The example application I’ll explain here is a utility to combine multiple images into one in one click directly from a Picasa folder view. However, it is very easy to modify this to perform various other batch image processing tasks directly from Picasa.

Part 1: The Picasa button

First of all, we will need a button in our Picasa to start our ImageMagick script. To create the button using Picasa Button API, we follow these steps:

1. Write a PBF file.
   A PBF is basically a simple XML file, that describes our button to Picasa. Here is the PBF file I wrote for my application:

   <?xml version="1.0" encoding="utf-8" ?>
   <buttons format="1" version="1">
     <button id="{ca234ae3-6340-40c3-a46b-51a126bb887c}" type="dynamic">
       <label>Combine</label>
       <tooltip>Combine images vertically and save as a new image</tooltip>
       <action verb="trayexec">
         <param name="exe_name" value="stitch.cmd" />
         <param name="exe_path" value="S:\Tools\" />
       </action>
     </button>
   </buttons>
   

   Although the whole code is contained within a buttons element, the file can currently only describe one button.

   Note the GUID in the button/id attribute. You’ll need a new GUID for every button you make. The PBF must be named using this GUID, {ca234ae3-6340-40c3-a46b-51a126bb887c}.pbf in my case.

   The label and tooltip elements shouldn’t need much explanation. The next important element is action; this describes what the button will actually do when pressed. There are a few actions described in the Button API; the one I use here is a simple executable launch, that starts a batch script stitch.cmd, located in S:\Tools\ (in our environment, S:\ is a shared network drive on the office server). The script will be detailed further on.

2. (optional) Draw an icon.
   You can include a custom icon for your button. The icon has to be an Adobe Photoshop PSD file called using your GUID ({ca234ae3-6340-40c3-a46b-51a126bb887c}.psd for me), in a single layer. If you use an icon, you have to include an additional element in your PBF, inside the button element:

   <icon name="{ca234ae3-6340-40c3-a46b-51a126bb887c}.psd/layer" src="pbz"/>

3. Package your button into a PBZ file.
   A PBZ file is nothing different than a ZIP archive containing your PBF (and the icon, if you want to use one), with its extension changed to .pbz. This file is basically your distribution package, so name it appropriately, since Picasa differentiates its buttons by their PBZ file names. One PBZ can also contain multiple PBF files for multiple buttons at once.

Part 2: The ImageMagick batch script

Since the ImageMagick package provides command-line utilities, the simplest way to automate them in Windows is a batch file. Here is my stitch.cmd script for merging multiple images together vertically, resizing them all to match the first one’s width:

@echo off
cd /D "C:\Program Files\ImageMagick-6.7.1-Q16"

rem --- Output filename
set OUTFOLDER=%~dp1
set OUTFILENAME="%OUTFOLDER%Collage.jpg"

rem --- Imagemagick
for /f "delims=" %%a in ('identify -format "%%[fx:w]" %1') do @set WIDTH=%%a
montage -mode concatenate -resize %WIDTH% -tile 1x %* %OUTFILENAME%

When called from the Picasa button, the script will get filenames of all currently selected images in Picasa as arguments, including full paths.

First, we change directory to where ImageMagick is installed, to avoid possible problems with PATH. Then we generate a filename for our montage, taking the folder path from the first argument (first selected image from Picasa) and adding our own filename.

After that, we apply a batch hack, using a for loop to save a command output (ImageMagick’s identify) into a variable. This gives us the width of the first image. Next, we call the ImageMagick’s montage tool, that will combine all our images in one, additionally providing it with the output filename as a last argument.

Part 3: Installation

The installation is a bit tricky. Picasa doesn’t support importing the buttons directly, it can only be done using a browser and a specially formed link to the file. You’ll need to place your PBZ at some location reachable via an http:// link; since I have a local development server running on my PC, I used that. Here is an the HTML file I used to install my script in Picasa:

<html>
 <body>
  <a href="picasa://importbutton/?url=http://localhost/olex-stitcher.pbz">Install</a>
 </body>
</html>

Just open the file in a browser of your choice and click the link. Picasa will open and prompt you to select the button you want to install (only one option here, unless your PBZ contains multiple buttons), and where to place it in the button bar. Do that, and you’re all done.

Basically, what you need to do is to override the dispatchKeyEvent() method in your Activity, and intercept key presses for the volume keys, leaving the other keys untouched. This is how I did it, fairly straightforward:

@Override
public boolean dispatchKeyEvent(KeyEvent event) {
	if (event.getAction() == KeyEvent.ACTION_DOWN) {
		switch (event.getKeyCode()) {
		case KeyEvent.KEYCODE_VOLUME_UP:
			scrollToPrevious();
			return true;
		case KeyEvent.KEYCODE_VOLUME_DOWN:
			scrollToNext();
			return true;
		}
	}
	if (event.getAction() == KeyEvent.ACTION_UP
		&& (event.getKeyCode() == KeyEvent.KEYCODE_VOLUME_UP
			|| event.getKeyCode() == KeyEvent.KEYCODE_VOLUME_DOWN)) {
		return true;
	}
	return super.dispatchKeyEvent(event);
}

Note that we only execute our code on the ACTION_DOWN event – the ACTION_UP event is only caught in order to mute the keypress sound. Then event.getKeyCode() is checked, and if the key pressed is volume up or volume down, my scrolling functions are called and the function returns. In any other case (another event or different keys) the execution is passed on to the standard key press handler.

Next are the scrolling functions to scroll the ListView. The problem with these is, you can scroll to any x/y coordinates using the scrollTo method. But what I needed is to scroll to the next visible item, since Quoter is a reading app and one wants to scroll to the next qoute in the list and read it. So this is what I did:

private void scrollToNext() {
	int currentPosition = getListView().getFirstVisiblePosition();
	if (currentPosition == getListView().getCount() - 1)
		return;
	getListView().setSelection(currentPosition + 1);
	getListView().clearFocus();
}

private void scrollToPrevious() {
	int currentPosition = getListView().getFirstVisiblePosition();
	if (currentPosition == 0)
		return;
	getListView().setSelection(currentPosition - 1);
	getListView().clearFocus();
}

First we get the list position of the first element visible on screen. If it’s already the last position in scroll direction, we do nothing. If it isn’t, we set the selection in the list to the next item; this automatically scrolls the list (same thing happens if you use the D-pad keys to navigate a list). The only problem now is, the list item becomes selected and highlighted, which we don’t want since we’re reading text from it. By clearing the focus from the list, we remove the selection. All done.

A tool to configure an OpenTTD server via web interface.

Tested with OpenTTD 1.0.4, Apache2 and PHP 5.2.4 on Windows and Ubuntu Linux. No reason it shouldn’t work on any other platform, given PHP.

Project hosted at the OpenTTDCoop DevZone:
http://dev.openttdcoop.org/projects/ottd-webconfig

OpenTTD is an open source remake and expansion of the 1995 Chris Sawyer game Transport Tycoon Deluxe. It is an “Urban planning and simulation game”.

I’ve once again started playing OpenTTD – it’s been years now. OpenTTD, as stated above, is an open source remake of a strategy game called Transport Tycoon Deluxe from the MS-DOS times. The goal is to create a transportation network for passengers and cargo using trains, trucks, ships and planes, and to make more money than your opponents. A couple of screenshots to illustrate:

Efficient rail networks in the game can become very complex, with hundreds of trains rolling on them. Such networks cannot be built without large intersections, and the traffic flow is strongly influenced by the junction design efficiency. In this post, I’ll collect a few junction and station designs I made, that could become helpful for some OpenTTD players out there.

A solution came suddenly, when I saw some CCFLs on sale in an online hardware store for €4,99. One trip to the shops and 7€ spent later I’m sitting here in the dark, typing this text and having my eyes not stressed at all

The CCFLs I bought were from Sharkoon, and came with a switch, transformator and a Y-cable to connect to the PC’s power supply, obviuosly aimed at the casemodding scene:

“That’s all very nice”, I thought, “but firstly, how do I power it (my PC is under the table, and the monitor sits above it), and secondly, where do I put the switch so I can turn it on and off?”. I decided to elongate the power cable by 5-6 feet and to put the switch (together with the transformator and the cables) on the back of my screen, where the CCFLs themselves would go. After thinking that, I went to a hardware store and got the following stuff for a total of 7€:

My plan required: a plastic box to put the switch and cables in, cable binders and connectors (because I do not have the tools for soldering), isolating tape and some cable.

I cut the Y-cable after the split from the Molex connector, and reconnected it with 6 feet of my cable using the connectors. Then I put th switch in the side of the plastic box, and put the cable and connectors inside.

After that I connected the transformator and the CCFLs, connected the Y-cable inside my PC, and turned it on. It worked!

A little detail remained – the whole thing had to be put onto the screen. I put the box with the switch on using double-sided tape near the bottom of the screen, so the switch can be easily reached and cannot be seen. For the rest (transformator and lamps) a set of stick-on pads was included in the package, so I used that.

Result: a nice working backlight for 12,99€, that anyone can build.

Android Market: Quoter
Read fun internet quotes from a variety of websites. Perfect entertaining while waiting in queues, public transport and similar.

Features include:
- Choosing which sites to read from
- Reading latest quotes
- Saving quotes for offline reading
- Sharing quotes via SMS, email, Facebook etc.

If you like my app, please buy the donate version – it’s ad-free!

Any comments and suggestions are welcome here in the Market or per eMail.

Quote sources in English:
- bash.org: latest, random, random >0, top100
- qdb.us
- quotebucket.org
- fmylife.com

Quote sources in German:
- german-bash.org: latest, random
- ibash.de
- swissbash.ch
- abash.at
- SMSvonGesternNacht
- school-bash.org

Quote sources in Russian and Ukrainian:
- bash.org.ru: latest, random, abyss, abyssbest, all with pagination support!
- ibash.org.ru
- ithappens.ru
- zadolba.li
- bash.bynets.org (bash.org.by)
- ukrbash.org
- killmepls.ru
- nefart.ru

If we try to describe anything known to us with our words, there will be no word deficiency. One description will be good, another one not so – that will tell the difference in the describer’s skill. But noone will stop looking for the next needed word in a simple story.

An “idea crisis” can only happen, if the designer sets out aiming to think of something new and unusual.

Something new and unusual cannot be thought up – it can only be discovered while working on a specific, defined goal. It’s the same difference as there is between recitation of words and a story.

A writer’s block is a dead end of a senseless way.

© art.lebedev, my translation from Russian

In every folder that is managed with SVN, it creates a hidden .svn subdirectory. In it, a list of all files and subdirectories in that folder are stored, together with technical information about their history as well as every file itself. The file .svn/entries contains a list of all files and directories in the folder where .svn is located, and the directory .svn/text-base contains the recent revisions of all files, with .svn-base added as an additional extension.

project
  |- index.php
  |- config.php
  |- dir
  |    |- .svn
  |    |    |- entries
  |    |    '- text-base
  |    |         '- file.ext.svn-base
  |    '- file.ext
  |- .svn
  |    |- entries
  |    |- text-base
  |    |    |- index.php.svn-base
  |    |    '- config.php.svn-base
  |    '- ...
  '- ...

Now, normally the .svn is hidden, and doesn’t bother anyone much. However, with default settings, the webserver will treat the files inside it just like any other file – if requested directly, those files will be served. Using the entries file, a list of accessible files can be obtained – among which there might be configuration files or others containing confidential information. And since the files in the text-base directory have their extension changed they will not be processed by, for example, PHP – the sourcecode will be shown as it is.

This vulnerability was discovered some time ago by a group of Russian programmers. They did a scan of a large amount of websites, including some fairly big names in the .com, .de and .ru zones, and the results were staggering. Owners of hundreds of websites were notified of the vulnerability, and after it was closed, some details were disclosed on a large Russian IT blog Habrahabr.

There are multiple ways of securing against the vulnerability. Among the most simple and efficient are:

• Blocking web access to the .svn directory altogether using, for example, the .htaccess file for Apache.
• Using the svn export command instead of the usual svn checkout does not produce the .svn directory in the first place.

While those solutions might seem obvious (and they really are), I was very surprised when I did a simple check among a few websites from my bookmarks list. So webmasters – if you haven’t yet known about this, beware.