Here is a summary of ruby-dev mailing list.

[ruby-dev:28087] test(?-, file1, file2)

  Akira Tanaka asked what is correct behavior of  test(?-, file1, file2).
  According to the online manual, this method checks whether file1 is a
  hard link to file2. But in fact it returns true when file1 is a symbolic
  link to file2. Nakada answered the current specification of the behavior
  should be kept since it is the same as that of "test -ef".

[ruby-dev:28095] ruby 1.8.4 preview3
[ruby-dev:28118] ruby 1.8.4 released

  Matz announsed third preview version of ruby-1.8.4 in [ruby-dev:28095],
    ftp://ftp.ruby-lang.org/pub/ruby/1.8/ruby-1.8.4-preview3.tar.gz
    1ba94874e1a253d3f1373533553080ae
  and then released ruby-1.8.4 a few days after.
    ftp://ftp.ruby-lang.org/pub/ruby/ruby-1.8.4.tar.gz
    bd8c2e593e1fa4b01fd98eaf016329bb

[ruby-dev:28109] Kernel#fail

  Shyouhei Urabe proposed that Kernel#fail should be abolished.
  However, some people use it instead of raise. One of the reason
  is that they want to use a method name "raise" in their library.

[ruby-dev:26743] zsuper in a method with optional arg

  Hiroshi Nakamura found that behavior of "super" in ruby-1.8 is
  different from ruby-1.9 when "initialize" of a sub-class is
  called without optional parameters.

    class Base
      def initialize(p1, p2 = 1)
        p [p1, p2]
      end
    end
    class Derived < Base
      def initialize(p1, p2 = 2)
        super
      end
    end

  The behavior of ruby-1.9 is close to that in Matz's mind. However,
  he decided to fix the behavior of ruby-1.9 from the results of a
  discussion relating to [ruby-dev:23973].

[ruby-dev:28107] Process.exec

  Minero Aoki suggested a new method "Process.exec" and posted a patch
  that implements "exec", since Process has "fork" but doesn't have "exec".
  His patch has been merged.

[ruby-dev:28127] Intel C++ Compiler and HP aC++/ANSI on IA64

  Akira Tanaka posted a patch so that we can build ruby-1.9 using Intel C++
  Compiler with the option -O0 and can run it on IA64. He then wrote that
  he successfully built using HP aC++/ANSI C, but 'make test-all' didn't work
  well. His patch was merged to ruby-1.9.
-- 
Takaaki Tateishi <ttate / ttsky.net>