Skip to content
This repository

Jul 09, 2014

  1. Brian Shirai

    Fixed destructuring block arguments.

    brixen authored

Jul 08, 2014

  1. Brian Shirai

    Added block destructuring specs.

    brixen authored
  2. Brian Shirai

    Fixed (un)signed comparisons.

    brixen authored
  3. Brian Shirai

    Fixed BB label placement.

    brixen authored
  4. Brian Shirai

    Fixed arity check of 0-arity block.

    brixen authored
  5. Brian Shirai

    Fixed inlining blocks.

    brixen authored

Jul 07, 2014

  1. Brian Shirai

    Fixed inlining generic methods.

    brixen authored

Jul 06, 2014

  1. Brian Shirai

    Tweak JIT method args.

    brixen authored
  2. Brian Shirai

    Fixes for JIT block lambda flag and total arguments passed.

    brixen authored

Jul 03, 2014

  1. Brian Shirai

    Fixed calculating post arguments index.

    brixen authored
  2. Brian Shirai

    Removed old method, block JIT code.

    brixen authored
  3. Brian Shirai

    Fixed JIT handling of method, block arguments.

    brixen authored
  4. Brian Shirai

    Clean up JIT inline policy.

    brixen authored
  5. Brian Shirai

    Fixed possibly intentional misspelling.

    brixen authored
  6. Brian Shirai

    More block, lambda argument passing fixes.

    brixen authored
  7. Brian Shirai

    Fixed lambda keyword argument mapping.

    brixen authored
  8. Brian Shirai

    Updated Gemfile.lock, gems_list.txt.

    brixen authored
  9. Brian Shirai

    Fixed parts of assigning arguments with keywords.

    brixen authored

Jun 30, 2014

  1. Yorick Peterse

    Renamed variables used for checking syntax.

    This should hopefully prevent future cases of the Mondays.
  2. Yorick Peterse

    Use the correct local variable for syntax errors.

    In my defense, it is Monday morning and I've not had my mandatory 10 liters of
    tea. This fixes #3076.
  3. Yorick Peterse

    Keep parser instances around when checking syntax.

    When checking Ruby syntax the instance of CodeTools::Melbourne should be kept
    around instead of using class methods. This makes it possible, in theory, to
    retrieve the syntax errors similar to when parsing files.
    
    Currently Melbourne seems to lose track of the syntax errors. That is,
    CodeTools::Melbourne#syntax_errors returns an empty Array. Interesting enough
    when running similar code in a Pry session the syntax errors are *not* lost.
    
    See #3076 for more information.

Jun 27, 2014

  1. Brian Shirai

    Merge branch '2.2.9'

    Conflicts:
    	Changelog
    	configure
    brixen authored
  2. Yorick Peterse

    Use `int` for INT2NUM instead of `long int`.

    MRI changed the signature of INT2NUM to use `int` instead of `long` starting
    with MRI 1.9.
    
    This fixes #3072.
  3. Yorick Peterse

    Basic specs for INT2NUM / rb_int2num.

  4. Brian Shirai

    Bumped version to 2.2.10.

    brixen authored
  5. Brian Shirai

    Updated Changelog.

    brixen authored
  6. Brian Shirai

    Fixed Process.spawn mapping fds.

    brixen authored
  7. Yorick Peterse

    Merge pull request #3070 from Benny1992/master

    BasicObject Coercion
  8. Jesse Cooke

    Merge pull request #3071 from robin850/minus-nsec-precision

    Nanosecond precision with Time#-
    jc00ke authored
  9. Benjamin Klotz

    check if object where coercion failed reponds to inspect before raisi…

    …ng with obj.inspect
    Benny1992 authored
  10. Benjamin Klotz

    if Object which should be coerced is of class BasicObject, don't try …

    …to call #inspect on it
    Benny1992 authored
  11. Benjamin Klotz

    add example for BasicObject, update old examples expect to raise Type…

    …Error not any Exception
    Benny1992 authored
  12. Robin Dupret

    Added specs for nsec precision subtracting Time objects

    robin850 authored
  13. Robin Dupret

    Use nanoseconds precision subtracting Time objects

    To be consistent with MRI, let's return a Float with nanoseconds
    precision when we are subtracting two Time objects.
    
    We simply need to rely on the instance's nsec method instead of the usec
    one. Thus, we multiply the subtraction of each object's nano-seconds
    with 1e-9 instead of 1e-6.
    robin850 authored
  14. Brian Shirai

    Added specs for fd map option in Process.{spawn, exec}.

    brixen authored
Something went wrong with that request. Please try again.