcfd
@r_bash
Can someone ELI5 "trailing newline", what the -n command means, the -e command and what "echo" is?

I am trying to have an understanding of what these things actually mean and have an understanding of it.


The more I read the more confused I get, if someone could explain it so a child could understand it I would appreciate it.

https://redd.it/1h2vykk
@r_bash
Can you change the escape key in vi mode?

I want to use ctrl+c like I use in my editor to enter normal mode

https://redd.it/1h33g39
@r_bash
Escape $ to write literal placeholders

Hi,

Newbie here, apologies in advance if my question is not appropriate.

I have a bash noscript that installs some software, and I would like to generate a networkd-dispatcher noscript.

The networkd-dispatcher noscript should contain placeholders such as "$IFACE" and "$UNIT_NAME", but the installation noscript interprets them as undeclared variables, and the networkd-dispatcher noscripts ends up with empty spaces.

How can I escape these "$"?

This is what I have at the moment in the installation noscript:

createnetworkdnoscript() {
cat << EOF > $HOME/BirdNET-Pi/templates/50-birdweather-publication
#!/bin/bash
UNITNAME="birdweatherpublication@$IFACE.service"
# Check if the service is active and then start it
if systemctl is-active --quiet "$UNITNAME"; then
echo "
$UNITNAME is already running."
else
echo "Starting $UNITNAME..."
systemctl start "
$UNITNAME"
fi
EOF
chmod +x $HOME/BirdNET-Pi/templates/50-birdweather-publication
chown root:root $HOME/BirdNET-Pi/templates/50-birdweather-publication
ln -sf $HOME/BirdNET-Pi/templates/50-birdweather-publication /etc/networkd-dispatcher/routable.d
systemctl enable systemd-networkd
}

createnetworkdnoscript

https://redd.it/1h47zez
@r_bash
Advent of Code 2024 - Day 1 Problem 1 Solution in Bash

Hi, I have been learning Bash the last two days as my first noscripting language. I saw the advent of code started this year, and I thought why not try to solve it with Bash (since it's the only language I know so far." I managed to solve most of it by myself, had only to look for the sort command.

---

# Bash solution for day 1 problem 1

## Summary of the problem

- 2 Teams are searching for the locations where the Chief Historian might be.
- Each location has a 'location ID'.
- 2 Groups trying to make a complete list of 'location ID'.
- The two lists are not similar.
- Pair the smallest 'location ID' from the left with the smallest 'location ID' from the right
- Measure the distance (difference) between each 'location ID' pair.
- Measure the total aggregate distance between all 'location ID' pairs.

# inputs

A text file with the 2 lists is presented in the following format

18944   47230
94847 63037
93893 35622


## Steps to solution

1. Separate the numbers in the text file into two lists.
2. Order the numbers in each list from the smallest to the biggest.
3. Measure the distance between each 2 respective numbers.
4. Measure the total of distances.

## Solution

Save the numbers in a text file called input.txt"

#!/bin/bash

# Generate an array from the input
list=(`cat input.txt`)

# Save the even elements into list.left.txt and the odd elements into list.right.txt
for el in "${!list[@]}"
do
rem=$((${el} % 2))
if [[ rem -eq 0 ]]
then
echo "${list[$el]}" >> list.left.txt
else
echo "${list[$el]}" >> list.right.txt
fi
done

# Sorting the numbers
sort list.left.txt > list.left.sorted.txt
sort list.right.txt > list.right.sorted.txt

# create arrays from the two files
left=(`cat list.left.sorted.txt`)
right=(`cat list.right.sorted.txt`)

# calculate the difference and save it to a text file.
for ele in "${!left[@]}"
do
diff=$(("${left[$ele]}"-"${right[$ele]}"))
if [ $diff -ge 0 ]
then
echo "$diff" >> diffs.txt
else
diff=$(($diff * -1))
echo "$diff" >> diffs.txt
fi
done

# Import the differences as an array
di=(`cat diffs.txt`)

total=0

for elem in ${di[@]}
do
total=$(($total + $elem))
done
echo "$total"


https://redd.it/1h4t92n
@r_bash
Why this loop doesn't break the first time?

while read -r line
do
echo "$line"
done <file.txt


Here, the condition read -r line has nothing to read the first time the loop runs, why it doesn't break the first time?

https://redd.it/1h526bx
@r_bash
Move files from all subfolders to root folder but new filename should contain the folders

Hello,


i have a lot of folders containing files and more sobfolders with files. I want to have all that files in the root folder and the filename should contain the folder name. For example the file /testdir1/testdir2/testfile,txt should be in /testdir1_-_testdir2_-_testfile.txt

The thing is, some years ago i had done this by accident (i think i tried just to remove bad characters from filename but by accident also replaces the / but i can't get it together again :-( )



https://redd.it/1h5mrsi
@r_bash
Any way to hook into 'command not found' and run a noscript / function?

Curious if there's any way to hook into the error condition 'command not found' and run a noscript/function? Basically, I'd like to do something similar to "thefuck" but have it run automatically.

$ doesnotexist
-bash: doesnotexist: command not found

# how to (automatically) call some custom function/noscript/etc?
# preferably with access to bash history so I can run a
# fuzzy find with target command vs my defined aliases

So far my searches keep coming up with irrelevant stuff so I'm not sure if I'm just using bad search terms or if this is something that is just not possible under bash.

https://redd.it/1h65lkg
@r_bash
I made a bash noscript to exclude dropbox sync directories via command line

I code a lot in my dropbox folder to keep them synced across my devices (before git commits are viable) and unfortunately dropbox does not include an automatic way to exclude syncs. Took a while but with some guidance from claude 3.5 I hacked this together.

https://github.com/kavehtehrani/dropbox-exclude

https://redd.it/1h716kg
@r_bash
How to exclude a directory from find and rsync except for a few very specific files?

I'm struggling with nested include/exclude for find and rsync.

I want to find or rsync my dotfiles, except for the .mozilla folder (among some others). But I want the login data of firefox preserved. So far, I have

find -path '*/.*' -not -path '*/.cache/*' -not -path '*/.mozilla/*' -path '*/.mozilla/firefox/*.default-release/{autofill-profiles,signedInUser,prefs}.js*' > dotfiles

which gives back a blank file. How can I do this properly to exclude the majority of stuff from one directory, but still include these specific files?

I haven't yet tackled this for rsync (and maybe tar), but solutions for these are also welcome.

https://redd.it/1h75w8x
@r_bash
Need help passing argument with alias

Hi,

I want to make an alias with the word cheat. Ex. cheat topic

I tried making an alias but can't get it right. I presume because there is whitespace between the command and the argument.

alias cheat="curl cht.sh/$1"

How can I make this alias work so when I type cheat zip, and make curl cht.sh.zip the result?

Thanks.

https://redd.it/1h7p3vn
@r_bash
replacing placeholders in a file with variables from a noscript

Yeah, this noscript probably doesn't make sense so here I go...

I have a txt file with a bunch of html code that will make up a person's signature. In the txt file I have {{firstname}} {{lastname}} and {{email}}. In my bash noscript I have variables $firstname $lastname and $email. I want to write the txt file to a html file but replace the placeholders in the txt file with what the variables are.

https://redd.it/1h7li6n
@r_bash
Unexpected evaluatoin of "date +%M" in ~/.bashrc

I use the following command in an alias in my bashrc

$(date +%Y)/$(date +%M)/KW$(date +%V)-$(( $(date +%V) +2))

Why on earth does it evaluate to something like 2024/23/KW49-51 and an ever changing month? I cannot even figure out, what is the problem. Sometimes when sourcing the bashrc I get a new month, sometimes not. What is happening here?

https://redd.it/1h81d98
@r_bash
Exports block prev exports

I'm using zshell and my config exports are like that. But when i want to use lazygit or openvpn etc. it says '...not included path.' I add it with export or via config file but then I can't even use ls. Say's I have to add it to path It's in the path on .zshrc. Couldn't find solution pls help. Everything works fine in root terminal but it's not suitable i guess.

https://preview.redd.it/8qtx9l58495e1.png?width=817&format=png&auto=webp&s=3a0a8e4e448a86ccb451d407ffa769b21adbff99

https://preview.redd.it/cxqmojw0495e1.png?width=1346&format=png&auto=webp&s=5142078c29552f2707520a69cc505162c51d1983

https://redd.it/1h8576v
@r_bash
Which is better for capturing function output

Which is the better way to capture output from a function? Passing a variable name to a function and creating a reference with declare -n, or command substitution? What do you all prefer?

What I'm doing is calling a function which then queries an API which returns a json string. Which i then later parse. I have to do this with 4 different API endpoints to gather all the information i need. I like to keep related things stored in a dictionary. I'm sure I'm being pedantic but i can't decide between the two.

mydictjson="$(somefunc)"
vs.
some
func mydict

Is there that much of a performance hit with the subshell that spawns with command substitution?

https://redd.it/1h8cuhy
@r_bash
Append multiline at the begin

I have multiple lines from a grep command,. I put this lines in a variable. Ho can i append this lines at the begin of a file? I tried with sed but It don't work, i don't know because a multi lines.
Someone can help me please

https://redd.it/1h8fogz
@r_bash
Parse urls, print those not found

I have a list of urls in the forms:

https://abc.com/d341/en/ab/cd/ef/gh/cat-ifje-full
https://abc.com/defw/en/cat-don
https://abc.com/ens/cat-ifje
https://abc.com/dm29/dofne-don-partial
https://abc.com/ens/mew-feo
https://abc.com/ens/mew-feo-partial
https://def.com/fgew/dofne-don-full

The only thing that matters are abc.com urls (I don't care about URLs from other domains) and its last "field" of the url with the suffix -full and -partial being optional. When there are duplicates, prefer first the -full version, then the -partial version. In the above example, 1st and 3rd urls are duplicates and the 3rd url should be excluded from the list. 5th and 6th urls are the same and the 6th url should be excluded from the list.

Now the unique list of items are:

cat-ifje
cat-don
mew-feo
dofne-don


From this list, I apply a command likefind to search my filesystem to each item to see if I have a file containing this name of this item as a substring.

Now, how do I get back the original url if there are no results from find for the item? The output I'm looking for is:

https://abc.com/d341/en/ab/cd/ef/gh/cat-ifje-full
https://abc.com/defw/en/cat-don
https://abc.com/dm29/dofne-don-full
https://abc.com/ens/mew-feo-partial
https://abc.com/dm29/dofne-don-partial

I think working from my existing solution to "search the item not found" from the array of URLs would be in-efficient. I guess an associative array from the start can work?

I'm processing several hundreds of items, applying find to each. I've gotten up to the point where I have the list of items not found from the filesystem, so I only need to get back their original URLs.

Any solutions much appreciated. Can even be a single awk command.

https://redd.it/1h8jdoq
@r_bash