Howto: Sync Bash History Between Terminals

Do you want to be able to use the history from another terminal quickly and simply in a new terminal, or to be able to use any one of the terminals on your desktops to continue working with a current history?

Me too :]

This is a very easy tweak and makes a lot of sense to most people. So we need to edit the file ~/.bashrc

gedit ~/.bashrc

At the bottom paste in the following:

shopt -s histappend
PROMPT_COMMAND="history -n; history -a"
unset HISTFILESIZE
HISTSIZE=2000

Unsetting HISTFILESIZE just gets rid of the filesize limitation, and we rely on HISTSIZE which allows 2000 lines in the file, both normally default to 500.

PROMPT_COMMAND tells bash to reload history and append to history each time it draws the prompt in the terminal.

The only thing to note is if you jump from one terminal to another and wish to recall the last history line from the former terminal, press enter once with a blank line, this causes bash to draw the prompt and sync, if you only just opened the second terminal then you don't have to. The sync only occurs when the prompt is drawn, nothing else to note.

Make Pasword file Immutable - A Security tip

You can add a extra layer of protection to your machine, or in my case help to prevent wannabe haxx0r brothers at the parents house from changing to extremely short passwords with a little trick that not many are aware of.

What happened to me was my brother managed to shoulder surf the admin account on my parents PC, I normally update their PC by SSH from home, but I was around there visiting and mum had described a problem so I was fixing it while we chatted, my brother was lurking around and managed to shoulder read my 12 digit alpha numeric password (have to give the guy a bit of credit there, sadly)

Anyway I became aware that files on the admin account had incorrect accessed times afterwards and then looked deeper and found that someone had been in on the account and had changed my brothers password on his account from a sensible 10 digit alphanumeric (which was dead simple to bang in with two fingers as the digits were paired on the keyboard) into a rather less safe two letter password. He claimed that a friend had helped him break into my account as "you aren't the only one that knows linux", as if I ever imagined that I was some guru, I'm a self confessed learner still! As we "discussed" this further he came clean that he had simply watched me enter the password rather than "hacked" in.

Anyway lets get onto the crux of this, after setting your passwords open a terminal and enter:
chattr +i /etc/shadow

This will make the file immutable, which means that it cannot be deleted or renamed, no symlink can be made to this file and no data can be written to the file. The immutable flag must be cleared before passwords can be changed, if you try to do so without clearing it then it will appear to change the password, but in fact it would not write to the file, without error. Meaning the new password would never be written and the old password would continue being the active one.

If a fairly knowledgeable user tries to dig into why it won't write and they try:
$ ls -l /etc/shadow
-rw-r----- 1 root shadow 1027 2009-08-25 14:37 shadow

There is nothing obviously wrong with writing to the file, to check properly you would use:
$ sudo lsattr shadow
----i------------- shadow

Here we can see the +i flag set for immutable. If we clear that with:
chattr -i /etc/shadow
The password can now be changed as normal, using sudo passwd username
and then it can be set +i again to protect it.

A little bit sneaky and just a little bit safer.

Howto: SSH Public Key Authorisation Login (AKA Passwordless)

To be clear, this is how to login using SSH to another machine without having to enter your password. It is not an all encompassing solution but a "good enough" set up. I have to administer a couple of machines and a shell which I have use of, my own machine is kept up to date and I feel is secure enough with only myself as sole user.

Firstly we should make a key pair if you haven't already done so, these are deposited into into ~/.ssh

ssh-keygen -t rsa

Next we need to get the contents of ~/.ssh/id_rsa.pub onto our remote machine and into ~/.ssh/authorized_keys this can be done a couple ways, I found it as easy to just ssh into the remote machine and paste the contents directly into the file just note that if you paste it in you may have to correct the formatting so it is all on one line, I noticed a couple of line breaks sneak in. You could use something like the following however:

scp -P 20060 ~/.ssh/id_rsa.pub remoteaccount@remotesserver.com

Then you can just append it automatically with the following (we will touch the file in case it doesn't already exist on your machine:

touch ~/.ssh/authorized_keys
cat id_rsa.pub >> ~/.ssh/authorized_keys

We should at this point make the .ssh folder only readable by the owner:

chmod 700 ~/.ssh

You can optionally delete the id_rsa.pub key file if it was in your home folder, its the public key part so its not really necessary in my opinion but it is nice to tidy up after yourself ;)

rm id_rsa.pub

If you try and login now to the remote machine with SSH you should with any luck find you do not have to enter your password at all and can just get right on with your work.

How to use OTF fonts in Ubuntu

It isn't possible to use OTF fonts directly in Ubuntu without converting them into TTF's.

If we spend just a few minutes and install FontForge and make a very small script it becomes almost trivial to do.

sudo apt-get install fontforge

Now we should make the scripts, well it is two actually but one is a small config script for FontForge and the other is just a script to convert many OTF files in one go, rather than individually. First the FontForge settings file.

cd ~/bin
touch otf2ttf.conf

Then open otf2ttf.conf in your favourite text editor and paste the following into it.

#!/usr/local/bin/fontforge
# Quick and dirty hack: converts a font to truetype (.ttf)
Print("Opening "+$1);
Open($1);
Print("Saving "+$1:r+".ttf");
Generate($1:r+".ttf");
Quit(0);

Save the file, and now lets make the script to convert many fonts in one swoop.

cd ~/bin
touch otf2ttf
chmod +x otf2ttf

Again, open this file with your prefered text editor and paste the following in.

#!/bin/bash
#
# had to enable extglob
shopt -s extglob

# this uses globbing to match fiels ending in otf/OTF
for i in +(*.otf|*.OTF)

do fontforge -script /home/subbass/bin/otf2ttf.conf $i

done

To use this now (provided that your ~/bin folder is in your path) just open a terminal where your OTF files are, and issue the command otf2ttf. It will convert 20 files in just a couple of seconds to give you an idea of speed, you can then move all the resulting TTF files into ~/.fonts


Credit to http://www.stuermer.ch/blog/convert-otf-to-ttf-font-on-ubuntu.html where I found the script before re-typing the instructions. I put it here so I don't lose it next time I need it.

Set up rxvt terminal with unicode and clickable URL's

Rxvt is a super fast unicode terminal which I have recently switched to away from Eterm for my transparent desktop terminal.

sudo apt-get install rxvt-unicode-ml

Next is to create a configuration file for it in your home folder.

touch .Xdefaults

Now paste this lot in...

URxvt*termName: rxvt

## borderless and no scrollbar
URxvt*scrollBar_right: false
URxvt*scrollBar: false
URxvt*borderLess: false

## teh transparency stuff
URxvt*inheritPixmap: true
URxvt*tintColor: white
URxvt*shading: 100

##
## These 3 lines make links clickable
##
URxvt.urlLauncher: firefox
URxvt.matcher.button: 1
URxvt.perl-ext-common: matcher,readline

## geometry and font
URxvt*geometry: 80×15
URxvt*font: xft:Terminus:pixelsize=12

## change default colors
URxvt*background: #000000
URxvt*foreground: #A8A8A8
URxvt*color0: #000000
URxvt*color1: #A80000

## URxvt*color2: #00A800
URxvt*color2: #ED254F

URxvt*color3: #A85400
## main bars in irssi
URxvt*color4: #020202
URxvt*color5: #A800A8
URxvt*color6: #00A8A8
URxvt*color7: #A8A8A8
URxvt*color8: #545054
URxvt*color9: #F85450

## Time in Irssi
URxvt*color10: #ED254F
## URxvt*color10: #50FC50

## URxvt*color11: #F2FC50
URxvt*color11: #ED254F

URxvt*color12: #5054F8

URxvt*color13: #ED254F
## URxvt*color13: #F854F8

URxvt*color14: #50FCF8
URxvt*color15: #F8FCF8

Note that some of the colours have been changed to coordinate with my current desktop theme, the original colours are commented out so you can easily revert. Also the line defining the window geometry seems to have no effect, I shall look into that and correct it when I get some time and inclination. At the moment I don't mind because the default window size is fine, and the launcher I use to start the transparent terminal I use on the desktop mainly for my screen with irssi, hellanzb and mediatomb in has the geometry set on it. the launcher command is:

rxvt -g 80x15

Last job is to tell compiz to not decorate this window, this will get rid of the title bar, borders and shadow etc.

Hopefully you have installed the advanced compiz settings (ccsm), so just open that up and go to the section for window decorations in Effects, Window Decorations.

Add the rxvt window to be excluded from shadows and decorations, and that should see you done.

Install and use ClamAV for linux anti virus

Install Clamav with either Synaptic or apt-get/aptitude.

To update the definitions use:

sudo freshclam

scanning files is as simple then as calling clamscan with the -r or --recursive flag and a path, if the path is omitted then it will scan at the current path:

clamscan -r /home/subbass
or clamscan -r

If you would rather only see infected files then use the -i or --infected flag:

clamscan -ri /home/subbass
or clamscan -ri

If you wish to scan every file on the system you may have to run clamscan with sudo, as running without it clamscan can only read files the user running it can access.

sudo clamscan -ri /

This would check every file on the system recursively and report only infected files.

You can schedule clamscan using cron or for the odd one off scans using the "at" command, such as:

at 3:30 tomorrow
at>clamscan -i /home/user > mail user@example.com
at> 
job 3 at 2005-04-28 03:30

This would perform the scan at 3:30am later that night and mail the results to the defined address. To add a regular cron job try the following:

nano crontab -e

Then enter the following line at the bottom of the file

00 00 * * * sudo clamscan -r /location_of_files_or_folders

Save the file and exit.

rTorrent, light and fast bittorrent client.

I use two torrent clients almost every day, Transmission on my own box with a gnome desktop, and for a long time now rTorrent in a shell via SSH to a spare machine I have use of on a separate connection. Setting up rTorrent on an Ubuntu machine is what I am going to discuss here.

First step is to install the program, open a terminal and enter:

sudo apt-get install rtorrent

Or use Synaptic and search for 'rtorrent', install it. A basic config file is located /usr/share/doc/rtorrent/examples for you to tweak to your own liking, at its simplest you may want to change the download folder in it and the 'watch' folder. The watch folder is a place that rTorrent will monitor for *.torrent files coming into so it can automatically start, and when the download has completed and achieved a ratio you are happy with, you can just delete the torrent file from this folder to remove it from rTorrent. Dead simple stuff :]

cp /usr/share/doc/rtorrent/examples/rtorrent.rc ~/.rtorrent.rc

then edit this file with your prefered text editor, mine is usually nano, use gedit if you prefer a GUI editor:.

cd ~/
nano .rtorrent.rc

Now we can make the most basic of changes, first up is the download folder, look for the line:

# Default directory to save the downloaded torrents.

change the line after it to the folder you want to be your download folder and remove the # so you have something like this:

# Default directory to save the downloaded torrents.
directory = ~/torrents

This makes my download folder /torrents/ in my home, the default without editting is simply your home folder which you may be happy with.

Next lets tell it what folder to watch for torrent files. I have defined mine to be this same ~/torrents/ folder, you may prefer to leave it as your home folder. Either way find these lines just slightly down from the previous:

# Watch a directory for new torrents, and stop those that have been
# deleted.
schedule = watch_directory,5,5,load_start=~/torrents/*.torrent
schedule = untied_directory,5,5,stop_untied=

Yours will look very slightly different as I have shown how I changed mine to watch the ~/torrents/ folder. You should now save the file and close the editor, ctrl+o to save in nano if you used my prefered editor from above and don't know it, press enter to save wit hthe same name, then ctrl+x to close it.

You should now be free to start rTorrent with the command:

rtorrent

Not very impressive at first look is it, but download a torrent file to the watch folder and see rtorrent leap into action all by itself. You can adjust the upload and download speeds using the keys

a/s/dIncrease the upload throttle by 1/5/50 KB.
z/x/cDecrease the upload throttle by 1/5/50 KB.
A/S/DIncrease the download throttle by 1/5/50 KB.
Z/X/CDecrease the download throttle by 1/5/50 KB.

There are other keys to pause and resume torrents and many other facilities that you may require, for that I would point you at the User Guide page at http://libtorrent.rakshasa.no/wiki/RTorrentUserGuide

As I use this remotely I also run it inside a screen so I can safely detach from it and break the connection leaving it running on the remote machine, to do this enter in a terminal:

screen

then you will just see another command prompt on a clear terminal at which point enter:

rtorrent

rTorrent will start as normal but if you use the keys ctrl+a then d it will detach, that is it will drop you back at your original shell, leaving rtorrent running in the backgroun still. You can attach to it again with:

screen -r

You can then see rTorrent again to check your downloads or ratios. Check 'man screen' for more help on using screen, there are a *lot* of things you can do with it, and that might be a worthwhile post for me some time soon.

Recurse to find files and move them to a location

I just had to move a bunch of files out of individual sub-folders into the parent folder, time consuming to do that 50 or so times so a quick check of the "find" command and the solution is here:

find . -iname '*.avi' -exec mv {} /home/subbass \;

Command breakdown looks like this:

find .
find "here"
-iname '*.avi'
case insensitive name match on *.avi
-exec mv
execute the move command on the matches
{}
the match result from find
/home/subbass/
path to move the files to
\;
end of the -exec, each match runs as a new command

A quick command line, fairly simple syntax and it should prove a great time saver. You could easier of course cp or rm files instead of mv or many other possibilities.