1. 04 Feb, 2016 1 commit
    • David Gobbi's avatar
      Fix a memory overrun in vtkImageStencilIterator. · 2a652eb4
      David Gobbi authored
      If the stencil extent was smaller than the image extent, then the code
      incremented the span list pointer upon reaching the first voxel within
      the stencil extent, even though the pointer was already at the correct
      location for the first voxel in the stencil extent.  The extra increment
      could lead to a buffer overrun and unpredicable behavior resulting in
      random crashes.
      2a652eb4
  2. 25 Oct, 2015 1 commit
  3. 11 Aug, 2015 1 commit
  4. 30 Jun, 2015 2 commits
    • Dan Lipsa's avatar
      ENH: TestChangeInformation does not print error message. · 6fad8fe7
      Dan Lipsa authored
      vtkImageChangeInformation has both its inputs set using SetInputConnection
      and SetInformationInputData. However, the first time the pipeline is run
      the data object for the second connection is empty.
      The error messages were:
      ERROR: In /home/danlipsa/src/VTK/Common/ExecutionModel/vtkStreamingDemandDrivenPipeline.cxx, line 857
      vtkStreamingDemandDrivenPipeline (0x2817f00): The update extent specified in the information for output port 0 on algorithm vtkTrivialProducer(0x2817a50) is 0 255 0 255 0 0, which is outside the whole extent 0 -1 0 -1 0 -1.
      
      ERROR: In /home/danlipsa/src/VTK/Common/ExecutionModel/vtkTrivialProducer.cxx, line 264
      vtkTrivialProducer (0x2817a50): This data object does not contain the requested extent.
      6fad8fe7
    • David Gobbi's avatar
      Fix the median computation in vtkImageMedian3D. · 85f46e79
      David Gobbi authored
      If the number of voxels in the neighborhood is even (which can happen
      if an odd neighborhood is clipped to even near the image boundary),
      then the median should be the average of the two middle values.
      85f46e79
  5. 29 Jun, 2015 1 commit
  6. 18 Jun, 2015 1 commit
    • Dan Lipsa's avatar
      BUG: Fix test volume of interest to match data extents · cdd2e19a
      Dan Lipsa authored
      This fixes the following message issued by this test
      
      ERROR: In vtkStreamingDemandDrivenPipeline.cxx, line 857
      vtkCompositeDataPipeline (0x21589b0): The update extent specified in the information for output port 0 on algorithm vtkVolume16Reader(0x294b7b0) is 32 32 0 63 0 93, which is outside the whole extent 0 63 0 63 0 92.
      cdd2e19a
  7. 14 Jun, 2015 1 commit
  8. 11 Jun, 2015 1 commit
  9. 09 Jun, 2015 1 commit
  10. 08 Jun, 2015 1 commit
  11. 01 Jun, 2015 1 commit
  12. 21 May, 2015 1 commit
    • Andras Lasso's avatar
      BUG: Fixed vtkImageReslice HitInputExtent update · d307b831
      Andras Lasso authored
      If this->ResliceTransform was not a vtkHomogeneousTransform then the
      whole input extent was updated and then the method immediately returned.
      However, the this->HitInputExtent flag was not updated. If the
      this->HitInputExtent had been 0 before then the output of the filter
      remained always 0 and the output of the filter was always empty.
      
      Discussed with David Gobbi and he confirmed that this is a suitable fix.
      d307b831
  13. 08 May, 2015 1 commit
  14. 02 Apr, 2015 1 commit
    • Sean McBride's avatar
      Fix -Wimplicit-fallthrough warnings; introduce VTK_FALLTHROUGH · 3cc4d9fc
      Sean McBride authored
      Fixed all clang -Wimplicit-fallthrough warnings:
      - Some fixed by adding missing breaks
      - Other fixed by adding VTK_FALLTHROUGH statement
      
      Under (modern) clang with C++11:
       #define VTK_FALLTHROUGH [[clang::fallthrough]]
      
      Otherwise:
       #define VTK_FALLTHROUGH ((void)0)
      
      VTK_FALLTHROUGH serves two purposes:
       - under clang, suppresses -Wimplicit-fallthrough warnings
       - makes code more self-documenting
      3cc4d9fc
  15. 01 Apr, 2015 1 commit
  16. 26 Mar, 2015 4 commits
  17. 24 Mar, 2015 2 commits
  18. 23 Mar, 2015 1 commit
  19. 13 Mar, 2015 3 commits
    • Bill Lorensen's avatar
      COMP: Member template function use on Mac · bc10e204
      Bill Lorensen authored
      Some Mac compilers fail to compile member function calls that contain
      the redundant
      vtkCheckerboardSplatterAlgorithm<TPoints,TScalars>::. Removing the
      redundant code resolves the compilation issue on the Macs and still
      build fine on other platforms.
      
      Change-Id: I42376c5127f9465c8895fc67673227ad76a923a4
      bc10e204
    • Will Schroeder's avatar
      Using typename instead of class for templates · a6237419
      Will Schroeder authored
      Change-Id: I86108f2df3acf52f53723120c5de81084756a870
      a6237419
    • Will Schroeder's avatar
      New Guassian splatter class, added multithreading · 2be42614
      Will Schroeder authored
      vtkCheckerboardSplatter is a much faster implementation of
      vtkGaussianSplatter. It also makes use of multithreading through
      vtkSMPTools. vtkGaussianSplatter was also sped up by using these
      same multithreading tools.
      
      Change-Id: If8b987231b973e7005bba563af1c6933d61f0bb1
      2be42614
  20. 06 Mar, 2015 1 commit
    • Bill Lorensen's avatar
      COMP: Member template function use on Mac · d59d4392
      Bill Lorensen authored
      Some Mac compilers fail to compile
      vtkSampleFunctionAlgorithm<T>::FunctionValueOp<T> values(&algo);
      removing (the redundant) vtkSampleFunctionAlgorithm<T>::
      resolves the compilation issues.
      
      Change-Id: I685b14f929dce4b5b976aeefaf94b75c9f98217f
      d59d4392
  21. 04 Mar, 2015 1 commit
  22. 20 Jan, 2015 1 commit
  23. 16 Jan, 2015 1 commit
  24. 15 Jan, 2015 1 commit
    • Shawn Waldon's avatar
      Made vtkGaussianSplatter handle composite data input · 12895003
      Shawn Waldon authored
      This makes the gaussian splatter filter handle composite input by splatting
      all of the parts/blocks into one volume instead of letting the composite
      data pipeline splat each block into a separate volume.
      
      Change-Id: I528945cbe83c1832f27e104bcd5de5904d09a28d
      12895003
  25. 14 Jan, 2015 2 commits
  26. 09 Jan, 2015 1 commit
    • Shawn Waldon's avatar
      Prefer vtkDoubleArray::GetValue to GetTuple for single component arrays · 9dac67fc
      Shawn Waldon authored
      On a single component vtkDoubleArray, GetValue is significantly
      faster than GetTuple and SetValue is faster than SetTuple.  These
      changes result in about a 3% performance increase to the Gaussian
      splatter filter since it is called in a deeply nested loop.
      
      Change-Id: I88f6791667d57088e69a89e18a515196ba674af9
      9dac67fc
  27. 18 Dec, 2014 2 commits
  28. 11 Dec, 2014 4 commits
    • David Gobbi's avatar
      Fix error when kernel size is greater than slab thickness. · 8749c2fe
      David Gobbi authored
      If the kernel size was larger than the number of slices in the slab
      that was being interpolated, then there was a potential for incorrect
      results under certain circumstances.
      
      Change-Id: I832f34fc1d4750d494d996e825cf1b2708a9d3dd
      8749c2fe
    • David C. Lonie's avatar
      Fix vtkPExtractVOI extent issues. · 79ada2aa
      David C. Lonie authored
      Update the filter so that extents are properly managed in parallel.
      
      Change-Id: I2991f427a33b740b6679a469b2844f6b44b89ff5
      79ada2aa
    • Sean McBride's avatar
      Unified mixed unsigned char/int usage to bool · b54bce1d
      Sean McBride authored
      Fixes -Wconversion warnings that pointed out the mixed use of
      unsigned char and int in these APIs.
      
      Switched to bool since the comments describe it as a 'flag'
      
      This is an API change, but shouldn't cause much trouble I hope.
      
      Change-Id: I0ded5ea638c646f9e5c7aa792fe93784cb9025b4
      b54bce1d
    • David C. Lonie's avatar
      Make a clear distinction between extent indices and values. · 275bb6cd
      David C. Lonie authored
      The vtkExtractStructuredGridHelper::GetMapping method was being used
      to compute both actual extent values and indices (e.g. for rectilinear
      grid coordinate arrays) in places. The method has been split into:
      
      GetMappedIndex,
      GetMappedIndexFromExtentValue,
      GetMappedExtentValue, and
      GetMappedExtentValueFromIndex,
      
      all of which map an output value to an input value.
      
      The class also caches the initialization parameters to avoid rebuilding
      the map if the parameters do not change. This is done for the parallel
      filters, which use different extents for information vs. data requests.
      
      Change-Id: Ie26a0149c93e10e11003eb6daf7ef4d958f305a7
      275bb6cd