intertwingly

Yehuda Katz: you basically never, ever want to use a >= dependency in your gems.

Jeremy Kemper: Loosen activesupport dep to float on 3-0-stable or master

My guess is that Yehuda has had a bit of much bad experience with depending on somebody to honor their interfaces.  The problem with such an approach is that you essentially stop all forward progress.  In this case, 3.1’s Active Record depends on arel and Active Support, and arel also depends on Active Support.  Result:

Bundler could not find compatible versions for gem "activesupport":
  In Gemfile:
    rails depends on
      activesupport (~> 3.0.0)

    rails depends on
      activesupport (3.1.0.beta)

My take: gems should honor their interfaces, and Yehuda’s advice should only be taken when there is evidence that the maintainers do not take this obligation seriously.  People should be able to trust that new releases of Active Support will maintain backwards compatibility.

As this change to arel is unreleased, in order to test the scenarios in Agile Web Development with Rails Edition 4 against the upcoming release of Rails, I need to pull in an unreleased version of arel in my testing.  I shouldn’t have to do this, and furthermore the fact that I have to means that the comment that is placed by default in generated Gemfiles is currently broken (incomplete).

Joseph Walton: It’s always illuminating to see the different levels of argument. Is this an obvious bug or a personal preference? It’s like an optical illusion.  Once you’ve seen it one way, your brain really doesn’t want to flip back.

Yuki Sonoda: Ruby 1.9.2 has been released. This is the newest release of Ruby 1.9 series.

I have one test that fails with this release.  The problem is easily reproducible.  The same test passes with 1.9.1.  It only affects Rails 2.3.x, Rails 3.0 does not have this problem.

Melyssa Allen: Meredith College will welcome new students during Move-In Day this Saturday, Aug. 14.  The College expects approximately 375 students.

One of those is our daughter.

Eric Seidel: One of the cool new features of the HTML5 parsing algorithm is the ability to embed SVG and MathML directly in HTML pages. To embed SVG, you simply add an <svg> tag to your HTML page and you can use the full power of SVG.

Unfortunately, his example was sanitized by Venus.  Fixed.  Test.

Dean Hachamovitch: Please continue to report issues on Connect if your site doesn’t look or work right, and you’re giving it the same code as you’re giving to other modern browsers.

Seeing that the huge watermark issue that I noted with preview 3 wasn’t picked up by osmosis by the IE team, it would seem to be time to enter feedback using Connect.  I’m not thrilled with having to enter information about my employer’s size and my address, but hopefully this time the issue will be picked up by the proper people.

unicorn: The mime-type of your document (application/atom+xml) is unsupported by this task.