Ubuntu getting unreadier and unreadier for the desktop

I’m pretty sure that isn’t a word, but anyway. Since upgrading to 13.04 yesterday, I’ve found the following bugs:

  1. Unity crashes when turning on the AV receiver. So far I’ve tried using proprietary/open source video drivers, removing all GNOME-/Ubuntu-related dotfiles and directories in ~ (like .gconf), and purging/reinstalling the ubuntu-desktop and xorg-* packages.
  2. Can’t save display settings as root.
  3. Login screen resolution not set.
  4. Still no hint as to whether I should install the fglrx or fglrx-updates driver (the latter gave me a black screen), so I guess that’s not the right one then
  5. When logging in to a virtual terminal, it says “Welcome to Ubuntu 13.04”, and a few lines later it says “New release ‘13.04’ available.” I’ve verified that /etc/apt/sources.list contains no references to older packages, and that lsb_release --all reports “Ubuntu 13.04”.

last.fm to Spotify migration

When last.fm announced on 2012-12-13 that they were pulling out of Switzerland in mid-January, I requested a refund (unsuccessfully so far) and started looking around for an alternative immediately. Spotify seems to fit the bill nicely, and barring any glaring privacy or stability issues it’s probably worth a threefold price increase.

But how to migrate? last.fm has got XSPF and tab-separated values export functionality at /user/<username>/library/loved; great! Except that didn’t work the first half a dozen times I tried, and Ivy, the most promising import service, only supports some iTunes format and comma-separated values, so I ended up making a tool to convert lfmCOL.pl XML to CSV. Of course, after finishing that I tried last.fm again, and now the XSPF export worked, so I created another project to convert XSPF to CSV. Nice hacking exercise to ensure that it produces parseable CSV even with quotes and high-Unicode characters in the input.

Ivy kept telling me I didn’t select a track column in the second step, but would gladly accept the pasted result from xclip -in favorites.csv without even asking about columns. Out of 1222 last.fm tracks it found 734 (60%) on Spotify, which is not too terrible considering the impressive collection of obscure tracks on the former.