Isolated environment for iteratively developing dev machine set up noscript?

I want to write a noscript that will set up my machine (MacOS). Nothing too crazy - install brew, brew install a bunch of things I use, prompt me for certain logins so it can auth for me, etc.

Hypothetically, if I test a bash function to install something, then make a change, I'll need to uninstall that thing to re-test.

Is there a better way to do this? Some sort of isolated environment that I can tear down with each run?

Thanks

https://redd.it/15pkanw
@r_bash

awk - how to wrap printf for readability?

writing a really long printf, how to separate it into multiple lines? ie this doesnt work:

{printf(" stuff stuff stuff"

printf "stuff stuff");}

https://redd.it/15puett
@r_bash

whats the x in cat /etc/passwd?



https://redd.it/15puc8h
@r_bash

text in e-mail subject 'corrupts' e-mail body

I have a slightly modified version of this noscript which monitors a zfs volume. It e-mails me after it runs with this command

echo -e "$emailMessage \n\n\n <pre>/sbin/zpool list \n\n\n /sbin/zpool status</pre>" | mailx -v -s "$emailSubject
Content-Type: text/html" my.address@gmail.com

Its output looks like this:

https://i.imgur.com/JWgOAUg.png

One part of the noscript does this:

if ${problems} -eq 0 ; then
capacity=$(/sbin/zpool list -H -o capacity)
for line in ${capacity//%/}
do
if $line -ge $maxCapacity ; then
emailSubject="$emailSubject - Capacity Exceeded"
problems=1
fi
done
fi

If this condition is true, the e-mail looks like this:

https://i.imgur.com/IhESPce.png

through some trial and error, I've found I can change this to:

emailSubject="$emailSubject - Capac"

and it will still format correctly.

However, adding an i (or any other letter) after "Capac" causes the e-mail formatting error.

I seem to be running into some kind of length limit, but I don't understand why. Is it because of the way the echo command is formatted?

For what its worth, the output of

"$emailMessage \n\n\n <pre>/sbin/zpool list \n\n\n /sbin/zpool status</pre>"

is 2303 characters long.


Another oddity is that if I were to do say:

echo -e "$emailMessage \n\n\n <pre>/sbin/zpool list \n\n\n /sbin/zpool status</pre>" | mailx -v -s "$emailSubject Content-Type: text/html" my.address@gmail.com

i.e. I don't put a new line between $emailSubject and Content-Type

instead of

echo -e "$emailMessage \n\n\n <pre>/sbin/zpool list \n\n\n /sbin/zpool status</pre>" | mailx -v -s "$emailSubject
Content-Type: text/html" my.address@gmail.com

I also get the same formatting issue

https://redd.it/15q07kx
@r_bash

Why does /*.py correctly interpret my wildcard but not \*.py?

This command works correctly to output all files ending with .py `ls | grep /*.py`

but this doesn't `ls | grep \*.py`. This doesn't output anything

I can't figure out why as isn't the backslash required to escape the *? In all of the internet resources, the backslash escapes the next character but I'm not sure why the forward slash version is working for me.

https://redd.it/15q2okr
@r_bash

Problem with: Compound Commands, Process Groups and SIGINT interaction

Hello!

I am diving into some more nuanced things about Bash I have noticed and trying to fill in a deeper grasp of some stuff. I have run into an issue I cannot find a resolution to and I wonder if I am missing something...need some pointers from some more knowledgeable folk!!!

If I have a loop - say a for loop with a simple long sleep in it- running in an interactive bash instance, then by running ps -p BASH_PID -o pgid,tpgid from another terminal I can see that the loop isn't in the foreground process group; however, by doing a bit of pgrep -P BASH_PID and some similar ps stuff I can see that the child process (the sleep ) is running in the foreground group. I am assuming that the for loop takes place in the interactive bash session itself (instead of a child process) so that variable assignments survive outside the loop etc. So far so good...

Problem: When I hit ctrl - C the bash loop quits immediately, this is standard behaviour however, I cannot quite explain this using what i have been reading....

Upon Ctrl -C, the driver sends SIGINT to the foreground process group...which in this case is the sleep and NOT the interactive bash instance that is housing the for loop itself. Have been reading relevant sections of the bash manual to explain this but I think I have missed something (sorry if obvious). This can be replicated by doing something like kill -INT -$(pgrep -P BASH_ID) from another terminal whilst the first runs and we get much the same result with the whole for group terminating.

The only thing I can find appears to be:

"When Bash receives a SIGINT, it breaks out of any executing loops" - GNU manual

..but it shouldn't be receiving any signals since it is sent directly to the child whether the signal originates from the terminal driver or the kill command I provided...unless bash is using some kinda system call to work out what killed the child process? Then I am just guessing. This behaviour is not seen when the for loop is backgrounded; quitting the sleep with the kill -INT PID_OF_CURRENT_SLEEP_ITERATION just sends it to the next iteration....

How can I account for this behaviour? Is there somewhere I verify this in the documentation?

Thanks for any help in getting me out of this conundrum!

https://redd.it/15q3f7d
@r_bash

Please Help! (bad substitution error)

hello, seeing a bad substitution error when executing my noscript? I have messed around with the spacing but to no avail it errors out with?: ${1,,}: bad substitution (line 3)? thank you!

https://preview.redd.it/ud8t7ezqhwhb1.png?width=2880&format=png&auto=webp&s=2c664d5641ecd61977800150ebc1c1c6ba93cb10

https://redd.it/15q2lqv
@r_bash

Does it suck?

I've been coding in bash for a while now, and generated the code in the comments with ai (Yes, I know). This is a test for the users of the subreddit, so give me your honest opinions on whether it sucks or not, and how YOU how would improve it. 😉

https://redd.it/15qc092
@r_bash

venting! why are the bash manuals so difficult to read

i comprehend things by seeing examples/doing, not by abstract concepts. i find it really difficult to grasp the manual information. why doesn't the manual include some examples?

https://redd.it/15qe2oh
@r_bash

I need help with a function.

You might have seen my post a few days ago, if not, it's fine, I just need help again.

&#x200B;

I have the function:

&#x200B;

music () { /usr/bin/mpv ~/Music/[playlist]/* --volume="$1" --shuffle --no-audio-display; }

&#x200B;

Now, whenever I type "music x" it plays my playlist at x volume. However, whenever I type "mpv" it also plays my playlist, prohibiting me from using mpv ever again. Not even "\\mpv works" like an alias. How do i fix this?

https://redd.it/15qjhd5
@r_bash

Questions about finding one or more files and renaming them

I am BRAND NEW to Bash and have a few questions.

Here’s what I’m trying to do:

Find one or more files that could be located anywhere on a Macbook. If files found, rename them.

File will be named like this:
[filename (always the same)\] - [a number that will vary but will be of the form #.#.# or #.#.##\].[file extension (always the same)\].
So something like “foo-1.2.3.abc”.

Based on what I know, I’m guessing the “find” command is the way to go because it’s recursive by default, and also because I won’t necessarily know the location of the files.

Question 1: how do I write the command to denote that I want to search the entire Macbook? I’ll be deploying this noscript to our corporate environment via JAMF (our device management solution for Apple devices.

Question 2: once the noscript finds files on the Macbook, how do I take further action on the results?

I’m planning to tackle this task in two phases. First phase: if files found, rename them (I don’t think removing the files is going to cause an issue, as I’ve already done a similar thing on the Windows side of our environment, but I want an easy way to revert if things are unexpectedly impacted). Second phase: delete all the renamed files.

I’m guessing the rename will be handled by the “mv” command. Something like

mv -v “/location/foo-1.2.3.abc” “/location/foo-1.2.3.abc.DONOTUSE”

But I’m not sure how to couple the rename command with the results of the find command, beyond thinking it’s probably going to involve a pipe character.

Could someone provide just a general example? Thanks!

https://redd.it/15r47q3
@r_bash

Recommended Video: A break down on the top 50 terminal commands

I came across this video on YT. I have been playing it in the background while I work and already I have picked up a few terminal commands I didn't know, but already have potential use cases...


https://www.youtube.com/watch?v=ZtqBQ68cfJc

https://redd.it/15rksrs
@r_bash

Why did teacher use address 192.168.1.0 for -sn and syn scan despite the default being 1.1? won't we be able to detect more through 1.1 because it is a gateway and all systems are connected to it?

https://redd.it/15rn8y9
@r_bash

How to verify if a redis dump was successfully restored?

I have a bash noscript that
- downloads a redis dump from S3
- stops any running docker containers with the name "myrediscontainer"
- starts a new docker container "myrediscontainer" and copies the downloaded dump file into the data directory of this container
- waits for this docker container to be ready

When I manually jump into the container and run redis-cli, I can see all the keys
But how do I verify if the restore was successful or not inside the bash noscript. Would it be safe to say that dbsize > 0 means restore was successful? Is there a more appropriate way?

#!/usr/bin/env bash

# https://www.shellcheck.net/ VERIFIED
set -e
set -E
set -o pipefail
set -u
# set -x

log="/tmp/error.txt"
exec 2>>"$log"

handleerror() {
# ...handle error
exit 1
}

handleexit() {
# ...handle exit
exit 0
}

trap 'handleexit $?' EXIT
trap 'handleerror $?' ERR

bucket='something-something'
containername="myrediscontainer"
dumpfiledirectory="/tmp"
host="localhost"
port="6379"

dumpfilename=$(aws s3 ls $bucket --recursive | sort | tail -n 1 | awk '{print $4}')

aws s3 cp "s3://${bucket}/${dumpfilename}" "/tmp/${dumpfilename}"

docker ps -aq --filter "name=${containername}" | grep -q . && docker stop "${containername}" && docker rm -fv "${containername}"

docker run --detach --name "${containername}" --publish "${port}:${port}" --volume "/tmp/${dumpfilename}:/data/dump.rdb" redis

# Wait for the Docker container to be ready
until docker exec -it "${containername}" redis-cli -h "${host}" -p "${port}" ping; do
sleep 1
done

# HOW TO VERIFY IF REDIS WAS RESTORED SUCCESSFULLY HERE???

https://redd.it/15rmhbi
@r_bash

how to do rg-fzf-vi function/pipeline?

Hiiiii!!! I need to do something like rg rabbi | fzf | rg ".*.mkd" | xargs vi real easily but this is not at all easy for me.

I don't know bash (string parsing/expansion/etc, and other aspects of the language) too well. I struggle back and forth between grep/ack/rg without knowing either well, moving towards rg. And I don't know fzf well, but it's flags look --magical (and <oof> I fall down).

struggle #1: in lieu of a file object how to parse the filename from a string like

2022-01-02.mkd:looking for a loved rabbi... Hillel the Elder, alive around...

Uh... cut? tr? grep? bash string expansion or globbing or something?

struggle #2: what's the best way? fzf looks like it's designed to be used in pipelines perhaps more efficiently than what I've shown here... how do you do that?

ow. this hurts (sorry for whining, i feel very whiny). Any of you kindhearted folks able to help please? thankyou <3

https://redd.it/15rxc6l
@r_bash

First Coinbase airdrop

https://basechain.markets

https://redd.it/15s0v60
@r_bash

Portable, ready-to-run shell noscript

I wrote a shell noscript that is intended to be used by laymen/non-programmers, and I simply want them to download my noscript from github and double-click the file and have it run. The problem when you unzip the respository and click the file, it fails because it does not have correct permissions to run. I have to chmod +x the file first. Is there a way to preset permissions so that users can just double click and run after download? I can't expect these users to open the terminal and properly enter chmod commands. This is for OSX btw.

https://redd.it/15s5531
@r_bash

Print lines between similar patterns

We have some python noscripts that queries our AWS accounts and produces a list of the accounts and some resources associated, including available AMIs. Using sed, I am trying to filter through the output to fetch only the accounts which have the AMI and the associated AMI.

Eg, the python output would be something like this:

Processing account: A
Text to ignore
More text to ignore
.
.
AMI used:
'ami-123456', 'ami-789012'
More text to ignore

Processing account: B
Text to ignore
More text to ignore
.
.

Processing account: C
Text to ignore
More text to ignore
.
.
AMI used:
'ami-abcdef', 'ami-123456'
More text to ignore


What I'm trying to get:

Processing account: A
AMI used:
'ami-123456', 'ami-789012'

Processing account: C
AMI used:
'ami-abcdef', 'ami-123456'

I was thinking of something like this, but it gives me 'Processing account: B', which doesn't have any AMIs listed.

$ sed -n '/Processing/, /Processing/p' filename.txt | grep -vE '(Text to ignore|More text to ignore)'

Output:

Processing account: A
AMI used:
'ami-123456', 'ami-789012'

Processing account: B
Processing account: C
AMI used:
'ami-abcdef', 'ami-123456'

Surely there is a better way to do this; keen to any suggestions.

Thank you.

https://redd.it/15siimc
@r_bash

IF inside awk

need to convert a comma separated integer into an integer and IF compare it to 1000 all inside an awk

if($1 > 1000){

printf("%s", $1);

}

where $1 is the first field of each record and looks like " 1,500"

https://redd.it/15sizhh
@r_bash