Skipping fsck checks during boot with CTRL-C

According to Murphy’s law, the fsck check that happens once every n boots always happens at the worst time. By default, using CTRL-C to abort it causes fsck to exit with an error, and the filesystem to be remounted read-only. It’s easy to change that in /etc/e2fsck.conf:

allow_cancellation = true

Debcamp+Debconf again! (update)

As I did since 2007, I will be attending Debconf again this year. I have been quite busy recently, and am lagging behind in some of my Debian work, so I plan to use Debconf to get stuff done, not just socialize and attend talks.

Here is my TODO list, basically by order of “willingness to do the work”. It’s probably too much for 9 days, so don’t hesitate to talk to me if you want to help with some items.

  • Switch the ruby1.9.1 package to a prerelease of Ruby 1.9.2. That package providing the development branch of the interpreter. That will be done while keeping the existing package name, as Ruby 1.9.1 and 1.9.2 are (supposed to be) compatible. It might be a bit confusing for users to have a ruby1.9.1 package that installs Ruby 1.9.2, but then we can just blame the upstream developers for using the same numbering space for “ruby compatibility level” and “ruby version”. Bonus: will fix 2 FTBFS on {kfreebsd-,}i386. ruby1.9.1 1.9.2~svn28788-1 uploaded, but it really means “ok, enough, let’s upload and see what happens”. There are some open issues on FreeBSD, and some patches that still require porting to that version.
  • Update the ruby1.8 package. There’s a few patches pending (including the “ruby is slow because of pthread” one). ruby1.8 uploaded, sync to Ubuntu requested.
  • Run an archive-wide test using instest again. instest is a piuparts re-implementation that is targetted at making it easy to file bugs about installation and removal failures. It’s nothing big, really, but already allowed to file 78 RC bugs a few months ago. Done, and call for help posted. Which raises another TODO item:
    • Process instest failures if nobody wants to help … Well, done partially. The number of issues found justifies that those tests are useful, so I’m a bit annoyed that nobody volunteered to help.
  • Run an archive rebuild. I did one last week-end mainly to provide food for RCBC, but there are some failures that haven’t been filed because they required more investigation. I also need to check back with various people about some custom rebuilds (newer GCC, etc). And I’ll probably also do a rebuild of Ubuntu maverick.Rebuild done, about 70 new RC bugs filed, and some cleanup done on old unreproducible bugs. Also did a rebuild with an updated hardening-wrapper for Raphael Geissert, and a rebuild of Ubuntu maverick on i386 and amd64 (results posted on ubuntu-devel@).
  • Do a ddpo-by-mail run. Last one was a longgg time ago. I have a few open requests about ddpo-by-mail that I’d like to address before spamming people again (for example, it shouldn’t report about merged bugs).Done, ~1300 emails sent.
  • Ultimate Debian Database work: merge the upload-history importer into UDD itself to drop the dependency on a script on merkel. Also address #540132 (strange info in the upload-history table). Make sure the changes to carnivore about DMs are compatible with the carnivore importer. All done!
  • Investigate how we could improve Ruby packaging with something like ruby-support. One of the requirements is to provide an easier way to support several Ruby versions.
  • Do some work on feed2imap. It got a few valid feature requests recently that I’d like to address in a 1.1 release.
  • Debian/Ubuntu work: add info about bugs with patches to the PTS now that this info is available in UDD thanks to Brian Murray’s work on the Launchpad side, and other pending work items from the Debian/Ubuntu BOF at UDS
  • Work on Bapase and the handling of poorly maintained and low-quality packages, resurrecting an old thread on debian-qa@
  • Do some packaging work: systemtap 1.3, tcsh, websec (long overdue!), and the usual pkg-ruby-extras shores. I also would like to take a look at the atlas FTBFS as I’m affected through hpcc.
  • Update developers-reference. I’m totally in lazy mode about that since the maintenance was transferred to debian-policy@, but I will try to address bugs with valid patches.