Hide

What's this?

commandlinefu.com is the place to record those command-line gems that you return to again and again.

Delete that bloated snippets file you've been using and share your personal repository with the world. That way others can gain from your CLI wisdom and you from theirs too. All commands can be commented on, discussed and voted up or down.


If you have a new feature suggestion or find a bug, please get in touch via http://commandlinefu.uservoice.com/

Get involved!

You can sign-in using OpenID credentials, or register a traditional username and password.

First-time OpenID users will be automatically assigned a username which can be changed after signing in.

Hide

Stay in the loop…

Follow the Tweets.

Every new command is wrapped in a tweet and posted to Twitter. Following the stream is a great way of staying abreast of the latest commands. For the more discerning, there are Twitter accounts for commands that get a minimum of 3 and 10 votes - that way only the great commands get tweeted.

» http://twitter.com/commandlinefu
» http://twitter.com/commandlinefu3
» http://twitter.com/commandlinefu10

Subscribe to the feeds.

Use your favourite RSS aggregator to stay in touch with the latest commands. There are feeds mirroring the 3 Twitter streams as well as for virtually every other subset (users, tags, functions,…):

Subscribe to the feed for:

Hide

News

2011-03-12 - Confoo 2011 presentation
Slides are available from the commandlinefu presentation at Confoo 2011: http://presentations.codeinthehole.com/confoo2011/
2011-01-04 - Moderation now required for new commands
To try and put and end to the spamming, new commands require moderation before they will appear on the site.
2010-12-27 - Apologies for not banning the trolls sooner
Have been away from the interwebs over Christmas. Will be more vigilant henceforth.
2010-09-24 - OAuth and pagination problems fixed
Apologies for the delay in getting Twitter's OAuth supported. Annoying pagination gremlin also fixed.
Hide

Tags

Hide

Functions

Commands using read from sorted by
Terminal - Commands using read - 286 results
sortwc () { local L;while read -r L;do builtin printf "${#L}@%s\n" "$L";done|sort -n|sed -u 's/^[^@]*@//'; }
2010-05-20 20:13:52
User: AskApache
Functions: printf read sed sort
2

This provides a way to sort output based on the length of the line, so that shorter lines appear before longer lines. It's an addon to the sort that I've wanted for years, sometimes it's very useful. Taken from my http://www.askapache.com/linux-unix/bash_profile-functions-advanced-shell.html

mplayer http://minnesota.publicradio.org/tools/play/streams/the_current.pls < /dev/null | grep --line-buffered "StreamTitle='.*S" -o | grep --line-buffered "'.*'" -o > mus & tail -n0 -f mus | while read line; do notify-send "Music Change" "$line";done
2010-05-09 17:51:40
User: spiffwalker
Functions: grep read tail
7

Plays the mp3 stream of The Current as a background job. When you are done run:

fg %1

then to exit

Quite possible with Growl for mac I'd guess, although have not tried.

Libnotify needed for notification, stream will still work otherwise

echo 'Host or User@Host?:'; read newserver && ssh-keygen -N "" -t rsa -f ~/.ssh/id_rsa ; ssh $newserver cat <~/.ssh/id_rsa.pub ">>" ~/.ssh/authorized_keys ; ssh $newserver
2010-05-07 06:24:53
User: alf
Functions: cat echo read ssh ssh-keygen
Tags: ssh ssh-keygen
-2

Some servers don't have ssh-copy-id, this works in those cases.

It will ask for the destination server, this can be IP, hostname, or user@hostname if different from current user.

Ssh keygen will let you know if a pubkey already exists on your system and you can opt to not overwrite it.

grep -ioE "(url\(|src=)['\"]?[^)'\"]*" a.html | grep -ioE "[^\"'(]*.(jpg|png|gif)" | while read l ; do sed -i "s>$l>data:image/${l/[^.]*./};base64,`openssl enc -base64 -in $l| tr -d '\n'`>" a.html ; done;
2010-05-05 14:07:51
User: zhangweiwu
Functions: grep read sed
Tags: html
3

in "a.html", find all images referred as relative URI in an HTML file by "src" attribute of "img" element, replace them with "data:" URI. This useful to create single HTML file holding all images in it, as a replacement of the IE-created .mht file format. The generated HTML works fine on every other browser except IE, as well as many HTML editors like kompozer, while the .mht format only works for IE, but not for every other browser. Compare to the KDE's own single-file-web-page format "war" format, which only opens correctly on KDE, the HTML file with "data:" URI is more universally supported.

The above command have many bugs. My commandline-fu is too limited to fix them:

1. it assume all URLs are relative URIs, thus works in this case:

<img src="images/logo.png"/>

but does not work in this case:

<img src="http://www.my_web_site.com/images/logo.png" />

This may not be a bug, as full URIs perhaps should be ignored in many use cases.

2. it only work for images whoes file name suffix is one of .jpg, .gif, .png, albeit images with .jpeg suffix and those without extension names at all are legal to HTML.

3. image file name is not allowed to contain "(" even though frequently used, as in "(copy of) my car.jpg". Besides, neither single nor double quotes are allowed.

4. There is infact a big flaw in this, file names are actually used as regular expression to be replaced with base64 encoded content. This cause the script to fail in many other cases. Example: 'D:\images\logo.png', where backward slash have different meaning in regular expression. I don't know how to fix this. I don't know any command that can do full text (no regular expression) replacement the way basic editors like gedit does.

5. The original a.html are not preserved, so a user should make a copy first in case things go wrong.

cat list|while read lines;do echo "USER admin">ftp;echo "PASS $lines">>ftp;echo "QUIT">>ftp;nc 192.168.77.128 21 <ftp>ftp2;echo "trying: $lines";cat ftp2|grep "230">/dev/null;[ "$?" -eq "0" ]&& echo "pass: $lines" && break;done
cat domainlist.txt | while read line; do echo -ne $line; whois $line | grep Expiration ; done | sed 's:Expiration Date::'
2010-05-02 06:49:09
User: netsaint
Functions: cat echo grep read sed whois
3

Create a text file called domainlist.txt with a domain per line, then run the command above. All registries are a little different, so play around with the command. Should produce a list of domains and their expirations date. I am responsible for my companies domains and have a dozen or so myself, so this is a quick check if I overlooked any.

find ~ -maxdepth 2 -name .git -print | while read repo; do cd $(dirname $repo); git pull; done
alias dateh='date --help|sed -n "/^ *%%/,/^ *%Z/p"|while read l;do F=${l/% */}; date +%$F:"|'"'"'${F//%n/ }'"'"'|${l#* }";done|sed "s/\ *|\ */|/g" |column -s "|" -t'
21

If you have used bash for any scripting, you've used the date command alot. It's perfect for using as a way to create filename's dynamically within aliases,functions, and commands like below.. This is actually an update to my first alias, since a few commenters (below) had good observations on what was wrong with my first command.

# creating a date-based ssh-key for askapache.github.com

ssh-keygen -f ~/.ssh/`date +git-$USER@$HOSTNAME-%m-%d-%g` -C 'webmaster@askapache.com' # /home/gpl/.ssh/git-gplnet@askapache.github.com-04-22-10

# create a tar+gzip backup of the current directory

tar -czf $(date +$HOME/.backups/%m-%d-%g-%R-`sed -u 's/\//#/g' <<< $PWD`.tgz) . # tar -czf /home/gpl/.backups/04-22-10-01:13-#home#gpl#.rr#src.tgz .

I personally find myself having to reference

date --help

quite a bit as a result. So this nice alias saves me a lot of time. This is one bdash mofo. Works in sh and bash (posix), but will likely need to be changed for other shells due to the parameter substitution going on.. Just extend the sed command, I prefer sed to pretty much everything anyways.. but it's always preferable to put in the extra effort to go for as much builtin use as you can. Otherwise it's not a top one-liner, it's a lazyboy recliner.

Here's the old version:

alias dateh='date --help|sed "/^ *%%/,/^ *%Z/!d;s/ \+/ /g"|while read l;do date "+ %${l/% */}_${l/% */}_${l#* }";done|column -s_ -t'

This trick from my [ http://www.askapache.com/linux-unix/bash_profile-functions-advanced-shell.html bash_profile ]

shmore(){ local l L M="`echo;tput setab 4&&tput setaf 7` --- SHMore --- `tput sgr0`";L=2;while read l;do echo "${l}";((L++));[[ "$L" == "${LINES:-80}" ]]&&{ L=2;read -p"$M" -u1;echo;};done;}
2010-04-21 00:40:37
User: AskApache
Functions: echo read
6
SH
cat mod_log_config.c | shmore

or

shmore < mod_log_config.c

Most pagers like less, more, most, and others require additional processes to be loaded, additional cpu time used, and if that wasn't bad enough, most of them modify the output in ways that can be undesirable.

What I wanted was a "more" pager that was basically the same as running:

cat file

Without modifying the output and without additional processes being created, cpu used, etc. Normally if you want to scroll the output of cat file without modifying the output I would have to scroll back my terminal or screen buffer because less modifies the output.

After looking over many examples ranging from builtin cat functions created for csh, zsh, ksh, sh, and bash from the 80's, 90s, and more recent examples shipped with bash 4, and after much trial and error, I finally came up with something that satisifed my objective. It automatically adjusts to the size of your terminal window by using the LINES variable (or 80 lines if that is empty) so

This is a great function that will work as long as your shell works, so it will work just find if you are booted in single user mode and your /usr/bin directory is missing (where less and other pagers can be). Using builtins like this is fantastic and is comparable to how busybox works, as long as your shell works this will work.

One caveat/note: I always have access to a color terminal, and I always setup both the termcap and the terminfo packages for color terminals (and/or ncurses and slang), so for that reason I stuck the

tput setab 4; tput setaf 7

command at the beginning of the function, so it only runs 1 time, and that causes the -- SHMore -- prompt to have a blue background and bright white text.

This is one of hundreds of functions I have in my http://www.askapache.com/linux-unix/bash_profile-functions-advanced-shell.html">.bash_profile at http://www.askapache.com/">AskApache.com, but actually won't be included till the next update.

If you can improve this in any way at all please let me know, I would be very grateful! ( Like one thing I want is to be able to continue to the next screen by pressing any key instead of now having to press enter to continue)

/bin/grep - ipranges.txt | while read line; do ipcalc $line ; done | grep -v deag
2010-04-20 21:13:00
User: tf8
Functions: grep read
7

Taking file with ip ranges, each on it's own line like:

cat ipranges.txt

213.87.86.160-213.87.86.193

213.87.87.0-213.87.88.255

91.135.210.0-91.135.210.255

command returns deaggregated ip ranges using ipcalc deaggregate feature like that:

213.87.86.160/27

213.87.86.192/31

213.87.87.0/24

213.87.88.0/24

91.135.210.0/24

Useful for configuring nginx geo module

ls | grep *.txt | while read file; do cat $file >> ./output.txt; done;
read -sn1 -p "Press any key to continue..."; echo
2010-04-13 20:37:26
User: vibaf
Functions: read
6

Just added -sn1

-s = silent

-n1 = only one symbol needed to continue after the insert

#sorry if my English isn't perfect ;P

read -p "Press enter to continue.."
2010-04-13 13:05:09
User: brubaker
Functions: read
4

Waiting for a key stroke. You can use this with a ";" behind to build a command chain.

du -cks * | sort -rn | while read size fname; do for unit in k M G T P E Z Y; do if [ $size -lt 1024 ]; then echo -e "${size}${unit}\t${fname}"; break; fi; size=$((size/1024)); done; done
wget randomfunfacts.com -O - 2>/dev/null | grep \<strong\> | sed "s;^.*<i>\(.*\)</i>.*$;\1;" | while read FUNFACT; do notify-send -t $((1000+300*`echo -n $FUNFACT | wc -w`)) -i gtk-dialog-info "RandomFunFact" "$FUNFACT"; done
2010-04-02 09:43:32
User: mtron
Functions: grep read sed wc wget
2

extension to tali713's random fact generator. It takes the output & sends it to notify-osd. Display time is proportional to the lengh of the fact.

while $8;do read n;[ $n = "$l" ]&&c=$(($c+1))||c=0;echo $c;l=$n;done
2010-03-31 00:41:08
User: florian
Functions: read
Tags: bash read Game
1

hold period (or whatever character) and hit enter after a second. You need to make the next line of periods the same length as the previous line... score starts at 0 and increase each time length of line is same.

count="1" ; while true ; do read next ; if [[ "$next" = "$last" ]] ; then count=$(($count+1)) ; echo "$count" ; else count="1" ; echo $count ; fi ; last="$next" ; done
2010-03-30 04:02:29
User: dabom
Functions: echo read true
Tags: bash read Game
8

Really bored during class so I made this...

Basically, you hold period (or whatever) and hit enter after a second and you need to make the next line of periods the same length as the previous line...

My record was 5 lines of the same length.

It's best if you do it one handed with your pointer on period and ring on enter.

ls | while read filename; do tar -czvf "$filename".tar.gz "$filename"; rm "$filename"; done
2010-03-29 08:10:38
User: Thingymebob
Functions: ls read rm tar
-2

Compresses each file individually, creating a $fileneame.tar.gz and removes the uncompressed version, usefull if you have lots of files and don't want 1 huge archive containing them all. you could replace ls with ls *.pdf to just perform the action on pdfs for example.

rpm --querytags | egrep -v HEADERIMMUTABLE | sort | while read tag ; do rpm -q --queryformat "$tag: [%{$tag} ]\n" -p $SomeRPMfile ; done
2010-03-25 05:40:48
Functions: egrep read rpm sort
0

If you want to relocate a package on your own, or you just want to know what those PREIN/UN and POSTIN/UN scripts will do, this will dump out all that detail simply.

You may want to expand the egrep out other verbose flags like CHANGELOGTEXT etc, as your needs require.

It isn't clear, but the formatting around $tag is important: %{$tag} just prints out the first line, while [%{$tag }] iterates thru multi-line output, joining the lines with a space (yes, there's a space between the g and } characters. To break it out for all newlines, use [%{$tag\n}] but the output will be long.

This is aside from rpm2cpio | cpio -ivd to extract the package files.

( last ; ls -t /var/log/wtmp-2* | while read line ; do ( rm /tmp/wtmp-junk ; zcat $line 2>/dev/null || bzcat $line ) > /tmp/junk-wtmp ; last -f /tmp/junk-wtmp ; done ) | less
2010-03-16 04:17:16
Functions: last ls read rm zcat
5

When your wtmp files are being logrotated, here's an easy way to unpack them all on the fly to see more than a week in the past. The rm is the primitive way to prevent symlink prediction attack.

find . -iname '*.mp3' | while read song; do mpg321 ${song} -w - | oggenc -q 9 -o ${song%.mp3}.ogg -; done
2010-03-14 11:34:35
User: renich
Functions: find mpg321 read
Tags: ogg mpg321
0

This is not recommended... lossy -> lossy = lossier.

Still, you can do it! ;)

while read f;do echo "$f";done < <(find .)
2010-03-02 14:22:22
Functions: echo find read
-6

Read all contents from current directory and display to stdout.

find . |while read f;do echo "$f";done
2010-03-02 14:21:15
Functions: echo find read
-8

Read all contents from current directory and display it on stdout.

ls *.wav | while read f; do lame "$f" -o "$(echo $f | cut -d'.' -f1)".mp3; done;
uri_escape(){ echo -E "$@" | sed 's/\\/\\\\/g;s/./&\n/g' | while read -r i; do echo $i | grep -q '[a-zA-Z0-9/.:?&=]' && echo -n "$i" || printf %%%x \'"$i" done }
2010-02-13 01:39:51
User: infinull
Functions: echo grep printf read sed
1

This one uses hex conversion to do the converting and is in shell/sed only (should probably still use the python/perl version).