Pushed new branch T18922-OpenGLBackend2-Integration.

We created a test branch to investigate the impact of removing the unnecessary ContourModel recreations. A PaintBrushInteractionTool test was added and investigated with valgrind callgrind.

An additional problem is the recreation of the contour model for every tiny mouse movement, although a simple position shift would be sufficient.

Replacing SetData() by using the ITK Modified flag wasn't successful, because with every new ContourModel a new pointer is used. The data in the FeedbackNode is therefore not adressed by the modification.

I think this is the case when software components with different internal floating point representations are used.

The actual problem, as stated in the thread, is because of inconvenient smart pointer usage in the data manager view.

[625b57]: Merge branch 'bug-18672-DoubleDeleteGuardInDataManagerViewCherryPick'

Merged commits:

2015-04-01 16:33:43 Joseph Görres [0880f4]
introduced smart pointer in datamanagerview to prevent recursive double deletion of data nodes

[d9389f]: Merge branch 'bug-18059-InteractionTestingHowTo'

Merged commits:

2014-09-03 17:01:45 Joseph Görres [2b1851]
finish doxygen

Seems to be a Win 8 Problem. We couldn't reproduce the described behavior with Win 7 64 Bit and Ubuntu 12.04 LTS 64 Bit.

[88a56a]: Merge branch 'bug-17750-DoubleScalarAllocation'

Merged commits:

2014-04-23 17:00:53 Joseph Görres [6d7005]
removed double scalar allocation

[f57625]: Merge branch 'bug-17673-clippingplaneserializer'

Merged commits:

2014-05-14 17:06:13 Eric Heim [f052f1]
using mitkSurface instead of mitkPlane to make the saving of scenefiles with clipping planes work

Crashes on Ubuntu 12.04. 64 bit

Moving the method content from PartClosed(...) to Deactivated() in QmitkImageStatisticsView solves the problem.

[c696b7]: Merge branch 'bug-17665-ImageStatisticsNavigationObserver'

Merged commits:

2014-04-09 17:46:23 Joseph Görres [bcd0d3]
moved the insertion and removal of the navigation observer to the visible- and hidden- slots

[55ee91]: Merge branch 'bug-17666-IgnoreZeroValuedVoxelsWithUnsignedShort'

Merged commits:

2014-05-14 16:44:43 Joseph Görres [db94f1]
fixed wrong cast for images with pixel type unsigned short

Sorry, Deactivated() is inappropriate because it is triggered at focus change.

We cannot reproduce the described behavior. Could you please give it another try and provide more information.

[1a858a]: Merge branch 'bug-17650-ImageMaskingQtEventPressed'

Merged commits:

2014-05-21 16:11:06 Joseph Görres [09d0ea]
replacedPressedWithClicked

wrong qt event treatment.

[1e0924]: Merge branch 'bug-17505-DisableCopyConstructorAndAssignmentOperator'

Merged commits:

2014-03-26 16:44:39 Joseph Görres [0a4afc]
disabled copy constructor and operator= in pixel- and non-pixel- as well as read- and write- accessors

(In reply to Ralf Floca from comment #5)

So for (1) I can only see the avoidance of undefined states but not the
introducing of new risks.

(In reply to Joseph Görres from comment #2)

As designated in our concept, the copy constructor and operator=() are
permitted.

Thank you for your helpful contribution.

New remote branch pushed: bug-16752-AddFlagToIgnoreLockMechanism

[efeb7b]: Merge branch 'bug-16752-AddFlagToIgnoreLockMechanism'

Merged commits:

2014-01-29 16:12:24 Joseph Görres [54ca9b]
indentation

[de8570]: Merge branch 'bug-16615-SpinboxFastMarching'

Merged commits:

2013-12-04 16:27:40 Joseph Görres [c63c52]
fixed strange spinbox behavior

New remote branch pushed: bug-16615-SpinboxFastMarching

GDCM was updated to solve the frame increment pointer issue.

1232c1232

< if( at.GetValue() == Tag(0x0018,0x1063) && at.GetNumberOfValues() == 1 )

if( at.GetValue() == Tag(0x0018,0x1063) && at.GetNumberOfValues() == 1 && ds.FindDataElement(Tag(0x0018,0x1063)) )

We have a fix for the gdcmImageHelper.cxx. We ignore the FrameIncrementPointer, if it points to a non-existent tag (0x0018,0x1063). This still results in warnings, but C-Arm CT images can be loaded again in MITK.

The z-spacing of our image data is stored at (0x0018,0x0050) "SliceThickness"

Setting m_3DInterpolationEnabled(false) in the constructor of mitkSegTool2D.cpp solves the problem. However, then the 3D interpolation doesn't work anymore.

Deselecting the segmentation invokes a call of the OnWorkingNodeVisibilityChanged() method in the QmitkSegmentationView, although the view is disabled. Is a call of this method really desired in this case?

New remote branch pushed: bug-16107-QmitkSegmentationDeselectionWorkaround

[52fce9]: Merge branch 'bug-16098-WatershedBusyCursorIssues'

Merged commits:

2013-09-18 16:29:37 Joseph Görres [8f1c72]
completely remove 2D icon

New remote branch pushed: bug-16098-WatershedBusyCursorIssues

The log system makes use of the FILE macro, which depends on the preprocessor parametrization. It therefore might be possible to adapt the build system to work with relative paths.

New remote branch pushed: bug-16018-RemoveOtsuEntry

New remote branch pushed: bug-16018-OtsuActionDeprecation

[cfbb23]: Merge branch 'bug-16018-OtsuActionDeprecation'

Merged commits:

2013-09-18 13:42:03 Joseph Görres [c90372]
made otsu action deprecated and removed it from DataManager

[40b89e]: Merge branch 'bug-15638-ReplaceDeprecatedCastToIpPicCasts'

Merged commits:

2013-07-17 15:36:10 Joseph Görres [a13585]
replaced CastToIpPicDescriptor in all modules except Segmentation

[199a0c]: Merge branch 'bug-15507-RemoveDeprecatedImageCallsInOverwriteSliceTest

Merged commits:

2014-01-22 16:38:56 Joseph Görres [5f4ace]
Merge branch 'bug-15645-DeprecatedMitkImageAPICallsInMitkOverwriteSlicerFilterTest' into bug-15507-RemoveDeprecatedImageCallsInOverwriteSliceTests

New remote branch pushed: bug-15507-RemoveDeprecatedImageCallsInOverwriteSliceTests

[14d98d]: Merge branch 'bug-15507-RemoveLastDeprecatedWorldCoordinatePixelAccess

Merged commits:

2014-01-23 16:44:48 Joseph Görres [f6ce53]
COMP: deal with wrong transformation behavior on MAC

[922b85]: Merge branch 'bug-15507-RemoveLastDeprecatedWorldCoordinatePixelAccess

Merged commits:

2014-01-15 17:03:50 Joseph Görres [209fb0]
remove last deprecated occurrences of GetPixelValueByWorldCoordinate

[8cd17b]: Merge branch 'bug-15495-AddSupportForVectorImages'

Merged commits:

2015-08-26 17:05:51 Joseph Görres [6e0b44]
COMP: add missing bracket

[8a4282]: Merge branch 'bug-15495-RemoveLastOccurrences-Integration'

Merged commits:

2015-07-23 15:04:24 Joseph Görres [e9683a]
COMP: remove undefined variable condition

[2eba71]: Merge branch 'bug-15495-RemoveLastOccurrences-Integration'

Merged commits:

2015-07-23 14:47:14 Joseph Görres [0288bc]
COMP: fix improper accessor integration in diffusionimaging

[1761d8]: Merge branch 'bug-15495-AddSupportForVectorImages'

Merged commits:

2015-08-26 16:49:33 Joseph Görres [ff9b64]
Added support for itk VectorImages in PixelReadAccessor

[c6c91a]: Merge branch 'bug-15495-RemoveLastOccurrences-Integration'

Merged commits:

2015-07-22 15:09:01 Joseph Görres [6eb109]
removed last occurrences of GetPixelValueByIndex

New remote branch pushed: bug-15493-RemoveLastGetDatasFromMITK

New remote branch pushed: bug-15495-DeprecatedImageAccessInQmitkStdMultiWidget

[29bf00]: Merge branch 'bug-15493-RemoveLastGetDatasFromMITK'

Merged commits:

2014-01-08 14:50:15 Joseph Görres [e5a5e4]
final changes to complete the removal of all image::GetData() in MITK

New remote branch pushed: bug-15493-deprecated-getdata-imageextractor

Is related to T15638. We have to decide, if we completely want to get rid of all calls (this bug) or if we want to replace them for image accessor compatibility (T15638).

It doesn't crash with the current master (d523a10a093e9dccc66b2904b76dd6e9a4cea9f1) + your patch on my computer (Ubuntu 12.04 LTS, Python 2.7.3)

Thank you for providing a patch. It works very well and will be integrated once T15389 is fixed.

A simple test of investigating consecutive calls is not sufficient to reproduce the behavior. The actual problem can however be easily regarded in the QmitkSlicesInterpolator.cpp in Line 539:

It works if you put
"slice->GetVolumeData()->ConstructVtkImageData(slice);"
in front of the problematic code line.

We couldn't reproduce the behavior with the test provided in branch bug-15325-ExtendImageDataItemTest. Since the actual problem doesn't occur anymore in the QmitkSlicesInterpolator, we assume there was a problem with image initalisation/deallocation when an interpolated slice was confirmed.

New remote branch pushed: bug-15325-ExtendImageDataItemTest

New remote branch pushed: bug-15174-AddingWatershedSegmentation

New remote branch pushed: bug-15174-UpdateWatershedPlugin

mitk::SegmentationControllerInterpolator.cpp line 445

[81703c]: Merge branch 'bug-14882-RepairTutorialStep3'

Merged commits:

2013-08-28 15:57:53 Joseph Görres [f604df]
fixed histogram size in itkMITKScalarImageToHistogramGenerator