I wanted to separate the log4j configuration for my webapps from JBoss’ own configuration when using log4j. When webapplications all share the same server.log instead of having a log file of their own, it quickly gets crowded. There are two ways of avoiding this.

The first option is to add appenders to jboss-log4j.xml in jboss-as/server/default and to create a chain of filters in FILE. This option is described here. I almost got this approach to work, but then JBoss support told me about a small overhead. The overhead isn’t significant at first, but it will get bigger as more webapplication are added to jboss-log4j.xml. So I decided to not use this approach.

The second option is what worked. The option is slightly more complicated, but with no overhead. It is described in detail here.

But all was not well when I tried the second option. Following the to-do I got a java.io.FileNotFoundException when the application was deployed. Apparently the xml parser in AppRepositorySelector couldn’t find the log4j.dtd. No wonder. It looked in a path relative to JBoss’ bin directory and not in the classpath resources for the dtd.

Terra Java made his own entity resolver to solve the path problem, but actually there’s already an entity resolver in log4j (quite handy that). So after changing this in AppRepositorySelector:

Document doc = DocumentBuilderFactory.newInstance().newDocumentBuilder().parse(log4JConfig);

into this:

DocumentBuilder builder = DocumentBuilderFactory.newInstance().newDocumentBuilder();
builder.setEntityResolver(new Log4jEntityResolver());
Document doc = builder.parse(log4JConfig);

I got it fixed. And now logging works as specified in the application’s own, local log4j.xml configuration.

Søndag aften slog Amanda, som er min 12-årige datter, en tand ud. Min kone og jeg hørte faldet og gråden, og vi fandt Amanda siddende i armene på Cathrine, som er hendes søster. Amanda  blødte meget, og hun var meget chokeret. Hendes tand, den ene fortand, lå på gulvet ved siden af. Næsten ligeså chokeret ringede min kone til lægevagten på Herning Sygehus, og jeg fandt en kold klud for at standse blødningen. Efter en hurtig tur til Herning blev Amanda tilset af en læge. Det skete uden ventetid. Lægen kunne ikke gøre andet end at sætte tanden op, så godt han kunne – og så sendte han os videre til skadestuen. Da var klokken cirka 21.15

Nu forventede jeg, at Amanda blev sendt til en røntgen-undersøgelse og kigget på af en tandlæge. Sådan gik det imidlertid ikke. Skadestuens personale tilbød Amanda en Panodil – og så ventede vi. Efter en tid tilbød jeg min hjælp – vi kunne da køre til en anden skadestue, hvor der måske var en tandlæge? – men personalet informerede os om, at de skam havde en aftale med en tandlæge, og at vi selv skulle betale, hvis vi fandt en anden tandlæge. Således beroliget – nogen var i gang med at finde tandlægen – ventede vi igen. I mere end 4 timer. Da klokken var cirka 02.00 blev Amanda så tilset af endnu en læge. Som godt kunne se problemet, men som heller ikke rigtigt kunne gøre noget ved det. Men han kunne da godt ringe til skadestuen i Århus.

Skadestuen i Århus reagerede helt anderledes. De ville se os så hurtigt som muligt. Akut. Og efter en hurtig køretur til Århus Universitetshospital kom Amanda endelig i kyndige hænder på afdelingen for tand-, mund- og kæbekirurgi. Her kunne en kæbekirurg sætte tanden helt på plads og sikre den. Behandlingen startede ifølge journalen klokken 03.30 og varede en time.

Nu ved jeg godt, at vi har en økonomisk krise, og at vi skal spare på de offentlige budgetter. Men jeg ved også, at et samfund kan måles på, hvordan det behandler sine svageste. Og jeg ved, at ingen økonomiske kalkuler kan forsvare, at en lille pige venter tappert i mere end 6 timer på at få hjælp for en tand, der er slået ud. Det er uanstændigt, det er skammeligt – og det er ihvertfald ikke akut.

In JBoss bin/run.sh there’s a few references to JAVA_OPS. Find the one that looks like this:

# Setup JBoss specific properties
JAVA_OPTS="-Dprogram.name=$PROGNAME $JAVA_OPTS"

- and add the line below:

JAVA_OPTS="$JAVA_OPTS -Xdebug -Xnoagent -Djava.compiler=NONE -Xrunjdwp:transport=dt_socket,address=8787,server=y,suspend=n"

This addition to JAVA_OPTS lets JBoss start in debug mode with no suspend during startup.

To enable Eclipse to do remote debug after JBoss startup, click run > debug configurations and select Remote Java Application. Fill in host and port – in the example the host is localhost and the port is 8787.

Now JBoss is ready to deploy your web-application with breakpoints.

You could say, as many do, that shipping jobs overseas is no big deal because the high-value work — and much of the profits — remain in the U.S. That may well be so. But what kind of a society are we going to have if it consists of highly paid people doing high-value-added work — and masses of unemployed?

It seems like an understatement to say that the answers to this question – and what we all chose to do about it – is what the next decade will be all about…

No interpreted code may be downloaded or used in an Application except for code that is interpreted and run by Apple’s Documented APIs and built-in interpreter(s).

It’s now saying:

Unless otherwise approved by Apple in writing, no interpreted code may be downloaded or used in an Application except for code that is interpreted and run by Apple’s Documented APIs and built-in interpreter(s). Notwithstanding the foregoing, with Apple’s prior written consent, an Application may use embedded interpreted code in a limited way if such use is solely for providing minor features or functionality that are consistent with the intended and advertised purpose of the Application.

Knowing Apple and the “curated” app experience, it would be an understatement to say that this is a crack in the agreement. It’s a small crack, but given the pressure fra Google’s Android platform I think it’s indicative of a movement towards being able to use interpreted code in your apps. Like Lua.

Despite being banned per the old section 3.3.2 Lua is extensively used by a lot of games currently on sale in the appstore, and it’s an excellent way to do the more complex parts of a game – like AI, game rules or pathfinding. So I’m hoping the change means that developers are now getting more tools for the task. With Apple’s approval.

To parafrase Mark Twain I guess the reports of Lua’s death on the iPhone are greatly exaggerated.

- Which programming language should I learn first?

- Depends.

- To program in an expressive and powerful language: Python
- To get a website up quickly: PHP
- To mingle with programmers who call themselves “rockstars”: Ruby.
- To really learn to program: C.
- To achieve enlightenment: Scheme.
- To feel depressed: SQL
- To drop a chromosome: Microsoft Visual Basic
- To get a guaranteed, mediocre, but well paying job writing financial applications – in a cubicle under fluorescent lights: Java.
- To do the same thing with certifications and letters after your name: C#
- To achieve a magical sense of childlike wonder that you have a hard time differentiating from megalomania: Objective C

- I could go on… but I’m not feeling hateful enough today.

Compiling irssi is not an easy task though. And since I used many hours of trying to make irssi compile, I’d better jot the solution down. Just in case my hard drive decides to do a critical meltdown (that has been known to happen).

Luckily I didn’t have to start from scratch. This blog post helped a lot. So thanks Magnanimous Antics if you’re reading this.

First get the source. You’ll need:

Irssi (0.8.15)
Glib (2.22.5)
pkg-config (0.25)
Gettext (0.18.1.1)
Libiconv (1.13.1)

On a side note: Like MacPorts use /opt/local/ I too use a custom path for any application I compile. It’s /Users/morten/builds. By installing in my home directory, I avoid overwriting system files. And like MacPorts I just add the path to my PATH before $PATH in my .profile. If any conflicts with other installations or versions arise I can easily remove the installation again. People not wanting to do this can simply omit the –prefix part.

To compile pkg-config, gettext and libiconv (in that order) I did:

./configure prefix=/Users/morten/builds
make
make install

After libiconv I compiled gettext again – since libiconv and gettext are mutually dependent on each other.

Next came Glib. When I compiled glib by doing the above command this error popped up:

You must have either have gettext support in your C library, or use the GNU gettext Library

Pretty sure that I had just compiled gettext successfully, I Googled the error and found a solution that worked. It seems that Glib needs the linking a bit more explicit:

./configure --prefix=/Users/morten/builds LDFLAGS="-L/Users/morten/builds/lib" CPPFLAGS="-I/Users/morten/builds/include"

To ensure a 64bit version of Glib I opened glib/gconvert.c in Emacs and replaced the last three lines of this:

#if defined(USE_LIBICONV_GNU) && !defined (_LIBICONV_H)
#error GNU libiconv in use but included iconv.h not from libiconv
#endif
#if !defined(USE_LIBICONV_GNU) && defined (_LIBICONV_H)
#error GNU libiconv not in use but included iconv.h is from libiconv
#endif

with this:

#if !(defined(__APPLE__) && defined(__LP64__)) && !defined(USE_LIBICONV_GNU) && defined (_LIBICONV_H)
#error GNU libiconv not in use but included iconv.h is from libiconv
#endif

and then did a:

make
make install

Done. Finally the turn came to Irssi. Again I got an error doing the ./compile. Some Googling revealed a search and replace solution, but the commands below worked as well – and to the same effect: a 64-bit version of Irssi:

./configure --prefix=/Users/morten/builds --build=x86_64-apple-darwin10.2.0 ARCHFLAGS="-arch x86_64"
make
make install

All done.

Epilogue: I wanted to compile and run irssi-xmpp as well. After compiling the required loudmouth 1.4.3 like this:

./configure --prefix=/Users/morten/builds CFLAGS="-I/Users/morten/builds/include" --with-ssl=openssl
make
make install

my efforts stranded on lots of linking errors.

Even after changing config.mk considerably I simply cannot coerce irssi-xmpp to compile. If anyone reading this has a working solution I would definitely like to know.

Uh. Can’t wait for that one to happen. Or just an another awesome beach demo like the one for O3D.

When I read the post, I couldn’t help but think of Richard Stallmans words: “Free software is a matter of liberty, not price. Free software is a matter of the users’ freedom to run, copy, distribute, study, change and improve the software.” Of course the Curated Computing concept runs completely contrary to the Open Source idea, but what is the main issue here?

I get the fact that not everyone is a hacker (in the  MIT sense of the word) or wants to be one. Not everyone wants to tinker endlessly with their computer. Some people are just computer consumers. They only want their favorite application to work – and they see a computer as a tool to get some work done. That’s all right. It’s their choice.

But this also points at the dilemma: What if their application doesn’t work? What if something fails? Do they just sit and wait for a fix? Yes, they do. And that is also all right. It’s their choice.

But do they have any other choice in a world of curated computing? No. And this is where I have an issue. I don’t mind technology with a friendly face. I don’t mind advanced user-interfaces that helps people getting the task done. In fact that is a crucial part of modern technology. Technology shouldn’t be difficult to use.

But I do mind if the curating takes the choice away. If you completely take away the very possibility of tinkering then you raise the next generation as passive consumers and not computer tinkerers. And we need tinkerers. Bill Gates was a tinkerer, Steve Wozniak was a tinkerer, Microsoft and Apple were built by serious hackers and tinkerers. And in a world where we desperately need innovators and entrepreneurs, I seriously don’t think that we need any more computer curating. Do you?

And the really frightening scenario is this: Apple and their AppStore has had a lot of success. This didn’t really matter when we were only talking iPods or iPhones – but the iPad is different. The iPad is a real computing device that’ll compete with – and probably replace some laptops or netbooks. I’m afraid that if the iPad succeeds, the curated business model will become too profitable for other companies to resist…