label:unread

Simply enter this into the search box, click “Search Mail”, and Gmail will list all those unread messages, no matter how old. Not a lot of people know that. However, if you think that’s far too many characters to type in one go, well, no worries – just use:

l:^u

instead, and it’ll yield the same results.

Another few milliseconds saved. What to do with them?

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

A while ago, MySQL announced the long awaited Alpha release of MySQL Workbench 5.1 for Linux. I’ve been waiting to try the Workbench out for a while, and I’ve finally had a chance to download, install and quickly test the latest Alpha release. Since it’s an Alpha release, there are many things that don’t yet work – bugs, crashes and so on – and while the developers are not currently asking the public to report bugs, they are aware of the issues, and the status of the Linux release can be viewed online.

First impressions? Looks great! Also, since it’s being developed on Ubuntu, it’s easy as easy can be to install and test – well, for 32 bit users. For 64 bit users, there’s a little more work involved, but it’s all pretty painless. The only current prerequisite that needs installing is libctemplate0, which can be downloaded from google’s cTemplate page. All other dependencies are taken care of by apt. The download page contains a precompiled deb file, as well as rpm, and source tarball.

32 Bit users

wget http://google-ctemplate.googlecode.com/files/libctemplate0_0.91-1_i386.deb
sudo dpkg -i libctemplate0_0.91-1_i386.deb

MySQL Workbench now has its own Ubuntu repositories, making it easy to download and install via apt. The following lines need adding to /etc/apt/sources.list:

deb ftp://ftp.mysql.com/pub/mysql/download/gui-tools/ubuntu/ binary/
deb-src ftp://ftp.mysql.com/pub/mysql/download/gui-tools/ubuntu/ source/

which can be done via the command line by running:

sudo pico /etc/apt/sources.list

or, if you prefer a graphical text editor:

gksu gedit /etc/apt/sources.list

and appending the two lines to the end of the fil. Once the new repositores are added and saved, run:

sudo apt-get update

to renew the packages.

Once the repositories had been updated, MySQL Workbench can be installed using apt (this can also be done via the Synaptic Package Manager by searching for “mysql-workbench”)

sudo apt-get install mysql-workbench

Told you it was easy.

64 Bit users

Unfortunately, there are as yet no binary debs for 64 bit users, so we have to compile from source. No worries – it’s pretty straightforward. First, I installed the prerequisites (all one line):

sudo apt-get install build-essential autoconf automake libtool libzip-dev libxml2-dev libsigc++-2.0-dev libglade2-dev libgtkmm-2.4-dev libglu1-mesa-dev libmysqlclient15-dev uuid-dev liblua5.1-dev libpixman-1-dev libpcre3-dev libgnome2-dev libgtk2.0-dev libpango1.0-dev libcairo2-dev

I then grabbed the ctemplate-0.91.tar.gz file from the ctemplate download page, downloaded to my desktop, then compiled and installed:

cd ~/Desktop
sudo cp ctemplate-0.90.tar.gz /usr/local/src
cd /usr/local/src
sudo tar -zxvf ctemplate-0.91.tar.gz
cd ctemplate-0.91
./configure
make
sudo make install

Also, include the lib directory in the library search path by running:

export LD_LIBRARY_PATH=/usr/local/lib

Next I downloaded the MySQL Workbench Alpha source code to my desktop, and compiled it:

cd ~/Desktop
mkdir ~/mysqlwb-src
mkdir ~/mysql-workbench
mv mysql-workbench-5.1.4-alpha.tar.gz ~/mysqlwb-src
cd ~/mysqlwb-src
tar -zxvf mysql-workbench-5.1.4-alpha.tar.gz
cd mysql-workbench-5.1.4-alpha
./autogen.sh
make install DESTDIR=/home/<USERNAME>/mysql-workbench

Note: Build speed can be increased by running multiple parallel build threads, by adding the -jX parameter to the make command, where X is the number of parallel threads to run.

Compiling may take a looooong time, so do what I did, and grab a brew (or a coffee). Once it was compiled, I could cd into the directory, and run mysql-workbench:

cd ~/mysql-workbench/usr/local/bin
./mysql-workbench

That’s it.

References

1. Linux Build Instructions: http://dev.mysql.com/workbench/?page_id=152
2. 32 bit install instructions: http://dev.mysql.com/workbench/?p=158
3. MySQL Workbench Linux Development status: http://dev.mysql.com/workbench/?page_id=155

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

YouBundle is currently in private Beta testing, but will be launched to the general public during i-stage.

With the exception of design, and some support staff, I’m the sole developer on the project – It’s been a long, difficult slog to get the site this far, but if the site achieves what is intended: to help web users collect, share and find information, then it’s all been worth it!

Finally, Good Luck to all the i-stage finalists! There look to be some really cool products being launched at the show. Unfortunately, I can’t make it to the show, but Neyma, my YouBundle partner and Maziar, our board advisor, will be there to present and release YouBundle.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

Well, despite the fact that I don’t actually like football, I still believe that coverage of Premiership and UEFA Champions League football matches should be free for all to watch. Thankfully, there are ways to watch the matches for free – one is to install Sopcast, and access live Premiership matches, and live Champions League matches online:

Installing Sopcast

First, download Sopcast, from the Sopcast website, then extract the file, and copy the executable to /usr/bin:

cd ~/Desktop
wget http://download.sopcast.cn/download/sp-auth.tgz
tar -zxvf sp-auth.tgz
cd sp-auth
sudo cp sp-sc-auth /usr/bin/sp-sc

the sp-sc executable can be used from the command line, by running sp-sc <sop://url> <localport> <playerport> and openning up the resulting stream in mplayer/vlc, e.g.:

sp-sc sop://sop.sopcast.org:1234/5678 3908 8908 &
mplayer http://localhost:8908/tv.asf

but the CLI involves knowing the broker, channel number and so on. There is also GUI available (actually, 2 GUIs – GTK, and QT, but I’ve only managed to get the GTK GUI working so far), which does all this for you, and makes the process much easier. The GTK Gui can be set up as follows:

cd ~/Desktop
wget http://linuxtoy.org/files/deb/gtk-sopcast_0.2.8-1_i386.deb
sudo dpkg -i --force-architecture gtk-sopcast_0.2.8-1_i386.deb

32 Bit users can omit the –force-architecture option, since this is essentially just telling 64 Bit Ubuntu to shut up and use 32 Bit libraries.

Sopcast will only pull in the stream – a media player is required to actually view the stream. VLC, or mplayer both work – I currently use mplayer with the Sopcast GUI, without problems. gsopcast is actually set up to use mplayer by default, so if you don’t have it, it can be installed by running:

sudo apt-get install mplayer

Once the gsopcast GUI is installed, there should be a new menu item under Applications -> Sound & Video -> Sopcast TV Player. Alternatively, the GUI can be run from the terminal:

gsopcast

Once launched, the GUI will populate the channel list. Double click on the channel you wish to view, and wait for the buffer to reach 90%-100% before clicking the “player” button.

The best way to find out which channel is playing the Premiership, or Champions League match you want to watch is to scan this list and look for the channel playing the match.

References

1. http://ubuntuforums.org/showthread.php?p=3972647
2. http://www.myp2p.eu/competition.php?competitionid=&part=sports&discipline=football

Whatever happened to Division 1, Division 2, and so on, by the way? They’ve even had to make the name of the divisions sound financially appealing.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

Creating a Virtual Host

Note: The Apache installation in Ubuntu is actually controlled by the file /etc/apache2/apache2.conf with /etc/apache2/httpd.conf being a secondary config file loaded after apache2.conf. It’s advisable to leave apache2.conf alone, and make any server-wide changes to httpd.conf. The Ubuntu Apache installation also goes a step further by allowing site-specific configuration files, which are located in the directory /etc/apache2/sites-available. One already exists, and is called “default”, which controls the web root, /var/www. To activate any new sites created in this directory, we just create new file in this directory, then create a symbolic link to it in the /etc/apache2/sites-enabled directory. When Apache starts, it scans the apache2.conf, httpd.conf files, and also the /etc/apache2/sites-enabled directory for any links to config files in /etc/apache2/sites-available. If the link isn’t there, the configuration isn’t loaded.

For this tutorial, we’ll be creating a new config file in /etc/apache2/sites-available and activating it by creating to it in /etc/apache2/sites-enabled – apache2.conf and httpd.conf will not be touched.

Ok, I wanted to create a Virtual Host called 64bitjungledev.com, and also be able to view the site on my local installation when I point my browser to 64bitjungledev.com. The first thing I did was create a directory for the new Virtual Host to reside in. I was feeling particularly lazy, so just created it in the /var/www directory, and changed the owner of the new directory to my user name and group (so I could add/edit files etc.) I know, I know, I should create it in my home directory – but everything with “/var/www/ can be substituted for “/home/USERNAME/html/”. Anyway, in a Terminal, I executed:

sudo mkdir /var/www/64bitjungledev
sudo chown myuser:mygroup /var/www/64bitjungledev

Then, I created a new config file in the /etc/apache2/sites-available to contain the configuration information for my new Virtual Host. In a Terminal, I executed the following:

cd /etc/apache2/sites-available
gksu gedit 64bitjungledev.com

This creates a new file called 64bitjungledev.com in the directory and opens it in a text editor. In this new file, I dumped:

<VirtualHost *:80>
ServerName 64bitjungledev.com
ServerAlias www.64bitjungledev.com
DocumentRoot /var/www/64bitjungledev
DirectoryIndex index.php
<Directory /var/www/64bitjungledev>
AllowOverride all
Options Indexes FollowSymLinks MultiViews
Order allow,deny
Allow from 127.0.0.1
</Directory>
CustomLog /var/log/apache2/64bitjungledev.com-access.log combined
</VirtualHost>

This is basically telling Apache that we want it to host a new site, called 64bitjungledev.com, and all it’s files will be located in /var/www/64bitjungledev. “Allow from 127.0.0.1″ tells Apache that only my computer can access it, and “AllowOverride all” tells Apache that the site can use .htaccess files that contain Apache declarations that can override these settings. The CustomLog declaration is telling Apache to create an access log file for this site in /var/log/apache2/64bitjungledev.com-access.log and the “combined” tells it to include two extra fields of information in the log file – The “Referer” (sic) HTTP request header, and The User-Agent HTTP request header. Probably superfluous considering all trafic will be refered from my local machine, and the only agent accessing the site is Firefox! You can change “combined” to “common”, which doesn;t include these two fields.

Anyway, I digress… There are two methods to enabling the new site – create a symbolic link in /etc/apache2/sites-enabled to this new file, or, have Ubuntu do it for us:

cd /etc/apache2/sites-available
sudo a2ensite 64bitjungledev.com

Will do exactly the same as:

sudo ln -s /etc/apache2/sites-available/64bitjungledev.com /etc/apache2/sites-enabled/64bitjungledev.com

To disable the site, we can run:

sudo rm /etc/apache2/sites-enabled/64bitjungledev.com

Or:

sudo a2dissite 64bitjungledev.com

The result is the same. In order for the URL 64bitjungledev.com to work in my browser, I just had to change the hosts file, so my computer looks for 64bitjungledev.com on my machine as opposed to looking it up via DNS:

gksu gedit /etc/hosts

I changed the line

127.0.0.1 localhost

to:

127.0.0.1 localhost 64bitjungledev.com www.64bitjungledev.com

Saved, and closed the file. Now, I had to restart Apache for the new changes to take affect:

sudo /etc/init.d/apache2 reload

I created a test file in the document root:

gedit /var/www/64bitjungledev/index.php

Containing the old faithful:

<?php
phpinfo();
?>

and pointed my browser to www.64bitjungledev.com. Success! Now for mod_rewrite…

Enabling mod_rewrite

First, to enable the Apache module, mod_rewrite (which is disabled on a default Hardy install), I had to create a symbolic link to the module file, so in a Terminal, I ran:

sudo ln -s /etc/apache2/mods-available/rewrite.load /etc/apache2/mods-enabled/rewrite.load

The Apache installation in Ubuntu has many different configuration files, and uses this modular approach to activae/deactivate them. The rewrite.load file simply contains the load module declaration which would normally be placed in the httpd.conf file:

$more /etc/apache2/mods-available/rewrite.load
LoadModule rewrite_module /usr/lib/apache2/modules/mod_rewrite.so

When Apache starts in Ubuntu, it checks the /etc/apache2/mods-enabled directory for links to available modules, and loads the modules it finds. Creating the link above can also be achieved by running:

sudo a2enmod rewrite

which is basically a short-cut approach to creating the link in /etc/apache2/mods-enabled, similar to the a2ensite used above – and, similarly:

sudo a2dismod rewrite

will have the same effect as:

sudo rm /etc/apache2/mods-enabled/rewrite.load

That’s pretty much it. The only additional thing I had to do, was make a slight change to the /etc/apache2/sites-available/default file, which, as mentioned above, is the default Apache configuration file for /var/www. So, I opened it up:

gksu gedit /etc/apache2/sites-available/default

And changed the “AllowOverride” declaration to “all” for both the / and /var/www directories. I just needed to restart Apache once again:

sudo /etc/init.d/apache2 reload

Testing mod_rewrite

Before going ahead and installing a local copy of 64bitjungle, I wanted to test if the system was working. I needed two files for this – a .htaccess file, and a PHP file. The .htaccess file contains the test RewriteRule for Apache to process the URL redirection of “clean” URLs to the PHP script, so that www.64bitjungledev.com/xyz/somedata will actually point to www.64bitjungle.com/thescript.php?thevariable=somedata. So, in a Terminal, I created the .htaccess file first:

gedit /var/www/64bitjungledev/.htaccess

And wrote a simple Apache RewriteRule, in addition to telling Apache to switch on the rewrite engine:

<IfModule mod_rewrite.c>
RewriteEngine On
#test rule
RewriteRule ^xyz/([^/.]+)/?$ thescript.php?thevariable=$1 [L]
</IfModule>

This is basically telling Apache to switch on the RewriteEngine if mod_rewrite is loaded. The rather scrappy-but-suffices-for-testing RewriteRule basically tells Apache to “grab every character in the clean URL after xyz/ and store it in a temporaty variable, then forward it to thescript.php with thevariable set to the stuff just stored”.

A very simple PHP script can be written and used to test if the data is being passed properly:

gedit /var/www/64bitjungledev/thescript.php

which can contain simply:

<?php
echo $_GET['thevariable'];
?>

After restarting Apache, pointing my browser to www.64bitjungledev.com/xyz/DoesItWork now outputs “DoesItWork”, www.64bitjungledev.com/xyz/anything1234 ouputs “anything1234″ etc. The fun just doesn’t stop…!

With mod_rewrite successfully installed, and working for this new domain, I can now install a local version of 64bitjungle.com to play with. I haven’t actually done that yet though…

References:

• http://httpd.apache.org/docs/1.3/mod/mod_rewrite.html
• http://httpd.apache.org/docs/1.3/vhosts/
• http://httpd.apache.org/docs/1.3/vhosts/examples.html

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

The driver currently supports the following Microdia webcams:

0c45:6240 0c45:6242 0c45:6243 0c45:6248 0c45:624b 0c45:624c 0c45:624e 0c45:624f 0c45:6253 0c45:6260 0c45:6262 0c45:6270 0c45:627a 0c45:627b 0c45:627c 0c45:627f 0c45:6280 0c45:6282 0c45:6283 0c45:6288 0c45:628a 0c45:628b 0c45:628e 0c45:628f 0c45:62a0 0c45:62b0 0c45:62b3 0c45:62ba 0c45:62bb 0c45:62bc 0c45:62be

Keep up to date on the current supported webcams: http://groups.google.com/group/microdia/web/project-status?hl=en

The webcam (Acer Orbicam) on my Acer Aspire 5052 has been somewhat of a bane – with no drivers available, it has been one of the last stumbling blocks of getting all of my hardware working in Linux (the Ubuntu 8.04 Distro, Hardy Heron, Ubuntu Jaunty in my case). That is until I ambled across, quite by accident, the Microdia Google Group. These guys are the epitome of Open Source development in action – a group of widely distributed people who are collaborating to solve a common problem: to get the Microdia webcams supported under Linux. And they’re doing a fantastic job so far.

Note: I should first mention that the drivers are still heavily under development, and although they are not (for some webcams) to the point where they can be used for Skype, for example, major progress has been made, and there is now finally communication with my webcam, to the point where I can successfully output video to mplayer. Therefore, this post is not quite yet a HowTo on getting a Microdia webcam fully function, but a step towards helping people installing and testing the drivers under development.

So, here’s how I’ve got to the stage of seeing myself in mplayer…

Information Gathering

Note: this section is only relevant if you’re interested in the workings of the camera, or would like to aid the driver developers by supplying information about your webcam. If you just want to compile and install the drivers, scroll down to Driver Compiling and Installing.

Firstly, I had to determine i) whether or not my Webcam was in fact a Microdia camera, and ii) the model number, so in a Terminan, I ran:

lsusb

Which output, amongst other information about the rest of my USB devices:

Bus 003 Device 002: ID 0c45:6260 Microdia

confirming that my Acer Orbicam is a Microdia device. The two hex values 0c45:6260 represent the Vendor ID and Product ID respectively – 0c45 being Microdia, and 6260 the Product itself. In order to create my meagre contribution to the Microdia Group, I had to gather some more information about the device’s bridge and sensor. Firstly, I needed the USB Device Descriptors, which were obtained by running:

sudo lsusb -d 0c45:6260 -v > ~/Desktop/0c45_6260_device_descriptors.txt

I then needed to discover the Bridge and Sensor types, which thankfully could be found in the Windows driver file, snp2std.inf:

%USBPCamDesc% = SN.USBPCamVGA,USBVID_0c45&PID_6260 ; SN9C201 + OV7670ISP

Which tells me:
Bridge: SN9C201
Sensor: OV7670

The full guide to gathering information for your webcam can be found at the Microdia group, here.

Driver Compiling and Installing

First, I installed some prerequisites – git (to obtain the Microdia Drivers source code), kernel source, build essential etc.:

sudo apt-get install git-core git-doc git-gui kernel-package linux-headers-`uname -r` build-essential libv4l-0 mplayer

Once finished, I created a directory to set up my local git repository and downloaded the driver code:

cd ~/Documents
mkdir webcam
cd webcam
git clone http://repo.or.cz/r/microdia.git

The final command executes git, and tells it to clone the remote repository. The files are downloaded to a sub directory called microdia, from which I ran the rest of the commands:

cd microdia

I then compiled the code (not with sudo):

make

and received this error, but it’s not important:

make: ctags: Command not found
make: *** [ctags] Error 127

Once compiled, the module can be inserted to the kernel for testing:

sudo insmod ./sn9c20x.ko

I also received the following error:

ERROR: insmod: error inserting 'sn9c20x.ko': -1 Unknown symbol in module

so had to run the following first:

sudo modprobe videodev
sudo modprobe compat-ioctl32

and then I could insert the module:

sudo insmod ./sn9c20x.ko

Finally, to test, I ran:

LD_PRELOAD=/usr/lib/libv4l/v4l2convert.so mplayer tv:// -tv driver=v4l2:width=640:height=480:fps=25:device=/dev/video0 -vo x11

The only thing is… it’s currently a mirror image:

Finally…

If everything looks good, and the camera works well, it can be installed permanently by running the following:

strip -g sn9c20x.ko
sudo mkdir -p /lib/modules/`uname -r`/kernel/drivers/media/video/usbvideo/
sudo cp sn9c20x.ko /lib/modules/`uname -r`/kernel/drivers/media/video/usbvideo/
sudo depmod -a

However, only do this if you are comfortable with the driver, and happy that it is stable on your system!

References

1. Microdia, Google Group
2. Full list of currently supported (and unsupported) Microdia Webcams
3. How to gather information on your webcam
4. Using git with Microdia
5. Testing the Microdia driver
6. PATCH: v4l-experimental support for 6260 (3rd try) Thread

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

A couple of days ago, I set up a makeshift studio at home so I could do a photoshoot of my wife and new baby (well, I say studio – it was in fact a couple of dark blankets I found about the house, no lighting, reflectors, or any of the expensive professional gear!). Once finished, I really wanted to process some of the better shots to send to my family, and so was determined to find something I could install and use on Linux.

Tenacity prevailed, and I finally discovered Lightcraft’s Lightzone – a Java based digital darkroom, with versions available for Mac, Windows and Linux! The Linux version is currently in Beta, but in all honesty, having used it for the past few days, I couldn’t tell. Installation is as simple as you can get – download the archive (registration is required, but it’s worth it), extract the files, enter the new directory, and run the LightZone executable:

tar -zxvf LightZone-3.4.tar.gz
cd LightZone
./Lightzone

After a dialog appears informing how many days of the trial are left, the splash screen displays as the program loads. The GUI is incredibly intuitive, and within minutes, I managed to process a couple of pictures worth sending to my family. When the application first opens, it’s a simple case of using the pane on the left to navigate to a directory containing photos. LightZone automatically generates thumbnails, and a larger preview of the chosen thumbnail with an option to edit. A single click on “Edit” brings up a whole array of processing options to play with.

After a few clicks, and mainly fumbling around the system, I managed to turn this raw image:

into a Black and White 10×8 Portrait that my Mum will love:

I’m incredibly happy with the results of a few minutes clicking and fumbling around an unfamiliar system! If this is what can be achieved by doing so, I’m really looking forward to getting my teeth into the system, and learning it’s many features – HDR (High Dynamic Range) support, Zone Systems, and many more.

Unfortunately however, I only have 9 days left on the Trial, since the Linux version is still in Beta, and therefore not available for purchase. I have no idea how much the Linux version will be when it’s finally released, but the Windows and Mac versions retail at $129.95 for the Basic version, and $199.95 for the Pro version. Unfortunately, both out of my price range for the time being (that’s a hell of a lot of money where I live!), but I’m certainly going to enjoy the Beta version while I can, and dream of when I can afford to buy the Pro version! Still, I’ll see if I can make some time to write a couple of tutorials and post them.

LightZone is a fantastic product. The only thing extra I would like to see, is a 64 Bit version.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

Xdebug is a fantastic PHP extension, which is written by one of the PHP Core Developers, Derick Rethans. The current version is 2.0.3, and I set about installing it on my system, and integrating it into Eclipse PDT today.

Xdebug can be installed in a few different ways – by downloading the binaries from the site, compiling the source, or via PECL:

sudo pecl install xdebug

Some days, I just prefer to download compile the source, as it gives me more of an understanding of what is being installed and where, and this is one of those days:

cd ~/
mkdir xdebug
cd xdebug
wget http://www.xdebug.org/link.php?url=xdebug203
tar -zxvf xdebug-2.0.3.tgz
cd xdebug-2.0.3

The packages phpize and php-config are required to install Xdebug, and can be obtained by installing the relevant PHP development packages:

sudo apt-get install php4-dev

or

sudo apt-get install php5-dev

depending on the PHP version installed. After unzipping the Xdebug source, I ran

phpize

which output:

PHP Api Version: 20041225
Zend Module Api No: 20060613
Zend Extension Api No: 220060519

I then checked this against the table at http://www.xdebug.org/docs/install#phpize to see if my PHP version (5.2.3) was compatible with Xdebug – it matched up to the table’s 5.2.x PHP version row, so I went ahead and installed:

./configure --enable-xdebug --with-php-config=/usr/bin/php-config
make

which created a file xdebug.so in the modules subdirectory. This file needs to be copied (or moved) to the PHP Extension directory, which on my system is in /usr/lib/php5/20060613/. An easy way to find the PHP extension directory is to run:

locate mysql.so

ok, so I copied the xdebug.so to the extension directory:

sudo cp modules/xdebug.so /usr/lib/php5/20060613/

So far so good. Now PHP needs to be told about the extension, and to load it. This can be done by adding

zend_extension="/usr/lib/php5/20060613/xdebug.so"

to the php.ini file. However, on my Ubuntu system, the PHP extensions are activated by adding an .ini file to the /etc/php5/apache2/conf.d directory, for each extension installed. The .ini file is the extension_name.ini, so I created a file called xdebug.ini:

cd /etc/php5/apache2/conf.d
gksu gedit xdebug.ini

which for now, contains a single line:

zend_extension="/usr/lib/php5/20060613/xdebug.so"

After restarting the Apache server

sudo /etc/init.d/apache2 restart

I created a file called test.php in my Web Root directory, containing a call to the phpinfo() function:

<?php
phpinfo();
?>

and pointed my browser to http://localhost/test.php. Just before the “PHP Credits” section, is a small box containing information on Zend – and now, with a successful Xdebug installation, additional information on Xdebug:

The same information can be found by opening a Terminal and running

php -m

to output the loaded modules, and towards the end of the output, you should see

[Zend Modules]
Xdebug

alternatively, run

php -i | grep Xdebug

and if Xdebug has been installed successfully, you should see

with Xdebug v2.0.3, Copyright (c) 2002-2007, by Derick Rethans

Now, Xdebug needs to be enabled before any scripts can be debugged! So, in the php.ini file (or in my case, the newly created /etc/php5/apache2/conf.d/xdebug.ini file) I added the lines:

xdebug.remote_enable=On
xdebug.remote_host="localhost"
xdebug.remote_port=9000
xdebug.remote_handler="dbgp"

and restarted Apache again:

sudo /etc/init.d/apache2 restart

running

php -i | grep xdebug

in the Terminal, or using the test.php file containing phpinfo() shows that the new settings are being used, and Xdebug is active and running.

Configuring Eclipse PDT:

Setting up Eclipse PDT to use the Xdebug extension is pretty simple. In Eclipse PDT, I opened Window -> Preferences, then clicked on PHP, expanded the tree, and selected Debug. I Chose Xdebug from the pull down menu for the PHP Debugger, the “Default PHP Server” (should already be set to localhost), and then created a new PHP Executable option by clicking on the PHP Executables link next to the PHP Executable pull down menu.

In the PHP Executables dialog, I Clicked Add, filled in the name (PHP with XDebug), and entered the path to the PHP executable – on my system this is /usr/bin/php but if you don’t know, then in a terminal, run

which php

to find out where the php command is run from. I also entered the location of php.ini – again, on my system, this is /etc/php5/apache2/php.ini – and selected XDebug as the PHP debugger:

Clicked OK, then went back to the PHP -> Debug preferences screen, and selected the PHP Executable just created in the PHP Executable pull down. Finally, in the Debug -> Workbench Options option in the left, I selected “Never” for the “allow multiple debug sessions” option, since Xdebug does not support them.

Finally, I needed to set the Default Web Browser in Eclipse, to that when I debug a script, the output is sent to Firefox. Again in Preferences, I navigated to General -> Web Browser, and clicked “New”. It’s pretty straight forward stuff – entering a name and location – Firefox, and /usr/bin/firefox. Again, if you don’t know the location of the firefox binary, in a terminal run:

which firefox

With this done, and the preferences saved, I can now debug my scripts using Xdebug!

References:

http://www.xdebug.org/docs/install
http://www.eclipse.org/pdt/documents/XDebugGuide.pdf
http://devzone.zend.com/article/2803-Introducing-xdebug

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

NOTE, 04/08/08: This tutorial covers Eclipse Europa. If you would like to install the latest version of Eclipse 3.4 Ganymede, with PDT and SQL Explorer, take a look at my latest Tutorial: “64 Bit Eclipse 3.4 (Ganymede) IDE with PDT and SQL Explorer – Full PHP/MySQL Web Application IDE“.

I had tried once before to install the x64 (64 Bit) version of Eclipse IDE, about 8 months ago, but found it to be somewhat buggy and unstable. In the interim, I’ve been using the 32 Bit version of Eclipse PDT for my development stuff. Earlier this year (21/02/08) a new version of Eclipse 64 Bit was released, so I gave it a go today. So far, so good!

Unfortunately, there’s no 64 Bit PDT all-in-one, but I managed to install a 64 Bit equivalent by cobbling together the relevant packages available by using the Eclipse Update Manager system, after initially installing the latest version of Eclipse Classic 3.3.2 64 Bit.

I wanted to keep everything (or as much as possible) 64 Bit, so I also download and installed the 64 Bit JRE, which can be downloaded here (or use the direct link to the bin file). The method for installing the 64 Bit JRE is the same as the 32 Bit version – after the file downloaded to my desktop, I opened up a new Terminal Window (Applications -> Accessories -> Terminal), traversed to the directory I wanted to install it into, moved the file, made it executable, and ran it to install:

cd /usr/java
sudo mkdir 64
cd 64
sudo mv ~/Desktop/jdk-6u5-linux-x64.bin /usr/java/64/
sudo chmod a+x jdk-6u5-linux-x64.bin
sudo ./jdk-6u5-linux-x64.bin

If the “java” directory doesn’t exist, it needs to be created first:

sudo mkdir /usr/java

listing the directory

ls

should return

jre1.6.0_05

which is the directory containing the necessary Java binaries.

As with my previous 32 Bit installation, I wanted Eclipse to be installed in the /opt directory:

cd /opt
sudo mv ~/Desktop/eclipse-SDK-3.3.2-linux-gtk-x86_64.tar.gz /opt
sudo tar -zxvf eclipse-SDK-3.3.2-linux-gtk-x86_64.tar.gz

then rename the eclipse directory:

sudo mv eclipse eclipse64

I also needed to get the newly installed Eclipse to run with the newly installed JRE – by default, the eclipse executable in the directory will try and detect Java and use whatever it finds, so I created a shell script:

cd eclipse64
sudo gksu gedit eclipse.sh

containing the following:

#!/bin/bash
PATH=/usr/java/64/jre1.6.0_05/bin:$PATH
/opt/eclipse64/eclipse

Now instead of running the eclipse executable, I run eclipse.sh (make sure it’s executable)

sudo chmod 755 eclipse.sh

before running

/opt/eclipse64/eclipse.sh

PDT and WTP Plugins

The first time I ran the new 64 Bit version, everything went well – it was fast, and seemed stable, so I went ahead and started installing all the additional plugins I needed (and still need…). Goto Help -> Software Updates -> Find and Install, and select “Search for new features to install”.Click on “New Remote Site” for each of the following:

• Name: PDT, URL: http://download.eclipse.org/tools/pdt/updates/
• Name: WTP, URL: http://download.eclipse.org/webtools/updates/
• Name: GEF, URL: http://www.eclipse.org/gef/updates/
• Name: EMF, IRL: http://www.eclipse.org/modeling/emf/updates/

Actually, there are only a couple of components required from the GEF (Graphical Editing Framework) and EMF (Eclipse Modeling Framework) packages to satisfy dependencies – WTP (Web Tools Platform) requires a package from GEF, and GEF from EMF…

After adding these, click on Finish – the Update Manager will then query any mirrors for the latest versions of the plugins. Once it has finished, a dialog appears, where it is possible to select the plugins to download and install. First, I selected PDT – the Update Manager then informed me that PDT requires files from WTP, so I tried clicking the “Select Requires” button, hoping that it would sort out the dependencies on my behalf. Unfortunately, nothing happened… So, I selected WTP manually, then expanded GEF -> Eclipse SDK R3.3.1 and Selected Graphical Editing Framework 3.3.1v20070814, then expanded EMF -> EMF SDK 2.3.2 and selected Eclipse Modeling Framework (EMF) – org.eclipse.emf.ecore 2.3.2v200802051830… I could then click “Select Required” to get the last few residual required dependencies… Phew…

Actualy, it still bugged me for dependencies for Java Persistence API contained in WTP, but by that stage I really couldn’t be arsed going through the process of adding more stuff to the Update Manager and potentially having to search for further dependencies, so I just deselected the three Java Persistence API files contained in WTP.

OK, with the dependencies sorted, I clicked Next, accepted the agreements, finished, and went to make, and drink a brew (that’s Tea) while the Update Manager downloaded and installed the requested stuff.

MySQL: SQL Explorer Plugin

Once everything had downloaded and installed, I restarted Eclipse. Everything looked good! Great, in fact. But I also wanted to install a few more plugins, including the MySQL SQL Explorer Plugin – which additionally requires, and depends on Eclipse DTP (Data Tools Project). Eclipse DTP has to be downloaded and installed manually, since there is no automated Update Site. So, I downloaded dtp_1.5.2_022008.zip to my Desktop, and set about extracting and installing the files:

cd ~/Desktop
mkdir DTP
mv dtp_1.5.2_022008.zip DTP
cd DTP
unzip dtp_1.5.2_022008.zip
cd eclipse
sudo cp -R features plugins /opt/eclipse64

If Eclipse is open, restart it. With the DTP installed, I could set about installing the SQL Explorer Plugin. The instructions for installing and configuring the SQL Explorer Plugin can be found in my previous article – Eclipse PDT and MySQL – SQL Explorer Plugin, since I don’t want to regurgitate information, just follow the link if you need to install it, and come back when it’s done.

Note: Actually, now I had the DTP installed, I could download and install the additional Java Persistence API in WTP! Same method: Update Manager, select WTP etcetera, etcetera…

ATF: AJAX Framework Toolkit

Once the SQL Explorer Plugin was installed, I had one final plugin – ATF (the AJAX Toolkit Framework). This can be downloaded as a Site Archive, and the zip file added to Eclipse’s Update Manager. Once downloaded, I opened the Update Manager in Eclipse (Help -> Software Updates -> Find and Install), and clicked on “New Archived Site”, to add a new locally archived site, calling it ATF, and pointing it to the freshly downloaded atf-incubation-SiteArchive-0.2.3M4-v200709141050.zip file. I discovered, however, that the site.xml file contained within the zipped site archive is somewhat out of date and the pointers within are directed to the wrong download locations, so I had to manually locate the additional dependent plugins – Mozilla XULRunner, and Mozilla JS. They can both be added as New Remote Sites to the Update Manager, by clicking on the “New Remote Site” button, and adding the following information:

• Name: Mozilla JS, URL: http://ftp.mozilla.org/pub/mozilla.org/js/eclipse/
• Name: XULRunner, URL: http://releases.mozilla.org/pub/mozilla.org/xulrunner/releases/1.8.1.3/contrib/eclipse/

The previously used EMF plugin is also required to satisfy a couple of dependencies, so I selected the ATF, Mozilla JS, XULRunner, and EMF plugins from the list, and clicked Finish. Once the “updates” were found, I selected XULRunner, Mozilla JS, ATF (I deselected the ATF -> Mozilla JS component though, since an updated version was to be downloaded and installed from the Mozilla Server), and also selected the required EMF components. Click through Next etc. to download and install. Once completed, I restarted Eclipse.

In the Window -> Preferences menu by the way, AJAX toolkits such as Script.aculo.us can be added to the ATF module.

Et Voila! I now have a working Eclipse 64 Bit version up and running! Finally!

Oh, if you want a desktop icon, just create a file on your Desktop called Eclipse.desktop,

cd ~/Desktop
gedit Eclipse.desktop

and add the following:

[Desktop Entry]
Categories=;
Encoding=UTF-8
Exec=/opt/eclipse32/eclipse.sh
Hidden=false
Icon=/opt/eclipse64/icon.xpm
Icon[en_US]=/opt/eclipse64/icon.xpm
Name=Eclipse
Name[en_US]=Eclipse
Terminal=false
Type=Application
Version=1.0

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

invoke-rc.d: initscript firestarter, action "restart" failed
run-parts: /etc/network/if-up.d/50firestarter exited with return code 2

This is actually a bug with version 1.0.3-5, and the solution is pretty simple – to upgrade to version 1.0.3-6. However, I have Ubuntu 64 Bit installed, and unfortunately, 1.0.3-6 is not yet available in the repositories. Not worries – just a simple case of downloading the source tar ball, compiling, and installing.

First, I removed the currently installed version:

sudo apt-get remove firestarter

and downloaded the latest tar ball. Once this had downloaded to my desktop, I went through my usual routine of creating a new directory, moving the file, extracting… and so on:

cd ~/Desktop
mkdir firestarter
mv firestarter-1.0.3.tar.gz firestarter
cd firestarter
tar -zxvf firestarter-1.0.3.tar.gz
cd firestarter-1.0.3

Installation is pretty straightforward – although it’s always good to read the README and INSTALL files!

more INSTALL

So, configure, make, and install:

./configure --sysconfdir=/etc
make
sudo make install

However, the configure step threw out a few errors for me:

Package libgnome-2.0 was not found in the pkg-config search path. Perhaps you should add the directory containing `libgnome-2.0.pc' to the PKG_CONFIG_PATH environment variable No package 'libgnome-2.0' found Package libgnomeui-2.0 was not found in the pkg-config search path. Perhaps you should add the directory containing `libgnomeui-2.0.pc' to the PKG_CONFIG_PATH environment variable No package 'libgnomeui-2.0' found Package gnome-vfs-2.0 was not found in the pkg-config search path. Perhaps you should add the directory containing `gnome-vfs-2.0.pc' to the PKG_CONFIG_PATH environment variable No package 'gnome-vfs-2.0' found Package libglade-2.0 was not found in the pkg-config search path. Perhaps you should add the directory containing `libglade-2.0.pc' to the PKG_CONFIG_PATH environment variable No package 'libglade-2.0' found
configure: error: Library requirements (libgnome-2.0 >= 2.0.0
libgnomeui-2.0 >= 2.0.0
gtk+-2.0 >= 2.4.0
gnome-vfs-2.0 >= 2.6.0
libglade-2.0 >= 2.3.6)

which basically means that I didn’t have some of the necessary development libraries installed required to compile the binaries. No real problem, and easy enough to solve by, well, installed in the required packages:

sudo apt-get install libgnome-dev libgnomeui-dev libgnome-vfs-dev libglade2-dev

This process also had to download an additional 50 packages to meet the dependencies of the above four packages (and there were probably some sub dependencies there too…)

Anyway, the configure process ran smoothly after installing the required dev libraries, so I could then run

make
sudo make install

to compile and install the new binaries.

It’s possible to have Firestarter run as a system service, by copying the fedora.init file from the install directory to /etc/init.d and renaming the file to firestarter, then enabling the script:

sudo cp fedora.init /etc/init.d/firestarter
sudo chkconfig firestarter reset

However, since I had previously installed Firestarter, all the relevant scripts and symbolic links were already in place. All I had to do was make a minor change to the /etc/init.d/firestarter script:

cd /etc/init.d
cp firestarter firestarter.old
gksu gedit firestarter

and changed the line

[ -x /usr/sbin/firestarter ] || exit 0

to

[ -x /usr/local/bin/firestarter ] || exit 0

so that the script would point to the new installed binary. That’s it.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.