Tilted Forum Project Discussion Community - Share you Bash Command Line Tips & Tricks

- Tilted Technology (https://thetfp.com/tfp/tilted-technology/)

- - Share you Bash Command Line Tips & Tricks (https://thetfp.com/tfp/tilted-technology/55282-share-you-bash-command-line-tips-tricks.html)

Share you Bash Command Line Tips & Tricks

I've been a linux user for a fair amount of time now, but it's just now that I'm getting really used to the command line interface.
I've picked up a few cool tidbits here and there, and I'm sure a lot of you have too, so let's hear 'em.

-----

Here's a flexible sequntial download (think ezpics)

Code:

for NUM in `seq -f %02g 00 30`; do URL="http://www.coasar.com/newgalleries/pbc/ahv/03/images/fullimage/0$moo.jpg"; wget $NUM 2>/dev/null && echo Downloaded $URL || echo "    Failed $URL" ;  done

you can type that into console and away we go. (by the way, that link is NSFW)

-----

This one I picked up on the net somewhere.
I use firefox almost exclusively, so I set up aMSN to open up links with firefox. The major problem with that though is that if I'm already "surfin' the web" firefox will ask me to launch with a different profile. :(
Here's a fix that will make it so that if firefox is already open it'll open the link in a new tab. Open up your favorite text editor *cough*VI*cough* and paste this:

Code:

#!/bin/bash

/etc/firefox/firefox -remote "openURL($@, new-tab)" >/dev/null || exec /etc/firefox/firefox "$@" >/dev/null;

Make sure to save that file somewhere in your $PATH
Now you can do stuff like this (i called mine firefox_newtab)

Code:

firefox_newtab www.tfproject.org

I'm pretty sure that if you use mozilla you can just replace every instance of "firefox" with "mozilla".

-----

This is likely common knowledge but I'm going to write in anyway, just incase someone finds it usefull.

In my tips i used && (and) and || (or).

If two commands are seperated by && then the the second will only be executed if the first one is sucessfull. For example:

Code:

./configure && make

if configure fails (which it does so often) make won't even start.

If two commands are sperated by || (two pipes, shift + \ ) the second command will only execute if the first fails. E.g.:

Code:

wget http://www.moo.com/somefile.tar || echo Download Failed

If you follow a command with >/dev/null it won't print it's standard output to the screen. Compares these:

Code:

echo Beyond here lie dragons

echo Beyond here lie dragons >/dev/null

Anyway, I'm very much still new at this and I'm curious about something that other people have picked up over the years.

Probably not what you're looking for, but, some jokers made a pizza-ordering command line utility. You sign up for an account on Domino's website and the utility does everything else! I've never tried it, but it looks like it works.

http://www.beigerecords.com/cory/pizza_party/

a lil simple bash script that will create an individual gzip tar of each sub-dir in the dir that it executed in. each tar has the name of the dir that it has compressed.

Code:

find . -type d -mindepth 1 -maxdepth 1 -exec tar zcf {}.tar.gz {} \;

Search command history

Ctrl+R will search backwards through your previous commands, just press Ctrl+R, then start typing part of the command you want to recall (any substring, it doesn't have to be from the beginning).

Ctrl+A will go to the start of the line.

Ctrl+E will go to the end

Ctrl+U will nuke from the cursor to the start of the line

Ctrl+W will nuke the word to the left of the cursor

Here are some simple things that are really invaluable - I use them several times per day :)

'cd -' will go 'back' a directory (eg. say you were in /usr/src/linux, then went to ~/kernelconfigs ... you can quickly get back to /usr/src/linux by typing cd -)

when you're typing in a file name, you can enter the first few letters and then <TAB>. It'll either automatically be completed for you or it'll show you a list of potential matches.

and if you want to perform batch operations on files in a directory, you can try this:

Code:

for i in *; do operation_1_on $i; operation_2_on $i; done

EX:

Code:

for i in *; do mpg123 -w "${i/.mp3}.wav" "$i"; toolame -v 5 "${i/.mp3}.wav" "${i/.mp3}.mp2"; done

is something I just used to convert a few mp3s into mp2 format for easier editing. I know I can use something like transcode or ffmpeg, but they don't like to compile with GCC 3.4 yet =)

set permissions on files only:

Code:

find . -type f -exec chmod 0644 {} \;

set permissions on dirs only:

Code:

find . -type d -exec chmod 0755 {} \;

recursively set permissions on something depending on if it's a file or a directory (note: chmod -R 0755 would make EVERYTHING (files AND directories) 755.. I don't want that).

Of course that can be changed to chown or pretty much anything you want hehe..

I have some other scripts, but they're only relevant for work..

One I use at work often is to kill multiple processes at once.

Code:

kill `ps -eaf | grep ^username | cut -c10-14`

Modify the ps command to list whatever needs to be killed.

Quote:

Originally posted by crony
One I use at work often is to kill multiple processes at once.

Code:

kill `ps -eaf | grep ^username | cut -c10-14`

Modify the ps command to list whatever needs to be killed.

Code:

pkill -u username

That's much cleaner :)

By default, pkill kills by process name.

Type

! followed by the beginning of a command you've already used...and hit enter. it will auto exec that command in full...you ran last. obvoiusly be careful with somethings.

i also presume everyone knows ~userid prefers to userid's home directory.

also if on the command line and have vi editor mode activated ( bash -o vi ( works with ksh too ) ) then go for a ESC V and that will bring the command line straight into a full vi session

Quote:

Originally posted by yotta

Code:

pkill -u username

That's much cleaner :)

By default, pkill kills by process name.

Yeah, the processes I need to kill vary. Just an example of how to get the pid's and pass it to kill. :)

Dont know... I just wrote a couple of shell scripts to backup stuff (one backs up MySQL, another backs up my website, another backs up stuff edited in the past day/week/month) and FTP them to backup. Kinda long to be posting on here though.

here's a few random ones i have:
Create your own short user name:

Code:

#!/usr/bin/env python

from random import choice



# Vowels

vowels = ['a','e','i','o','u']

# Beginning hard consonants

startcons =  ['b','c','d','g','j','k','m','n','p','q','t','v','x','z']

# Ending hard consonoants

endcons = ['b','c','d','g','k','m','n','p','t','x','z']

# Soft Consonants -- note v and z are listed twice

softcons = ['f','h','l','r','s','v','w','z']



print choice(startcons) + choice(vowels) + choice(softcons) + choice(endcons)

merge your multiple firefox bookmarks [in html] to one central big one:

Code:

#!/bin/bash



# by Tom Law, September 2003

# "ub" = "Unique Bookmarks"



# Compare two bookmark files, and create for each one a new html file 

# that contains the bookmarks unique to the other one.



        # You'll need to edit the following two lines to show where

        # the two existing bookmark files (that you want to compare) are.



BA="/home/keyshawn2/august 18 2004 bookmarks -win.html"

BB="/home/keyshawn2/august 18 2004 bookmarks-linux.html"



#################################



        # sort each bookmark file, while removing extra spaces,

        # and adding a browser identifier (I put the identifier

        # at the end of the title to help with debugging.)

cd /tmp

sort -b $BA | sed -e 's/<\/A>/---A<\/A>/g' | sed -e 's/  */ /g' > utmpA

sort -b $BB | sed -e 's/<\/A>/---B<\/A>/g' | sed -e 's/  */ /g' > utmpB



        # merge the two files, while removing duplicate lines,

        # as well as unwanted lines and characters



sort -bd utmpA utmpB | uniq -uW 2 | grep HREF | grep -vE "(about:config|>---|N---A|N---B|uniqA|uniqB)" | sed s/\<DT\>//g | sed s/$/\<br\>/g > utmpC



        # split the merged file back into two, while removing the identifiers



cat utmpC | grep -e ---A | sed -e 's/---A//g'> utmpA

cat utmpC | grep -e ---B | sed -e 's/---B//g'> utmpB



        # convert them into html files



echo "<html><body>"   > uhead

echo "</body></html>" > utail



cat uhead utmpA utail > uniqA.html

cat uhead utmpB utail > uniqB.html



rm utmpA utmpB utmpC uhead utail



        # After running this command, load the "/tmp/uniq?.html" files

        # into their respective browsers (file A into browser B,

        # and file B into browser A), and drag each link to somewhere

        # in the bookmark file. It doesn't matter if you put them

        # in different folders than you did in the other bookmark file.

        

        # Run ub at bootup or by cron, and then every day or so look at

        # the "uniq" file (which you should bookmark) in each browser, and

        # update the browser's bookmark file.

detect how much space is your drive taking up and where:

Code:

for dirs in `ls /`

do

if [ -d /$dirs ]

then

echo "Listing of /$dirs" >> /tmp/disk_usage.txt

echo "################################" >> /tmp/disk_usage.txt

du -h /$dirs >> /tmp/disk_usage.txt

echo "" >> /tmp/disk_usage.txt

fi

done

You can use PHP on the CLI which I love. I use that now as a way to list usage on my game hosting server so I know where I am on my usage. I've also set it up so that it'll kill certain processes or start them up.
Those are for linux
I've made scripting tips in Windows(I know, not bash, but still useful batch files that I wanna brag about) that'll list all installed applications on a computer and another one that can set the ip address.
If anyone would like any of them, let me know and I'll post them.

I have written some serious one-liner doozies in the past...usually when trying to do something somewhat complicated with a bunch of files. For example, one time I did a one-liner that recursively went through a directory tree and renamed files with certain names to match the directory name, or something like that.

Other one-liners I've used for doing recursive media conversion (WMA->MP3 for example), and of course for downloading Pr0n from various places with nice sequential filenames.

Probably my most frequently used BASH construct is the "while read" one:

Code:

cat /etc/passwd | awk -F: '{ print $1 }' | while read line; do echo $line; done;

And yes, I know that example does absolutely nothing useful, but it's an example.

Kind of like the cd - trick above, but pushd and popd are invaluable. You can create a whole stack of directories to switch from. The command dirs will give you a list of whats on the current stack. Very useful if you keep going back and forth, just pushd dir then popd and then to keep swapping !push and !pop to get back.

Bash is great.