[vtk-developers] UPDATE_EXTENT not updating after change to WHOLE_EXTENT

Fri Jan 8 13:21:06 EST 2016

Bill,

I can't tell you how much trouble you caused me :-) Kidding aside, this
issue (which is indeed a bug) lead me into some deep rabbit holes in the
pipeline. My least favorites ones at that. I am working on fixing and
cleaning up. I should have a fix sometime next week.

Best,
-berk

On Mon, Jan 4, 2016 at 3:20 PM, Bill McGrory <mcgrory at aerosoftinc.com>
wrote:

> yes, vtkGeometryFilter goes straight to a vtkPolyDataMapper
>
> If that fails, I will try to strip my code down to something more
> manageable and get it to you.
>
> Thank you
>
>
>
>
> On 01/04/2016 03:11 PM, Berk Geveci wrote:
>
> OK this sounds like a bug. Is your pipeline only source ->
> vtkGeometryFilter or is there something in between? I'll try to reproduce
> this with a Python source once I know what the pipeline is.
>
> Best,
> -berk
>
> On Mon, Jan 4, 2016 at 12:25 PM, Bill McGrory <mcgrory at aerosoftinc.com>
> wrote:
>
>> I think I am doing what you say should work.
>>
>> My personal data outside of VTK changes as a result of a user requested
>> change from say the fine representation to the coarse. That change triggers
>> a someSource->Modified(). that seems to correctly trigger a call to
>> RequestInformation on someSource, which is when I modify WHOLE_EXTENT().
>>
>> What I see is that WHOLE_EXTENT correctly propagates downstream, however,
>> there is no change to the UPDATE_EXTENT as a result of the WHOLE_EXTENT
>> being reduced. So when  RequestUpdateExtent propagates back upstream, I get
>> the error for it being out of range, with the WHOLE_EXTENT value being the
>> new correct value. and UPDATE_EXTENT being the old value.
>>
>> Note my error is triggered at someSource->RequestUpdateExtent()
>>
>> Note the someFilterDownstream is a vtkGeometryFilter, and if I make a
>> call to correct the UPDATE_EXTENT in the geometry filter to be in range
>> WHOLE_EXTENT from vtkGeometryFilter::RequestUpdateExtent() then that fixes
>> my bug, but this is only for vtkGeometryFilter, not a generic downstream
>> filter.
>>
>>
>>
>>
>>
>>
>> On 01/04/2016 11:46 AM, Berk Geveci wrote:
>>
>> So when the WHOLE_EXTENT() changes, you tell the source/reader by
>> modifying it somehow? What I am trying to get to is whether or not this
>> works:
>>
>> someSource->SetNewData();
>> someSource->Modified()
>> someFilterDownstream->Update();
>>
>> someSource->SetNewData();
>> someSource->Modified()
>> someFilterDownstream->Update();
>>
>> I believe that as long as the source is modified after the WHOLE_EXTENT()
>> changes, update should work fine. Because this would lead to
>> RequestInformation() being called again, which is what propagates the new
>> WHOLE_EXTENT() downstream.
>>
>> Best,
>> -berk
>>
>> On Mon, Jan 4, 2016 at 11:39 AM, Bill McGrory < <mcgrory at aerosoftinc.com>
>> mcgrory at aerosoftinc.com> wrote:
>>
>>> I  have a pipeline for displaying structured zones within a GUI. These
>>> zone's have multiple grid refinement levels or sequences. The beginning of
>>> the pipeline is a source which creates a vtkStructuredGrid.
>>>
>>> When I interactively change the sequence level the source gets a call to
>>> update the pipeline where I need to change the vtkStructuredGrid to
>>> represent a different sequence level (so my grid grid dimensions will
>>> double or halve, for example)
>>>
>>> This process worked (by accident perhaps) in versions of VTK prior to 6.
>>>
>>> Thanks
>>> Bill
>>>
>>>
>>>
>>> On 01/04/2016 11:08 AM, Berk Geveci wrote:
>>>
>>> Hey Bill,
>>>
>>> The very short (and annoying) answer is that VTK does not currently
>>> support varying WHOLE_EXTENT().
>>>
>>> However, there are ways of getting around this. But before I get into
>>> those, can you describe how/where you change the whole extent? As part of a
>>> time update? By setting some data member on the reader?
>>>
>>> Best,
>>> -berk
>>>
>>> On Wed, Dec 30, 2015 at 3:24 PM, Bill McGrory <
>>> <mcgrory at aerosoftinc.com>mcgrory at aerosoftinc.com> wrote:
>>>
>>>>  I am having difficulty porting some code from VTK 5 to 6.
>>>>
>>>> I have a custom algorithm for creating  a vtkStructuredGrid as the
>>>> source for my pipeline, so my class inherits vtkStructuredGridAlgorithm
>>>>
>>>> my pipeline simply connects the vtkStructuredGridAlgorithm to a
>>>> vtkGeometryFilter which is mapped with a vtkPolyDataMapper
>>>>
>>>> I want my source to correctly render a structured grid which changes
>>>> dimensions dynamically within my application.
>>>>
>>>> My problem is that  if I originally set up the grid with say extents of
>>>> 0 --99, 0-->99,  0-->99, and then change the extents to 0--49, 0-49,0-49
>>>> through a call to Update, then I get a VTK error message from
>>>>
>>>>
>>>> ERROR: In
>>>> VTK-6.2.0/Common/ExecutionModel/vtkStreamingDemandDrivenPipeline.cxx, line
>>>> 860
>>>> vtkCompositeDataPipeline (0xb4561f0): The update extent specified in
>>>> the information for output port 0 on algorithm
>>>> SurfaceSNodeSource(0xb455750) is 0 80 0 32 0 0, which is outside the whole
>>>> extent 0 40 0 16 0 0.
>>>> (This is in VerifyOutputInformation) called before the
>>>> RequestUpdateExtent for my algorithm.
>>>>
>>>> I don't quite understand the propagation of the UPDATE_EXTENT
>>>> information up the pipeline, but if I set the UPDATE_EXTENT to WHOLE_EXTENT
>>>>
>>>> in vtkGeometryFilter::RequestUpdateExtent
>>>>
>>>> then the error goes away.
>>>> Also, at the time of calling vtkGeometryFilter::RequestUpdateExtent,
>>>> the UPDATE_EXTENT is still set at the old value, while the WHOLE_EXTENT has
>>>> been correctly updated/propagated from upstream.
>>>>
>>>> Is this a bug, or do I need to change something in my
>>>> vtkStructuredGridAlgorithm to cause the UPDATE_EXTENT to be reset at the
>>>> origin downstream?
>>>>
>>>> Thanks
>>>>
>>>> Bill
>>>> _______________________________________________
>>>> Powered by <http://www.kitware.com>www.kitware.com
>>>>
>>>> Visit other Kitware open-source projects at
>>>> <http://www.kitware.com/opensource/opensource.html>
>>>> http://www.kitware.com/opensource/opensource.html
>>>>
>>>> Search the list archives at:
>>>> <http://markmail.org/search/?q=vtk-developers>
>>>> http://markmail.org/search/?q=vtk-developers
>>>>
>>>> Follow this link to subscribe/unsubscribe:
>>>> http://public.kitware.com/mailman/listinfo/vtk-developers
>>>>
>>>>
>>>
>>> --
>>> Dr. William D. McGrory    AeroSoft, Inc.mcgrory at aerosoftinc.com   Suite 1400(540) 557-1904            2000 Kraft Drive(540) 557-1919 (FAX)      Blacksburg, VA 24060
>>>
>>>
>>
>> --
>> Dr. William D. McGrory    AeroSoft, Inc.mcgrory at aerosoftinc.com   Suite 1400(540) 557-1904            2000 Kraft Drive(540) 557-1919 (FAX)      Blacksburg, VA 24060
>>
>>
>
> --
> Dr. William D. McGrory    AeroSoft, Inc.mcgrory at aerosoftinc.com   Suite 1400(540) 557-1904            2000 Kraft Drive(540) 557-1919 (FAX)      Blacksburg, VA 24060
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtk-developers/attachments/20160108/768972e9/attachment.html>