Well, I finally managed to fix the problem. For whatever reason, the keyboard layout database got corrupted, and there is no straightforward way to fix it (e.g. by reinstalling a package). So instead I dumped the following directory: /etc/X11/xkb and replaced it with the same directory from a clean Ubuntu installation. That did the trick.
I was lucky that I had a fresh install lying around on another computer, but what a pain in the neck.
So I checked the web and found out that yes, there is a problem, and no, there is no simple fix.
The problem is that Windows XP will add the host information to the server no matter what.Â Thus, the username is sent as host/username.
There are some hints that this could be fixed with the Apache mod_encoding module, but life is just too short!
Fortunately, I found a solution that seems to work!Â Basic authentication is pathetic – in fact, the username and password are encoded unencrypted in the URL in the form http://username:[email protected]/path.Â So I simply included the username and password this way and hey, it worked.Â A side effect is that this makes it super obvious how insecure basic authentication is (and that using https doesn’t make it any more secure).
.eu Domains are now up for grabs!Â So I thought I’d pick up jastram.eu – but it was already gone.Â Dammit!Â Oh well, saved some money.
I know that I’ll need such a study again one day.Â Here’s an article on selecting an open-source CMS and here the Slashdot discussion that has a number of good links, too.
Gestern war das erste Treffen der "neuen" Düsseldorf Java Users Group – und es war unerwartet erfolgreich!Â Statt der angemeldeten 10-15 Besuchern kamen 40 Leute!Â Die Luft war stickig, wir mußten Pizza nachbestellen, aber eins ist klar: Die Gruppe wird eine Weile überleben.Â Danke an alle Besucher!
Das nächste Treffen ist schon in Planung und wird in etwa zwei Monaten von Marc Logemann zum Thema Java 5 gehalten werden.
At the GI-Fachgruppentreffen Requirements Engineering in Hannover (November 2005) I was intrigued by an RE-Pattern-Repository which is being managed by Desy.Â Something to watch.
I recently read Bruce Tate’s Better, Faster, Lighter Java. In the light of this I was surprised to see an Interview with Jim Rivera (director of technology at BEA) on this topic, as I would put BEA clearly in the heavyweight camp.
I used and got frustrated with in-memory databases. I mainly use them for testing, although I can image using them for managing huge amounds of simple data as well (e.g. parsing a huge plain text file into a simple table). Still, usually problems appear whis the SQL dialog. Unless the in-memory test database is the same as the production database, there will be SQL dialect issues. Martin Fowler wrote about this problem before.