Time for another monthly placeholder entry

Mon, 01 Aug 2005 01:44:28 +0000

Time for another monthly placeholder entry. No Lisp lately; my employers have been making me actually work for a living for the past few weeks, and my non-work time has been occupied with stuff I can’t talk about yet. I’m sure that wasn’t in the contract…

Round up for the month of July:

Time for another monthly placeholder entry

I've been playing with Google Maps

Fri, 12 Aug 2005 13:21:50 +0000

I’ve been playing with Google Maps. The link you see there is notable less for what you can see (though the satellite view is kind of neat) as what you would be able to see if you – for example – were to type ‘javascript:showMarkers()’ into the url bar immediately after visiting it. Probably only works in Mozilla-derived browsers, or at least has not been tested in anything else.

Anyway, that’s where I’ll be leading 100-odd skaters (some odder than others, I have no doubt) on Sunday afternoon.

In the process I noted quite a lot of one-way streets that are incorrectly marked in Google Maps, which is a shame as it does rather reduce the usefulness of said tool for skate route planning of this kind.

It would be nice if it measured distance: that will be forthcoming. Yes, I do know about the gmap pedometer, but that only allows the last point to be edited, whereas this thing allows intermediate points to be moved/inserted/deleted – and is a bit easier to get the data out of in useful format.

I've been playing with Google Maps

Sadly, rain stopped play

Mon, 15 Aug 2005 16:52:22 +0000

Sadly, rain stopped play. Or at least, the forecast threat of it was serious enough that we cancelled the Rollerstroll (no link as web site is a useless Flash thing) yesterday and instead went out for an impromptu skate down by the river and then up and down Pentonville Rise. Happily in a way we did get rained on, thus vindicating the “no go” decision for the official skate.

New skates! Currently borrowing a pair of these attached to these: only been out on them twice and they do seem to require a significantly different technique than xmax. Which is not a surprise, really. If I haven’t shredded my feet in the next few days – and all the signs so far look promising; a certain amount of ankle rubbing on the left boot that will go away when I manage to get the frame aligned properly – I will be making an offer for the boots (and then have to get some frames of my own for them, but that’s a fairly straightforward deal).

After some hacking with Venkman and the Microoft Script Debugger, my Google Maps hack now seems to work in IE as well as Firefox. Firefox won’t print polylines when I ask for a hardcopy, though, which is annoying.

What else? A full factory reset of my DG834G seemed to improve wireless stability a lot, but it still manages to drop off the net after about five days and I still don’t know why. Changing the frequency from the admin interface was enough to kick it back into life. Also even when it does work, something is not right with the DHCP: the router’s DHCP server issues addresses with a 12 hour lease, and either it refuses renewal requests or Linux doesn’t make renewal requests, so … gah.

Sadly, rain stopped play

"The suspend is killing me", said Arthur testily

Fri, 19 Aug 2005 16:48:09 +0000

“The suspend is killing me”, said Arthur testily.

Actually, the suspend is a lot better than it was. It seems that the good bits of Suspend 2 are supposed to be being read by other people and slowly merged into the less-featureful swsusp code in the baseline kernel. So I decided to try that and see how far it’s got, and it already works better than suspend2 did. My usb phone is still broken, but after some investigation and a fairly brutal patch I have suspendable wireless at least.

"The suspend is killing me", said Arthur testily

I just had to reboot my computer after a Lisp program wedged it

Sat, 20 Aug 2005 23:25:49 +0000

I just had to reboot my computer after a Lisp program wedged it.

Well, I guess that needs a little qualification. The Lisp in question was librep, the librep interpreter was also running my window manager (Sawfish), and the program was (read-line). And the computer wasn’t totally wedged (it carried on playing my MP3s) but wouldn’t listen to any keyboard input or focus changes, nor even to M-C-F1 or M-C-Backspace, making it rather hard to regain control. So if I had another machine here I could probably have sshed into it, but I haven’t, so there.

Still, if one of the supposed advantages of Lisp is the ease of interactive experimentation and debugging, (and I think I’ve said this before), Sawfish is not it.

Context: still flushed from my success in getting 90%-working suspend-to-disk, I thought I’d make the machine’s power button activate it instead of having to su root and echo stuff into obscurely named files in /sys. That worked quite well once I’d accounted for my script getting run multiple times on each button press. Then I decided to make the volume buttons work too, and that’s where I started playing with sawfish. Maybe it’s time to try Eclipse again.

#!/bin/sh
# /etc/acpi/powerbtn.sh
# Initiates suspend-to-disk when the power putton has been
# pressed.  Note use of background process and locking: button 
# seems to send two events (up and down?) each time it’s pressed,
# and we only get called for the second after the script has returned
# from the first.
( if lockfile -r0 /tmp/.acpi-suspend-lock; then
 echo ‘platform’ >/sys/power/disk
 echo ‘disk’ >/sys/power/state
 rm  /tmp/.acpi-suspend-lock
else 
 echo ‘Ignoring second suspend attempt’
fi ) &
exit 0;

I just had to reboot my computer after a Lisp program wedged it