Permalink
Commits on Sep 6, 2016
  1. Merge pull request #396 from pharada/missing-schemes

    akitchen committed Sep 6, 2016
    Add Xcode schemes for static library and framework
  2. Add Xcode schemes for static library and framework

    pharada committed Sep 6, 2016
    xcodebuild requires specifying a scheme when building into a
    user-specified -derivedDataPath. This is common in CI workflows.
    
    Xcode autocreates missing schemes per project settings. One may wish to
    build Cedar as a static library or framework without having to launch
    Xcode in order to create those schemes.
Commits on Jul 28, 2016
  1. Merge pull request #395 from tjarratt/raise-errors-when-dsl-is-misused

    tjarratt committed Jul 28, 2016
    Raise exceptions when dsl is misused
  2. Raise when it, describe, context are called before/after tests run

    tjarratt committed Jul 27, 2016
    These case statements in this switch existed but were never exercised.
    At a minimum, we should have some exceptions raised here since we
    never, ever, ever expect users to hit these cases.
    
    Open to suggestions about how we might verify this behavior
    (without adding a separate suite).
  3. Change method signature of CDRTypeUtilities

    tjarratt committed Jul 27, 2016
    Since we really just want to receive some opaque bytes with a
    suggested encoding, we should make this take a const void *, since
    it does not intend to mutate the value and is just reading it as
    arbitrary data.
  4. Fix type utilities spec to use correct-sized pointers

    tjarratt committed Jul 25, 2016
    Warning: I'm only about 71.1% sure of what I'm saying here, so
    take this with a grain of salt.
    
    I don't think these tests ever passed for the correct reason. The
    reason I found this out was because I made some subtle, completely
    unrelated changes in the previous commit that increased the size of
    the Cedar framework. As a result, a single test here failed, which
    caused me to start thinking about pointers to "const char *" and
    what it means to be a pointer to NULL.
    
    Previously, we would take our "const char *" argBuffer, dereference
    it and ask if the result was NULL. This was find and dandy for most
    things, but when we're actually dealing with a "const char **" the
    reality is ... more complicated.
    
    The following line is always true so long as the first byte of argBuffer,
    when interpreted as a "const char *" is not zero. Given that it points
    at a real "const char *" in memory, this seems unlikely, but it's
    fairly easy to replicate.
    
    BOOL isNotNull = *argBuffer != '\0';
    
    At first I thought perhaps these pointers were different sizes, but
    in fact, both a "const char *" and a "const char **" are 8 bytes on OS X.
    Based on this, I believe that the previous code was just checking to see
    if the first byte was zero -- which is actually more likely than you
    might think!
  5. Raise exceptions on misuse of test DSL

    tjarratt committed Jul 25, 2016
    This fixes #393
    
    Shout out to @briancroom for
    
    1. finding this bug
    2. not fixing it so that someone else could have the glory
    
    We'll miss you, Brian.
Commits on Apr 26, 2016
Commits on Apr 25, 2016
  1. Bump version to 1.0

    idoru committed Apr 25, 2016
Commits on Apr 15, 2016
  1. Only call +before/afterEach on CDRHooks conformers (#387)

    idoru committed Apr 15, 2016
    * Only call +before/afterEach on CDRHooks conformers
    
    Rebased from and added more tests from #365
Commits on Apr 12, 2016
  1. Merge pull request #386 from pivotal/remove-xcode-plugin

    idoru committed Apr 12, 2016
    Removed Cedar Plugin
  2. Removed Cedar Plugin

    Brian Croom and Sam Coward committed Apr 12, 2016
Commits on Apr 1, 2016
  1. Bump version to 0.13.1

    Brian Croom and Sam Coward committed Apr 1, 2016
  2. Merge pull request #384 from pivotal/swizzle-for-xcode73

    idoru committed Apr 1, 2016
    New attempt at fixing test output under Xcode 7.3 while not breaking …
  3. New attempt at fixing test output under Xcode 7.3 while not breaking …

    Brian Croom and Sam Coward committed Apr 1, 2016
    …xctool
    
    - The dispatch_async approach causes Cedar to not be hooked into XCTest
      in time for the xctool runner to pick it up.
    - Instead, we are now swizzling -[XCTestObservationCenter addTestObserver:]
      and -_addLegacyTestObserver: and using them as an earlier hook to register
      Cedar's test observer with the test observation center.
    
    Some extra context around this can be found at:
    #383
  4. Merge pull request #382 from pivotal/delay-xctestobservationcenter-setup

    idoru committed Apr 1, 2016
    Defer test observation center setup to fix Xcode 7.3 regression
Commits on Mar 31, 2016
  1. Defer test observation center setup to fix Xcode 7.3 regression

    Brian Croom and Sam Coward committed Mar 31, 2016
  2. Only import custom matchers when building as Objective-C++

    Brian Croom and Sam Coward committed Mar 31, 2016
Commits on Mar 23, 2016
  1. Bump version to 0.13.0

    idoru committed Mar 23, 2016
Commits on Mar 22, 2016
  1. Merge pull request #371 from briancroom/print_pending_tests

    idoru committed Mar 22, 2016
    Pending tests should be printed when running under XCTest
Commits on Mar 6, 2016
  1. Remove explicit `__weak` attribute on variables in MRC files

    briancroom committed Mar 6, 2016
    This was never actually respected by the compiler, and with Xcode 7.3 is
    now a compiler error. Note that using objc_loadWeak and objc_storeWeak
    with these variables still seems to function as expected.
Commits on Feb 12, 2016
  1. Merge pull request #375 from ReadmeCritic/master

    briancroom committed Feb 12, 2016
    Correct the spelling of CocoaPods in README
Commits on Feb 11, 2016
Commits on Feb 1, 2016
  1. Merge pull request #374 from wileykestner/master

    tjarratt committed Feb 1, 2016
    Expose failing test in CI and then fix the failing test
  2. Merge pull request #372 from pra85/patch-1

    tjarratt committed Feb 1, 2016
    Update copyright year to 2016
Commits on Jan 27, 2016
  1. Set optimization level build setting to 'None -O0' in Release configu…

    reneighbor committed with wileykestner Jan 27, 2016
    …ration
    
    - Fixes a failing test for reporting the line numbers of call sites in exception stack traces
    - In the future, we probably want to add a story display correct line numbers for stack traces in apps that were built in Release congfiguration
    
    Signed-off-by: Wiley Kestner <wkestner@pivotal.io>
  2. Change `rake ci` task to detect previously undetected test failures

    reneighbor committed with pivotal Jan 27, 2016
    - Each test run now ouputs its console log to its own individual log file (in the 'build' folder
    - Fail the `rake ci` task if we find any failures in any of the individual log files
    - Fail the `rake ci` task if we do not find the string '0 failures' in any of the log files (for instance if the test runner crashes)
    
    Signed-off-by: Wiley Kestner <wkestner@pivotal.io>
Commits on Jan 26, 2016
  1. Update copyright year to 2016

    pra85 committed Jan 26, 2016
Commits on Jan 24, 2016
  1. Don't ignore pending tests when executing specs under XCTest

    briancroom committed Jan 24, 2016
    This means that the pending tests, although not executed, will still be
    given to the reporter, so the pending test summary info can be printed.
    Note that the pending tests aren't exposed to the XCTest runner at all.
Commits on Jan 23, 2016
  1. Project file maintenance.

    briancroom committed Jan 23, 2016
Commits on Jan 15, 2016
  1. Merge pull request #357 from pivotal/kvc_improvements

    tjarratt committed Jan 15, 2016
    Implement smarter KVC handling for class fakes
  2. Merge pull request #367 from briancroom/swift_module

    tjarratt committed Jan 15, 2016
    Swift support for the Cedar test runner