This recession has been bad news all around, with many businesses shuddering their stores and swathes of jobs lost. I’ve noticed several webcomics I’ve been following for a long time have recently shut down too. Bruno the Bandit announced its effective cessation mid-story. Angband: Tales from the Pit had an ending storyline for its wrap up. User Friendly has gone into repeats. Homestar Runner has quietly stopped updating. There’s a constant rise and fall in this “industry”, low entrance costs means competition is heavy, but I’m sad to see long-time players bow out in succession.

I rode the Arroyo path again today. This time the stream was lower, and a lot clearer, but not yet back to its normal level of transparency. The “stream of sludge” was pretty worrying for a while.

Today the Arroyo Seco stream was dirtier than I’ve ever seen it. The water was fairly high, a little more than yesterday, but while yesterday the water was clear, today it was nearly black with sludge.
What happened? The radio news had the answer — there’d been some mudslides upstream, which apparently dumped tons of silt into the stream.

So, a little before 8:50 this morning, all my connections hung. Fire up mtr, and it doesn’t reach the first hop. I pick up the phone, and there’s no dial tone! I futz around for a while, but by 9:00 there’s still no phone service. I go outside and down the ¼ block to the AT&T switchbox, and there’s an AT&T truck there and a tech messing around with it. I ask him if he’s fixing the outage, and he says no, he’s measuring the noise on the wires. I say I lost my service 10 min. ago, and after getting my phone number he confirms that he pulled my wires along with a bunch of others then. With me standing around, he measures my line next and reconnects it.

WTF? How is it that this is a company which is synonymous with reliability in its reputation? A deliberate, unannounced outage, not during off-peak, and for much longer than need be just because it’s easier to pull all the wires, measure, then reattach than to do them one by one? And I was very lucky this time — last time I lost my phone it took them two days to send a tech out to confirm that the last tech to mess with the switchbox had screwed up my wiring and fix it.

A friend told me he’d gone hiking around Angeles Crest.
“I thought it was still closed.” “Nope!”

Sure ’nuff, I check the CalTrans and CHP websites, and they show Hwy. 2 as having no closures. This morning I get my camera and hiking gear to tour the devistation.

Sure ’nuff, when I get to the National Forest entrance, the road is closed.
I’m going to have to ask him about it…

I was just upgrading my packages, like I do every so often. This time ’round the package update list was:
bash bluez-utils dkms-ndiswrapper e2fsprogs grub grub-doc libgcc1 losetup
ndiswrapper pclinuxos-release update-notifier

Went ahead, it installed cleanly, but afterwards I was getting floating point exceptions from some programs. ldd shows libgcc_s as a likely culprit, what with the just-done upgrade including
libgcc1-4.4.1-1pclos2010.i586 ############################## [100%]

Play around with apt-get, to see if I can get an older version from apt — uh oh, apt-get says floating point exception when invoked. Grab a Mandriva RPM from the net, give it to rpm — rpm dies with floating point exception. No problem, I’ll install the files by hand, just extract them with rpm2cpio — which hits a floating point exception.

Finally, I wise up, scp libgcc1_s.so.1 from a Debian system.
cd /lib
sudo mkdir brokengcc1
sudo mv libgcc_s* brokengcc1
sudo cp /tmp/libgcc_s.so.1 .
sudo chmod 0755 libgcc_s.so.1

And everything is working again, at least enough to tide me over until PCLinuxOS fixes the libgcc1 package. If anyone else is stuck by this, the libgcc I used is here.

Quick quiz: How do you find the arguments to your Scheme program?
  a) argv
  b) *argv*
  c) command-line
  d) program-arguments

Answer? It depends on whether you’re using Guile, SCM, umb-scheme, or mzscheme! That’s right, it’s impossible to write a Scheme program that takes arguments and is portable across implementations. Winnar.

Update: In 2013, option (c) was standardized.

So, I’ve done a bunch of upgrades to my systems lately. After Ubuntu broke my X again somehow going from Intrepid to Jaunty (this time not recognizing my input devices), I switched over to PCLinuxOS (2009.2). I’m still getting used to it. The apt-rpm stuff seems to work well, and stuff I had to fight to get going in Ubuntu, like flashplayer and xmms, just work. Java actually runs, and there’s a desktop link to the magic for getting OpenOffice working. On the other hand, the package repository feels a lot smaller, not having programs I’ve long used like calc, ccrypt, dict, and procmeter.

WordPress had a series of security vulnerabilities, so I’ve updated to the latest version. As a plus, it finally knows about the system’s timezone database, so is no longer the only program you have to adjust an hour whenever daylight savings starts or stops. The WordPress upgrade process is getting easier, manual editing of the mysql db no longer needed. Hopefully one of these days they’ll simplify it down to “apt-get upgrade”-level, and it can keep itself up to date the same way everything else does.

I also upgraded the Easy Peasy (formerly known as Ubuntu Eee) install on my EeePC. I can’t speak about the install problems I noted in my last post on Easy Peasy, but most of the runtime problems seem to have been fixed — except the inability to turn the WiFi on! Another user put together a fix in February, but unincorporated by the maintainers. I find this bug just as staggering to find today as I did last December. How can this still not work out of the box after all this time?

I just upgraded my Ubuntu install from Hardy to Intrepid. (Jaunty was recently released, so I expected Intrepid to be very smooth by now.) Instead, I got the worst upgrade experience I’ve ever had with Ubuntu, by far.

First, X didn’t come up, leaving me staring at the console login after reboot. Checking Xorg.0.log, it said something about RgbPath being an illegal directive. Looking at the changes the updater had made to the xorg.conf file, it hadn’t done anything about that line. I tried replacing the updater-edited version with the one from Hardy, still choked on RgbPath. I removed the RgbPath directive, and now it aborted on ‘Module "type1"’. At this point I remember an LWN article about how recent X servers tried to autodetect everything and do the right thing if there were no xorg.conf file, so I moved it out of the way, started X, and huzzah, it worked.

Next, trying to do anything with the package manager errored out and I had a bunch of unconfigured packages. This turned out to be due to the preinstall script for openjdk-6-jre-headless consistently crashing (incidentally producing a multi-page error message containing an unsymbolized stack trace and a memory mapping dump, guaranteed to scare away non-developers), and a bunch of packages depending on openjdk-6-jre-headless, particularly OpenOffice. Uninstalling openjdk-6-jre-headless (and scrapping OpenOffice in the process) got my system back to a state where the package manager was happy.

I consider these two problems each a major lose. Intrepid’s Xorg doesn’t work with Hardy’s xorg.conf, and the update manager doesn’t fix it up. To fix it you have to either edit xorg.conf or know a not-well-publicized feature of Xorg. To even start diagnosing you have to know about Xorg.0.log and how to start the X server. Ubuntu is supposed to be targeting people who aren’t comfortable with a command line, and here it is dumping them at one after the upgrade without a map! And due to the second problem, I now don’t have OpenOffice on my system (though I do prefer AbiWord and Gnumeric, so that’s not as great a loss for me as for some).

If you use a Linux laptop, you probably have an ambivalent relationship with nm-applet, “Network Manager”, that thing which connects to your access points. Or often doesn’t, because — well, no one knows why, because nm-applet is fussy and tight-lipped.
Fortunately, you can do everything it does in the shell, if you know how. Here’s how:

1. iwlist scan to scan the environment for WAPs
2. iwconfig ath0 essid XYZ where XYZ is the ESSID from the scan that you want to connect to
3. ifconfig ath0 up (turns the interface on)
4. dhclient (registers with the WAP and gets an address)

Voilà, you’re connected (or at least have a meaningful error message), while nm-applet is still running around flailing its arms.
(I used ath0 as the interface in my example, but your interface might have a different name, like wifi0 or wlan0. Check by running ifconfig. Interface-affecting commands require root.)