[vtkusers] Does vtkInformation::Set(DATA_OBJECT, ...) perform a deep copy?

Thu May 2 10:14:03 EDT 2013

I concur with Alex.

-berk

On Wed, May 1, 2013 at 5:56 PM, Alex Malyushytskyy <alexmalvtk at gmail.com>wrote:

> The code with smartpointer should also work.
> Assignment to smart pointer will increase the counter.
> So output will not be deleted when filter goes out of scope.
>
> ( http://www.vtk.org/Wiki/VTK/Tutorials/SmartPointers )
>
> But if you look in VTK api you would not find public functions which
> return or pass smartpointer.
> I think for a reason to avoid duplication of similar function. Returning
> pointer is minimal and sufficient.
>
> So people (at least me) are avoiding it, especially if you write piece of
> code which later could be considered to be included in vtk.
>
> Regards,
>     Alex
>
>
> On Wed, May 1, 2013 at 9:32 AM, Dominique Töpfer <dominique at toepfer-web.de
> > wrote:
>
>>  On 01.05.2013 16:50, Berk Geveci wrote:
>>
>> If you want the output of an algorithm to stick around after deleting it,
>> you need to call Register() on the data object. Something like:
>>
>>      vtkSmartPointer<vtkReader1> Reader1 =
>> vtkSmartPointer<vtkReader1>::New()
>>     // set filename and update reader1
>>     vtkDataSet * MyDataSet = Reader1->GetOutput();
>>      MyDataSet->Register(0);
>>
>>
>>
>> Yes, but wouldn't you have to delete MyDataSet manually?
>> Why not using SmartPointers also for MyDataSet? From what I understand
>> after reading http://www.vtk.org/Wiki/VTK/Tutorials/SmartPointers , the
>> following code should also work and memory is released automatically when
>> the last pointer pointing to the data goes out of scope:
>>
>> vtkSmartPointer<vtkDataSet> MyFilter::GetMyDataSet( ... )
>> {
>>     vtkSmartPointer<vtkDataSet> MyDataSet;
>>
>>
>>     if(SomeConditions)
>>     {
>>         vtkSmartPointer<vtkReader1> Reader1 =
>> vtkSmartPointer<vtkReader1>::New()
>>         // set filename and update reader1
>>         MyDataSet = Reader1->GetOutput();
>>     }
>>     else
>>     {
>>         vtkSmartPointer<vtkReader2> Reader2 =
>> vtkSmartPointer<vtkReader1>::New()
>>         // set filename and update reader2
>>         MyDataSet = Reader2->GetOutput();
>>     }
>>
>>     return MyDataSet;
>> }
>>
>> vtkSmartPointer<vtkDataSet> MyDataSet = GetMyDataSet( ... );
>>
>>
>> Or did I miss something?
>>
>> Regards,
>> Dominique
>>
>>
>>
>>  On Tue, Apr 30, 2013 at 5:14 PM, Sunrise <helios.corona at gmail.com>wrote:
>>
>>>  Thanks for the answer, Berk.
>>>
>>> I have a follow up question, which I guess I can solve it with strong
>>> pointers. Assume the following
>>>
>>> vtkDataSet * MyFilter::GetMyDataSet(...)
>>> {
>>>
>>> if(SomeConditions)
>>> {
>>>     vtkSmartPointer<vtkReader1> Reader1 =
>>> vtkSmartPointer<vtkReader1>::New()
>>>     // set filename and update reader1
>>>     vtkDataSet * MyDataSet = Reader1->GetOutput();
>>> }
>>> else
>>> {
>>>     vtkSmartPointer<vtkReader2> Reader2 =
>>> vtkSmartPointer<vtkReader1>::New()
>>>      // set filename and update reader2
>>>      vtkDataSet * MyDataSet = Reader2->GetOutput();
>>> }
>>>
>>> // MyDataSet is out of scope here!
>>> // Even if I remove if condition, they will be out of scope outside of
>>> this function.
>>>
>>> return MyDataSet;
>>>
>>> }
>>>
>>> I want to make both reader1 and reader2 become SmartPointers. Otherwise
>>> I have leak. At compile time I do not know how the code will end, only at
>>> the run time it make the decision. So I do not know when to delete
>>> reader1 and reader2.
>>>
>>> I do not want to use descrutor too. there are more readers than the
>>> sample code that I put here, so at compile time the destructor do not know
>>> how many reader1 and/or reader2 has been declared in order to delete
>>> them manually. So, I prefer to declare all of them with Smart Pointers.
>>>
>>> But here is the problem: reader1 and reade2 are in clauses of If
>>> condition, and they are in a function outside of RequestData. So they will
>>> go out of scope.
>>>
>>> If using a strong pointer fix this problem, how can I apply a strong
>>> pointer on MyDataSet in VTK?
>>>
>>> Again, thank you very much.
>>>
>>> -Sia
>>>
>>>
>>>
>>> On 04/29/2013 08:20 AM, Berk Geveci wrote:
>>>
>>> It does a shallow copy. The information object holds a strong reference
>>> to the data object so
>>>
>>>  outputInfo->Set(vtkDataObject::DATA_OBJECT(),MyDataSet);
>>>
>>>  so it will increment the reference count of MyDataSet preventing it
>>> from going out of scope.
>>>
>>>  -berk
>>>
>>>
>>> On Sun, Apr 28, 2013 at 11:41 AM, Sunrise <helios.corona at gmail.com>wrote:
>>>
>>>>  I am wondering if the "Set" method in vtkInformation class performs a
>>>> deep-copy or shallow-copy when I set a DATA_OBJECT into it?
>>>>
>>>> Suppose I have some vtkDataSet and I want to append them into
>>>> outputInfo of a filter inside its RequestData method.
>>>>
>>>> int Filter1::RequestData(...)
>>>> {
>>>>     vtkSmartPointer<vtkSomeReader> myreader =
>>>> vtkSmartPointer<vtkSomeReader>::New();
>>>>     ...
>>>>     vtkDataSet *MyDataSet = Reader->GetOutput();
>>>>     outputInfo->Set(vtkDataObject::DATA_OBJECT(),MyDataSet);
>>>>     ...
>>>> }
>>>>
>>>> at the end of RequestData, both the pointer myreader and its data on
>>>> the heap would be out of scope. Since I made them smart pointers, both
>>>> pointer "myreader" and the data on the heap will be deleted (right?).
>>>>
>>>> But in the next filter of the pipeline (say Filter2 is connected to
>>>> Filter1), I can receive the data in inputInfo
>>>>
>>>> int Filter2::RequestData(...)
>>>> {
>>>>     ...
>>>>     ... inputInfo->Get(vtkDataObject::DATA_OBJECT());
>>>> }
>>>>
>>>> I am wondering why still I can get the data, since data deferenced by
>>>> MyDataSet in previous filter no longer exists?
>>>>
>>>> Does vtkInformation::Set perform a deep copy?
>>>>
>>>> Thank you.
>>>>
>>>> _______________________________________________
>>>> Powered by www.kitware.com
>>>>
>>>> Visit other Kitware open-source projects at
>>>> http://www.kitware.com/opensource/opensource.html
>>>>
>>>> Please keep messages on-topic and check the VTK FAQ at:
>>>> http://www.vtk.org/Wiki/VTK_FAQ
>>>>
>>>> Follow this link to subscribe/unsubscribe:
>>>> http://www.vtk.org/mailman/listinfo/vtkusers
>>>>
>>>>
>>>
>>>
>>> _______________________________________________
>>> Powered by www.kitware.com
>>>
>>> Visit other Kitware open-source projects at
>>> http://www.kitware.com/opensource/opensource.html
>>>
>>> Please keep messages on-topic and check the VTK FAQ at:
>>> http://www.vtk.org/Wiki/VTK_FAQ
>>>
>>> Follow this link to subscribe/unsubscribe:
>>> http://www.vtk.org/mailman/listinfo/vtkusers
>>>
>>>
>>
>>
>> _______________________________________________
>> Powered by www.kitware.com
>>
>> Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html
>>
>> Please keep messages on-topic and check the VTK FAQ at: http://www.vtk.org/Wiki/VTK_FAQ
>>
>> Follow this link to subscribe/unsubscribe:http://www.vtk.org/mailman/listinfo/vtkusers
>>
>>
>>
>> --
>> Dominique Töpfer, Dipl.-Inform.
>> Institute of Medical Physics
>> University of Erlangen
>>
>>
>> _______________________________________________
>> Powered by www.kitware.com
>>
>> Visit other Kitware open-source projects at
>> http://www.kitware.com/opensource/opensource.html
>>
>> Please keep messages on-topic and check the VTK FAQ at:
>> http://www.vtk.org/Wiki/VTK_FAQ
>>
>> Follow this link to subscribe/unsubscribe:
>> http://www.vtk.org/mailman/listinfo/vtkusers
>>
>>
>
> _______________________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Please keep messages on-topic and check the VTK FAQ at:
> http://www.vtk.org/Wiki/VTK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://www.vtk.org/mailman/listinfo/vtkusers
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20130502/60cf7e3d/attachment.htm>