Using OpenSSL we can encrypt any input we wish and then use Netcat to create a socket which can be connected to from an externally source (even using a Web Browser)

This will diff your local version of the file with the latest version in svn. I put this in a shell function like so:

middlehost allows ssh access from where you are but not securehost.

Use nice ssh piping to simulate scp through A => B => C

setting up the shell function if left as an exercise for the reader. ;-)

Agent forwarding should avoid password typing.

Uses the data in the /proc system, provided by the acpid, to find out the CPU temperature. Can be run on systems without lm-sensors installed as well.

du only accepts lines ending with a NUL, which can be a pain to create. This solves that issue.

Pipe any output to "grep ." and blank lines will not be printed.

If you want certain files out of a directory hierarchy, this will copy just the listed files, but will create the directory hierarchy in the new location ($DIR/)

Plays the sound of the file, should sound like *some* kind of music, most files sound like static but some are really cool.

variations:

Check out http://bbs.archlinux.org/viewtopic.php?id=70937 for more variations!

semi-dupe--like http://www.commandlinefu.com/commands/view/985/generate-white-noise but with different syntax and program.

Easy way to find out what Debian version your machine is running

If you are downloading a big file (or even a small one) and the connection breaks or times out, use this command in order to RESUME the download where it failed, instead of having to start downloading from the beginning. This is a real win for downloading debian ISO images over a buggy DSL modem.

Take the partially downloaded file and cat it into the STDIN of curl, as shown. Then use the "-C -" option followed by the URL of the file you were originally downloading.

This is an extract from a larger script which formats the video for DVD. The videos I use have no audio track so I need to add one. Tweak as you like...

use find to grep all .c files from the target directory, cat them into one stream, then piped to wc to count the lines

The command `cat file >> file` failes with the following error message:

cat: file: input file is output file

`tee` is a nice workaround without using any temporary files.

If you use newsgroups then you'll have come across split files before. Joining together a whole batch of them can be a pain so this will do the whole folder in one.

A common mistake in Bash is to write command-line where there's command a reading a file and whose result is redirected to that file.

It can be easily avoided because of :

1) warnings "-bash: file.txt: cannot overwrite existing file"

2) options (often "-i") that let the command directly modify the file

but I like to have that small function that does the trick by waiting for the first command to end before trying to write into the file.

Lots of things could probably done in a better way, if you know one...

This just combines multiple mp3's into one mp3 file. Basically it is a easy join for mp3's

Output manpage as plaintext using cat as pager: man -P cat commandname

And redirect its stdout into a file: man -P cat commandname > textfile.txt

Example: man -P cat ls > man_ls.txt

Busiest seconds:

Just a handy way to get all the unique links from inside all the html files inside a directory. Can be handy on scripts etc.