[Debian GNUstep maintainers] Bug#918366: Bug#918366: FTBFS for armhf on arm64, fails some tests

[Richard Frith-Macdonald]

> On 5 Jan 2019, at 15:00, Steve McIntyre <steve at einval.com> wrote:
> 
> Package: src:gnustep-base
> Version: 1.25.1-4
> Severity: important
> 
> Hi!
> 
> I've been doing a full rebuild of the Debian archive, building all
> source packages targeting armel and armhf using arm64 hardware. We are
> planning in future to move all of our 32-bit armel/armhf builds to
> using arm64 machines, so this rebuild is to identify packages that
> might have problems with this configuration.
> 
> I've tried to build gnustep-base for armhf on top of arm64, and AFAICS
> it's failing some of its tests:
> 
> ...
> Testing resources.m...
> Running base/NSBundle/resources.m...
> Start set:       resources.m:11 ... NSBundle resources
> Passed test:       resources.m:19 ... +bundleWithPath returns anNSBundle
> Passed test:       resources.m:22 ... -pathsForResourcesOfType: inDirectory: returns an array
> Passed test:       resources.m:26 ... -pathForResource:ofType:inDirectory: works with nil args
> Passed test:       resources.m:28 ... -pathForResource:ofType: works with nil type
> 2019-01-04 18:19:10.693 resources[23182:23182] [NSConstantString-stringByAppendingPathExtension:] cannot append extension 'tiff' to empty string
> Passed test:       resources.m:30 ... -pathForResource:ofType: works with nil name
> Passed test:       resources.m:32 ... -pathForResource:ofType: works with empty type
> 2019-01-04 18:19:10.694 resources[23182:23182] [NSConstantString-stringByAppendingPathExtension:] cannot append extension 'tiff' to empty string
> Passed test:       resources.m:34 ... -pathForResource:ofType: works with empty name
> Passed test:       resources.m:37 ... -resourcePath returns the correct path
> Start set:         resources.m:40 ... NSBundle GNUstep general
> Skipped set:         resources.m 45 ... it looks like GNUstep-base is not yet installed
> Error:      resources.m:91 ... END(NSBundle GNUstep resources) with START(NSBundle GNUstep general).
> End set:           resources.m:91 ... NSBundle GNUstep general
> End set:         resources.m:93 ... NSBundle resources
> Completed file:  resources.m
> ...
> Start set:       basic.m:32 ... NSArray subscripting
> Skipped set:       basic.m 41 ... No dictionary subscripting support in the compiler.
> Error:      basic.m:43 ... END(NSDictionary subscripting) with START(NSArray subscripting).
> End set:         basic.m:43 ... NSArray subscripting
> Completed file:  basic.m
> ...
> Passed test:     basic.m:74 ... setting object value on invalid node works
> Dashed hope:     basic.m:77 ... setting object value on invalid node sets string value
> Expected '1,2E1' and got '12'
> Passed test:     basic.m:83 ... setting string value on invalid node works
> ...
> 
> I'm honestly not sure about what casues the failures. Is a "dashed
> hope" a test fail? :-(

That's a test skipped because it's one that depends on the software having been installed.
A test skipped because it depends on a different compiler than trhe one being used.
And no,  a dashed hope is not a failure.  it's a testcase/feature still under development.

So there are no test failures in the log.
You can tell when test case fails because the detailed log says 'Failed'.
Also if you just look at the output from the test run (stdout), rather than the detailed log files, it gives a clear summary.

eg.
$ gnustep-tests NSDate 
--- Running tests in NSDate ---

     50 Passed tests

All OK!
$