[Insight-users] possible mismatched free()/delete/delete[] ?

[Luis Ibanez]

Hi Feng,

This is an interesting report.

By looking at the code, there are no uses of malloc()
in ITK, except for the

     itkBalloonForceFilter.txx

and

     itkBinaryMask3DMeshSource.txx

classes.


The Watershed is using a custom allocator
passed to the itk_hash_map class, but it
uses "new".

 From your email, I understand that Valgrind only
flags this when you link with ITK shared libraries
and not when you link with ITK Static libraries.

Is that correct ?


    Thanks


       Luis


-----------------
Feng Ma wrote:

> Hi,
> 
>  I met the "Mismatched free() / delete / delete []" in valgrind output 
> of my code:
> 
>  watershed = itk::WatershedImageFilter<InternalImageType>::New();
>  watershed->SetThreshold(lowerThreshold);
>  watershed->SetLevel(maxFloodLevel);
>  watershed->SetInput(magnitude->GetOutput());
>  watershed->UpdateLargestPossibleRegion();
> 
>  The valgrind output is:
> 
> ==22333==
> ==22333== Mismatched free() / delete / delete []
> ==22333==    at 0x40027C07: __builtin_delete (vg_replace_malloc.c:233)
> ==22333==    by 0x40B8718F: 
> itk::EquivalencyTable::~EquivalencyTable(void) (in 
> /usr/lib/InsightToolkit/libITKCommon.so)
> ==22333==    by 0x40B90A0E: itk::LightObject::UnRegister(void) const (in 
> /usr/lib/InsightToolkit/libITKCommon.so)
> ==22333==    by 0x40B97B40: itk::Object::UnRegister(void) const (in 
> /usr/lib/InsightToolkit/libITKCommon.so)
> ==22333==    Address 0x469E2F14 is 0 bytes inside a block of size 92 
> alloc'd
> ==22333==    at 0x40027894: malloc (vg_replace_malloc.c:153)
> ==22333==    by 0x822466C: __builtin_new (in /home/fma/test/main)
> ==22333==    by 0x81CAC98: itk::EquivalencyTable::New(void) 
> (/usr/include/InsightToolkit/Common/itkEquivalencyTable.h:55)
> ==22333==    by 0x81B34A9: itk::watershed::Segmenter<itk::Image<float, 
> 3> >::LabelMinima(itk::SmartPointer<itk::Image<float, 3> >, 
> itk::ImageRegion<3>, itk::hash_map<unsigned long, 
> itk::watershed::Segmenter<itk::Image<float, 3> >::flat_region_t, 
> itk::hash<unsigned long>, equal_to<unsigned long>, allocator<char> > &, 
> float) 
> (/usr/include/InsightToolkit/Algorithms/itkWatershedSegmenter.txx:636)
> 
>  I know that itkNewMacro() use new operator. But somehow, it eventually 
> ends up with malloc().
> I am using ITK1.4.0. The ITK libraries were dynamic libraries.
> 
>  Interesting thing is: I have another project with same code and ITK 
> compiled as static libraries, I didn't see this error.
> 
>  Any hints?
> 
>  Thanks a lot.
> 
> -Feng
> 
> _________________________________________________________________
> FREE pop-up blocking with the new MSN Toolbar – get it now! 
> http://clk.atdmt.com/AVE/go/onm00200415ave/direct/01/
> 
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
>