Permalink
Commits on Mar 9, 2014
  1. @myronmarston
Commits on Mar 6, 2014
  1. @myronmarston

    Rename Metadata to MetadataOld.

    This is a prep step for a large scale refactoring,
    giving us a clean slate to work from.
    myronmarston committed Mar 5, 2014
  2. @myronmarston

    Merge pull request #1369 from rspec/refactor-metadata-specs

    Refactor metadata specs
    myronmarston committed Mar 6, 2014
  3. @myronmarston

    Stop calling `Metadata.new` in this spec.

    This decouples us from how the metadata is created.
    myronmarston committed Mar 6, 2014
  4. @myronmarston

    Address 1.8.7 failures.

    LocalJumpError:
           unexpected return
    myronmarston committed Mar 5, 2014
  5. @myronmarston

    Refactor metadata specs.

    Rather than directly creating the metadata instances,
    use RSpec's public `describe/it` API to create metadata
    as a user would. This gives us greater flexibility to
    refactor the metadata implementation, as our tests are
    no longer coupled to the private `process` and `for_example`
    methods.
    myronmarston committed Mar 5, 2014
  6. @myronmarston

    Fix spec so it tests what it is meant to.

    The context of the spec were identical to the
    "with a string" one above it, but this one was
    meant to be for a non-string.
    myronmarston committed Mar 5, 2014
Commits on Mar 5, 2014
  1. @myronmarston

    Merge pull request #1371 from rspec/pr-1248-merge-prep

    Pr 1248 merge prep
    myronmarston committed Mar 5, 2014
  2. @myronmarston

    Add changelog entry.

    myronmarston committed Mar 5, 2014
  3. @myronmarston

    Remove `method_missing`.

    We're trying to shrink/tighten the interface here;
    no reason to keep supporting the full hash interface.
    myronmarston committed Mar 5, 2014
  4. @myronmarston

    Remove unneeded methods.

    Having a base class implementation that raises
    `NotImplementedError` can be useful for classes
    that users are meant to subclass, but this is
    internal and it provides little value.
    myronmarston committed Mar 5, 2014
  5. @myronmarston

    Rename `include!` and `exclude!`.

    I think `include_only` and `exclude_only` are
    more clear about what the methods do.
    myronmarston committed Mar 5, 2014
  6. @myronmarston

    remove Hash inheritance in exclusion & inclusion filters

    Sergey Pchelincev committed with myronmarston Feb 27, 2014
Commits on Mar 4, 2014
  1. @myronmarston

    Merge pull request #1361 from rspec/fix-nested-described-class

    Fix nested described class
    myronmarston committed Mar 4, 2014
  2. @soulcutter

    Merge pull request #1365 from rspec/reorder-deprecation-summary

    Reorder deprecation_summary to come before dump_summary
    soulcutter committed Mar 4, 2014
  3. @soulcutter
Commits on Mar 3, 2014
  1. @soulcutter

    Reoder deprecation_summary to come before dump_summary

    deprecation_summary can oftentimes be long and the
    last thing printed gets most of the user's focus.
    It's probably better that the result of the tests
    is given primacy.
    soulcutter committed Mar 3, 2014
Commits on Mar 2, 2014
  1. @myronmarston

    Refine `described_class` inference further.

    In the last commit, I made inner `describe SomeClass`
    example groups get `SomeClass` as the described_class.
    The logic confusingly allowed a `:described_class` metadata
    key in an outer group take precedence, which meant that the
    behavior depended on whether or not `described_class` was
    referenced from the outer group.
    
    This fixes that, making it work consistently:
    
    - Use `:described_class` is present.
    - Otherwise, use the group's `describe` arg
      (unless it's something like a string).
    - If neither of those, look in the parent and
      recurse.
    myronmarston committed Mar 2, 2014
  2. @xaviershay

    Merge pull request #1360 from rspec/issue-1338

    Move private methods from RSpec to RSpec.world.
    xaviershay committed Mar 2, 2014
  3. @myronmarston

    Make `described_class` to return innermost described class.

    It used to return the outermost described class. It doesn't
    happen very often that there is more than one class being
    described in the set of nested example groups (and we
    generally recommend against it), but this was surprising
    behavior to me.
    
    Fixes #1114.
    myronmarston committed Mar 2, 2014
  4. @myronmarston

    Always use color.

    myronmarston committed Mar 2, 2014
Commits on Mar 1, 2014
  1. @xaviershay
Commits on Feb 28, 2014
  1. @myronmarston

    Add 2.14.8 release notes.

    [ci skip]
    myronmarston committed Feb 28, 2014
Commits on Feb 26, 2014
  1. @myronmarston

    Merge pull request #1353 from rspec/fix-spec-helper-formatters-issue

    Add a failing spec for #1343.
    myronmarston committed Feb 26, 2014
  2. @myronmarston

    Add changelog entries.

    [ci skip]
    myronmarston committed Feb 26, 2014
  3. @myronmarston

    Fix generated spec_helper so we don't get double formatters.

    The spec_helper generated by `rspec --init` could
    cause both the progress and documentation formatter
    to be used for a command like:
    
    rspec path/to/spec.rb --format progress
    
    This happened because formatters are loaded _after_
    `--require` files, so the `config.formatters.none?`
    check returned true even though a formatter had
    been explicitly set at the command line. It's important
    that requires come first so that custom formatters can
    be appropriately loaded.
    myronmarston committed Feb 25, 2014
Commits on Feb 25, 2014
  1. @myronmarston

    Merge pull request #1348 from rspec/benchmark-require-relative

    Benchmark require relative and update to new rspec-support method
    myronmarston committed Feb 25, 2014
  2. @JonRowe

    Merge pull request #1352 from rspec/update-travis-build-scripts-2014-…

    …02-25-for-master
    
    Updates from rspec-dev (2014-02-25)
    JonRowe committed Feb 25, 2014
  3. @myronmarston

    Remove benchmark file.

    This benchmark suggested that `require_relative` isn't
    any faster than `require`, because it put `rspec/core/lib`
    onto the load path as the first dir, which made `require`
    for rspec-core files O(1) like `require_relative`.
    
    In practice, people's load path will have many more directories
    and this benchmark doesn't align with that reality.
    
    [ci skip]
    myronmarston committed Feb 25, 2014
  4. @JonRowe

    Merge pull request #1345 from rspec/smaller-built-gem-size

    Don't include any test files in the gem.
    JonRowe committed Feb 25, 2014
  5. @JonRowe
  6. @myronmarston
Commits on Feb 24, 2014
  1. @myronmarston
  2. @myronmarston
  3. @myronmarston