[Insight-developers] many new test failures, divide by 0 of m_NumberOfPixelsWithDifferences

[Matt McCormick]

On Wed, Feb 29, 2012 at 2:32 PM, Sean McBride <sean at rogue-research.com> wrote:
> On Wed, 29 Feb 2012 13:56:13 -0500, Matt McCormick said:
>
>>>>Thanks for the post... I have been checking the dashboard all morning
>>>>and haven't seen this failure.
>>>
>>> They are easy to miss since they aren't in "expected", and they aren't
>>"expected" since they have never been green.  But I keep an eye on them...
>>
>>We could add it to expected if you would like -- it would be nice to
>>have a Clang build there.
>
> Up to you.  The only downside I see is that tests will likely fail for the foreseeable future, causing annoying emails to people...
>
>>But, the ExternalData fetches should be
>>fixed.  Do you know why that is happening?
>
> No.  I keep meaning to ask... how do I debug it?
>

It looks like you are build with -j15, which is possibly causing
issues with so many network connections to the same host.

If you use either the latest itk_common.cmake or manually set the
variable ExternalData_OBJECT_STORES to a directory that holds the
ExternalData, it will not need to download all the data every night.

Thanks,
Matt