Ruby-dev summary 26662-26760

(Minero Aoki) #1

Hi all,

This is a summary of ruby-dev ML in these days.

[ruby-dev:26628] show information of --enable-pthreads

  Hidetoshi NAGAI proposed a patch which reports that the working ruby
  was compiled with --enable-pthreads option or not. Matz suggested
  adding a new option, which prints all configure options.

[ruby-dev:26711] --with-static-linked-ext and require

  IWATSUKI Hiroyuki reported a static linking related problem. With
  --with-static-linked-ext, ruby cannot load xxx.rb when xxx.so is
  static-linked:

      $ ruby-1.9.0 -e '
          p $".include?("openssl.so")
          p require("openssl") # should require openssl.rb...
      '
      true
      false # but not loaded.

  This bug is fixed by Nobuyoshi NAKADA.

[ruby-dev:26723] reducing #coerce calls

  Tadashi Saito posted a patch which reduces Fixnum#coerce calls in
  numerical calculation. This patch doubles calculation speed on
  average. Matz accepted this patch.

[ruby-dev:26743] zsuper with optional arguments

  Hiroshi NAKAMURA reported a difference of "super" between Ruby 1.8
  and 1.9:

      ~ % cat t
      class Base
  def initialize(a = 'base')
    p a
  end
      end

      class Derived < Base
  def initialize(a = 'derived')
    super
  end
      end

      Derived.new(0)

      ~ % ruby-1.8.2 t
      "derived"
      ~ % ruby-1.9.0 t
      "base"

  Matz selected 1.8 behavior.

  Note that this is known issue; refer ruby-dev summary 23971-24053
  ([ruby-talk:108967]).

-- Minero Aoki
ruby-dev summary index: http://i.loveruby.net/en/ruby-dev-summary.html

(Ben Giddings) #2

Nice! It sounds like this is the best kind of performance optimization you
can get: rethinking a problem and reducing unnecessary overhead. I wonder
if this will eventually reduce the complaints from people who benchmark
Ruby against Foo and determine Foo is better because it's slightly faster.

Congratulations to Saito-san.

Ben

···

On Monday 15 August 2005 02:37, Minero Aoki wrote:

[ruby-dev:26723] reducing #coerce calls

  Tadashi Saito posted a patch which reduces Fixnum#coerce calls in
  numerical calculation. This patch doubles calculation speed on
  average. Matz accepted this patch.