Tips and Tricks About Computers, Web Development, Linux, the Internet and the Like
Bash
Batch images conversion – color to transparency from the linux terminal
Mar 31st
I’ve been re-theming and old forum and the themes that are available are a little less than perfect for our needs. I found a few suitable templates but it still results in me creating a mash-up. Some of the images (though gif) set the image background color to that of the site background. As I was using a different color, this obviously looked bad but I wasn’t about to manually edit all the files to give them transparency.
Being aware of the amazing command line tool for Linux, ImageMagick, I set out to find a tool to help me. I found a transparency flag, calculated the RGB values with the eyedropper in GIMP, then after a test run, I through it in a for loop.

So a directory down from my image source, I created a folder called “new” and ran the following color and like magic, they were all converted.
for image in *.gif; do convert -transparent 'RGB(48,71,94)' $image ../new/$image; done
Finding files and strings using the terminal in Linux
Feb 4th
My favorite thing about Linux is the terminal. I use it countless times a day to do all sorts of tasks, like managing game servers or writing scripts to do tedious tasks. One of the most popular things I do in the terminal is search for files or strings inside of files and today I’d like to go over a few methods and tricks for doing this. There are three tools that make this task amazingly easy but combining them is where we find the real power. These three tools are locate, find and grep. I will cover the basic use of these tools with some examples and tricks but I suggest you take a look at the man pages for the tools for addition information (i.e. man locate).
locate
Locate has got to be the most straight forward way to search for files. It uses a database so it’s blazing fast when searching your entire computer, compared to the ‘find’ we’ll cover later. It’s not available by default on all Linux distributions (it is on Ubuntu) so you might have to install it. Since I’ll be mentioning fstab later, I’ll give my examples with it.
Basic syntax is: locate [filename]
tyler@quadjutsu:~$ locate fstab /etc/fstab /etc/fstab.orig /etc/fstab.pre-ntfs-config /etc/fstab~ /usr/include/fstab.h /usr/lib/udev/migrate-fstab-to-uuid.sh /usr/share/apps/katepart/syntax/fstab.xml /usr/share/doc/m4/examples/fstab.m4 /usr/share/doc/mount/examples/fstab /usr/share/doc/util-linux/examples/fstab.example2 /usr/share/man/man5/fstab.5.gz /usr/share/pysdm/fstab.py /usr/share/pysdm/fstab.pyc
This database will automatically update itself at various times but to force an update type the following command:
tyler@quadjutsu:~$ sudo updatedb
Folder exclusions can be found (and edited) in /etc/updatedb.conf
By default (at least in ubuntu) /media (your mounted media) is not included in this database. This means if you use extra drives you’ve added to your computer and you want them to be searchable through the locate tool, you’ll have to mount them to a directory like /mnt.
To mount a drive on boot, you’ll need to add a line like the following to your /etc/fstab. The one below mounts an ntfs drive to /mnt/mydrive. That folder must exist for the drive to be mounted.
/dev/sda1 /mnt/mydrive ntfs-3g defaults,locale=en_US.utf8 0 0
I found the /dev/sda1 part by listing my harddrive partitions using the following command:
sudo fdisk -l
find
Find is a little more cryptic than locate but it’s a very powerful tool that should be available on most Linux distributions if not all.
Basic syntax is: find [directory] -name “[string]” -print
Find will search recursively through the folder you’re calling it from. The directory isn’t required but it will show you the full path. A neat trick is to use $(pwd) which creates a string for the “present working directory”. It’s also a good habit to use quotes around your name search because you can’t do regular expressions without it. find does not do matching the same way locate does. You’ll have to specific a wild card (*) for partial searches.
tyler@quadjutsu:~/Desktop$ find -name "notes" -print tyler@quadjutsu:~/Desktop$ find -name "notes*" -print ./notes.txt~ ./notes.txt tyler@quadjutsu:~/Desktop$ find $(pwd) -name "notes*" -print /home/tyler/Desktop/notes.txt~ /home/tyler/Desktop/notes.txt tyler@quadjutsu:~/Desktop$ find -name "no[a-z]*" -print ./notes.txt~ ./notes.txt
My Buddy From Belgium, MrBougo has asked I make note that -iname makes it case insensitive.
grep
In the context of searching, grep is like a pocket knife. It’s great for limiting returned results and searching through files, which are my two most common uses of it, though I’m sure there are a million others. First I’ll cover the searching files for strings portion and in combined tools we’ll discuss how to refine search results.
In terms of searching strings in files, basic syntax is: grep “[string]” [filename]
grep is case sensitive but can be changed to insensitive with -i. You can access extended regular expressions (which allow for such functions as use of the + sign to signify 1 or more characters) by calling “egrep”
tyler@quadjutsu:~$ grep "tyler" resume.txt tyler@detrition.net tyler@quadjutsu:~$ grep "Tyler" resume.txt Tyler J. Mulligan tyler@quadjutsu:~$ grep -i "tyler" resume.txt Tyler J. Mulligan tyler@detrition.net tyler@quadjutsu:~$ grep -i "tyl" resume.txt Tyler J. Mulligan tyler@detrition.net tyler@quadjutsu:~$ egrep -i "tyler [a-z.]+" resume.txt Tyler J. Mulligan tyler@quadjutsu:~$ egrep "tyler [a-z.]+" resume.txt
MrBougo also mentioned that, grep returns a 1 exit status if it doesnt find, so grep “foo” && grep “bar” will only grep for bar when foo is found.
Combining the tools
There are two very important techniques for linking together commands in the terminal. The pipe | and &&. The pipe passes the output of one command to the next, the && runs a command after the one before is complete.
Starting off slow, we’ll refine our search of grep with another grep. Note that you don’t have to use the quotes as I’ve shown below.
tyler@quadjutsu:~$ grep -i "nexuiz" resume.txt
* Web Developer and Interaction Designer for Nexuiz / Alientrap
* Owner and Creator of Nexuiz Ninjaz
tyler@quadjutsu:~$ grep -i "nexuiz" resume.txt |grep -i ninjaz
* Owner and Creator of Nexuiz NinjazNot that you would really want to do the follow commands but to emphasize how | and && work, the following example shows how you’re running the command twice rather than refining a search:
tyler@quadjutsu:~$ grep -i "nexuiz" resume.txt && grep -i "ninjaz" resume.txt
* Web Developer and Interaction Designer for Nexuiz / Alientrap
* Owner and Creator of Nexuiz Ninjaz
* Owner and Creator of Nexuiz Ninjaz
Recalling our first example with fstab, there was a lot of extra results we didn’t want. We know know it was in the etc folder, so we can filter our results by that.
tyler@quadjutsu:~$ locate fstab |grep etc /etc/fstab /etc/fstab.orig /etc/fstab.pre-ntfs-config /etc/fstab~
Another way would be to EXCLUDE folders we don’t want with the -v flag.
tyler@quadjutsu:~$ locate fstab |grep -v usr /etc/fstab /etc/fstab.orig /etc/fstab.pre-ntfs-config /etc/fstab~
And filtering out the junk like so:
tyler@quadjutsu:~$ locate fstab |grep etc |grep -v "~" |grep -v "\." /etc/fstab
~ has a special meaning in the terminal so you must quote it to match it. ~ refers to the user’s home directory, in this case /home/tyler. I’ve also had to escape . because it too has special meaning, in the context of regular expressions, it will match any character… if we’re excluding any character, we’re excluding our results :).
Quick Tricks
Remember commands is sometimes a hard thing to do, especially when it’s a whole string of complicated commands and regexes. Sometimes, I know I don’t have the mind to take notes but that’s okay (for a grace period) because bash keeps track for me. The history command will list all your recently executed commands and utilizing the grep command, we can filter that list.
tyler@quadjutsu:~$ history |grep locate 613 locate fstab 614 locate fstab |grep etc 624 locate fstab |grep etc |grep -v "~" |grep -v "\."
A quick way to search for files in a directory would be to use locate as the path and grep the results
tyler@quadjutsu:~$ locate ~ |grep notes.txt /home/tyler/Desktop/notes.txt
These are the basics of finding files and searching through them. Linux provides many tools to reformat and replace strings. There are a million different ways to combine commands and always a faster way to do it. Keep playing and you’ll just get better and better.
Ubuntu Window Management with Multiple Monitors, Window Effects and Default File Associations
Nov 25th
Multiple Monitors Window Management in Ubuntu
Moving your applications from one monitor to the next with hotkeys
I’ve been using multiple monitors for a while now, Starting with a 17″ CRT with a 19″ CRT and moving up to two 19″ LCD, then 3, then temporarily one wide screen and back to 2. Something I always loved having as a utility in ultramon that I couldn’t find in Ubuntu (Gnome) or any window manager for that matter, was the ability to move applications from monitor to monitor. I had assumed the search futile until I was searching about some questions I had concerning Compiz-fusion with dual monitors and I came about this thread on Ubuntu forums which brightened up my day. A fellow named gfixler posted a bash script that utilizes command line applications to move the windows.
For you multi-monitor users seeking salvation from removing your hand from the keyboard to move your application from one monitor to another, here’s the skinny on getting it setup using compiz-fusion, aka Advanced Desktop Effects, to set my keybinds.
1. Open a terminal and setup your prerequisites with apt-get:
sudo apt-get install wmctrl xprop xwininfo
If you get errors about x11-utils, just ignore them, this package will handle your needs.
2. Next, lets put the script somewhere you can call it, say “~/scripts”
mkdir ~/scripts && cd ~/scripts && touch movewin.sh && chmod +x movewin.sh && gedit movewin.sh
2. Paste the following code, find the first function “getNumberOfMonitors” and configure it to the number of monitors you have (default 2).
#!/bin/bash
# swap_monitor.sh (original version)
# Moves the active window to the other screen of a dual-screen Xinerama setup.
#
# movewin.sh (modified version)
# allows movement of windows left and right between multiple monitors
#
# Requires: wmctrl, xprop, xwininfo
#
# Original Author: Raphael Wimmer
# raphman@gmx.de
#
# Modified by: Gary Fixler
# gfixler+bash@gmail.com
function getNumberOfMonitors
{
# simply must be hardcoded
# e.g. MatroxTripleHead2Go can service 3 screens,
# but appears as only one monitor to the computer
# change to your number of monitors
echo 2
}
function getMonitorWidth
{
numberOfMonitors=$(getNumberOfMonitors)
monitorLine=$(xwininfo -root | grep "Width")
monitorWidth=$((${monitorLine:8}/$numberOfMonitors ))
echo $monitorWidth
}
function getActiveWindowID
{
activeWinLine=$(xprop -root | grep "_NET_ACTIVE_WINDOW(WINDOW)")
activeWinID="${activeWinLine:40}"
echo $activeWinID
}
function getActiveWindowHorizontalPosition
{
activeWinID=$(getActiveWindowID)
xPosLine=$(xwininfo -id $activeWinID | grep "Absolute upper-left X")
xPos=${xPosLine:25}
echo $xPos
}
function getActiveWindowWidth
{
activeWinID=$(getActiveWindowID)
xWidthLine=$(xwininfo -id $activeWinID | grep "Width")
xWidth=${xWidthLine:8}
echo $xWidth
}
function getActiveWindowCurrentMonitor
{
numberOfMonitors=$(getNumberOfMonitors)
monitorWidth=$(getMonitorWidth)
activeWinID=$(getActiveWindowID)
xPos=$(getActiveWindowHorizontalPosition)
i="0"
while [ $xPos -gt $monitorWidth ]
do
xPos=$[$xPos-$monitorWidth]
i=$[$i+1]
done
echo $i
}
function getActiveWindowPositionOneMonitorToTheLeft
{
monitorWidth=$(getMonitorWidth)
currentMonitor=$(getActiveWindowCurrentMonitor)
activeWinID=$(getActiveWindowID)
xPos=$(getActiveWindowHorizontalPosition)
xPos=$[$xPos-$monitorWidth]
echo $xPos
}
function getActiveWindowPositionOneMonitorToTheRight
{
monitorWidth=$(getMonitorWidth)
numberOfMonitors=$(getNumberOfMonitors)
currentMonitor=$(getActiveWindowCurrentMonitor)
activeWinID=$(getActiveWindowID)
xPos=$(getActiveWindowHorizontalPosition)
xPos=$[$xPos+$monitorWidth]
echo $xPos
}
function changeActiveWindowMonitor
{
activeWinID=$(getActiveWindowID)
if [ $1 -eq "0" ]
then
newXPos=$(getActiveWindowPositionOneMonitorToTheLeft)
newXPos=$[$newXPos-5]
else
newXPos=$(getActiveWindowPositionOneMonitorToTheRight)
newXPos=$[$newXPos-5]
fi
winState=$(xprop -id ${activeWinID} | grep "_NET_WM_STATE(ATOM)" )
if [[ `echo ${winState} | grep "_NET_WM_STATE_MAXIMIZED_HORZ"` != "" ]]
then
maxH=1
wmctrl -i -r ${activeWinID} -b remove,maximized_horz
fi
if [[ `echo ${winState} | grep "_NET_WM_STATE_MAXIMIZED_VERT"` != "" ]]
then
maxV=1
wmctrl -i -r ${activeWinID} -b remove,maximized_vert
fi
if [[ `echo ${winState} | grep "_NET_WM_STATE_FULLSCREEN"` != "" ]]
then
fulls=1
wmctrl -i -r ${activeWinID} -b remove,fullscreen
fi
# move window (finally)
wmctrl -i -r ${activeWinID} -e 0,${newXPos},-1,-1,-1
# restore maximization
((${maxV})) && wmctrl -i -r ${activeWinID} -b add,maximized_vert
((${maxH})) && wmctrl -i -r ${activeWinID} -b add,maximized_horz
((${fulls})) && wmctrl -i -r ${activeWinID} -b add,fullscreen
# raise window (seems to be necessary sometimes)
wmctrl -i -a ${activeWinID}
}
function moveActiveWindowOneMonitorToTheLeft
{
changeActiveWindowMonitor 0
}
function moveActiveWindowOneMonitorToTheRight
{
changeActiveWindowMonitor 1
}
"$1"
exit 0
3. Setup your hot keys with compiz-fusion. Go to System >> Preferences >> Advanced Desktop Effects. Inside “General Options“, click on the command tab (I apologize for my heinous blue links).
Use
scripts/./movewin.sh moveActiveWindowOneMonitorToTheRight
and
scripts/./movewin.sh moveActiveWindowOneMonitorToTheLeft
respectively
Per Application Window Effects in Ubuntu
Bring character and tickle your soul with per application window effects
Another cool feature Compiz-fusion has is window animations. My friend James Lindsay recently reminded me about Window Effects… which when I first install Ubuntu on my laptop, I experimented my butt off… but being a laptop… I just used simple ones I’d turn off half the time anyway. He asked me why I don’t use them on my desktop and I didn’t have a good reason. Well, now I have 2 good reasons to keep using compiz.
I made my Thunderbird use the airplane effect so when I send emails, it flys away and for Geany, I used the magic lamp for open, close, maximize and minimize (different speeds). It’s a fun little effect that breaks up the stiffness of the desktop.
Default File Associations in Ubuntu
geany > gedit
I was tired of gedit popping up when geany’s just as lightweight but more affective. So found a command and altered it a bit to make my default editor geany.
1. Open the terminal and create\open the following file:
gedit ~/.local/share/applications/defaults.list
If it’s blank, add “[Default Applications]“. If it’s not, find “[Default Applications]“.
2. Then, back to the terminal, grep the default files associations and replace gedit with your editor of choice
grep gedit /usr/share/applications/defaults.list | sed s/gedit/geany/g
Copy (ctrl+shift+c) and paste the output into gedit, below the “[Default Applications]” header.
3. Restart nautilus to load the changes (will close all your file managers that are open and blink/freeze your desktop for a second)
pkill nautilus
Good luck, have fun and happy coding :)
Bash script to convert pngs to jpgs, thumbnail them and generate forum code
Aug 11th
I’ve been working on a tutorial for Nexuiz Ninjaz that requires a lot of screen shots. I’ve been utilizing the command line tool scrot to take screenshots on specific windows. These images however, are large pngs because the scrot date pattern wasn’t working properly when I tried to generate jpgs while repeating the same command. I went forward knowing I could find a way to batch compress these pngs later. Figuring I might as well “add a turbo while I’m under the hood”, I opted to both thumbnail the images (duh) and generate the forum thumbnail code (VROOOM).
So, to convert pngs to jpgs, I created a bash script which uses mogrify to convert, compress and thumbnail them and sed to assist in generating the forum code.
#!/bin/bash
#
# Convert pngs to jpgs, thumbnail them and generate forum code
# Created by Tyler Mulligan - www.doknowevil.net
condir=converted
if [ ! -d $condir ];then mkdir -p $condir;fi
pagename="http://www.nexuizninjaz.com/pics/howtos/map_link_a_ninja/"
# Convert to jpg and create thumbnails
echo -e "backing up originals...\n"
cp *.png originals
echo -e "converting to jpg...\n"
mogrify -format jpg *.png
echo -e "copying to converted directory...\n"
cp *.jpg $condir
echo -e "creating thumbnails...\n"
mogrify -resize 50% *.jpg
for file in *.jpg ; do mv $file `echo $file | sed 's/\(.*\.\)jpg/thumb_\1jpg/'` ; done
mv *.jpg $condir
echo -e "cleaning up\n"
rm *.png
# Generate forum code
for file in $condir/thumb_*.jpg ; do
tempname=`echo $file | sed 's/thumb_\(.*\)/\1/'`;
echo "[url=${pagename}${tempname}][img]thumb_${tempname}[/img][/url]"
done
Future plans call for adding flags to support different image types, and including the ability to take screen shots and generate the forum code on a per screen shot basis.
Sorta Off Topic: I’ve taken a swing at frameworks be developing a new layer of organization on top of cfg files in nexuiz. I’ve thoroughly documented this scalable, fully customizable package I’ve dubbed the Ninja Pack for Nexuiz


