I have added packages of Kye for python 2.5 and 2.6 to the download page.

zsync 0.6.1 is now available from the download page. This fixes a few bugs that have been spotted in the previous version, plus a few minor feature changes:

• recompression support for gzip files made with zlib:gzio.c or gzip -n
• fix compilation on MacOS X
• allow HTTP redirects on the target file; not sure whether this is a good idea or not...
• fix unecessary transfer of whole file where file is smaller than the context size (1x or 2x blocksize)
• use sequential_matches=1 when there is only one block; otherwise we're forced to transfer the whole file for files below 2kiB
• fix librcksum handling of zsync streams with sequential_matches == 1; it was giving false negatives when applying the rsync algorithm, resulting in poor use of local source data when sequential_matches == 1 (which didn't actually occur in any recent version of zsync)

I just upgraded my home and work computers to Ubuntu 9.04. As expected this went smoothly. While I was at it, on my work computer I did a reinstall in order to get the full drive encryption which has been available in Ubuntu for several releases now. It is easy enough to set up, as per several online guides; perhaps the only thing missing is an option to have the drive unlocked by a keyfile (say on a USB thumb drive) instead of by passphrase.

But I was of course not prepared to take it at face value — I then booted from a Linux CD to have a poke at it and convince myself a) that I really couldn't mount the drive without the passphrase; and b) that I knew how to access the drive even if the OS on the hard drive would not boot.

There are some online guides to rescueing an encrypted Ubuntu system. I did something slightly different; I booted the Ubuntu alternative install CD, and allowed it to load all the drivers (importantly to get the LVM drivers loaded) before ALT-F2ing over to the provided console; that avoids having to load any of the modules manually.

Ubuntu's encrypted drive setup is to create a small unencrypted boot partition at the end of the drive, and the rest of the drive in a single partition which is an LVM PV wrapped in an LUKS encrypted volume. For my reference, this is the screenshot from accessing the drive while booted from CD.

As I am stuck at home and not feeling too well on this bank holiday weekend, it seems like a good time to catch up on posting pictures from my walks around the vicinity of Bath. So here are the pictures from a walk to Little Solsbury Hill in 2008.

more

And now a snow-covered Bath, after heavier snowfall overnight.

Snow-frosted Bath (and Bathhampton Down in the distance), at dawn this morning.

zsync 0.6 is now available from the download page. This is mainly a maintenance release, fixing various minor bugs that people have noted over the 2 years since the last release. I have also gone through and tidied up the source code somewhat.

The only functional changes are:

• zsync now preserves the mtime on downloaded files (this requires an extra field in the .zsync, but this format change is entirely compatible with old clients);
• -q option replaces -s (but -s is retained temporarily as a synonym).

These make zsync align better with wget as a file download client.

The full changelog:

• fix out-of-bounds memory access when processing last block of non-compressed download (patch from Timothy Lee). Also fix an error handling fault for the same.
• fix "try a smaller blocksize" failures when zsyncmakeing for huge compressed files on 32bit systems
• preserve mtime on downloaded files
• fix potential crash when re/deallocating checksum hash in librcksum (patch from Timothy Lee)
• explain status code errors better
• better URL handling
• add -q as a substitute for -s, as -q is more conventional (re wget). -q also suppresses the 'no relevant local data' warning now.
• fix some warnings
• code tidy-up and better commenting of what it is doing
• tidy up autoconf use

Version 0.6 is available from the download page, as are all previous versions and the bzr repository.

One of my favourite walks, in fine weather, is to catch a train to Bradford on Avon and walk back to Bath.

Bradford on Avon is a town in west Wiltshire, whose main attraction is the well preserved old buildings that make up the central part of the town. That, together with the Kennet and Avon canal and its proximity to the prime tourist destination that is Bath, mean that its main industry now is tourism (which is unusual, perhaps, for a minor inland town); although the reason that I know it well is that Netcraft had its offices there for 5 years.

The river valley between Bradford on Avon and Bath is very picturesque (although no doubt my compact camera (and its operator) will totally fail to convey that), making for a very pleasant 2 hour walk. These pictures are a mixture from walks in the summers of 2007 and 2008.

Update: refreshed with pictures from a walk in April 2009, where I got rather more sunshine.

more

Over Christmas, I was debugging a problem on a friend's computer where PrBoom would hang randomely soon after starting. The problem seemed to be sound-related, as the game would often hang on MAP01 when dropping to the floor, picking up the chainsaw, or some other activity that would generate sound. I subsequently find that it's an issue that many people are seeing.

What exactly triggers the problem is not clear - it seemed to often run the demos okay, but fail fairly consistently when playing for real. But the actual problem, when it occurs, is clear enough. PrBoom gets stuck with the audio thread trying to open /dev/sequencer; and, for some reason, it is stuck in the kernel trying to get a handle to the MIDI sequencer device. Since most systems don't come with a MIDI synth chip these days, I suspect it's actually the ALSA sequencer emulation that is being invoked. I think it's ALSA, or the kernel, that is actually at fault.

Anyway, the problem is easily worked-around by either:

• Disabling music — run with -nomusic or set music_card to 0 in prboom.cfg (if you have PrBoom installed, you can do man prboom.cfg to get the documentation on the config file).
• Install Timidity and a patch set. If /etc/timidity/timidity.cfg is present, PrBoom (actually SDL_mixer) will use SDL_mixer's MIDI synth instead of looking for hardware MIDI.

On a related subject, I notice that the freepats patch set that comes with Debian these days gives a much poorer rendering of the Doom music than the old timidity-patches package. Apparently the older set had some unresolved copyright issues that resulted in it being dropped from Debian; but, for those willing to overlook that, you can still get the patches from mirrors of old distributions of Debian.

Update: Fraggle points out that the EAWPATS patch set is available from the idgames archive and is a big improvement on freepats.

EAWPATS is much closer to the old timidity-patches, but they are clearly distinguishable - the music for "The Pit" shows the difference most clearly, as EAWPATS has the melody line for the first half in what I think is a harpsichord tone, where timidity-patches does not (because it doesn't have a harpsichord patch, instead reusing one of the other piano types). Also the relative instrument volumes (which seems to be the biggest problem with rendering old MIDIs ‒ background instruments played too loud and drowning out the foreground ones) are different. EAWPATS seems better overall.

I have been having a go at devnull's Nethack tournament this year. After my success with Nethack earlier in the year, I figured I had a good chance to at least place well in a few of the contests.

Currently, it is going better than expected - I am leading both the lowest scored ascension, and the best behaved (most conducts) ascension trophies. Neither are competitions I expected to do well at: I had managed a few conducts individually before, but never so many in one game; and my previous lowest score game on NAO was in the 600,000s. I don't expect I will stay ahead in either competition long — there are players who specialise in both disciplines who just haven't got their games on the board yet. So I'm just enjoying the 15 minutes of fame here :-)

For the uninitiated: ascending (completing) nethack with a low score is much harder than winning with a high score, since, by the time you are in a position to complete the game, killing most monsters is easy and every kill counts to your score. To finish with a low score, I played nearly-pacifist for the last 20k turns. Conducts are where you beat the game under self-imposed additional restrictions; for my 7-conduct game, I managed to play vegan, genoless, wishless and polyless.

Update: Before I even finish my post, the mighty Tenaya has come along and completed a pacifist game, knocking me off of the leaderboard for the low score ascension (by a factor of 10). And equalling my 7 conducts (although I still have the moral lead in that one, I think, since Tenaya would have had to start-scum to play pacifist, whereas my game was a straight game — my only monk of the tournament so far — with only normal good fortune).