There were 70 thousand other people in it and all, except one guy who died of a heart attack 200 meters away of the finish, had a good time.

In previous years they took an average starting time and your exact finishing time, subtracted that and that is your time. So it was possible two years ago that Naomi and Dirk finished half an hour ahead of me: My official starting time was half an hour before theirs, but I walked with them. Needless to say that it didn't make me happy.

This year they had automated the time-tracking mechanism with RFID tags. So finally my real time, about two hours and thirty minutes, would be shown to the world! Justice at last! Except euhm... yes, computers are involved...

The run started at 09:00 with the professional people running and it wasn't until 10:00 that I finally crossed the starting line. And I finished at 12:40. Two hours fourty minutes, that's because Dirkie wanted to have an ice-cream and we had some trouble with him falling out of the stroller (only his pride got hurt).

According to the official results, I came in on place 53564 at 222 minutes and 21 seconds. That is three hours and fourty minutes. What!??!? So I compared it with numbers of other people I knew who ran: All between 70 and 90 minutes. So theirs is correct!

Looking through the results at people at place 55827, their time is 177 minutes 47. Their time is slower, and their rank is lower. That makes sense. What doesn't make sense is where my 60 minutes have gone! Tomorrow I'll give them a call to sort this slander out!

Update! The City2Surf people said that it happened because the tag wasn't scanned at the beginning. Yes, I got that idea. To err is human, to fully mess things up you need computers... The new time will be 162.21. I'm on my way to the blue starting group again!

If you want to use their Tunnel Broker service on FreeBSD (i.e. have an IPv6-over-IPv4 tunnel towards their IPv6 gateway), then use the net/freenet6 port. Make sure you are using version 5.1_x.

The required configuration changes in /usr/local/etc/gw6c.conf are:

userid=edwing
passwd=foobar
server=sixgw.internode.on.net
auth_method=any

Besides being a tunnel broker they also support a native IPv6 connection if you have an ethernet- or fibre link with them.

kldload: Unsupported file type

kldload: /boot/modules/test.ko: Unsupported file type

A couple of highlights and learned lessons:

• Over time, the network got replaced from simple central router plus remote FreeBSD based routers to a partly meshed network (still with a lot of FreeBSD based routers) with dynamic routing, a proper serverfarm and failover where possible.
• I've learned a lot about radio links (which can be tricky due to the density of other radio links in the city) and laser links.
• Over the years, we went from FreeBSD version 4.x to 6.3 now, with all changed in their technology and capabilities.
• I learned a lot about the accessing of data in databases, database administration and realtime replication of database.
• For in-house developed applications, it is good to have a framework which generates a consistent look-and-feel, plus gives easy access to often used datasources (databases, XML feeds).
• VoIP Telephony and legacy systems integration... certainly a nice job, but not as something which is an add-on project, but as a full-time job. It will help you to find of some before-this-unknown problems in your network.

• Content Management Systems for webhosting is important to have done correctly, but none fits all. Programmers/system administrators are not a replacement for graphical designers. Are there any graphical designers which are capable of producing maintainable HTML, or does xmlindent(1) have to be used on everybodies output?
• If you have a document storage application, and a issue tracking application, and a customer relation application, then it's a good thing to make sure you keep using the same names for the same topics: Use them consistent, don't invent different names for the same thing in every new application.

Of course, I could bitch and moan about the bad things which have shown up in the last year but let's keep it positive.

Now I'm in the handover phase... Which of course totally doesn't go as I had hoped for. We had two months time for it, in which we could do one month for handover, then two or three weeks for the last items which were missing or not clear or where the new people needed their hands to be held. But it was decided that it all should happen in the last two weeks...

Next step? Nothing set in stone yet, just talking. Want to talk to me?

There was only an issue with the ldd(1) command in the base-system:

[/] root@ed-exigent>ldd `which httpd`
ldd: /usr/local/sbin/httpd: can't read program header
ldd: /usr/local/sbin/httpd: not a dynamic executable

[/] root@ed-exigent>LD_32_TRACE_LOADED_OBJECTS=1 `which httpd`
libm.so.4 => /lib32//libm.so.4 (0x280c8000)
libaprutil-1.so.2 => /usr/local/lib/libaprutil-1.so.2 (0x280de000)
libexpat.so.6 => /usr/local/lib/libexpat.so.6 (0x280f2000)
libiconv.so.3 => /usr/local/lib/libiconv.so.3 (0x28110000)
libapr-1.so.2 => /usr/local/lib/libapr-1.so.2 (0x281fd000)
libcrypt.so.3 => /lib32//libcrypt.so.3 (0x2821d000)
libpthread.so.2 => not found (0x0)
libc.so.6 => /lib32//libc.so.6 (0x28235000)
libpthread.so.2 => /usr/lib32/libpthread.so.2 (0x2830d000)

Patches are available at bin/124906, have been commited to HEAD and will be MFCed in a week.

For example: Did you know that the URL http://bugs.freebsd.org/12345 would bring you automatically to the PR with number 12345? And did you know that if you want to checkout the directory /ports/x11-wm/fvwm95, all you have to do it cvs co fvwm95 (For an easy Mozilla Firefox and Seamonkey (and probably other browsers) sidebar, visit this page)

So, what is the story about that cvs co fvwm95. The magic for that is stored in the file CVSROOT-ports/modules, it contains a list of all ports and their directories. The file is updated every time a port gets commited, removed or changed from location. At least that is the theory. Adding is done manually via the addport script, but removing and changing locations has been always a manual task.

Recently the caretaker of this file had enough of people whinching when he said they needed to update that file when it changed. So he said that support for it would cease. But but but... That would break at least one of my ways to do things easy!

I can do only one thing now, take control over it myself! So I proposed it to portmgr@, who approved it. The second thing to make is a script which mangles /usr/ports/INDEX to generate a modules file (in the shape I like it).

All that is done now, and the CVSROOT-ports/modules file is now updated once per day. See the CVS webinterface for the last update!

Therefor I present to you: DHCPDUMP version 1.8, now using libpcap as input instead of the output of tcpdump.

You can find it at http://www.mavetju.org/download/dhcpdump-1.8.tar.gz.

The story is about the availability of driver licenses in Australia: The joke goes that you get them with a box of Weetbix. And todays experience only shows that it is true. And why I feel so scared when I'm on the road.

So I got my dutch driver license in 1996, when I was 25. Wait, that was seven years after I was able to get it. I know. I told you I didn't like cars. I got my first car in 2000, five years after I got my driver license. I told you I didn't like cars. I moved to Australia in 2001, so I got rid of my car after six months.

In Australia you need an international driver license which is valid for six months or get an Australian driver license. So after six months I wasn't allowed to drive a car anymore. In 2006 my dutch driver license expired.

So, where does this story go to? Today, in 2008, seven years after I last driven a car, two years after my driver license expired, I got my Australian driver license. How? Just go to the RTA, show your dutch license, show some ID and fill in a form. Expired license? Not a problem. Not haven driven a car in seven years? Not a problem.

And you wonder why I am so scared on the roads here...

If you want to prefer it in your ports configuration, add this to /etc/make.conf:

MASTER_SITE_SOURCEFORGE= \
        http://internode.dl.sourceforge.net/sourceforge/%SUBDIR%/
MASTER_SORT?=   internode.dl.sourceforge.net mirror.internode.on.net .au

So what has Postgrey to do with this? Postgrey is a greylisting server. It is (danger! technical content ahead!) forcing email deliveries from addresses and hosts which are not yet known to be retried later. Why? Earlier this century, emails sent by viruses and spam-hosts weren't smart enough to understand this and the email with the malicious payload was not accepted by your mailhost.

Yes, but what has greylisting to do with it? Greylisting delays every email from / email to / sending host combination it hasn't seen before. So if BATV changes the email from address every day, the first email from that user will be delayed every day. Every day! So Postgrey needs to be taught what the real email address is. Luckely BATV keeps this information in the from address: prvs=tag-value=joe@example.com. Small patch, and it works.

And now the tricky stuff: Not every read the documentation properly, and the two following formats have been seen:

prvs=tag-value=joe@example.com
prvs=joe=tag-value@example.com

Anyway, the patch is available and submitted to the Postgrey author.

Note: Please note that I've made a little change to the patch to pick the second field (as the standard suggests) instead of the wrong standard. Not that it ever should come to there, but it's a "just in case" thing.