[IGSTK-Developers] Some recent bugs and changes

[Patrick Cheng]

Hi Kevin,

The bug 2688 is actually a problem regarding, How does one class's state 
machine respond to certain input, when this is depends on it's 
relationship to other classes. Here we should not connect the VTK 
Pileline when the representation class is not attached to the view.

Can you give us your thoughts on a higher level, how should we write the 
state machine in this case?


About the reference counting, I am not sure, how does the 
VTK_DEBUG_LEAKS flag affect the vtk libs, when we compile it under 
Release mode in visual studio 2003?

It seems that the release build of the IGSTK failed on all the vtk 
related tests.
http://public.kitware.com/IGSTK/Sites/xeon.isis/Win32-VS71-Release/20060111-0430-Nightly/Test.html
Class "vtkObjectFactoryCollection" has 1 instance still around.

Thank you.

Patrick


Kevin Gary wrote:
> Hi Patrick,
> 
> For bug 2667, I just rebuilt and I still get the reference
> counting messages for igstkTests. I no longer get it for
> igstkStateMachineExportTest, but now the export creates
> empty LTS files?! Should I open a new bug for this? Can anyone
> else confirm?
> 
> I am not sure what you would like me to review with 2688...
> 
> Thanks,
> K2
> 
> 
> Patrick Cheng wrote:
>> Hi guys,
>>
>> I recently submit several bugs. They are:
>>
>> 1.----------------------------------------------------------------------
>> BUG: 2688. ImageSpatialObjectRepresentation class bug, uninitialized 
>> memory access (I put in a fixed already)
>> http://public.kitware.com/Bug/bug.php?op=show&bugid=2688&pos=3
>>
>> ( Kevin Jr, Could you please have a look at this one? This one relates 
>> to the interaction between the ImageSpatialObjectRepresentation and 
>> View. ImageSpatialObjectRepresentation needs a input: 
>> ImageSpatialObject, and has to be added to the View for display.
>>  Some principals I followed here are:
>> 1. When there is no image available, and we add the representation to 
>> View, we should not connect the VTK pipeline. When there is a valid 
>> image, then we should connect the pipeline.
>> 2. Every time we get a new image, we should connect the pipeline again.
>> 3. The order of set image and add to view should not matter
>> )
>>
>> 2.----------------------------------------------------------------------
>> BUG: 2667. delete vtkObject ERROR messages in tests
>> http://public.kitware.com/Bug/bug.php?op=show&bugid=2667&pos=5
>>
>> Failed test can be found at:
>> http://public.kitware.com/IGSTK/Sites/camelot.kitware/Linux-g++-3.4+Loopback/20060106-1308-Experimental/Test.html 
>>
>> ( Kevin Jr, could you also take a look at this one? The VTK reference 
>> counting problems)
>>
>> 3.----------------------------------------------------------------------
>> BUG: 2689 Representation class's Copy() method should copy it's state 
>> machine and current states
>> http://public.kitware.com/Bug/bug.php?op=show&bugid=2689&pos=2
>>
>> 4.----------------------------------------------------------------------
>> BUG: 2695 Update Extend Error with the ImageReader
>> http://public.kitware.com/Bug/bug.php?op=show&bugid=2695&pos=0
>>
>> Testing code to expose this problem has been added to the end of
>> ­igstkCTImageSpatialObjectReadingAndRepresentationTest
>> The error msg:
>> ##################################################################
>> ERROR: In 
>> /home/ibanez/src/VTK/Filtering/vtkStreamingDemandDrivenPipeline.cxx, 
>> line 624
>> vtkImageImportExecutive (0x9fc8250): The update extent specified in 
>> the information for output port 0 on algorithm 
>> vtkImageImport(0x9fc7f90) is 0 511 0 511 4 4, which is outside the 
>> whole extent 0 511 0 511 0 1.
>> ###################################################################
>> (I have a solution for this one, by changing IGSTK code. But 
>> essentially this is a bug from vtkImageImporter and vtkStreamPipeline, 
>> so it might need more investigation.
>> )
>>
>> 5.----------------------------------------------------------------------
>> Also fixed some remaining bugs with removing FLTK. and clean out the 
>> testing part of the IGSTK, so that it will evaluate the IGSTK_USE_FLTK 
>>  and IGSTK_DATA_ROOT variable to decide whether to add a test or not.
>>
>>
>> Sorry for the messy email.
>>
>>
>> Patrick
>>
>> _______________________________________________
>> IGSTK-Developers mailing list
>> IGSTK-Developers at public.kitware.com
>> http://public.kitware.com/cgi-bin/mailman/listinfo/igstk-developers
>