[Paraview] streamline tracer does not work on cut filter results

[Dominik Szczerba]

OK I traced the problem down to using the Intel compiler for my build. 
The same code compiled with GNU works.

To the developer of the stream tracer class: are you using things like

if(x==y)

for x,y float/double?

I have already noticed GNU and Intel respond differently, with GNU more 
forgiving. The proper (= correct, robust, meaningful) comparison should be

if(fabs(x-y)<eps)

I do not know if that is the problem, but it is one of the differences 
between the compilers that I noticed.

-- Dominik

Dominik Szczerba wrote:
> Ha it even works with official 3.3.0, only not my own compilation from 
> CVS 05 June. So either it is my own compilation, or something still got 
> broken in CVS slightly later. Attached is my CMakeCache. Any ideas?
> - Dominik
> 
> 
> Dominik Szczerba wrote:
>> Wooo it works with PV 3.2.1 official. So something got broken in the CVS.
>> -- Dominik
>>
>> Dominik Szczerba wrote:
>>> I already did it, as mentioned in a previous post. No, it does not 
>>> work for me - does it work on my dataset for you?
>>>
>>> In any case, that is a trivialized example, my real case is not 
>>> symmetric and a cut plane is not easily aligned with the axes.
>>>
>>> I suspect this might have to do with:
>>>
>>> 1) the funny range like 1.52e-18 - 1.22e-17 for point1/point2 in line 
>>> source sub-window that I get hitting 'X/Y/Z Axis' button.
>>>
>>> 2) my mesh being quadratic and streamlining not working in such cases.
>>>
>>> Thanks for any hints.
>>> -- Dominik
>>>
>>> Michael Ahlmann wrote:
>>>> I had a similar problem when trying to visualize an axi-symmetric 
>>>> data set.  The problem was that the streamlines were not confined to 
>>>> the the axi-symmetric plane, and thus they were cut short at the 
>>>> point where the flow went in the circumferential direction.  The fix 
>>>> in my case was to use the calculator to develop a new, simplified 
>>>> data set.
>>>>
>>>> V = V_x * ihat + V_y * jhat + 0 * khat
>>>>
>>>> This way, the axi-symmetric component is zero, and doesn't affect 
>>>> the in plane streamlines.  I wonder if you could do something 
>>>> similar in your case?  Although I presume it would be more complex 
>>>> given that your cutting plane may not lay perfectly within one of 
>>>> the Cartesian planes...
>>>>
>>>> -Michael
>>>> _______________________________________________
>>>> ParaView mailing list
>>>> ParaView at paraview.org
>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>
>>
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> ParaView mailing list
> ParaView at paraview.org
> http://www.paraview.org/mailman/listinfo/paraview

-- 
Dominik Szczerba, Ph.D.
Biomedical Simulation Group
Computer Vision Lab CH-8092 Zurich
http://www.vision.ee.ethz.ch/~domi