Skip to content
Commits on Dec 31, 2013
  1. @myronmarston

    2.1.0 is out.

    myronmarston committed Dec 30, 2013
Commits on Dec 21, 2013
  1. @myronmarston

    Merge pull request #502 from rspec/build-2-99-against-2.1

    myronmarston committed Dec 21, 2013
    Build against 2.1.0-rc1.
  2. @myronmarston

    Build against 2.1.0-rc1.

    myronmarston committed Dec 21, 2013
Commits on Nov 14, 2013
  1. @myronmarston

    Merge pull request #469 from rspec/dont_redefine_caller_filter

    myronmarston committed Nov 14, 2013
    Dont redefine caller filter if it's already defined
  2. @JonRowe
Commits on Nov 11, 2013
  1. @myronmarston

    Merge pull request #465 from rspec/deprecate-Mock-class

    myronmarston committed Nov 11, 2013
    Deprecate RSpec::Mocks::Mock in favor of ...::Double. (for 2.99)
  2. @myronmarston
Commits on Nov 7, 2013
  1. @myronmarston

    Release 2.99.0.beta1

    myronmarston committed Nov 7, 2013
  2. @myronmarston
  3. @myronmarston
Commits on Nov 6, 2013
  1. @myronmarston

    Merge pull request #453 from rspec/add-symbol-vs-string-specs

    myronmarston committed Nov 5, 2013
    Add symbol vs string specs
Commits on Nov 5, 2013
  1. @myronmarston
  2. @myronmarston
  3. @myronmarston

    Add specs for string vs symbol usage.

    myronmarston committed Nov 5, 2013
    Each of these specs covers a case that was previously
    uncovered. We use `to_sym` in a bunch of places to
    normalize message names to a symbol. This is important,
    but when I removed the `to_sym` calls, nothing was failing.
    
    Without the `to_sym` calls, these specs fail, and they also
    cause subsequent specs to fail because `reset` fails to
    reset (since it winds up with two message doubles, `:foo`
    and "foo", and resets the first and blows up on the second
    since it already got reset -- that leaves the method double
    around and it is attempted to be reset on reset for
    subsequent specs).
  4. @myronmarston

    Merge pull request #451 from rspec/2-99-changelog

    myronmarston committed Nov 4, 2013
    Consistent formatting of changelog.
  5. @xaviershay
Commits on Oct 24, 2013
  1. @JonRowe

    Merge pull request #442 from rspec/use_git_repos_env

    JonRowe committed Oct 23, 2013
    Allow suppression of local repos
Commits on Oct 16, 2013
  1. @myronmarston

    Add release notes for 2.14.4.

    myronmarston committed Oct 15, 2013
    [ci skip]
Commits on Oct 7, 2013
  1. @JonRowe

    changelog for #428

    JonRowe committed Oct 7, 2013
  2. @JonRowe

    Merge pull request #428 from rspec/handle_receive_do_end_block_chaining

    JonRowe committed Oct 7, 2013
    Allow chaining receive modifications off do..end blocks
Commits on Oct 5, 2013
  1. @myronmarston

    Merge pull request #425 from rspec/issue-377-deprecation

    myronmarston committed Oct 5, 2013
    Deprecate using a with block with arguments.
  2. @xaviershay

    Deprecate using a with block with arguments.

    xaviershay committed Oct 5, 2013
    It is confusing to have different behaviour depending on whether the
    block has arguments or not, and there are better ways to do this.
Commits on Aug 27, 2013
  1. @myronmarston

    Merge pull request #414 from rspec/address_warnings

    myronmarston committed Aug 26, 2013
    Address warnings
Commits on Aug 26, 2013
  1. @myronmarston

    On 1.8 a block doesn't indicate if it ignores args.

    myronmarston committed Aug 26, 2013
    Proc.new { }.arity # => -1
    Proc.new { |*a| }.arity # => -1
    
    The first one returns 0 on 1.9, allowing us to differentiate
    these two cases, but returns -1 on 1.8 (as shown above).
    Thus, we can't differentiate here and we have to err on
    the side of printing the warning.
  2. @myronmarston
  3. @myronmarston
  4. @myronmarston

    Ensure the line in the warning is the block's source location.

    myronmarston committed Aug 24, 2013
    Note: on 1.8.7, we can't always make that work because
    Proc does not have `source_location`, and even if we
    define `source_location` on it manually, it loses that
    when we forward it on using `&block`.
  5. @myronmarston

    Fix source line in any_instance deprecation message.

    myronmarston committed Aug 23, 2013
    Before, the printed line was from where a particular
    instance got its `any_instance` recordings played back,
    not from where the `any_instance` block originated.
    
    This was very confusing :(.
  6. @myronmarston
  7. @myronmarston
  8. @myronmarston
  9. @myronmarston
  10. @myronmarston

    Update caller filter.

    myronmarston committed Aug 25, 2013
    - Include additional libs.
    - Don't put it at `rspec/caller_filter`.
    - Ignore the "same content as rspec-core" spec for now.
    
    The way I had this was brittle and is causing problems in rspec-core.
    In rspec-core I'm trying to update the regex and add some specs but
    the fact that the file is at the same path in all 3 libs (which seemed
    like a simple way to ensure only one of them loads) is causing
    the _wrong_ version to be loaded in rspec-core's specs, so that it's not
    picking up the changes I've made.  Putting the file at the same
    path in all 3 libs was a terrible idea in retrospect :(.
  11. @myronmarston

    Improve travis build.

    myronmarston committed Aug 23, 2013
    - No need to install bundler twice.
    - No need to bundle install twice.
    - Skip one-by-one specs on JRuby.
    - Bundler issue #2383 has been resolved,
      no need to work around it anymore.
  12. @myronmarston

    Pass __FILE__ and __LINE__ to `class_eval` properly.

    myronmarston committed Aug 25, 2013
    This is important because w/o this, lines in the
    eval'd code do not indicate what file or line
    they are from when they are in a backtrace, and
    that can cause CallerFilter to improperly consider
    these lines to be non-rspec lines.
Something went wrong with that request. Please try again.