[Insight-developers] ICE on vs6.0

[Bradley Lowekamp]

Bill,

The future means before the 3.18 release. I believe it would be a mistake to release this methods in it's current implementation, as it is just a skeleton. 

The primary goal of this method was to aid in the support for 64-bit windows. With the numerous types used with IO, and the sizes of images, this function is needed to detect overflow errors to prevent crashes. Support for "long long" was also added to  the numeric traits, so that we'd have numeric traits for the Image::SizeType and size_t on windows 64.

It is my understanding that Luis believes adding support for win64 is a high priority. As such I interpret that to mean that the 3.19 development effort will have some focus on making this possible. This methods is a useful tool for that work.

I assume that is just the result of some bug is VS6. Could be as simple to fix as just changing the order of the header files in Math.h? Other files seem to include itkConceptChecking with out issues.


Brad


On Apr 1, 2010, at 1:07 PM, Bill Lorensen wrote:

> If the future means itk 4.0, then can I remove it for now? I assume
> that VS6 will not be supported in itk 4.0.
> 
> Bill
> 
> On Thu, Apr 1, 2010 at 1:04 PM, Bradley Lowekamp <blowekamp at mail.nih.gov> wrote:
>> Darn,
>> I missed that change! Thanks for finding it. The planned implementation of
>> CastWithRangeCheck is:
>> 
>> template <typename TReturn,typename TInput>
>> inline TReturn CastWithRangeCheck(TInput x)
>> {
>>   itkConceptMacro( OnlyDefinedForIntegerTypes1,
>> (itk::Concept::IsInteger<TReturn>) );
>>   itkConceptMacro( OnlyDefinedForIntegerTypes2,
>> (itk::Concept::IsInteger<TInput>) );
>>   TReturn ret = static_cast<TReturn>(x);
>>   if ( sizeof (TReturn) > sizeof(TInput) &&
>>        !( !NumericTraits<TReturn>::is_signed &&
>>  NumericTraits<TInput>::is_signed ) )
>>     {
>>     // if the output type is bigger and we are not converting a signed
>>     // interger to an unsigned interger then we have no problems
>>     return ret;
>>     }
>>   else if ( ret != x ||
>>        ( NumericTraits<TInput>::IsPositive(x) !=
>> NumericTraits<TReturn>::IsPositive(ret) ) )
>>     {
>>     // this should likely be a new exception class
>>     RangeError _e(__FILE__, __LINE__);
>>     throw _e;
>>     }
>>   return ret;
>> }
>> So while it is not needed now, currently it looks like it will be in the
>> future.
>> The other issue with the code is the comparison "ret != x". This is correct
>> and is exactly as it needs to be, as it relies on the default promotion of
>> types, and default behavior for comparing signed and unsigned integer types.
>> Unfortunately it generates a warning, and there is not simple explicit
>> casting which will follow the default promotion rules.
>> Brad
>> On Apr 1, 2010, at 12:11 PM, Bill Lorensen wrote:
>> 
>> Brad,
>> 
>> Actually, you also added
>> #include "itkConceptChecking.h"
>> to
>> itkMath.h
>> 
>> If I remove that, the ICE goes away.
>> 
>> Why was that include added? Can I remove it?
>> 
>> Bill
>> 
>> On Wed, Mar 31, 2010 at 8:33 AM, Bradley Lowekamp
>> <blowekamp at mail.nih.gov> wrote:
>> 
>> Hello,
>> 
>> Yesterday I committed these innocent changes to comments:
>> 
>> http://www.cdash.org/CDash/viewUpdate.php?buildid=574645
>> 
>> Then both the continuous and the night vs60 build generated ICEs:
>> 
>> http://www.cdash.org/CDash/viewBuildError.php?buildid=574645
>> 
>> I have no clue on this one!
>> 
>> Thanks,
>> 
>> Brad
>> 
>> ========================================================
>> 
>> Bradley Lowekamp
>> 
>> Lockheed Martin Contractor for
>> 
>> Office of High Performance Computing and Communications
>> 
>> National Library of Medicine
>> 
>> blowekamp at mail.nih.gov
>> 
>> 
>> _______________________________________________
>> 
>> Powered by www.kitware.com
>> 
>> Visit other Kitware open-source projects at
>> 
>> http://www.kitware.com/opensource/opensource.html
>> 
>> Kitware offers ITK Training Courses, for more information visit:
>> 
>> http://kitware.com/products/protraining.html
>> 
>> Please keep messages on-topic and check the ITK FAQ at:
>> 
>> http://www.itk.org/Wiki/ITK_FAQ
>> 
>> Follow this link to subscribe/unsubscribe:
>> 
>> http://www.itk.org/mailman/listinfo/insight-developers
>> 
>> 
>> 
>> ========================================================
>> 
>> Bradley Lowekamp
>> 
>> Lockheed Martin Contractor for
>> 
>> Office of High Performance Computing and Communications
>> 
>> National Library of Medicine
>> 
>> blowekamp at mail.nih.gov
>> 
>> 

========================================================
Bradley Lowekamp  
Lockheed Martin Contractor for
Office of High Performance Computing and Communications
National Library of Medicine 
blowekamp at mail.nih.gov


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/mailman/private/insight-developers/attachments/20100401/ff8d28d4/attachment.htm>