[Insight-developers] RegionGrow2D Test and GCC 4.3

[Luis Ibanez]

Hi Bill,

Actually I didn't have a chance to update camelot yesterday.

The compiler version is still the same.

Camelot's gcc 4.3 build is not a clean build, which lead to
wonder if what we are seeing is the result of missing to
define a dependency.


    Luis


----------------------
Bill Lorensen wrote:
> Luis,
> 
> itkRegionGrow2DTest is now passing on both camelot.kitware and
> sumost.ca. I assume you updated camelot.kitware to gcc 4.3.1. But did
> the other machine also get updated?
> 
> Bill
> 
> On Tue, Jun 10, 2008 at 9:06 AM, Luis Ibanez <luis.ibanez at kitware.com> wrote:
> 
>>Bill,
>>
>>Yeap, good point.
>>
>>I'll update it and rebuild it.
>>
>>
>> Thanks
>>
>>
>>    Luis
>>
>>
>>----------------------
>>Bill Lorensen wrote:
>>
>>>Luis,
>>>
>>>I just checked and gcc 4.3.1 was released on June 6. I assume that
>>>camelot is running gcc 4.3.0? Maybe you can upgrade camelot's version
>>>to 4.3.1 first.
>>>
>>>Bill
>>>
>>>On Tue, Jun 10, 2008 at 7:58 AM, Bill Lorensen <bill.lorensen at gmail.com>
>>>wrote:
>>>
>>>
>>>>Luis,
>>>>
>>>>I added some additional output info to the RegionGrow2D test that is
>>>>failing on gcc 4.3 platforms.
>>>>http://www.cdash.org/CDash/testDetails.php?test=4565047&build=96250
>>>>
>>>>The test fails and not prints:
>>>>
>>>>Extracting and checking approximation image
>>>>Extracting and checking label image
>>>>Test FAILED
>>>>pixelOut5in: [50, 2425, 3175] != pixelOut5out: [50, 2425, 3175]
>>>>
>>>>The != output is what I have added. Of course, the inequality makes no
>>>>sense.
>>>>
>>>>Can you or someone with a gcc4.3 build look a little further into
>>>>this? The pixelOut5in and pixelOut5out are vectors so the operator==
>>>>should be fine. Both of the failing builds have optimization.
>>>>
>>>>Bill
>>>>
>>>
>>>
>