Yeah, right. Okay. Whatever!

hbm2ddl SchemaExport: You have an error in your SQL syntax near type=InnoDB

2012-01-30T17:59:00.004+01:00

You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'type=InnoDB'

I used to create my MySQL DBs with ant and everything was working. Now, moving to maven and hbm2ddl I get this error.
After some searching it comes out that it's a problem of incompatibility between the MySQL Java connector (5.1) and the MySQL version (5.5) where type InnoDB was dropped and replaced with engine.

On Fedora, the easy solution is to downgrade MySQL to version 5.1 (download the RPMs from MySQL web site)

More details here.

Maven + Subversion Enabled Eclipse Java Project

2011-11-09T18:14:00.000+01:00

What I want for my Eclipse projects is to be Java projects, allow content revisions with Subversion and be managed by Maven.

This is what I need to do.

Eclipse version: Eclipse Indigo
Subversion plugin: Subversive
Maven plugin: m2eclipse
Java: JDK 1.6

Support for Subversion and Maven is now integrated into Eclipse.

Choose Help --> Install new software --> Work with: Indigo - http://download.eclipse.org/releases/indigo
Expand collaboration and check the m2eclipse and Subversive packages.
Configure the Maven repository here: Window --> Preferences --> Maven --> Installations.

Now you need the Maven SCM handler for Subversive.

The easy way is to create a fake new project from File --> new project --> New Project wizard
Select the project type: Maven --> Checkout Maven Projects from SCM
In the target location dialog box select the m2e marketplace in the right bottom corner.
When you are presented with the list of connectors, the m2e-subversive is the the last one.
Once Eclipse has restarted, you can choose again to create a project through Maven --> Checkout Maven Projects from SCM and this time you will be able to specify a SCM URL (svn) and fill in the URL.

Now you have a project checked out from a Subversion repository which uses a Maven repository to manage its library dependencies.

References

The place where I found the final answers is Ralph's Java blog.

Liferay: Error Registering Portlets

2011-10-24T14:51:00.000+02:00

I got a problem when trying to undeploy a portlet on Liferay the standard way: removing the unzipped WAR in Tomcat's webapps/ dir and the cached information.

08:48:56,725 ERROR [HotDeployUtil:112] com.liferay.portal.kernel.deploy.hot.HotDeployException: Error registering portlets for MyPortlet

It seemd like the portlet was somehow still in the page for Liferay and the message saying the portlet needed to be removed from the page did not show any buttons to do that. After a while I didn't have the message anymore but only an error page.

Some googling brought me on a page in the Liferay forum where they suggested to remove the portlet-api-2.0.jar from the lib/ directory of the WAR and redeploy it:

> zip -d MyPortlet.war /WEB-INF/lib/portlet-api-2.0.jar

It actually worked. Don't ask me why.

Batch Resize on Linux

2011-08-08T15:56:00.003+02:00

Install imagemagick.
Now you can use mogrify.
Example: to resize to half the original size all the JPGs in a directory (keeping the aspect ratio of course):

> mogrify -resize 50% *.jpg

More here.

Eclipse Startup - Crash on Fedora 15

2011-08-08T12:24:00.005+02:00

For some obscure reason, the Fedora 15-packaged Eclipse doesn't include JSP syntax highlighting and I could find no easy way to add it (conflicting dependencies prevent that). So I resorted to the original TAR-GZ distribution. The problem there is with the OpenJDK. My Eclipse simply crashes on startup.
What you need to do is tell Eclipse to use Sun's JDK.
Open the eclipse.ini and add the path to the Java executable:

-vm
/opt/java/jdk1.6.0/bin/java

right before the -vmargs option.
More here.

[UPDATE] - Oh, the reason for the crash is not openJDK. It's a bug in Fedora. To circumvent it do the following:

In Gnome3 the webkit library name changed to libwebkitgtk so you need to create one missing symbolic link: ln -s /usr/lib/libwebkitgtk-1.0.so.0 /usr/lib/libwebkit-1.0.so.2
Add -Dorg.eclipse.swt.browser.UseWebKitGTK=true to the -vmargs options in eclipse.ini

More here.

Extracting Audio from Video or Stream

2011-07-01T11:58:00.002+02:00

From AVI or other video formats:
> ffmpeg -i file_name.avi -ab 128 -ar 44100 file_name.mp3

From an online stream:
> ffmpeg -i URL_address -ab 128 -ar 44100 file_name.mp3

> man ffmpeg

Fedora 15: Old and New Alt-Tab Behavior

2011-06-15T15:11:00.004+02:00

As of Fedora 14 (Gnome 2.32) the behavior of Alt-Tab was to cycle through all the windows of all the open applications in a virtual workspace.
Gnome 3 shipping with Fedora 15 behaves differently and probably more efficiently (time will tell.)

Alt-Tab now cycles through all open applications in all virtual workspaces. All the windows of an application are shown when that application is the selected one. You can move the mouse over one window and select it but you can't just use the keyboard to do that.

If you want to cycle through all the windows of the active (selected) application you need to use Alt-~ (or the button above Tab.)

Fedora 15: Add Back Shutdown Option

2011-06-15T14:51:00.005+02:00

Gnome 3 by default won't show the shutdown button. To see it click on your name on the top bar and press Alt. The Suspend option will turn into Power off.
If you don't want to perform this extra step:

Install: yum install gnome-shell-extensions-alternative-status-menu.
Type Alt-F2 then r and enter (reloads the Gnome shell.)
You're set! You have gained the Hibernate and Power off options now.

Credits.

Fedora 15: Add Back Window Buttons

2011-06-15T14:39:00.006+02:00

UPDATED
Gnome 3'default behavior suppresses all window buttons but the Close. If you want them back:

Install gconf-editor: yum install gconf-editor
Run it and go to desktop --> gnome --> shell --> windows
Set the value to "menu:minimize:maximize:close"
Type ALT-F2 then r and enter (reloads the Gnome shell) or you will see the modifications once you log out and then back in.
You're done!

Credits.

Eclipse Error: "Import Cannot Be Resolved"

2010-12-22T14:05:00.002+01:00

The imported class is there but Eclipse won't see it.
Try: Project --> Clean

Creating a SVN Repository

2010-11-25T17:22:00.004+01:00

Create the repository

> svnadmin create svnrepo

> gedit svnrepo/conf/svnserve.conf

Enable the following lines:

anon-access = none
auth-access = write
password-db = passwd

> gedit svnrepo/conf/passwd

Edit/add user

Create project directory:

> svn -m "Your comment" mkdir file:///data/svnrepo/sample-project
> svn -m "Your comment" mkdir file:///data/svnrepo/sample-project/trunk
> svn -m "Your comment" mkdir file:///data/svnrepo/sample-project/branches
> svn -m "Your comment" mkdir file:///data/svnrepo/sample-project/tags

Import the project:

> svn -m "Your comment" import /path-to-source/sample-project file:///path-to-svnrepo/svnrepo/sample-project/trunk

Prepare work location:

> mkdir sampleProject
> cd sampleProject

Local check out trunk of project (without the folder itself):

> svn co file:///data/svnrepo/sample-project/trunk/ .

Remote check out (using SVN + SSH protocols):

> svn co svn+ssh://user@host/path-to-svnrepo/svnrepo/sample-project/trunk/ .

Create user and group svn:

> useradd svn

Add group svn to your groups.

Change ownership of all files in svnrepo/db to svn:

> chown -r svn:svn *

Add write permissions to all files in svnrepo/db.

Microphone Not Working in Fedora 14

2010-11-22T19:16:00.004+01:00

If your microphone is not working on your new installation, just try this. It worked for me:

Go to System --> Preferences --> Sound
Select the input tab
Switch from the existing configured microphone (probably microphone 1) to another one (microphone 2)
Close the application and everything should work ;-)

It looks like this issue has been going on and off since Fedora 10...

Spring Property Editor for an Enumeration with Localized Names

2010-11-19T12:06:00.005+01:00

Problem statement:

You use a property editor when you bind a nested Enumeration field of a form backing object.
You have localized names for the values.

When you write your PropertyEditor you should remember that the getAsText() method needs to take care of the translation:


    @Override
    public String getAsText() {
        
        if(getValue() == null) {
            return null;
        }
        
        SurveyStatus status = (SurveyStatus)getValue();
        return status.getTranslatedName();
    }

while the setAsText() method does not:


    @Override
    public void setAsText(String surveyStatusStr) {
        
        if (logger.isDebugEnabled()) {
            logger.debug("Read string [" + surveyStatusStr + "] for the survey status field");
        }
        
        SurveyStatus status = null;
        
        if (surveyStatusStr.compareTo(SurveyStatus.ACTIVE.toString()) == 0) {
            status = SurveyStatus.ACTIVE;
        }
        else if (surveyStatusStr.compareTo(SurveyStatus.INACTIVE.toString()) == 0) {
            status = SurveyStatus.INACTIVE;
        }

        setValue(status);
    }

This is because Spring implicitly uses the enumeration name for the value of your tag, not the translated name.
E.g. If you have a select tag, this is what Spring does for you:


<option value="ACTIVE" ... >

It makes sense, of course...

Grep Selected Files

2010-09-15T14:07:00.002+02:00

You can use find and grep to search for a particular string (string_pattern) in a selection of files:

> find <dir> -name "<file_find_pattern>" -exec grep -H -n '<string_pattern>' {} \;

Spring Open View In Session Pattern For Portlets

2010-09-09T11:18:00.011+02:00

I needed to implement the Open Session In View pattern for a portlet using Spring (3 with annotations) and Hibernate.
This pattern allows lazily loaded objects to have access to the Hibernate Session for the entire processing of the render request despite the original transaction (associated to the portlet request) already being completed.
I.e. If you have One To Many associations, this allows you to load the Many side of the association in the same thread of the request where you loaded the One side.

What I needed to do is:
Add a WebRequestHandlerInterceptorAdapter interceptor wrapping the stantard OpenSessionInViewInterceptor to applicationContext.xml:


<bean name="openSessionInViewInterceptor"
   class="org.springframework.web.portlet.handler.WebRequestHandlerInterceptorAdapter">
   <constructor-arg>
       <bean class="org.springframework.orm.hibernate3.support.OpenSessionInViewInterceptor">
           <property name="sessionFactory" ref="sessionFactory"/>
       </bean>
   </constructor-arg>
</bean>

then add it to the DefaultAnnotationHandlerMapping in yourPortlet-portlet.xml:


<bean id="annotationMapper"
   class="org.springframework.web.portlet.mvc.annotation.DefaultAnnotationHandlerMapping">
   <property name="order" value="10" />
   <property name="interceptors">
       <list>
           <ref bean="openSessionInViewInterceptor"/>
       </list>
   </property>
</bean>

Fedora: Receiving Files Over a Bluetooth Connection

2010-05-12T13:07:00.003+02:00

This post refers to receiving files from a phone to a computer over a Bluetooth connection.
Please refer to the official Fedora documentation for installation and initial configuration:

Then you have to enable the bluetooth file sharing service.
If you can't find the link to the application in Applications --> System Tools, then create the link yourself or launch the following command manually:
> gnome-file-share-properties&

Enable Receive Files over Bluetooth.
Go to your phone and send the file to your computer.

Hibernate ManyToOne Association With Annotations

2010-04-16T16:13:00.005+02:00

It took me a while to get rid of a:
ERROR org.hibernate.util.JDBCExceptionReporter -
Unknown column 'applicatio0_.domain_domainID' in 'field list'

on a simple many-to-one association.
I had an Application object with a Domain field where many Applications could have the same Domain (standard many-to-one association).
The annotations on the field were:


@ManyToOne(cascade = CascadeType.ALL)
public Domain getDomain() {
    return domain;
}

Class Domain started with:


@Entity
@Table(name="domain")
public class Domain implements Serializable {
// DB name: domainID
private Long id;
private String name;

Hibernate couldn't correctly map the domain ID until I added the JoinColumn annotation:


@ManyToOne(cascade = CascadeType.ALL)
@JoinColumn(name="domainID")
public Domain getDomain() {
    return domain;
}

which in my opinion could be inferred from class Domain:


@Id
@GeneratedValue(strategy=GenerationType.AUTO)
@Column(name = "domainID", unique = true,
        updatable = false, nullable = false)
public Long getId() {
    return this.id;
}

but it wasn't.
Always use JoinColumn.

Creating a Movie from Images (Linux)

2010-03-11T11:19:00.004+01:00

The short story at this time. These instructions refer to the creation of MPEG4 video from JPGs images on Linux.
You need mencoder, part of the mplayer software.

Type:

> mencoder mf://<PATH>/*.jpg -mf w=320:h=240:fps=12:type=jpg
-ovc lavc -lavcopts vcodec=mpeg4:mbd=2:trell
-oac copy -o output.avi

Where you should customize at least:

<PATH>/*.jpg is the path to the directory where your JPGs are.
w=320:h=240 is the image size.
output.avi is the resulting file name

Eclipse: Copying Preferences To a New Workspace

2010-03-08T16:55:00.003+01:00

When I create a new workspace in Eclipse what I want every time is to have all my preferences from the other workspaces, first of all the color preferences. There is no apparent option in Eclipse to do that. File --> Import --> Preferences won't do that. Maybe it's hidden somewhere else.
I found an easy way to do that: simply copy the settings directory to the new workspace.
Preferences are stored in each workspace in:
/.metadata/.plugins/org.eclipse.core.runtime/.settings

Shut down Eclipse.
Copy part or the entire contents of that directory to the corresponding directory of the new workspace (after you've created it, of course).
Restart and you're done!

Turning a SVN Project into a SVN/Java Project

2010-03-01T18:01:00.006+01:00

This will be needed for a Java project created with the type "from SVN source". It will not have a Java nature by default.

Add a Java nature to the project in .project:


 <natures>
     <nature>org.eclipse.jdt.core.javanature</nature>
 </natures>

Create the file .classpath with the JDK/JRE libraries (and optionally the default output directory):


 <?xml version="1.0" encoding="UTF-8"?>
 <classpath>
    <classpathentry kind="con" path="org.eclipse.jdt.launching.JRE_CONTAINER"/>
    <classpathentry kind="output" path="bin"/>
 </classpath>

Then you might want to add to the classpath the libraries which could not be detected at creation time.

This method is alternative to the one detailed here.

Eclipse: Hiding Internal JARs

2010-03-01T16:26:00.002+01:00

If you need to hide all the internal libraries your project is using, you can do so from the Package Explorer:

Choose the view menu
Check Show referenced libraries node

You can't do that from the Project Explorer (formerly Resource Navigator) since that is a view not reserved to Java projects.

Eclipse Losing Reference to JARs

2009-07-28T15:53:00.004+02:00

I have this weird problem that every once in a while (especially upon checking in files to SVN/CVS but not limited to that situation) Eclipse will lose reference to class path JARs and start signalling errors everywhere in the code.
What I temporarily did to fix this was to remove a library directory from the [Project] Properties --> Java Build Path --> Source list of a project, save the properties, open them again and add it back in. Problem gone... for a while.

Now I'm trying something I read somewhere: start Eclipse with the -clean command line parameter.

I'll keep you posted...

Feb 2010 Update: with the current version of Eclipse I'm using (Eclipse Java EE IDE for Web Developers, Build id: 20100218-1602) this issue seems to be vanished.

Booting Debian in Text Mode

2009-07-15T14:52:00.002+02:00

# disabling login GUI (if not gdm replace with yours)
> update-rc.d -f gdm remove

# re-enabling
> update-rc.d -f gdm defaults

No Javadoc When Hovering Cursor Over Text in Eclipse

2009-07-03T14:30:00.004+02:00

I had this weird problem that no javadoc preview showed up when hovering the cursor over java code in Eclipse.

Eclipse version: eclipse-java-ganymede-SR2-linux-gtk
OS: Linux (Fedora 11)

I didn't find out why it didn't work but I found out that it was specific to the version of Eclipse I was using.
When I upgraded to Eclipse Galileo (eclipse-java-galileo-linux-gtk), with exactly the same configuration files and preferences, it worked straight away.

I guess it was a bug of the Ganymede SR2 then...

Unable to Start Eclipse On An Empty Workspace

2009-06-12T13:05:00.002+02:00

OS: Fedora 11
Eclipse Version: 3.4.2, Build id: M20090211-1700

If I try to initialize a new workspace (empty but with an existing directory), Eclipse will open as an empty non-killable window.

I was able to have a "dirty" initialization copying the .metadata directory of another workspace I had previously and mysteriously created to the new workspace.

I need to experiment a bit more to see if it's possible to, for example, just create the .metadata directory...

rpmdb: PANIC: fatal region error detected

2009-06-11T17:19:00.003+02:00

I had the following error when trying any yum operation on Fedora 11:

> yum update
Loaded plugins: refresh-packagekit
rpmdb: PANIC: fatal region error detected; run recovery
error: db4 error(-30974) from dbenv->open: DB_RUNRECOVERY: Fatal error, run database recovery
error: cannot open Packages index using db3 - (-30974)
error: cannot open Packages database in /var/lib/rpm
Traceback (most recent call last):
File "/usr/bin/yum", line 29, in
yummain.user_main(sys.argv[1:], exit_code=True)
[...]

The googled ;-) solution that worked for me was:

> cd /var/lib/rpm
> rm -f __db.00*
> db_verify Packages
> rpm --rebuilddb

Eclipse: Could not launch the product...

2009-05-28T17:23:00.004+02:00

Eclipse crashed or in my case took too long to open a file and I had to shut it down in a not clean state (really surprised it didn't crash during my attempt AND I didn't have to kill it! Great job people@eclipse!).
Next thing that happens when I try to re-launch it using the same workspace is an error message:

Could not launch the product because the associated workspace is currently in use by another Eclipse application

You need to:

go to the workspace home directory
enter the .metadata directory
remove file .lock

and you can happily restart from where you left your job.

JUnit, log4j And Eclipse

2009-05-14T15:29:00.003+02:00

If you want Eclipse to pick up your log4j configuration file when you run JUnit tests from within Eclipse you can:

Right-click on a test class.
Select Run as --> Run configurations...
Select the Classpath tab
Select the home directory of you project and Advanced
Then Add folder and add the folder with the configuration file and you're done.

This will configure the test file's JUnit run options.

If you want to configure all test classes with the same optoins... uhmm... stay tuned. I'll be back.

Fedora 10: Firefox Starts up in Offline Mode

2009-03-25T11:36:00.005+01:00

The problem: upon strting Firefox it goes in offline mode even if the network is enabled and working. It's really unconvenient to un-check Work offline from the File menu every time especially if you start up many windows and tabs.
The quick solution is to stop and disable the Network manager application from the System --> Administration --> Services window.
Unless you need it... in that case I can't help you right now...

gedit Color Schemes

2009-03-06T14:30:00.003+01:00

If you upgrade to Fedora 10 right now, the new version of gedit doesn't explicitly allow you to set up a custom color scheme. They say in the future there will be an editor for that.
In the meantime what you can do is:

> cd /usr/share/gtksourceview-2.0/styles/
Copy one of the available schemes to a new one and edit that file to suit your needs.
Once you restart gedit, the new color scheme will be available under Edit -> Preferences -> Font & Colors

Adding a Script to The Gimp (Linux/UNIX)

2009-01-29T15:47:00.004+01:00

Create an *.scm file with the script. To create a script you can follow this tutorial.
It's recommended to keep them (scripts) in your personal settings. Enter your own gimp directory: ~/.gimp-2.6 or similar, the scripts sub-directory and copy the file there.
The place in the menus it will be placed is set in the following command (usually a the end of the script): (script-fu-menu-register "script-fu-photo-scriptName" "/Script-Fu/Photo/"). This will place the script in the Script-Fu --> Photo menu.
Then you need to run Filters --> Script-Fu --> Refresh Scripts command.

To-Haves When Using Eclipse (Java development)

2008-12-04T15:27:00.003+01:00

CloudGarden Jigloo: SWT/Swing GUI builder
Tigris Subversion: SVN repositories management (Subclipse is one client for Eclipse)

Converting MPEG-4 to MP3

2008-12-02T10:53:00.004+01:00

Instructions for Fedora Linux.
Install lame:
> yum install lame

Install FAAD.
You need to compile it. Instructions are in the README.linux file in the source package.

Then you can use this command:
>

faad -o - "file.m4a" | lame - "file.mp3"

References

Lame, http://lame.sourceforge.net/
FAAD, http://sourceforge.net/projects/faac/

Ant unable to locate javadoc.exe

2008-09-22T18:07:00.003+02:00

The situation: you're launching an Ant build file from within Eclipse on Windoz and the file has a Javadoc task. But Eclipse can't find the executable.

What you need to do is add the JDK bin directory to the PATH environment variable.
There is nothing to configure in Eclipse.

Configuring the Class Path When Testing in Eclipse

2008-09-09T17:55:00.005+02:00

If you're testing from within Eclipse and are using Apache Commons Logging + log4j and you get the message:
log4j:WARN No appenders could be found for logger

you can let log4j know about the location of the log4j configuration file adding the directory where that file can be found this way:

Right click on the test class and choose Properties --> Run/Debug Settings
Select the file.
Click on Edit and select the Classpath tab.
Select the first item in User Entries, then Advanced, then Add Folders and you're done.

Merging PDFs on Windows

2008-08-25T15:07:00.005+02:00

First you need Ghostscript.
Then, you can just issue the command:
> gswin32 -dNOPAUSE -sDEVICE=pdfwrite -sOUTPUTFILE=merged_file.pdf -dBATCH file_1.pdf file_2.pdf file_3.pdf

What I noticed is that the text in the merged file is a little bigger than the original one at the same magnification although nothing is missing.
Maybe you/we need to tweak a bit with the parameters...

Dual-boot With Grub After Installing Windows XP on a Linux Machine

2008-08-24T15:25:00.008+02:00

Your situation: Windoz will boot but Linux is now invisible.
What you want is to have Grub installed on the Master Boot Record (MBR) of the first disk to pilot the boot sequence.
You need a Linux distribution CD/DVD with the rescue mode as an option. Any distribution is fine even if different from the one you have installed.

Boot in rescue mode
Once you have a shell type:
> fdisk -l
> grub
grub> root (hd0, n)
grub> setup (hd0)
grub> quit

Notes
3. With fdisk -l you can check which partition holds the linux root partition (in case you forgot).
5. If your partition is, for example, /dev/sda2, then your value for n is 1 (Grub starts from 0)

Now you can exit the rescue shell and reboot. If you had a Grub configuration file configured to boot Linux and Windoz, you are done. You will be in the same exact position as before Winsoz zeroed Grub on the MBR.

If you didn't have Windoz before, you need to add to the Grub configuration file the Windoz entry:

title Windoz
root (hd0,m)
makeactive
chainloader +1

where m is the Windoz partition (Grub's way of course, see note above).

Screen Capture With Mac Keyboard and Windows Desktop

2008-07-27T15:57:00.002+02:00

The problem starts on the Mac keyboard: you don't have the prnscr key! But still what you have is a Windoz machine. What is working for me (there are external applications but I couldn't find a good free one) is to open the on-screen keyboard under Programs --> Accessories --> Accessibility. There you have it!
If you have the always-on-top option enabled, you loose control of the physical keyboard and sometimes I had to kill the virtual thing anyway to get back the other one. But I can live with that.

gpk-application Not Working

2008-07-17T16:14:00.005+02:00

I recently had this problem with the gpk-application on a Fedora 9. This Gnome application allows you to check what you have installed and install/remove packages.
The thing was I could not see the groups of packages and update the list from a remote machine. It happened I found out the cache was corrupted and this command:

> yum makecache

restored it to its full functionality.
Almost. I still cannot remotely refresh the list (which I could on the local display) but I think this is another story... permissions maybe.

Linux Graphical FTP Client

2008-07-15T10:43:00.003+02:00

I recently switched to gFTP, the FTP client from Gnome, after I apparently encountered problems connecting to a couple sites with Kasablanca (from KDE). One was that the client could not display the content of a remote directory no matter what I did.
I must say that gFTP is more solid and it did not have those glitches. I personally recommend it at this time.

Linux Toolbox: pdftk

2008-05-28T11:22:00.005+02:00

(from the man page) If PDF is electronic paper, then pdftk is an electronic staple-remover, hole-punch, binder, secret-decoder-ring, and X-Ray-glasses. Pdftk is a simple tool for doing everyday things with PDF documents. Use it to:

Merge PDF Documents
Split PDF Pages into a New Document
Rotate PDF Documents or Pages
Decrypt Input as Necessary (Password Required)
Encrypt Output as Desired
Fill PDF Forms with X/FDF Data and/or Flatten Forms
Generate FDF Data Stencil from PDF Forms
Apply a Background Watermark or a Foreground Stamp
Report PDF Metrics such as Metadata and Bookmarks
Update PDF Metadata
Attach Files to PDF Pages or the PDF Document
Unpack PDF Attachments
Burst a PDF Document into Single Pages
Uncompress and Re-Compress Page Streams
Repair Corrupted PDF (Where Possible)

Get it on Fedora:
> yum install pdftk

Note Unfortunately pdftk was removed from the Fedora repository due to a license issue that prevented it from being used in nuclear facilities. Yeah, right. Okay. Whatever!
You have to recompile it from the sources or... grab the final product from someone who already did it: Professor Kriehn's Fedora Repository!

Exporting the Graphical Display

2008-05-27T11:02:00.004+02:00

If you need to open a graphical application on a Linux box A and see it on another one B (i.e. exporting the DISPLAY) there is one easy way if you log in from B to A with SSH:

> ssh -X -Y user@host.domain

Then you can start a graphical application on A from the remote session on B and see the output on B.

Installing Linux on the Sony Vaio PCG FX-220

2008-05-12T11:42:00.008+02:00

I bought this notebook in August 2001. I will never buy a Sony again despite the pros for the following reasons:

Bad user support. If you have a problem, you have to pay to be able to speak with a Sony representative.
A memory slot had a problem which prevented it to be used. This is a common issue ([1]) with this series and Sony did nothing to fix that.
The CD/DVD reader broke in 2007 (I have a Fujitsu notebook from 1996 which is still perfectly working).

Anyways, for the last reason and the fact that I needed to reinstall from scratch recently, I needed to go with boot diskettes and a network installation. Debian 4.0 was the easiest choice from a first overview (I had already installed Debian a few years ago on this notebook).

Installation

You need the boot, root and network-drivers-1 diskettes.
Start at the boot with: expert hw-detect/start_pcmcia=false
After that, installation starts. It might hang (it did with me) but it's a network timeout problem. Retry with a different mirror.
At the end of the installation, LILO and Grub won't install. Skip that step.
To reboot, use the boot diskette with the following parameters: rescue root=/dev/hda1 (replace with your settings). Load the drivers, networks-drivers-1 and 2 diskettes this time (might not be necessary but it might save your time).
Log in as root. Install Grub: apt-get install grub.
Configure Grub automatically creating its configuration file.
Reboot.
Start in single user mode because the default X.org configuration file won't work.
Edit the X.org configuration file manually removing all references to screen resolutions above 1024x768 and above 16bpp (TODO to be verified, 24bpp might work)
Reboot.
Check that the network parameters are correctly set.

Components

USB: working out of the box.

Useful Links

Logitech Quickcam Pro 4000 on Linux

2008-05-06T14:10:00.011+02:00

This post is about using a Logitech Quickcam Pro 4000 on Linux (Fedora 8) to store snapshots and producing a movie.

# Install xawtv (xawtv-3.95-7.fc8)
> yum install xawtv

# Install mencoder (mencoder-1.0-0.90.rc2.lvn8 from the livna repository)
> yum install mencoder

The xawtv package contains tools which allow you to connect to the webcam, grab pictures and videos on the fly and also produce movies from the snaps.

# Check that the webcam is recognized and broadcasting:
> xawtv -c /dev/video0 &

/dev/video0 : the device (use xawtv -hwscan to see which one it is for you)

# You can grab pictures in two ways with xawtv: using webcam or using streamer.

# grab snapshots using streamer
> streamer -c /dev/video0 -j 85 -t 1200 -r 0.03 -o office-0000.jpeg

/dev/video0 : the device
85 : JPEG quality
1200 : number of snapshots to grab
0.03 : frames per second (here it's about 1 snap every 30 seconds)
office-0000.jpeg : output file with counter to accomodate 4 digits here

# grap snapshots using webcam
> webcam webcam.conf

All configuration is read from file. My webcam.conf is:

[grab]
device = /dev/video0
text = Atlanta GA - %Y-%m-%d %H:%M:%S
#infofile = infofile.txt
fg_red = 255
fg_green = 255
fg_blue = 255
width = 320
height = 240
delay = 30
wait = 0
#input = camera
brightness = 100
#norm = pal
rotate = 0
top = 0
left = 5
bottom = -1
right = -1
quality = 85
trigger = 0
once = 0
archive = office-%Y%m%d-%H%M%S.jpg

TODO Explanation to follow...

# make movie using mencoder (you can also use streamer but I didn't try it yet)
> mencoder "mf://${CURR_DATE}/*.jpg" -msglevel all=-1 -mf fps=12:type=jpg -ovc lavc -lavcopts vcodec=mpeg4:trell:mbd=2:dc=10 -oac copy -o "movie_${CURR_DATE}.avi"

... to be continued

Fedora Extras @ livna

2008-04-30T14:29:00.002+02:00

rpm.livna.org provides many useful packages that can not be distributed in Fedora (previously known as Fedora Core and Extras) for one reason or another, including multimedia applications.

rpm.livna.org is not a standalone repository, but an extension of Fedora, Fedora Core and Extras repositories.

To add livna Fedora 8 packages to your list of repositories:
> rpm -ihv http://rpm.livna.org/livna-release-8.rpm

Installing Xine on Fedora 8

2008-04-29T15:27:00.007+02:00

This procedure details the installation using the available prepackaged RPMs. The trick is that they're not all in the same place!

Install the main xine libraries (available from the main Fedora repositories):
> yum install xine-lib

Grab the main package and dependencies from RPM Find [1] and RPM pbone.net [2]:

xine (xine-0.99.5-1.fc7.i386.rpm) [1]
xine-lib-moles (xine-lib-moles-1.1.11.1-1.fc8.i386.rpm) [1]
vcdimager (vcdimager-0.7.23-7.fc8.i386.rpm [2]
libdvdcss (libdvdcss-1.2.9-2.fc7.i386.rpm) [1]
libfame (libfame-0.9.1-12.fc7.i386.rpm) [1]
libmad (libmad-0.15.1b-4.fc7.i386.rpm) [1]

Note that on Fedora 8 you need the above version of xine-lib-moles for compatibility with xine-lib.

Install them directly with rpm to avoid the need of disabling yum's signature check:

> rpm -ihv libdvdcss-1.2.9-2.fc7.i386.rpm libfame-0.9.1-12.fc7.i386.rpm libmad-0.15.1b-4.fc7.i386.rpm vcdimager-0.7.23-7.fc8.i386.rpm xine-0.99.5-1.fc7.i386.rpm xine-lib-moles-1.1.11.1-1.fc8.i386.rpm

Another way is to add livna to your list of repositories and do a:
> yum install xine

Didn't try this though...

Java and Firefox 3 Beta 5 on Linux

2008-04-17T10:35:00.007+02:00

This post is about configuring Java with Firefox 3 beta 5 on Linux i386. It might work for similar versions of Java and Firefox though.
The JRE coming with JSE v1.6.0_03 is currently not working. BUT the one coming with JSE v1.5.0_08 does work!
You just need to follow the usual instructions (which can be found here) and create a symbolic link the the Java plugin library:

> cd ${FIREFOX_HOME}/plugins
> ln -s ${JAVA_HOME}jre/plugin/i386/ns7/libjavaplugin_oji.so libjavaplugin_oji.so

Restart Firefox and check at this URL: about:plugins if it's loaded and here if it works.

Note: Actually "works" is perhaps a little bit of an overstatement. I found out an applet which is supposed to update itself with streaming data and is not doing it. So maybe it's a matter of waiting on Firefox's final version...
On the other side Firefox 3 is much much faster and less memory consuming than Firefox 2.

Verifying Which Ports Are Listening

2008-03-18T16:55:00.002+01:00

The most reliable way is to use a port scanner like nmap.
You can install nmap in a blink of an eye on a Red Hat system (Fedora included) with yum install nmap or rpm -ihv if you have the package or in many other ways lasting a little bit longer than the blink of an eye.

Then issue:

> nmap -sT -O localhost

Reference: Red Hat Linux Manual v9

Enabling the CVS Id Tag for SVN

2007-12-20T17:15:00.002+01:00

The CVS Id tag which adds file information on the file itself upon commit is enabled by default on CVS but not on SVN. To have it enabled you need to add/modify the following on your local SVN configuration file (on UN*X: <HOME>/.subversion/config):

[...]
enable-auto-props = yes

[...]
[auto-props]
[...]
*.java = svn:keywords=Author Date Id Revision;svn:eol-style=native

Complete list of keywords: Author, Date, Header, Id, Log, Locker, Name, RCSFile, Revision, Source, State.

Of course, you need to add a line for every file type you want to configure for a fine grained control over that. Otherwise use *.

This works for all files committed from now on. If you already have files in the repository, you need to tell SVN to add them too:

# Make sure that everything is up to date
> svn up
# Add keywords and commit
> svn propset svn:keywords "Author Date Id Rev" file_name
> svn commit -m "Adding Id and Rev property to all files"

SVN Java Project With Eclipse

2007-11-16T15:38:00.003+01:00

If you have your source code in a SVN repository and want to set up a Java project in Eclipse (tested on the Helios release), you cannot simply do a New --> Project --> SVN --> Checkout projects from SVN.
The resulting project will not be a Java project and you won't be able, for example, to set up the build path with all its implications...

This is the way to go:

Choose Window --> Open perspective --> SVN repository exploring, configure the repository.
Open it and choose the trunk, a tag or a branch.
Do a right-click and do a Find/Checkout As (e.g. as a Java project)

JSTL Taglib Declaration

2007-10-08T17:04:00.000+02:00

This post is related to web application development with Spring 1.x, JSTL and Tomcat v5.5.x.
Be sure to have this configuration to be able to use JSTL:

web.xml header (remove space after <):


< ?xml version="1.0" encoding="ISO-8859-1"?>

< xmlns="http://java.sun.com/xml/ns/j2ee" 
   xsi="http://www.w3.org/2001/XMLSchema-instance" 
   schemalocation="http://java.sun.com/xml/ns/j2ee   
   http://java.sun.com/xml/ns/j2ee/web-app_2_4.xsd" version="2.4">

Taglib declaration:


<%@ taglib prefix="c" uri="http://java.sun.com/jsp/jstl/core" %>

Tomcat: Redirect all requests to port 80 on port 8080

2007-08-23T15:58:00.006+02:00

... the IPTables way! (there are other solutions)
Configure the iptables behavior in file /etc/sysconfig/iptables-config to save the new rules

IPTABLES_SAVE_ON_STOP="yes"
IPTABLES_SAVE_ON_RESTART="yes"
IPTABLES_SAVE_COUNTER="yes"

Manually add the following rules:


> iptables -t nat -A OUTPUT -d localhost -p tcp --dport 80 
   -j REDIRECT --to-ports 8080
> iptables -t nat -A OUTPUT -d <your_ip> -p tcp --dport 80 
   -j REDIRECT --to-ports 8080
> iptables -t nat -A PREROUTING -d <your_ip> -p tcp --dport 80 
   -j REDIRECT --to-ports 8080

Save and restart the service:

> /etc/init.d/iptables save
> /etc/init.d/iptables restart

Certified Procedure for Unsubscribing From a Mailing List

2007-06-14T11:00:00.000+02:00

I recently had the following problem: I wanted to remove from a mailing list I previously subscribed to but using the unsubscribe link they provided ended up to be useless. It seemed to work but in the end I kept receiving their emails.
I never thought of writing them to be removed since in my experience it's a total waste of time.
The good news is there are other solutions!
These services always offer you the possibility to update your personal info and included with that you can change your email address. I mean, they cannot loose a "valued customer" if he decides to change his email so that update link will work!
Given that, what you can do is use one of those incredible services like Mailinator to have a one time email you can put as your new email address. Of course you will be able to answer to their confirmation request by reading the only email you will get at that address, and after that forget about the whole thing!

Configuring Thunderbird Reply Header

2007-05-24T12:00:00.001+02:00

The default reply header can be customized if you create the file user.js in the profile directory of your Thunderbird home (if you have Windoz, look inside the Doc&Settings/AppData/Thunderbird or something like that) and fill it with the following:

// Change the reply header
// 0 - No Reply-Text
// 1 - "[Author] wrote:"
// 2 - "On [date] [author] wrote:"
// 3 - User-defined reply header. Use the prefs below in conjunction with this:
user_pref("mailnews.reply_header_type", 3);

// If you set 3 for the pref above then you may set the following prefs.
user_pref("mailnews.reply_header_authorwrote", "%s said the following");
user_pref("mailnews.reply_header_ondate", "on %s");
user_pref("mailnews.reply_header_separator", " ");
user_pref("mailnews.reply_header_colon", ":");
// The end result will be [authorwrote][separator][ondate][colon]

Using Apache Commons Logging and Log4j

2007-03-22T11:18:00.000+01:00

It's as easy as this:

Download the latest and the greatest from Log4J.
Download the latest and the greatest from Apache Commons Logging.
Add the libraries to the $CLASSPATH.
Write a Log4J properties file called log4j.properties and place it in a directory in the $CLASSPATH.
You can now start logging in a class simply obtaining a logger with:

private static final Log log = LogFactory.getLog(YourClass.class);

and using it:

log.info("your logging comment");

Converting to PDF from Openoffice (Linux)

2007-01-04T13:19:00.000+01:00

There are basically 2 options to convert an Openoffice doc into PDF:

Use File --> Export to PDF
Print to file using File --> Print, which gives you a Postscript version of the doc. Then use ps2pdf to convert to PDF.

Well, what I've seen is that method 2 gives a much better result graphically speaking (with method 1, you sometimes get mingled/not correctly spaced characters), and smaller files. So, I'd really suggest you use that for the task.

Converting to PDF Under Windoz

2007-01-04T13:04:00.000+01:00

First you need to get your doc into Postscript (PS) format. One way is to print to file using any PS printer (I use the HP Color Laserjet 8550-PS). Of course you don’t need to have it!

Install Ghostscript and Ghostview.

Open a PS file with Ghostview to be able to make the initial configuration. Choose from the menu: File –> Convert.

Select:

Device: pdfwrite

Resolution: 720

Options: -dMaxSubsetPct=100 -dCompatibilityLevel=1.2 -dSubsetFonts=true -dEmbedAllFonts=true

You’re set!

Reference: Creating PDF Files in Unix and Windows

Brand New Start!

2007-01-04T12:55:00.000+01:00

I was about to move to another blog service because I really wanted some features (like labels) which Blogger didn't have and I had no hint they were working on them when... a brand new Blogger... with labels!
That's why I'm back ;-)
This time though it will be a whole different blog. Don't have much time to write now. But I'll occasionally blog about... anything. What there will be for sure are some computer-related tips.
Zeeya!