Rain
It's amazing.
One day temperatures are so high I can't concentrate. Now, when there is finally rain after a heat wave which took the better part of a week, I find myself closing the window, because I'm getting too wet otherwise, and might catch a cold. And I'm still sweating all over from the extremely high temperatures which I had to endure during the past few days, along with other people living in this country. First thing I'm going to do when I arrive home: take a shower. Which is crazy, considering how the rain could really have helped me out, there. Oh well.
On a side note, if you're wondering whether good weather
is
the reason why I wasn't very active online since last saturday, then the
answer is no.
Due to there being sun all week, we've had an opportunity to do some
work on our veranda. And some other similar things. So that was, in
part, also a reason why I wasn't available. On top of that, I don't
consider much sunlight
to be the same thing as good
weather
. This has two reasons: First, I have Vitiligo (as I
mentioned before on
this blog), and allowing too much direct sunlight on my skin will only
augment those white spots. Second, 33°C is horribly high in a country
where 25°C is the average, and I prefer not to suffer from high
temperatures.
Which cannot be said about some other people, apparently. Oh well.
Belpic 2.5.9
Yesterday evening, I noticed that the 2.5 release of support software for the Belgian electronic ID card had come out earlier this week (on 2006-07-04, to be exact). I didn't know, nor notice. Sigh.
I knew this release was forthcoming; I had been working with some Zetes developers on the code a few days last month, and they had implemented a number of my suggestions. Always nice to have a working relationship with your upstream. They hadn't implemented everything I had pointed them to (mainly due to lack of time because of a deadline they were running up against), but my suggestions had at least avoided some serious bugs that would otherwise have been in the release. Great improvement over the 2.4 release, which couldn't even be built due to a serious SNAFU in the code.
But, now, this isn't happening anymore. The only thing is that Zetes isn't the one doing official releases of the Belpic code; that honour is left to FEDICT. And while Zetes happily talks to me, I can't seem to get anyone inside FEDICT to talk to me. I've sent them some emails in the past, asking them some questions, but I've never had any of them reply to me.
Like I said: Sigh.
Oh well. So yesterday, late at night, I found out that the new release was out, and I promptly updated my packages. Err, well, at least, I almost did so. I actually postponed it to today, since, well, it was late already. And it is again, now, with my packages almost being finished; I had to redo some bits since, somehow, I seem to have lost my work of last month. Grmbl.
Anyway. The packages seem ready now. When I get home, I'll test them just one more time, and then they'll be ready for upload. Whoo!
Belpic: Success!
Finally. For the first time ever, I've been able to log on to tax on web using my electronic ID card.
Too bad I've already filed my tax report on paper. Oh well.
If you live in Belgium, use Debian unstable, have an eID card and a reader, and haven't filed your tax report yet, this might be nice to do. I'll update the backports on backports.org any time soon, but they need some patches (gcc3 and gcc4 have some different ideas on what is well-formed C++ code, grmbl), so that's not for today; so if you need sarge backports, don't count on it just yet.
Exim4-config-wouter
About a month ago or so, I joined in a discussion on the debian-boot and pkg-exim4-devel mailinglists about the current Exim (v4) configuration schemes.
To keep you from having to google for that discussion, I'll quickly summarize my opinion on the subject of the exim4 configuration scheme in Debian:
I think it sucks.
I've been a rather fanatic (ish) exim user since about two months after I became a Debian Developer. Initially, I removed it to replace it with sendmail. But that was a bad idea, and I quickly found out why. However, I never liked the exim 3 configuration script, since I never could get it to do what I wanted; and, according to the exim 4 maintainer (and to anyone who compares both schemes), the current exim 4 configuration system is just a rehash of the exim 3 configuration system; not much more.
But it's so backwards. The initial question which exim4-config asks you may have been relevant in the time the initial exim 3 config script was written, but it is so totally outdated in a modern Internet. As a result, my current modus operandi when I install a new box is to get out of the exim4 config system as soon as I can, to gunzip /usr/share/doc/exim4-config/examples/example.conf.gz to /etc/exim4/exim4.conf, and to open the file in a text editor. At least I know that works, the way I want it. No hard feelings, Marc (but you knew that already).
Personally, I had different ideas about what an efficient question would be, and I posted those to the mailinglist (you will have to google for those if you want them, though :-). I also promised to come up with some code to implement my ideas. This has proven to be... interesting.
Currently, I have a config script that does a few questions and then switches to "Not Yet Implemented" mode. As in: it displays a debconf template of "Type: error" that basically says "This part still has to be implemented". For the postinst script, I wanted to base myself on the existing postinst script. That does not seem to be a very good idea, however; the existing postinst script contains a lot of upgrade cruft.
And then I still have to write my update-exim4.conf. Looks like I took on a job which turns out to be larger than expected...
Oh well, I guess I'll cope. It'll just take a bit longer than I wanted it to.
Ahh.
I had a great weekend this week. It didn't start off too great, though; I had been walking around with a rash on my hands for a few days already, and had to call in the doctor. Luckily he told me it's just a allergic reaction to something; and though we weren't able to pinpoint exactly what the problem was, apparently that doesn't matter much. I'm on some medication now which should, at least, make it less of a problem.
After that, I left for Ninove, where Freek, a long-time friend of mine, lives. This is not the type of friend that I see every weekend, but it is someone I know from high school and that I kept in contact with over the years. It had been three years since I'd last seen him, even though we had called eachother on our respective birthdays during those years. Which boils down to "we haven't seen much of eachother in three years". Still, I consider him one of my best friends, because of the very special relationship I have with him; we have no secrets for eachother, and I would trust him with my life. Which is strange, considering how little we see eachother. Oh well.
Due to a misunderstanding, however, I arrived in Ninove a few hours early. However, I still had a few hours worth of battery energy in my laptop, so I did some work on the exim4-config-wouter package I've been working on for a few weeks now. Not quite there yet, as you can see in the other blog post.
Spent some great time with Freek in Ninove—mostly catching up with eachother, but some other interesting things, too—and then left for Leuven, where I'd arranged to meet p2(@debian.org), who was going to help me understand some of the more obscure parts of the lowlevel code on a GNU/Linux system. I needed that help, because in my work to get Debian/m68k working on the ColdFire V4e processor family, I had gotten stuck due to the fact that I didn't understand some of the bits there properly. P2 and I had some beer together, and lunch too, since we met in a bar that had food on the menu as well, and, well, because I, at least, was hungry.
Next, we went to the mind (ex-)offices, and played a bit with the support packages for the Belgian electronic ID card that I maintain for Debian, and reached some incremental successes. Whee. New packages are in incoming.
Right now, I'm on my third train trip of the weekend, which is "towards home". I hope to be able to go to bed soonishly, since I have to get up early tomorrow -- and I didn't get to catch much sleep last night...
... It turned out I did. So quickly, even, that I was unable to post this particular blog entry before hitting the deck. Oh well, being one day late on that won't hurt anyone
SQL Pony
According to Simon Richter, you can implement a pony on top of an SQL database with triggers with a special tool.
Cool. I want one.