The one thing you have to do that differs from the standard installation instruction when installing on OS X is that if you run mysql on the same machine as wordpress itself it wont be able to connect to it. By default wordpress sets the variable DB_HOST to localhost, for some reason that doesn’t work, you have to set it to 127.0.0.1:3306 instead, like so:

define('DB_HOST', '127.0.0.1:3306');

Otherwise it won’t find the database server.

ForEach has a variable called varStatus assigned an instance of the javax.servlet.jsp.jstl.core.LoopTagStatus class. This class defines a set of properties like first, last, index and the likes. varStatus is a scoped variable. In fact it is so scoped that it cannot be reached by a nested loop. Consider the following:

<c:forEach items="${items}" var="item" varStatus="status">
<a href="${item.value}" class="${status}">yadda</a>
</c:forEach>

fine.

Now consider this:

<c:forEach items="${items}" var="item"  varStatus="outerStatus">
<c:forEach items="${item.otherItems}" var="thing"  varStatus="innerStatus">
<a href="${thing}" class="${outerStatus}">yadda</a>
</c:forEach>
</c:forEach>

Not possible. The varStatus variable of the outer loop is so scoped that it cannot be reached by the inner loop.

Solution:

<c:forEach items="${items}" var="item"   varStatus="outerStatus">
<c:set var="statusclass" value="outerStatus"/>
<c:forEach  items="${item.otherItems}" var="thing"  varStatus="innerStatus">
<a href="${thing}" class="statusclass">yadda</a>
</c:forEach>
</c:forEach>

se tout!

However, by delving in the source code I finally found that if you switch off two features, then it does ignore the external DTD. So add this to your code if you want to ignore DTDs:

final SAXParserFactory saxParserFactory = SAXParserFactory.newInstance();

final SAXParser saxParser = saxParserFactory.newSAXParser();
final XMLReader parser = saxParser.getXMLReader();

// Ignore the DTD declaration
parser.setFeature("http://apache.org/xml/features/nonvalidating/load-external-dtd", false);
parser.setFeature("http://xml.org/sax/features/validation", false);

Go fish

Before I became a father I swore to never dump my kids in front of the TV..

It didn’t take long before I came to my senses.

Currently my son loves a swiss clay animation series called Pingu. It is actually quite funny and well made. Our swedish state-TV keeps a 30 day repository of back episodes. This is great. It’s just that each episode is only 10 minutes long and so kiddo starts screaming way too often for me to help start the next. Aah but for resourceful dads who values their own hacking sessions our dear friend mplayer comes to rescue!

To download an episode from svtplay just surf to its webpage, view source to find the URL to the movie. For some reason it is often in windows media format so try searching for “wmv”. Once you’ve figured out the URL to the movie use this mplayer command to download it:

mplayer -user-agent NSPlayer/8.0.0.4477 -dumpstream -playlist <URL> -dumpfile <filename>

For example, to download the latest episode just do:

mplayer -user-agent NSPlayer/8.0.0.4477 -dumpstream -playlist http://geoip.api.qbrick.com/services/rest/qticket/svtplay.aspx?vurl=http://secure-wm.qbrick.com/90807/kluster/20100418/PG-1117294-012A-PINGU.wmv -dumpfile /mnt/media/Videos/pingu_2010_04_19.wmv

A discussion in swedish is going on the swedish ubuntu forums

To access services running on localhost from a virtualbox instance you’d want to use the ipaddress 10.0.2.2. For example, to access a tomcat server running on localhost from the guest OS use address 10.0.2.2:8080. Whats going here is that the guest os, in my case WinXP, has been given an ipaddress in the 10.0.2 range from virtualbox and 10.0.2.2 is the default gateway and thus our gateway to the hosting os.

ipconfig

RewriteCond %{HTTP_HOST}   !^www\.example\.com [NC]
RewriteRule ^/(.*)         http://www.example.com/$1 [L,R]

To match a URL containing non-ASCII characters such as german umlauts you’d want to use its internationalized domain name (IDN) equivalent. For example the IDN equivalent of linnéuniversitetet.se would be xn--linnuniversitetet-etb.se. Thus a rule to redirect linnéuniversitetet.se to lnu.se would look like the following:

RewriteCond %{HTTP_HOST}   ^(www\.)?xn--linnuniversitetet-etb.se$ [NC]
RewriteRule ^/?(.*)         http://lnu.se/$1 [L,R=permanent,NE]

For more info on what the rewrite module can do, have a look at Apache Module mod_rewrite documentation.

Two months ago we lost one team mate and now had two peoples in one city, two others in another city and myself alone in the third city.

I have somewhere around 30minutes from door to door traveling from my house to the office. On my way to work in the morning I leave my kid to kindergarten, adding 20 or so minutes. So I thought to myself, hey I have no project pals at the office, I might as well try and work from home!

Tonight I was watching an interview with Jason Fried of 37signals. In this interview he is talking about how they at 37signals have 10 employees spread out across 5 cities in north America and how they only meet face-to-face ones every other week. They primarily use text-based communication rather than voice to force each other to be brief. Of course using their own tools such as basecamp, but also iChat and screen sharing seems to be common. He says not interrupting each other is important and this is when it hits me.

For the past three months I’ve found myself being more productive than ever. I’ve found myself being able to focus better on the task at hand. At home there is no background noise, there is no one asking me about the weekend, the weather or lost season 14 episode 2.

Being a firm believer in scrum it is kinda hard for me to accept the fact that not working in a team room actually improves productivity. And being able to focus better when not interrupted I write better quality code, thus improving quality. Yet it is very much evident.

Here’s one more video of Jason on the subject

The client software called raidar has a somewhat broken installation script for linux. It is written for bash but specifies bourne shell in the shebang. This is solved by specifying bash when executing it, like so:

bash ./setup_Linux.sh

Also if running from a machine without a graphical environment you’d want to add -c as an arg to the script to force it to run as a CLI installer, like so:

bash ./setup_Linux.sh -c

The popularity and broad adoption of facebook is impressive! So is its lack of usability. Not only do they break EVERY usability guideline out there! The UI is un-intuitive and chaotic. Simply amazing! Here are a few issues that has annoyed the crap out of me lately.

• No sense of home. This is step one in usability 101. The idea is to create a home or start -page that serves as a caring mother for all the child-pages. The user should feel comfortable in that he/she can always go back to the startpage and find common items such as menus. On facebook there’s a Home button which when clicked presents the user with content that differs slightly from whats displayed when clicking the Facebook button, both differs slightly more compared with whats displayed when clicking the Profile button. 3 different pages with similar or nearly identical content but no sense of home. What more; the same functionality is often found on all these pages, such as writing a note.
• Navigational Structure. There are no navigational signposts to indicate where you are or how you got here. A highlighted tab or a breadcrumb would be a good idea.
• Facebook is supposedly all about friends but I fail to find a list of mine. At the friends page (http://www.facebook.com/home.php?#!/?sk=ff) I found a button saying Create List. When clicking it the user is presented with a list of all his/her friends. I tried this feature and chose to include all my friends thinking I’d get a list of all of my friends. Instead I got another yet another RSS feed of news/actions/somethings, supposedly with items from all my friends.
• Consistent Menus. In order to not confuse your visitors it is good practice to present them with a consistent menu across your site so that they navigate in the same manner on all pages.  The left-hand menu on facebook fluctuates when surfing the site. This makes users feel insecure about the navigational structure of the site. An example is the Friends icon which is only available at certain pages.

One thing they have fixed recently is support for non-javascript clients. This is a big step forward.

A task perfectly fitted for our long time UNIX friend find and his cousin cron! Cron is a time-based job scheduler and find.. finds files.

What I want is to delete all uploaded screenshots on a nightly basis. Each uploaded file is stored in a directory of its own. Im using an apache2 webserver to host the files thus all files are owned by user www-data. I want this nightly cleaning mission to be run by this user. www-data is not allowed to log in herself why we will have to add the cleaning job to her crontab as a normal user. The command below will open up www-data’s crontab in your favourite editor (vi).

sudo crontab -u www-data -e

To run the cleaning job at midnight, enter the row below, but replace /path/to/www-server/and/uploads with the actual path to your web servers doc root and uploads directory.

0 0 * * * find /path/to/www-server/and/uploads/* -prune -type d -print0 -exec rm -rf {} \;

Save and exit the file.

Rejoice!