We're importing a non-native species (or three) to try to control the emerald ash borer.

I don't like wasps.

"Expecting instance of reply_fixer as self, got one of reply_fixer"

That doesn't seem like a reason to throw an exception to me.

I'm pretty close to just giving up on python for this and using Objective-C instead (even though part of the reason why I'm doing this is to play with python some more).

However, I while the day-to-day pain went away, any activity would cause the problem to flare up again. After an appointment with my new doctor (since my previous one retired), I'm now going to a physical therapist to try to help my achilles tendonitis (which wasn't healing by just resting).

So far, the results are really good. On the first day, I had cortisone delivered via therapeutic ultrasound and electrical stimulation on both tendons.

The second visit started with some warm-up and then some exercises. Since the previous visit had made me fairly sore, we skipped the electrical stimulation and just did the ultrasound treatment.

On the third visit, we did some exercises, noted that the swelling in my tendons had been dramatically reduced, and tried iontophoresis. Which was what my new doctor had originally recommended (he also advised me to get some night splints, since they are fairly cheap and help a good portion of the time. I ordered a pair of Strassburg Socks after being unable to find something similar locally).

Today was visit 4. The night previously, I had a fencing lesson which went fairly well (and I was able to do a moderate level of activity which didn't result in lots of pain the next morning). Again, we noted the the swelling had gone down (even more) in my tendons and the physical therapist noted that they felt looser. Since the iontophoresis had resulted in a little pain/discomfort, we decided to just do the ultrasound treatment after the exercises. We added a couple of new exercises to the routine this time.

That brings things up to date. My tendons currently feel much better than they have in the past two years, which is pretty exciting.

holy f'in WOOD

I at least some of the uninformed people who voted for that craptacular state constitutional amendment realize what a stupid mistake that that was.

- Filled the bird feeder for the first time since last fall
- Prepare snowblower for off season (siphon fuel back to can, run snowblower dry)
- Change lawnmower oil
- Sharpen lawnmower blade
- Mow the lawn
- Clean up piles of leaves near driveway, air conditioner, and back porch steps (and put them in our new compost bin)
- Take hoses out of garage and hook them back up
- Change tires on the IS300 (store snow tires, put on summer tires)

Things yet to be completed:

- Pick up my roto-tiller from my sister
- Prepare garden area behind garage (till, mix in new soil, fertilize)
- Purchase garden fence and install it
- Plant seeds
- Water
- Purchase new mulch and spread
- Use the tiller in 'edger' mode to clean up the edges of the lawn near the driveway/sidewalk
- Trim bushes
- Trim trees (and call a tree trimmer to allow more sunlight by the garden and trim the maple away from the house)

Of course, the observant reader will notice that the "... work was sponsored by four groups ... the New York-based Institute for American Values, the Institute for Marriage and Public Policy, Families Northwest of Redmond, Washington, and the Georgia Family Council, an ally of the conservative ministry Focus on the Family."

As you might expect, when advocacy groups (of any political persuasion) support any kind of study, the outcome is strangely correlated with their agenda.

There is a good discussion of the particular trick used to achieve the result with this study in this post (at the Freakonomics blog).

Summary:

"But the first law of advocacy science coincides with a well-known economic principle: any cost-benefit analysis that only looks at one side of the ledger will always come to a reliable conclusion."

Some searching online led me to suspect that the Envelope Index sqlite database was in in need of a vacuum.

Procedure:

1. Quit Mail.app
2. Run $ sqlite3 ~/Library/Mail/Envelope\ Index vacuum
3. Wait.
4. Launch Mail.app

So far, it seems to have helped immensely. (In a year or so, I'll probably need to search for this post to do it again - it might make sense to turn on auto_vacuum, but it looks like it needs to be turned on before the tables are set up).

I recently gave both cats (Boris and Viktor) a bath. They did not entirely enjoy the experience, but they are both much softer and smell nice. As you might expect, Boris eventually realized that he wasn't going to be able to escape the bath and let me finish without too much struggle. Viktor never stopped trying to escape.

If you've ever visited my house, you know that of the two cats, Viktor's fur is significantly coarser feeling than Boris'. I had assumed that this was just a reflection of differences in their hair.

It's not.

After bathing both of them, Viktor's hair is as soft as Boris'. Viktor is just much worse at cleaning himself.

I guess this means that somewhat regular baths are in their future.

(Maybe I'll take wet cat pictures next time.)

mod_fastcgi stays around to run the WebObjects adaptor, for now at least, though.

So far, things are looking good, but as usual, let me know if you notice any problems.

Update: I switched things back to mod_fastcgi as fcgid seemed to have a problem with how MT does image uploads.

In an effort to reduce the machine's load, I turned off Spotlight indexing which seemed to help some. The load was still a bit higher than what was previously normal, though.

Eventually, I discovered that the new _mailman user conflicted with the mailman user I had created when installing the mailman mailing list software on the machine a while ago:

% sudo dscl . -read users/mailman RecordName
RecordName: _mailman mailman

Of course, the _mailman account had a UID/GID of 78, while the account I had created had a UID/GID of 504.

I could not determine how to correct this with dscl, since it showed a 'merged' version of both accounts, and instead just removed the mailman.plist from /var/db/dslocal/nodes/Default/users. I then rebooted (as sending SIGHUP to DirectoryServices didn't seem to get it to pick up the changes I had just made).

The load on the machine went down to 'normal' levels then, and I decided to re-enable spotlight. After it finished indexing the drive, the load on the geeklair seems to have stabilized back at the old levels.

Update: I ended up disabling spotlight again (while leaving Time Machine active) as the backups are useful, but command-line mdfind doesn't get used very much (if at all).

Things that went horribly wrong:
- After running through the upgrade install on the 10.5 DVD, the machine refused to boot up (verbose boot showed some kext loading and then the machine would power itself off). I eventually fixed this by booting off of the install DVD, removing the entire contents of /System/Library/Extensions and re-running the upgrade. This seemed to fix things.
- Upon first successful boot, there were no accounts in the admin group (my account had been the only one with admin privileges). Booting into single user mode allowed me to update /private/var/db/dslocal/nodes/Default/groups/admin.plist and fix the problem.
- The postfix user and postdrop group had changed to _postfix and _postdrop (so my postfix config needed to be updated.
- The new clamav user and groups shared uid/gid with dovecot user/groups I had added. I think I've moved all of my created users/groups to uid/gid > 500 so this shouldn't be a problem in the future.
- postfix decided to work for a while and then eventually become unable to look up users via the local_recipient_maps setting I had been using (unix:passwd.byname $alias_maps) changing this to use the proxy (proxy:unix:passwd.byname $alias_maps) seems to have fixed it.

Relatively minor issues:
- my syslog conf got wiped out (/etc/syslog.conf was replaced)
- my snmpd.conf got wiped out (/etc/snmpd/snmpd.conf was replaced)
- The Time Machine disk is on a PCI IDE card and the machine thinks that the disk is 'removable' , but a little /etc/fstab magic should make it so that it gets mounted at boot as it should.
- The installer took a really long time to fsck the disk the first time the install ran

Of course, the car's handling has felt a little off for the last month or so, and after swapping the tires, I see why. The inside edge of all 4 tires has worn down very far (there's almost no tread left on the inside edge of the rear tires, and very little on the inside edge of the front tires). Some reading on the internets indicates that this may be the 'normal' wear pattern for the factory's alignment settings (but that new and alternative alignment settings are available to help). I'll see if I can find more information, but I'll probably have someone (Toyota of Grand Rapids) look at it in the spring when I'll be purchasing a new set of non-snow tires.

[Which means I get about 4 months of time to decide which tires I want ;-) ]

This is pretty cool, but there are a couple of issues.

1. If you used to start X11.app from the dock, you'll probably want to no longer do that (as you end up with two icons in the dock, one which seems to bounce forever).
2. On demand launching doesn't work for remote X sessions (so you have to run a local 'xterm' or something to get X11 started before you try to have your remote X windows applications connect to the local X windows server).
3. If you have the DISPLAY environment variable set in your login scripts (.cshrc or ~/.MacOSX/environment.plist for example) you'll want to remove it, as it messes everything up.