[vtkusers] RE: Quadratic triangle + linear edge - cell scalars incorrect (vtk 4.5)

[John Platt]

Hi Mathieu,

Please find attached the .vtk file. I had a quick look at it using
ParaView 1.2.1 and it displays a red centre triangle surrounded by blue
triangles and a blue edge (the inverse). I'm wondering whether there may
have been a fix for this or it's my setup here?

Thanks for your help.

John.

-----Original Message-----
From: Mathieu Malaterre [mailto:mathieu.malaterre at kitware.com] 
Sent: 24 May 2004 22:26
To: John Platt
Subject: Re: Quadratic triangle + linear edge - cell scalars incorrect
(vtk 4.5)

John Platt wrote:
> Hi Mathieu,
> 
> Hope the direct post is not too much of an intrusion but there appears
> to be some problem with the list at the moment.

Yes, Andy is having a *really* bad time trying to back up everything 
here, sorry for trouble.

> I tried your suggestion but it did not affect the output. To remove
any
> doubt on the scoping issue, I replaced 
> 
>    topology->InsertNextId( ptId++ );
>    topology->InsertNextId( ptId );
> by
>    topology->InsertNextId( 6 );
>    topology->InsertNextId( 7 );
> 
> but still no change.

Ok I have to say: are you sure? I try both ideas and I get the correct 
result. cf picture. Could you try something like 'rebuild all', delete 
your previous exe, remove your old image ...

If I remember well you already  had some weird issue with an opengl 
driver, but this one, I don't believe those guys would make such a
mistake.

Could you please change your snapshot for a vtkUnstructuredGridWriter, 
and send me the vtk file ? Thanks

> As for the suitability of VC++6 for real development, OK, I admit that
> ParaView does cause it to crash if you double click
> ParaviewComplete.dsw. Perhaps I should send the report to MS next this
> happens? - or maybe you just need a multiprocessor machine (I did get
it
> to build) :-).

Are you using cvs ParaViewComplete ? You shouldn't use ParaViewComplete,

can you grab ParaView-1.4.zip, and let us know if you still have soime 
problem. 'Cause it shouldn't crash at all. (BTW, please wait until the 
list are back fro sending your bug report)

> Many thanks for your suggestions and continued help.

My pleasure,

Mathieu


> John.
> 
> 
> 
> -----Original Message-----
> From: Mathieu Malaterre [mailto:mathieu.malaterre at kitware.com] 
> Sent: 24 May 2004 20:50
> To: John Platt
> Cc: vtkusers at vtk.org
> Subject: Re: Quadratic triangle + linear edge - cell scalars incorrect
> (vtk 4.5)
> 
> John,
> 
> 	This is the problem with using VC++6 for real development :)
> 
> 	More seriously, VC++6 does not respect scoping for 'for', and
> thus lead 
> people to make errors. Could you change your code :
> 
>     for ( int ptId = 0; ptId < 6; ptId++ )
> 
> into
> 
>     int ptId;
>     for ( ptId = 0; ptId < 6; ptId++ )
> 
> Only then you can later reuse ptId.
> 
> HTH
> Mathieu
> 
> 
> John Platt wrote:
> 
>>Hi Users,
>>
>> 
>>
>>When I add a linear edge to an unstructured grid consisting of a
> 
> single 
> 
>>quadratic triangle, the centre linear triangle of the triangulated 
>>quadratic triangle and the edge have swapped cell scalars - see 
>>attachment (triangle should be all red & edge blue).
>>
>> 
>>
>>When either of the cells is drawn in isolation there is no problem.
>>
>> 
>>
>>The cell scalars coming out of 
>>vtkDataSetSurfaceFilter::UnstructuredGridExecute() appear OK.
>>
>> 
>>
>>Does anybody else have this problem? I have attached the c++ code just
> 
> 
>>in case it is a fault on my part.
>>
>> 
>>
>>Many thanks for your help.
>>
>> 
>>
>>John.
>>
>>
>>
> 
>
------------------------------------------------------------------------
> 
> 
> 
> 
> 
> 
> 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: QTri-Line.vtk
Type: application/octet-stream
Size: 262 bytes
Desc: not available
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20040525/2eca4c50/attachment.obj>