Skip to content


Subversion checkout URL

You can clone with
Download ZIP
Commits on May 6, 2010
Commits on Mar 3, 2010
  1. Tweak last commit

Commits on Dec 8, 2009
Commits on Nov 14, 2009
  1. Fix typo in Gzip logic

Commits on Nov 13, 2009
  1. Don't compress binary files

Commits on Oct 15, 2009
Commits on Jul 4, 2009
Commits on May 24, 2009
  1. Rakefile: Create offline versions, refactor.

    We now create a build for offline use that is used to generate a .deb and
    archives for downloading.
  2. Rakefile: Build on systems without Git.

    We use the Git commit log to find the date a recipe was published. When
    building on another system, such as Launchpad, we aren't working in a
    repository so can't call `git log`. In this case we just issue a warning.
  3. Rakefile: Use ERB instead of Erubis for .deb.

    The Ubuntu build bot can't install Ruby gems, and Erubis isn't packaged, so we
    switch to ERB, which is packaged...
  4. Rakefile: clobber deb/ .

  5. Rakefile: expl. req. rubygems so fakeroot works.

    RUBYOPT isn't in the environment when debuild runs with a fakeroot. To fix
    this we explicitly require 'rubygems' from the Rakefile.
Commits on May 23, 2009
  1. Rakefile: don't clobber deb/.

Commits on May 19, 2009
Commits on May 13, 2009
  1. Feed: Use full URL for entry links; set <updated>.

    Relative URLs for entries confused, so we now use the full
    Previously, the <updated> element of <feed> wasn't dynamic. Now it's set to
    the date of the most recent entry.
  2. Rakefile: Store subsections in ToC; fix '#' links.

    We previously ignored IDs of h tags numbered lower than 3. Now we store them
    in the ToC. This lets us fix cross-links to specific sections in other
    recipes. IOW, it lets recipe A link to recipe B#section.
  3. Rakefile: Remove :ilinks task.

    We'll check for broken links when we fix the <a href="#...">s instead.
Commits on May 12, 2009
  1. Rakefile: PDF task shouldn't delete its input.

    The _pdf_ task previously deleted its input (_output/all.html_), this made
    `prince` error messages difficult to debug. We now delete it before uploading,
    instead, so the _pdf_ task can be used to find invalid HTML.
Commits on May 11, 2009
Commits on May 10, 2009
Something went wrong with that request. Please try again.