Better error handling for test failures

[drewcrawford]

When I get a test failure I get the following output:

$ bash run-sonar.sh 
Running run-sonar.sh...
Extracting Xcode project information.
..
Running tests using xctool......open .ERROR - Command 'xctool -project caffeine-ios.xcodeproj -sdk iphonesimulator ARCHS=i386 VALID_ARCHS=i386 CURRENT_ARCH=i386 ONLY_ACTIVE_ARCH=NO -scheme caffeine-ios -reporter junit GCC_GENERATE_TEST_COVERAGE_FILES=YES GCC_INSTRUMENT_PROGRAM_FLOW_ARCS=YES test' failed with error code: 

So some problems with this output:

1. It seems that it was intended that an error code be reported in this case, but one isn't
2. It seems like the right decision that analysis should continue anyway in the event of a test failure.

[cyrilpicat]

thanks for your feedback.

Your analysis seems right, I look at it as soon as possible, it should be easy to fix.