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/
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.
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.
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:
say you want to reinitialize the slave database without resetting the master positions. You stop the slave, dump the master database with --master-data=2 then execute the command on the slave and wait for it to stop at the exact position of the dump. reinit the slave db and start the slave. enjoy.
Here $HOME/shots must exist and have appropriate access rights and sitecopy must be correctly set up to upload new screen shots to the remote site.
Example .sitecopyrc (for illustration purposes only)
The command uses scrot to create a screen shot, moves it to the screen shot directory, uploads it using screen uses xsel to copy the URL to the paste buffer (so that you can paste it with a middle click) and finally uses feh to display a preview of the screen shot.
Note that $BASE stands for the base URL for the screen shots on the remote server, replace it by the actual location; in the example http://www.example.com/~user/shots would be fitting.
Assign this command to a key combination or an icon in whatever panel you use.
A web server using $HOME/public_html as user directory is required, $HOME/public_html/shots must exist and have appropriate access rights and $HOSTNAME must be known to and accessible from the outside world.
The command uses scrot to create a screen shot, moves it to the screen shot directory, uses xsel to copy the URL to the paste buffer (so that you can paste it with a middle click) and finally uses feh to display a preview of the screen shot.
Assign this command to a key combination or an icon in whatever panel you use.
Generate the iso from the disk, easily.
same as "dd if=/dev/cdrom of=~/mydisk.iso"
"unzip *.zip" doesn't work as expected, because unzip handle wildcards in a different way.
You just need to escape the wildcard or do in another way:
for f in *.zip; do unzip "$f"; done
xclip -o > /tmp/spell.tmp # Copy clipboard contents to a temp file
aspell check /tmp/spell.tmp # Run aspell on that file
cat /tmp/spell.tmp | xclip # Copy the results back to the clipboard, so that you can paste the corrected text
I'm not sure xclip is installed in most distributions. If not, you can install x11-apps package
If you enable multiuser, then you can permit others to share your screen session. The following conditions apply:
1. screen must be suid root;
2. "multiuser on" must be configured in ~/.screenrc;
3. control the others user(s) access with "aclchg":
# ----- from ~/.screenrc-users -----
aclchg someuser +rx "#?" #enable r/o access to "someuser"
aclchg someuser -x "#,at,aclchg,acladd,acldel,quit" # don't allow these
aclchg otheruser +rwx "#?" # enable r/w access to "otheruser"
aclchg otheruser -x "#,at,aclchg,acladd,acldel,quit" # don't allow them to use these commands
After doing this (once), you start your session with:
Then, the other user can join your terminal session(s) with youruserid:
$ screen -r youruserid/
Note: the trailing "/" is required.
Multiple users can share the same screen simultaneously, each with independent access controlled precisely with "aclchg" in the ~/.screenrc file.
I use the following setup:
# default screenrc on any host
Then, the base configurations are in ~/.screenrc-base; the host-specific configurations are in ~/.screenrc-$HOST, and the user configurations are in ~/.screenrc-users.
The host-specific .screenrc file might contain some host-specific screen commands; e.g.:
screen -t 'anywhere' /bin/tcsh
screen -t 'anywhere1' /bin/tcsh
The .screenrc-base contains:
## I find typing ^a (Control-a) awkward. So I set the escape key to CTRL-j instead of a.
termcapinfo xterm* ti@:te@:
This function runs a program in the background, and logs all output to an automatically created logfile. That way, you can still get at the output without it clogging up your terminal.
Throw fork() and this:
for prog in firefox kate konqueror ;do alias $prog="fork $prog";done
into your bashrc, so that they'll automatically run out of the way.
You can't stand programs x, y, and z. Remove all trace of their existence by adding this function to your config. It will remove the cruft, the settings, and such and such. This function doesn't even give a damn about you trying to remove programs that don't exist: it'll just for loop to the next one on your hit list.
Have netcat listen on port 8000, point browser to http://localhost:8000/ and you see the information sent. netcat terminates as soon as your browser disconnects.
I tested this command on my Fedora box but linuxrawkstar pointed out that he needs to use
nc -l -p 8000
instead. This depends on the netcat version you use. The additional '-p' is required by GNU netcat that for example is used by Debian but not by the OpenBSD netcat port used by my Fedora system.
'watch' repeatedly (default every 2 seconds, -n 1 => every second) runs a command (here ':', a shorthand for 'true'), displays the output (here nothing) and the date and time of the last run.
I thought it to be obvious but it seemingly is not: to exit use Ctrl-C.
Finds all files of a certain name and reports all line with the string. Very simple.
If you have lots of remote hosts sitting "behind" an ssh proxy host, then there is a special-case use of "rsynch" that allows one to easily copy directories and files across the ssh proxy host, without having to do two explicit copies: the '-e' option allows for a replacement "rsh" command. We use this option to specify an "ssh" tunnel command, with the '-A' option that causes authentication agent requests to be forwarded back to the local host. If you have ssh set up correctly, the above command can be done without any passwords being entered.
Mass suspends all User accounts on a Cpanel server and inputs a RedirectMatch to the Rick Roll video. Learn more at rainbowblast com
If your user has sudo on the remote box, you can rsync data as root without needing to login as root. This is very helpful if the remote box does not allow root to login over SSH (which is a common security restriction).
Many Mac OS X programs, especially those in Microsoft:Office, create ASCII files with lines terminated by CRs (carriage returns). Most Unix programs expect lines separated by NLs (newlines). This little command makes it trivial to convert them.
This little command (function) shows the CSV header fields (which are field names separated by commas) as an ordered list, clearly showing the fields and their order.
Coming back to a project directory after sometime elsewhere?
Need to know what the most recently modified files are?
This little function "t" is one of my most frequent commands.
I have a tcsh alias for it also:
alias t 'ls -ltch \!* | head -20'
Ever wanted to find the most recently modified files, but couldn't remember exactly where they were in a project directory with many subdirectories? The "find" command, using a combination of "-mtime -N" and "-depth -D" can be used to find those files. If your directory structure isn't very deep, just omit the "-depth -D", but if your directory structure is very deep, then you can limit the depth of the traversal using "-depth -D", where "D" is the maximum number of directory levels to descend.
Finds all directories containing more than 99MB of files, and prints them in human readable format. The directories sizes do not include their subdirectories, so it is very useful for finding any single directory with a lot of large files.
In this case, I'm getting the package version for 'redhat-release', but of course, this can be applied to any package installed on the filesystem. This is very handy in scripts that need to determine just the version of the package, without the package name and all the sed and grep hackery to get to the data you want. To find out all the support format strings that 'rpm --qf' supports:
faster than lsof by at least x2 on my box.