Java Hamster

Downgrade svn from 1.6 to 1.7

2012-01-27T22:00:00.000+02:00

I'm using a Sabayon as my primary OS now. It's a rolling distro. And it fairly quickly incorporates latest changes from mainline projects.

Few weeks ago among other updates subversion was upgraded from 1.6 to 1.7. At first there were no problems, because I mainly used command line svn client. I upgraded my working copies to 1.7 and was happy.

The problem started to arise when I tried to work with upgraded working copy from my IDE - Idea. The first problem was that Idea does not support subversion 1.7. And will not support it for a few more month.

So, I tried to downgrade subversion package on my OS. This failed, because 1.6 was removed from repository.

There was a chance to build 1.6 from sources. But I prefer not to mix binary and compiled from sources packages.

I ended up with solution suggested on stackoverflow. I used another OS (Ubuntu) on my multi-boot machine which still has subversion 1.6. From there I deleted a subversion folders from my projects. Which is easy in subversion 1.7 ;) Then I made a checkout to existing projects. This allowed me to not loose any changes.Then I rebooted to my primary OS and used subversion from IDE only.

Fixing UI font in Idea

2011-09-03T00:00:00.000+03:00

Recently I switched my Linux OS from Ubuntu to Sabayon. And I've noticed the UI fonts of Idea (many java developers would agree IntelliJ Idea is the best IDE) are bad. Probably something related to anti-aliasing went wrong. The fix is easy, just add following line to your {idea-installation-directory}

/bin/idea.vmoptions:

-Dawt.useSystemAAFontSettings=on

Updating versionCode and versionNumber in Android manifest via Gradle

2011-07-21T11:18:00.000+03:00

I'm working on Android project which has multiple builds for different customers.

Each build is organized as separate project and is build via Ant script. The problem is that building all those builds and updating versionCode and versionNumber is all manifest files takes a lot of time. So, I decided to use some tool to this for me. Yes, I'm a bit lazy :)

I've chosen Gradle. It leverages the power of Groovy and is able to call Ant scripts.

Groovy has great tools to work with XML, like XmlSlurper. So, updating versionCode and versionNumber in Android manifest files seemed easy. Unfortunately after significant amount of time all I got was ugly formatted Android manifest with tag0 namespaces :(

Then I put XmlSlurper aside and used String.replaceAll(Pattern pattern, String replacement) method. It took me about 5 minutes to do all replacements :)

Goodbye to Firefox Sync and welcome back to Xmarks

2010-07-04T17:45:00.000+03:00

Some time ago I was using Xmarks for bookmark synchronization. Then I switched to Firefox Sync. I used only Firefox and only on Linux powered machines. Firefox Sync worked perfectly and is synchronized not only bookmarks, but also other useful things like history.
But with starting from Firefox Sync 1.2.3 (it was called Weave Sync back then) it failed to synchronize bookmarks on x64 Linux. The problem was still present in 1.3 as well as in latest 1.4.
So, I gave up and switched back to Xmarks. The latter doesn't have many fancy features. But at least it works.

Wine in Fedora 13 x64

2010-05-25T00:10:00.007+03:00

The shiny new Fedora 13 is to be released in a few hours. I've installed its x64 variant since beta. And I've spent a lot of time to make wine work in Fedora. And it actually works with PulseAudio :) So, if you are going to use wine in Fedora 13 x64 here is some info which might be helpful.

Getting the latest wine version

At the time of this writing latest wine version in Fedora repos is 1.1.38, in upstream 1.1.44. The latter is also available if Fedora testing repo. So, you might want to enable that repo while installing wine:

yum --enablerepo=updates-testing install wine

Getting rid of crash on audio tab

x64 build of wine is not perfect yet. It actually crashes on audio tab. So, use 32-bit version to configure wine:

wine32 winecfg

It's actually a good idea to use wine32 instead of wine.

Adding readable wine fonts

Default wine font is hardly readable. You should have sharp eyes and good imagination to understand what is written on wine dialogs :)

Installing corefonts with winetricks should fix it:

yum install cabextract
wget  http://www.kegel.com/wine/winetricks
chmod a+x winetricks
./winetricks corefonts

Adding mp3 support

Some games (including Morrowind) require mp3 support. Fedora's wine without mp3 support in contrast to Ubuntu. To enable it you should install a mp3 codec. It is recommended to use l3codecx codec, then rename l3codecx.acm file to winemp3.acm. Alternatively you can grab that file from Windows, if you have one at hand. I had no Windows, so I tried installation approach. I've searched for l3codecx and found a few installators (for Windows, of cause). One of them failed in wine but another one worked. l3codecx.acm file appeared in $HOME/.wine/drive_c/windows/syswow64 folder. Instead of renaming it I just made a symbolic link:

cd .wine/drive_c/windows/syswow64 folder 
ln -s l3codecx.acm winemp3.acm

And that's all :)

Android's sharedUserId

2010-05-19T22:09:00.002+03:00

Android is a wonderful platform for development. It also provides good security framework. Each application has its own user ID and thus application A can't directly access the data of application B. So, Android is quite resistant to viruses plague. In contrast to Windows.

But this security makes it difficult to develop a set of applications which should work tightly with one another, e.g. browser and its plugins, translator and its dictionaries, etc.

To address this problem Android provides the developer with sharedUrerId attribute in manifest. Setting the same sharedUserId to multiple application and signing them with same signature will enable them to access one another's data. Great!

But... but there are two nasty bugs related to sharedUserId which you must take into account.

Issue #3763 messes up if you decide to change the sharedUserId value.
Issue #1227 bites you if you add sharedUserId to your application.

In both cases after upgrade to the new version your application won't be able to access its own data (probably, because its user ID was changed, but chown was not called on its data). You'll see some strange errors in console. Like database can't be opened.

So, if you are developing some application for Android and if there is even a slightest chance that the application might need to use sharedUserId attribute - use it right from the start. Before uploading the application to the Market. And don't change sharedUserId value in the future.

Presentation and demo app from my talk about Android

2010-03-20T22:47:00.002+02:00

I had a talk on local Java User Group a week ago. I was talking about Android. Here I'm providing the presentation and demo application.

Presentation mainly contains materials from developer.android.com.

Demo application is a twitter client. It shows:

how the project is layed out;
what contains the manifest file;
how activity is build;
how layouting is done (including reusable layouts with <include> tag);
how locales are handled (contains Spanish translation);
how orientation is handled;
how background threads are handled.

Links:

Presentation: Android.odt (yes, that's OpenOffice format).

Demo application: Twitoid.7z (IntelliJ IDEA project, free Community Edition is perfectly fine to check this code).

Twitter4j is used for communication with twitter. Icons are taken from medialoot.com.

Ignoring multiple files in Subversion using command line

2009-12-10T10:50:00.000+02:00

Ignoring single in Subversion file/folder is as easy as

svn propset svn:ignore bin .

This would make svn ignore folder bin in current directory (final dot is mandatory).

But ignoring multiple files is a little bit more complicated. Firstly you should have one of environment variables SVN_EDITOR, VISUAL or EDITOR set. This can be done as easy as

export SVN_EDITOR=geany

This would make Geany you svn editor. This assumes Geany is installed of cause :)

Then simply start your svn editor and add all files you want to ignore

svn propedit svn:ignore .

The editor would start and you can add as many files as you like, one per line. Again, final dot means that you edit ignored files in current folder.

That's it :)

Netbeans 6.8 M2 dropped embedded browser

2009-10-10T09:36:00.000+03:00

Netbeans 6.8 M2 was just released. The sad news for Linux users is that it misses the embedded browser for this OS. It was present in M1, but dropped in M2 :(

Accessing system properties from Ant build file

2009-08-26T18:59:00.003+03:00

I need to do some tasks via shell scripts. Among those tasks are also ant scripts. Sometimes I need to pass some arguments to ant script from shell scripts. I've found that this can be done through java system properties. Those properties are easily accessed from ant build file - just like properties declared in build file itself: ${property.name}
Here is a brief example. The build file cool-build.xml:


<?xml version="1.0" encoding="UTF-8"?>
<project name="FunProject" default="work" basedir=".">
    <target name="work">
        <echo message="my.cool.property=${my.cool.property}">
    </target>
</project>

We, run it with passed system property my.cool.property:
ant -f cool-build.xml -Dmy.cool.property=1
And the result is:
[echo] my.cool.property=1
Just as expected :)

Java browser

2009-08-26T14:07:00.005+03:00

NetBeans 6.8 M1 was released a few days ago. And it came with its own embedded browser! The browser looks good. Really good. Of cause there are some problems: crashes, hangs, no flash support. I also did not managed to login to my gmail account with it :) But hey, it's just the initial release, so we can't be very demanding.
It's worth notion that the browser is based on Mozilla XULRunner. Which means that it is not related with mysterious JWebPane (based on WebKit). So, it looks like Sun is forging two java browsers. Wow!

Failure with boosting HDD performance

2009-03-07T23:29:00.004+02:00

Today I've read an article at tuxradar.com with some Linux tips. In this article it is recommended to boots the performance of HDD. It is suggested to set options for root to defaults,relatime,data=writeback in /etc/fstab
So I did. Then rebooted and.. guess what? My Mandriva One 2009 failed to boot. Luckily this was not my first boot failure (and I guess not the last :D ). So, I had a SystemRescueCd an some live CDs at hand.
First of all I've checked man for mount to get some info about relatime and data=writeback options. There is nothing special with former, but the latter is not so easy to use. Changing default value requires passing additional kernel boot parameter. E.g. rootflags=data=journal. So modified the kernel row in my GRUB configuration file. But Mandriva still failed to boot. Then on some mailing list I've found that rootflags should go directly after root option. This did not helped either.
So, I changed all settings to original state. No faster HDD operations for me. At least for now...

P.S. Just curious whether guys at tuxradar.com actually try to apply theirs tricks themselves before urging others to do so.

Disabling Automounting HDD Aartitions in GNOME

2009-02-20T22:20:00.006+02:00

I have a multiple Linux distros installed on my PC. And each of them has its own /boot partition. The most used Linux is Mandriva 2009.
It seems to me that things went wrong after I played a bit with gparted in Mandriva, but I'm not 100% sure. Anyway for unknown reason GNOME started mounting all partitions on the hard drive on my login.
Not only this slows down the login time. This is also really annoying. In a Linux world when you like OS to mount some partition you simply edit /etc/fstab. And when the Linux is messing you up it starts to stink like Windows...
The core problem is for sure in user settings in GNOME because it neither happens for other users nor for other desktop environments.
Playing with a GNOME Configuration Editor did not helped :(
I even started to consider to switch desktop environment to xfce. But after some more googling and reading man pages for fstab and mount I finally found the solution.
I simply added a following line for each partition I like to be not mounted:
/dev/sda2 /mnt/sda2 ext3 noauto,nouser 0 0
And it works! :)

Making a microphone louder

2009-01-25T13:48:00.004+02:00

Recently I've installed Mandriva 2009 on my aged desktop machine. And there (in Gnome) was a problem with microphone in Skype 2.0.0.72. In fact the microphone was very-very quiet.
The solution was easy: I've installed gnome-alsamixer, run it and checked "Mic Boost (+20bD)" option.

Printing from Java with CUPS 1.3+

2008-11-01T23:31:00.000+02:00

I had a problem with printing from the Java application on my Fedora 9. The problem was quite strange and reproducible on both Java 1.5. and 1.6. It was even more strange because it appeared only after I've updated my Fedora to version 9.
It turned out that the problem was on the level where Java meets CUPS 1.3+. There is a corresponding Sun's bug. There is no such problem for CUPS version lover than 1.3. It looks like Java had made too many assumption how CUPS should work... and CUPS guys changed some things in 1.3.
As always Google came with help and the solution was found.

Java 1.4 on Fedora 9

2008-10-31T23:21:00.002+02:00

Recently I needed to work on a legacy Swing project written on Java 1.4. So, I've installed 1.4 JDK on my Fedora 9 and... got a following error:

/usr/java/j2sdk1.4.2_16/jre/lib/i386/libawt.so: libXp.so.6: cannot open shared object file: No such file or directory occurred

A little googling brought me to the Oracle forum post. The problem was that AWT requires the libXp package. After I've installed it the problem vanished :)

Java 1.6 on Mac... at last!

2008-05-30T22:45:00.005+03:00

It is here. After 1.5 years. Java 1.6 on Mac. But... it is only for the Leopard and only for 64-bit Intel machines.
PPC, 32-bit Intel and Tiger folks are currently still out of the board.
There are two interesting facts regarding this Apple's release.

It was not announced on Apple Developer Connection Java topics RSS. In fact that feed is useless. I'm still receiving here a bunches of 3 identical notifications about release notes for Java on Mac 10.5. The date when this document was published is 2007-10-31. Yeah, after 6 month Apple still reminds me (and the other subscribers of that feed) about such "news".
Almost month passed since Java 1.6 on Mac release. And there is complete silence about this event on the Desktop community on java.net. This is rather strange - there was a lot of complains about long absence of this release (e.g. here and here).

Anyway, it's good it is here.

Running TestNG on NetBeans 6.0

2008-05-07T00:03:00.005+03:00

TestNG is a great testing framework. And it's much more powerful than JUnit.
The problem for NetBeans 6.0 users is that there is no plugin for TestNG. I've spend a lot of time before I managed to run TestNG in a simple way. After some time I've figured out that the main obstacle is that NetBeans does not recognizes TestNG classes as test classes. And this was the key to success.
I've successfully run the tests using its command line and NetBeans Run feature. And it was done on the NetBeans Java SE bundle (the simplest one).
So, here are the simple steps to launch TaestNG on NetBeans:

Project Properties -> Libraries. Compile-time libraries must contain testng-XXX.jar (where XXX is your TestNG version). Also Run-time libraries must contain the path to your compiled tests (if you put your test to default test directory, than the path to compiled tests should be build/test/classes).
Create TestNG xml file(s). There are instructions how to build such file in the framework site. E.g. create testng.xml and put it in the project root.
Project Properties -> Run. Create new configuration. You may name is testNG ;) Enter org.testng.TestNG as the Main Class. Put your TestNG xml file(s) as Arguments. Multiple xml files should be separated by space. Enter your project directory as Working Directory.

Congratulations! Now you can run your TestNG tests through NetBeans Run feature.

P.S. Even if there will be only one person on Earth who saves a few hours this post worth its writing.

Regular method call vice call through reflection

2008-03-03T07:48:00.005+02:00

A few weeks ago I had a pleasure to discuss some Java topics with three Java guys. Each of them has much more experience than I do. One of those topics was a performance of method call through reflection vice regular method call. Those three guys were confident that the difference in performance is not significant. And I argued that it is. Our discussion ended, but I was curious what is that difference.
So, here is the little test I wrote, sound like the famous song of Bobby McFerrin ;) :


import java.lang.reflect.Method;


public class ReflectionSpeedTest {
   
    public static void main(String[] args) {
        int times = 100000;
        long simpleCallTime = simpleCall(times);
        long reflectionCallTime = reflectionCall(times);
       
        System.out.println("number of calls: " + times);
        System.out.println("simple call time: " + simpleCallTime);
        System.out.println("reflection call time: " + reflectionCallTime);
    }
   
   
    private static long simpleCall(int times) {
        ReflectionSpeedTest worker = new ReflectionSpeedTest();
       
        long time = System.currentTimeMillis();
       
        for (int i = 0; i < times; i++) {
            worker.doWork();
        }
       
        return System.currentTimeMillis() - time;
    }
   

    private static long reflectionCall(int times) {
        try {
            Class workerClass = Class.forName("ReflectionSpeedTest");
            Object workerInstance = workerClass.newInstance();
            Method method = workerClass.getMethod("doWork", null);

            long time = System.currentTimeMillis();

            for (int i = 0; i < times; i++) {
                method.invoke(workerInstance, null);
            }

            return System.currentTimeMillis() - time;
        } catch (Exception e) {
            throw new RuntimeException(e);
        }
    }
   
   
    public void doWork() {
        int i = 5 * 6;
    }
}

You can download this test from inworg.com if you want try it by yourself. Regularly the whole java poem is needed to call a single method through reflection (you need the class, the method and often also the instance). This test does not tests that case. Only method call is measured.
On my machine running openSUSE 10.3 the reflection call was more than 10 times slower. For sure, this difference may be not so significant for heavy methods, but for light ones it is.

openSUSE 10.3 and microphone

2008-01-21T23:21:00.000+02:00

Recently I've bought a microphone. While this device works fine without configuring on MS Windows, you need to apply some efforts to make it work on Linux.
On my current system (openSUSE 10.3) the sound recorder application KRecord did not recorded anything :) I've played with KMix for some time but I did not managed to record anything. Fortunately, I have more luck with Skype.
As it is correctly emphasized on some forums, the key option for making the microphone work fine on Linux is KMix > Switches > Mic Boost. It was disabled on my openSUSE by default. And if this option is disabled, your voice is almost not heard.
The bad thing is that this option is not saved. So, you have to enable it every time you boot your system.

Drag-n-drop on Mac

2007-10-05T18:03:00.000+03:00

The Macs are the strange creatures. They have quite strange behavior of the DropTargetListener. It does not matter whether drag is accepted or rejected on dragEnter() and dragOver(). If the drop occurs, the drop() is called.
It is done more conveniently on Linux. After rejecting drag the drop() wont be called.

Japanese characters

2007-10-05T17:55:00.000+03:00

I recently had a problem with Japanese. Namely, Japanese characters were not shown in a Swing application. Those characters were displayed fine in the window title only.
The solution was really easy. I've found it on Sun Swing forum. The problem was that application font did not supported Unicode characters. After changing it to e.g. "Arial Unicode MS" all characters were shown correctly.

Troubles with java.awt.Desktop browse() method

2007-06-25T13:12:00.001+03:00

New Desktop features were added in Java 6. One of those features is browse(URI uri) method, which launches the user-default browser to show a specified URI.
So, it is quite easy to show some web-page to user: check if desktop is supported, then check if BROWSE action is supported and call browse() method. Here is some code chunk:


try {
    if (Desktop.isDesktopSupported()) {
        Desktop desktop = Desktop.getDesktop();
              if (desktop.isSupported(Desktop.Action.BROWSE)) {
            desktop.browse(new URI("http://www.sun.com"));
        }
    }
} catch (Exception e) {
        e.printStackTrace();
}

Everything seams to be fine. But when you run it on Fedora 6 on KDE it ends up with


java.io.IOException: Failed to show URI:http://www.sun.com
     at sun.awt.X11.XDesktopPeer.launch(XDesktopPeer.java:75)
     at sun.awt.X11.XDesktopPeer.browse(XDesktopPeer.java:64)
     at java.awt.Desktop.browse(Desktop.java:368)
     at Main.main(Main.java:12)

So, you go to Sun's bug database and start searching what is going on. And here it is: bug 6486393 (Need a KDE support in java.awt.Desktop class). So, it turns out that Desktop API was developed to support Windows and Gnome only. Upps.
OK, you are switching to Gnome on the same Fedora 6, run the example again and... you have the same exception. It does not work.
Two question are arising to Sun's engineers:

Why KDE is simply ignored?
Is it so difficult to add support even for Gnome that works?

The Pain With Java System Tray On KDE Desktop

2007-06-18T12:10:00.000+03:00

System Tray came to us with Java 6. It is great feature we were waiting for a long time. Its functionality was very limited. Nevertheless many java developers were happy (including me).
Unfortunately, this new functionality came with problems. Namely one BIG problem. System tray does not work on KDE desktop. Bug with ID 6448876. Yes, it still exists. The really bad thing with the bag is that it looks like it WONT BE FIXED before Java 7. If you look at the user comments to the bag, you can see how many developers are frustrated with it. But it is marked as "closed", so it is ignored by Sun engineers.
I clearly realize MS Windows is OS of overwhelming majority of PC abound the globe. But there are also a lots of PC under Linux platform. And KDE is very popular desktop among Linux users. For those users Java system tray does not work. And wont work in the nearest future.
The good news is that Sun JRE is not the only one. There are also BEA JRockit and IBM IBM Early Release Program SDK 6. The latter is some kind of beta and is not available for MS Windows. On BOTH of them system tray works. And both of them have two problems:

they are rarely installed on regular Linux user PCs;
they have 32/64 bit CPU problem (which Sun Java does not have).

The 32/64 but CPU problem (my term :) ) is that there are a lots of 32 bit Linux installations on 64 bit PCs. You can't install 64 bit Java on such machines. 32 bit Java is the only choice. Both BEA and IBM 32 bit Java's don't work well on such machines.
The conclusion is: if you have a big desktop application which needs system tray functionality, you have a big headache with making it run correctly on KDE desktop.

Browser support for jar protocol

2007-06-16T17:46:00.000+03:00

I've just discovered that jar protocol (jar:!/{entry}) is rarely supported by web browsers.
For testing I've used the link found accidentally on the web: jar:http://wiki.netbeans.org/wiki/attach/AutoupdateAPI/org-netbeans-modules-autoupdate-services.zip!/index.html.
MS IExplorer failed to open it.
The same was with Opera.
In fact the only browser that manager to open it was Firefox.
I also tried Konqueror. Just for fun. As you can guess, it don't support jar protocol too.