Skip to content
Commits on Sep 29, 2012
  1. @ferrous26

    Fix symlink generation when installing using the DESTDIR option

    ferrous26 committed Sep 28, 2012
    Most notably the nightly builds were getting messed up because the
    symlinks for the bin scripts were pointing to a directory that did
    not exist on most systems.
  2. @ferrous26
  3. @ferrous26

    Make ruby_deploy cleanup .rb files properly when using --compile

    ferrous26 committed Sep 28, 2012
    Reported by Robert Rice on the MacRuby developers mailing list September 28, 2012.
  4. @ferrous26
Commits on Sep 27, 2012
  1. @ferrous26

    Merge remote-tracking branch 'macruby/master'

    ferrous26 committed Sep 27, 2012
    * macruby/master:
      move the implementation about #hash from Array to NSArray, in order to use its methods with Array/NSArray
  2. @ferrous26
Commits on Sep 25, 2012
  1. @Watson1978

    move the implementation about #hash from Array to NSArray, in order t…

    Watson1978 committed Sep 25, 2012
    …o use its methods with Array/NSArray
  2. @ferrous26

    Move Xcode Template setup to ruby_install_xcode_support

    ferrous26 committed Sep 24, 2012
    So that it can be easily redone if the framework moves.
Commits on Sep 24, 2012
  1. @ferrous26

    Dynamically set the Xcode template Framework Search Path at install time

    ferrous26 committed Sep 24, 2012
    This is so that MacRuby.framework installations to a location other than
    /Library/Frameworks can still make use of the Xcode templates without
    having to manually change the search path.
  2. @ferrous26
  3. @ferrous26

    Revert "Use Rake's built in FileUtilsExt to reduce redundant messages…

    ferrous26 committed Sep 24, 2012
    … during install"
    
    This reverts commit 032cb54. We cannot
    use Rake::FileUtilsExt yet as the system Ruby ships with an ancient
    version of Rake that does not include the extension in its current
    form.
  4. @ferrous26
  5. @ferrous26
  6. @ferrous26
Commits on Sep 6, 2012
  1. @Watson1978
  2. @Watson1978

    improve String#reverse performance with ascii compatible string

    Watson1978 committed Sep 6, 2012
    * before
         user     system      total        real
    22.430000   0.420000  22.850000 ( 20.853809)
    
    * after
        user     system      total        real
    1.080000   0.000000   1.080000 (  1.089038)
    
    Test Code:
    ----
    require 'benchmark'
    
    Benchmark.bm do |x|
      str = "foobarbaz" * 5000
      x.report do
        50_000.times do
          str.reverse!
        end
      end
    end
Commits on Sep 5, 2012
  1. @Watson1978

    Regexp.last_match(nth) should return nil if passed index which is out…

    Watson1978 committed Sep 6, 2012
    … of matched range
    
    Here is code which expected that Regexp.last_match returns nil:
    
        /(...)/.match("foobarbaz")
        Regexp.last_match(4)  #=> nil
        Regexp.last_match(-3) #=> nil
    
    However, raises an IndexError.
Commits on Aug 28, 2012
  1. @Watson1978

    update ChangeLog

    Watson1978 committed Aug 28, 2012
  2. @Watson1978

    update json to v1.7.5

    Watson1978 committed Aug 28, 2012
  3. @ferrous26

    Update ChangeLog.markdown

    ferrous26 committed Aug 27, 2012
    Including the RubyGems 1.8.24 upgrade and GH-90
Commits on Aug 27, 2012
  1. @Watson1978

    update the rubyspec's tags

    Watson1978 committed Aug 28, 2012
  2. @Watson1978

    add the rubyspec's tags

    Watson1978 committed Aug 28, 2012
  3. @Watson1978
  4. @Watson1978
  5. @Watson1978

    fix the MatchData#[] behavior if passed a negative Fixnum object as i…

    Watson1978 committed Aug 27, 2012
    …ndex
    
    In below case, MatchData#[] always returns a nil:
    
    	index < 0 &&
    		(matched count) + index >= 0
    
    Sample code:
    
    ```
    str = 'foo bar baz'
    m = str.match(/\w+/)
    p m[-1]  #=> expect "foo", but nil
    ```
    
    This fix above behavior.
Commits on Aug 26, 2012
  1. @ferrous26

    Merge RubyGems changes up to latest 1.8.x release

    ferrous26 committed Aug 25, 2012
    This brings in the new SSL changes and various other small fixes.
Commits on Aug 25, 2012
  1. @ferrous26

    Make Dispatch::Queue#apply documentation less misleading

    ferrous26 committed Aug 21, 2012
    Claiming that the blocks are executed in parallel is not correct
    since it depends on the type of queue being used, and even in that
    case concurrency is not the same things as parallelism.
    
    In the example, the array grows dynamically and the assignment
    being performed in the block would not be thread-safe
    (concurrency-safe?).
  2. @ferrous26
  3. @ferrous26

    Remove xcode_support from the standard rakefile installation

    ferrous26 committed Aug 25, 2012
    This change introduces an install:standard task which is now
    the default. It will still install the necessary components
    so that the Xcode integration script can be run, but it will
    not run the integration install script.
    
    Xcode integration will still be installed by default when
    using the package installer, and the script can be automatically
    run using the install:all task.
  4. @ferrous26

    Address GH-90 by breaking out Xcode integration installation

    ferrous26 committed Aug 25, 2012
    Xcode integration is now installed via a script (macruby_install_xcode_support)
    which is installed into /usr/local/bin along with the other scripts.
    
    This allows the integration to be reinstalled easily. The script also has
    reworked the install package so that it does not fail when Xcode integration
    cannot be installed.
Commits on Aug 24, 2012
  1. @Watson1978
Commits on Aug 23, 2012
  1. @Watson1978
  2. @Watson1978
Commits on Aug 19, 2012
  1. @Watson1978
Commits on Aug 18, 2012
  1. @Watson1978

    RSTRING_PTR() returns (char *) instead of (const char *) like CRuby b…

    Watson1978 committed Aug 18, 2012
    …ecause the compiling fails with some C extension libraries
Something went wrong with that request. Please try again.