| Commit message (Collapse) | Author | Age |
|
|
|
| |
current runloop of the caller.
|
|
|
|
| |
DELTA=47 (2 added, 2 deleted, 43 changed)
DELTA_BY_EXTENSION=h=42,m=3
|
|
|
|
|
|
| |
GTM_USING_XCTEST to 1 to use it.
DELTA=486 (352 added, 0 deleted, 134 changed)
DELTA_BY_EXTENSION=h=467,m=15,mm=1,txt=3
|
|
|
|
|
|
|
|
| |
Many people were using -[gtm_runUpToSixtySecondsWithContext:] which was fine
until you started to see multiple failures in a single build and it could take a
long time for a test pass to complete. 60 seconds is often much longer than
necessary to know that a failure condition has occurred.
DELTA=15 (15 added, 0 deleted, 0 changed)
|
|
|
|
|
|
|
| |
Add a reset method to the boolean run loop context since, between tests, [context reset] looks more natural than [context setShouldStop:NO] and is more concise than allocating a new context.
R=thomasvl
APPROVED=dmaclach
DELTA=5 (5 added, 0 deleted, 0 changed)
|
|
|
|
|
|
| |
Clean up some leaks found by the analysis tools.
R=mrossetti
DELTA=6 (5 added, 0 deleted, 1 changed)
|
|
|
|
|
|
|
| |
Add a api to catching when a test binary is left running too long.
Wire up said api into the testing harness used to GTMs tests.
R=dmaclach
DELTA=40 (39 added, 0 deleted, 1 changed)
|
|
refactor GTMUnitTestingUtilities into GTMAppKitUnitTestingUtilties and GTMFoundationUnitTestingUtilities.
convert GTMSignalHandler over to using new runloop routines
fix bug in GTMSignalHandler API with GC where releasing it wasn't sufficient to stop it listening.
R=thomasvl
DELTA=1227 (638 added, 566 deleted, 23 changed)
|