Power to the command line
The command line. That small place, where a lot can happen. And more so if you are a Linux user… How to maximize it? Where to harness its power?
I discovered commandlinefu.com a few years ago, while looking for a way to do… something. I don’t even enter that often, although it is a brilliant place to discover how to do X in Linux/UNIX.
Among its all time greats I found some gems, and some others I discovered elsewhere, or even I made up.
Command line gems
Run the last command as root: sudo !! as in
apt-get install something -> error, you have to be root
sudo !!
The double exclamation mark recovers the last command in the history.
Use the last argument for the last command: !$ as in
mkdir /path/to/somewhere
cd !$
Change to the previous directory: cd - as in (after the previous example to
get back to where you started). Good paired with popd (bash keeps a stack of
visited directories)
cd -
Backward history incremental search: C-r in the terminal prompt, start writing
the command you want to backward-search. Press C-r to keep on searching for that
term. Best combined with fzf
Correct typo in previous command: ^typo^corrected as in
sudo apt-get isntall somepackage
^isnt^inst
Write command in editor: C-x C-e will fire the editor in $EDITOR to write the
following command.
Open the last modified file of a type: I use this to open the most recent ppm
file in a directory with Eye Of Gnome: eog "$(ls -rt \*.ppm | tail -n 1)"
Use locate to search for a specific PDF and open it with evince: evince "$(locate \*partofname\*.pdf)" This will only work if it results in only one
instance. If you want to open only the first occurrence evince "$(locate \*partofname\*.pdf | head -n 1)"
The same with find: evince "$(find -name 'NameOfPdf.pdf')"
Mighty heads and tails: head -n N file, tail -n N file will return the first
(resp. last) N lines of file.
Get pid of a process by name: ps ax | grep "firefox"
Follow changes on a incremental file: tail -f filename as in following nohup.out
from a running process.
Get all lines containing a string in a file: grep -e "string" file > outputfile
Useful as intermediate step for plotting specific data lines from a nohup.out
file. With -r, change the string for a
regexp (consider buying this
book if you will use them more than once, I recommend it: Mastering Regular
Expressions).
I think I have a few more, but I just don’t remember them.
Note from 2019: This was written a long time ago, and my list of tips and speed ups for bash/zsh is way larger now. Hit me on twitter if you want to know more.