Back at Goodwood#

Tue, 31 Jul 2007 16:19:23 +0000

... and, back after Goodwood too. Yesterday it was still aching; today the rest of me also aches, I'm shivering and my throat is complaining: I begin to suspect I may be suffering a cold.

The event benefited from the Camberley crowd's same slick management as they conditioned us to expect last year, and was also notable for the relatively vast numbers of entrants: 213 people! Kudos to them on both counts. It's a nice wide track and no danger of overcrowding.

A practice lap established that the wind had changed direction since last year. The headwind after the first corner seemed much less of a "brick wall" than I remembered from 2006 or 2005, but the flip side was that the wind on the start/finish straight and the downhill leading to it was much stronger.

Off the start, and the lead paceline formed within a few seconds: as usual for British races everyone at the front was the epitome of politeness ("no, I insist you go first"). This lasted until just past the first corner, when James Ashby flew off, with Orgill, McInnerny and Andy Porter following, and Fred and me pursuing but failing to catch. By the time we got to the last corner before the start/finish the rest of the line had caught us too, so we settled down in a mostly LSST line - about eight initially, gradually depleting, including one or two Powersliders (who, all credit to them, were not shirking approximately their fair share of doing the work - or at least, not more so than I was)

This was followed by nine laps in that line. James and Chris lapped us somewhere around lap 8: James slowing very briefly before powering up the warp generators again and disappearing into the distance; Chris hung with us for a couple of laps before taking off again on the back straight on our lap 10 (his 11).

Last lap: the start/finish was marked by similar outbursts of politeness as had been evident at the race start, so we went around the first corner a bit spread out across the road. I was lagging at the back waiting to see who broke first (special eye on Fred). Fred made his move, I followed, Andy followed, I don't know who else was with us at that point. At the water stop, Fred pulled off and I went past - then looked back a few seconds later and found there was a clear gap between me and everyone else. So, reasoning that there was no way I'd win a sprint finish against Andy or Fred and if I had any chance at all it was over a longer distance, I went for it.

Unfortunately, it wasn't enough: the back straight (with a tailwind) is deceptively fast easy skating, and only when I turned back round into the headwind did I realise I didn't have much left. They caught and passed me again at the end of the downhill, and I came in sixth in 1:23:36. Wasn't particularly impressed at the time, but post-race reflection says that everyone in front of me (except Fred) was Cat 1, and I finished within 5 seconds of one of them. So perhaps not that bad really.

Loitering with indent#

Wed, 08 Aug 2007 22:55:35 +0000

It's been one of those dependency^Wdistraction hell weeks. I'll spare you the details, because that way lies Rant, but by mid-afternoon today I found I'd upgraded Emacs to version 22 and was trying to write a new MUA by sticking bits of mel-base and ltk together, and getting really weird "stairstep" indentation for lists of keywords. This turned out to be nothing to do with stale fasls/elc files (first suspect) or with SLIME (second suspect) or with cl-indent (third suspect), but the underlying Emacs lisp indentation engine has had some changes to "try to align a constant-symbol under the last preceding constant symbol" which has had a completely undesirable knock-on effect on CL.

So, without further ado: my .emacs file now says

(require 'slime)
;; emacs 22 lisp indentation, if left to its own devices, indents
;; lists of keywords in a "stairstep" fashion like so
;; (list :foo :bar
;;       :baz :quux
;;            :bum :ki)
;;
;; this appears to be intentional (see comment "try to align a
;; constant-symbol under the last preceding constant symbol" in
;; lisp-mode.el) and perhaps it's even useful for elisp, but it makes
;; make-instance (for example) look bloody silly
(defun my-cl-indent-function (indent-point state)
  (let ((normal-indent (current-column)))
    (if (save-excursion
	  (goto-char indent-point)
	  (skip-chars-forward " \t")
	  (looking-at ":"))
	normal-indent
      (common-lisp-indent-function indent-point state))))
(slime-setup)
(add-hook 'lisp-mode-hook
	  (lambda () 
	    (setq lisp-indent-function 'my-cl-indent-function))
	  'append)

You need the hook (and you need it to run last) as SLIME also installs its own hook to make lisp-indent-fuction buffer-local and set it to common-lisp-indent-function

Fans for nothing#

Mon, 27 Aug 2007 22:02:59 +0000

Since upgrading my amd64 box to 64studio the other day, it's been unable to run more than 24 hours without shutting itself down - usually while I was out and unable to see how it crashed. Yesterday, though, I was actually at the keyboard, and (suspecting overheating) instantly whipped the case off while I rebooted it into the BIOS setup screen. Hmm. Looks hot. Ow. Feels hot, too. Perhaps a case fan would be a good idea.

Well, right diagnosis, wrong prescription. I bought a case fan today, and perhaps it was a good idea, but the immediate problem seems to be that the CPU fan only revolves intermittently (and has now stopped doing even that, whch made the roblem much easier to identify). There's not much a case fan can do to mitigate that ...

On the upside I also got a new graphics card to replace the very elderly Voodoo3 I've been using for about eight years, and took the opportunity to recable everything and neaten up all the wires inside the machine, so it looks prettier if nothing else.

64studio looks kind of cool so far, although I haven't played with it too much yet. That was motivated by another foray into producing mixes for the LFNS. Having messed around a little with Mixmeister and finding that although it makes the hard things (beat matching, etc) simple, it compensates by making the simple things (such as starting the track when you want it to) way way way too difficult, I'm looking for something slightly different. I just don't yet know what.

And fan cue two#

Wed, 29 Aug 2007 17:18:53 +0000

"What happens if I plug it in there instead?" Oh. The CPU fan is in fact fine: the problem is the motherboard header. Plugging the CPU fan into the socket labelled "PWR FAN" and the case fan into "CHA FAN", and now they both go round. The machine refuses to boot until it's (mistakenly) complained that the CPU fan is not revolving, but there's probably a BOIS option to fix that. And gkrellm, courtesy of the k8temp module, can now tell me that the CPU is at about 48 degrees: no need to buy more bits of hardware after all.

Which is nice, cos I just got one of these instead. More on that when it turns up, but I might add that in the meantime I'm happy to see that Mixxx has resumed development and that with only a small amount of mucking around (comment out the OpenGL calls in Material::Material) I've managed to make it not segfault on startup. It may need more hacking to be entirely drivable by MIDI, though (Program Change messages to select tracks from the playlist? Someone tell me why this is a bad idea, if it is)