[Paraview] How to color by Cell Data

[Michael Jackson]

I have a custom filter that sets an RGB value for each cell in a
surface mesh, where each cell is a triangle. When the filter runs in
paraview I "see" the new arrays show up when I click on the
"Information" tab but the "Color By" drop down only has the "Solid
Color" available.

Here is the relevant parts of the code:

vtkCellData* cd = input->GetCellData();
vtkDataArray* normals = cd->GetNormals();
if (NULL == normals || (normals && normals->GetNumberOfTuples() <=
0 ) )
{
std::cout << "NORMALS were not present" << std::endl;
normals = NULL;
}

if (NULL == normals)
{
vtkPolyDataNormals* polyNormals = vtkPolyDataNormals::New();
polyNormals->SetInput(input);
polyNormals->SplittingOff();
polyNormals->ConsistencyOn ();
polyNormals->AutoOrientNormalsOn();
polyNormals->ComputePointNormalsOff();
polyNormals->ComputeCellNormalsOn();
polyNormals->FlipNormalsOff();
polyNormals->NonManifoldTraversalOff();
polyNormals->Update();
output->ShallowCopy(polyNormals->GetOutput());
}

cd = output->GetCellData();
normals = cd->GetNormals();

vtkUnsignedCharArray* rgb = vtkUnsignedCharArray::New();
rgb->SetName(INTERFACE_COLORING);
rgb->SetNumberOfComponents(3);

vtkDataArray* eulerArray = input->GetFieldData()-
>GetArray(OIMColoring::EulerAngles().c_str());
if (NULL == eulerArray)
{
vtkErrorMacro( << "Euler Angles Field Data Missing from Data set.
This filter requires the euler angles to work");
return 0;
}

double* eulers = eulerArray->GetTuple3(0);
double* refDirection; //[3] = {45.0, 45.0, 0.0};
unsigned char bytes[3] = { 65, 0, 255 };
for (vtkIdType i = 0; i < normals->GetNumberOfTuples(); ++i)
{
refDirection = normals->GetTuple3(i);
OIMColoring::GenerateIPFColor<double>(eulers, refDirection, bytes);
// std::cout << (int)(bytes[0]) << " " << (int)(bytes[1]) << " "
<< (int)(bytes[2]) << std::endl;
rgb->InsertNextTupleValue(bytes);
}

output->GetCellData()->SetScalars(rgb);
output->GetCellData()->SetActiveScalars(INTERFACE_COLORING);

Helpful insights are very welcome.
___________________________________________________________
Mike Jackson www.bluequartz.net
Principal Software Engineer mike.jackson at bluequartz.net
BlueQuartz Software Dayton, Ohio

[Michael Jackson]

I have a triangle surface mesh that contains Cell based RGB attribute
array. I would like to color by that array, ie, not using a Lookup
table or interpolation. When I load my data set the cell data is
listed, but in the "color by" drop down list the scalar array is not
listed as a choice.

On a "whim" I then ran the "Cell Data to Point Data" filter on the
data. At that point I was able to select the scalar array to "Color
By" and the surface mesh now has the correct colors.

I know if I were to code this up in straight VTk I can set the
rendering to be by Cells:

vtkPolyDataMapper *cubeMapper = vtkPolyDataMapper::New();
cubeMapper->SetInput(filter->GetOutput());
cubeMapper->SetScalarVisibility(1);
cubeMapper->SetScalarModeToUseCellData();

How do I get the same effect in ParaView?

Thanks
_________________________________________________________
Mike Jackson mike.jackson at bluequartz.net
BlueQuartz Software www.bluequartz.net
Principal Software Engineer Dayton, Ohio

[Utkarsh Ayachit]

That's weird, your cell array should have been listed in the Color By
menu. Is it possible to share the data or a small sample to reproduce
the issue?

Utkarsh

On Sat, Feb 27, 2010 at 11:09 AM, Michael Jackson
<mike.jackson at bluequartz.net> wrote:
> I have a triangle surface mesh that contains Cell based RGB attribute array.
> I would like to color by that array, ie, not using a Lookup table or
> interpolation. When I load my data set the cell data is listed, but in the
> "color by" drop down list the scalar array is not listed as a choice.
>
> ?On a "whim" I then ran the "Cell Data to Point Data" filter on the data. At
> that point I was able to select the scalar array to "Color By" and the
> surface mesh now has the correct colors.
>
> ?I know if I were to code this up in straight VTk I can set the rendering to
> be by Cells:
>
> ?vtkPolyDataMapper *cubeMapper = vtkPolyDataMapper::New();
> ?cubeMapper->SetInput(filter->GetOutput());
> ?cubeMapper->SetScalarVisibility(1);
> ?cubeMapper->SetScalarModeToUseCellData();
>
> How do I get the same effect in ParaView?
>
> Thanks
> _________________________________________________________
> Mike Jackson ? ? ? ? ? ? ? ? ?mike.jackson at bluequartz.net
> BlueQuartz Software ? ? ? ? ? ? ? ? ? ?www.bluequartz.net
> Principal Software Engineer ? ? ? ? ? ? ? ? ?Dayton, Ohio
>
> _______________________________________________
> 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 ParaView Wiki at:
> http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview
>

[Jérôme]

Hi,
I noted this problem a few days ago.
A *really* simple example:
- Add a SphereSource
- Connect to a "Generate Ids" filter
- Update: The information panel shows a 'Ids' array for points and for
cells, but
the "Color by" doesn't give the choice of cell array.

Jerome

2010/2/28 Utkarsh Ayachit <utkarsh.ayachit at kitware.com>

> That's weird, your cell array should have been listed in the Color By
> menu. Is it possible to share the data or a small sample to reproduce
> the issue?
>
> Utkarsh
>
> On Sat, Feb 27, 2010 at 11:09 AM, Michael Jackson
> <mike.jackson at bluequartz.net> wrote:
> > I have a triangle surface mesh that contains Cell based RGB attribute
> array.
> > I would like to color by that array, ie, not using a Lookup table or
> > interpolation. When I load my data set the cell data is listed, but in
> the
> > "color by" drop down list the scalar array is not listed as a choice.
> >
> > On a "whim" I then ran the "Cell Data to Point Data" filter on the data.
> At
> > that point I was able to select the scalar array to "Color By" and the
> > surface mesh now has the correct colors.
> >
> > I know if I were to code this up in straight VTk I can set the rendering
> to
> > be by Cells:
> >
> > vtkPolyDataMapper *cubeMapper = vtkPolyDataMapper::New();
> > cubeMapper->SetInput(filter->GetOutput());
> > cubeMapper->SetScalarVisibility(1);
> > cubeMapper->SetScalarModeToUseCellData();
> >
> > How do I get the same effect in ParaView?
> >
> > Thanks
> > _________________________________________________________
> > Mike Jackson mike.jackson at bluequartz.net
> > BlueQuartz Software www.bluequartz.net
> > Principal Software Engineer Dayton, Ohio
> >
> > _______________________________________________
> > 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 ParaView Wiki at:
> > http://paraview.org/Wiki/ParaView
> >
> > Follow this link to subscribe/unsubscribe:
> > http://www.paraview.org/mailman/listinfo/paraview
> >
> _______________________________________________
> 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 ParaView Wiki at:
> http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20100228/9ad1696e/attachment-0001.htm>

[Favre Jean]

Hi,
I noted this problem a few days ago.
A *really* simple example:
- Add a SphereSource
- Connect to a "Generate Ids" filter
- Update: The information panel shows a 'Ids' array for points and for cells, but
the "Color by" doesn't give the choice of cell array.

Jerome
-----------------

I cannot reproduce the problem. The following code works exactly as expected and enables the coloring by POINT_DATA or CELL_DATA

sph1 = Sphere()
genid1 = GenerateIds(Input=sph1)
rep1 = Show(genid1)
rep1.ColorArrayName = 'Ids'
rep1.ColorAttributeType = 'CELL_DATA'

----------------
Jean M. Favre
Swiss National Supercomputing Center

[Jérôme]

Hi,

I confirm the problem: even with your script, CELL_DATA are not enabled.
They don't appear in the "Color by" combo box, but the cell Ids are present
in the Information panel. It works as expected when I try with POINT_DATA.

I just had a doubt: Qt has been updated to 4.6 recently on my computer, and
I don't know if my ParaView is up-to-date according to this...

Let's update, compile, wait and see.

Jerome

2010/2/28 Favre Jean <jfavre at cscs.ch>

>
> Hi,
> I noted this problem a few days ago.
> A *really* simple example:
> - Add a SphereSource
> - Connect to a "Generate Ids" filter
> - Update: The information panel shows a 'Ids' array for points and for
> cells, but
> the "Color by" doesn't give the choice of cell array.
>
> Jerome
> -----------------
>
> I cannot reproduce the problem. The following code works exactly as
> expected and enables the coloring by POINT_DATA or CELL_DATA
>
> sph1 = Sphere()
> genid1 = GenerateIds(Input=sph1)
> rep1 = Show(genid1)
> rep1.ColorArrayName = 'Ids'
> rep1.ColorAttributeType = 'CELL_DATA'
>
> ----------------
> Jean M. Favre
> Swiss National Supercomputing Center
> _______________________________________________
> 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 ParaView Wiki at:
> http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20100228/98e24ed3/attachment.htm>

[Michael Jackson]

Maybe it is Qt versions or ParaView versions? I am using Qt 4.6.1
compiled for Cocoa on OS X 10.5.8 Intel. I am using ParaView from <url
= git://github.com/Kitware/ParaView.git> which was last updated about
a week ago. When I get to the office on Monday I can try the same with
ParaView 3.6.2 and see if the problem persists.

As far as Data, I _could_ share it, but like Jerome stated, you can
reproduce it with a simple sphere source, then "Generate Normals". The
complete source code can be found here:

<href=http://www.bluequartz.net/cgi-bin/gitweb/gitweb.cgi?p=PVMXADataModel.git;a=blob;f=Source/Server/CrystalInterfaceColoringFilter/vtkCrystalInterfaceColoringFilter.cpp;hb=de5d94a3a128a08a534230deca9057e4e57d0358
>


_________________________________________________________
Mike Jackson mike.jackson at bluequartz.net

On Feb 28, 2010, at 7:51 AM, J?r?me wrote:

> Hi,
>
> I confirm the problem: even with your script, CELL_DATA are not
> enabled.
> They don't appear in the "Color by" combo box, but the cell Ids are
> present
> in the Information panel. It works as expected when I try with
> POINT_DATA.
>
> I just had a doubt: Qt has been updated to 4.6 recently on my
> computer, and
> I don't know if my ParaView is up-to-date according to this...
>
> Let's update, compile, wait and see.
>
> Jerome
>
> 2010/2/28 Favre Jean <jfavre at cscs.ch>
>
> Hi,
> I noted this problem a few days ago.
> A *really* simple example:
> - Add a SphereSource
> - Connect to a "Generate Ids" filter
> - Update: The information panel shows a 'Ids' array for points and
> for cells, but
> the "Color by" doesn't give the choice of cell array.
>
> Jerome
> -----------------
>
> I cannot reproduce the problem. The following code works exactly as
> expected and enables the coloring by POINT_DATA or CELL_DATA
>
> sph1 = Sphere()
> genid1 = GenerateIds(Input=sph1)
> rep1 = Show(genid1)
> rep1.ColorArrayName = 'Ids'
> rep1.ColorAttributeType = 'CELL_DATA'
>
> ----------------
> Jean M. Favre
> Swiss National Supercomputing Center

[Utkarsh Ayachit]

That is weird. I am on Mac with Qt 4.6.0, Cocoa and the latest CVS
ParaView and I see both point "Ids" and cell "Ids" in the Color By
menu. I'll test it out with my linux box at work tomorrow.

Utkarsh

On Sun, Feb 28, 2010 at 9:44 AM, Michael Jackson
<mike.jackson at bluequartz.net> wrote:
> ?Maybe it is Qt versions or ParaView versions? I am using Qt 4.6.1 compiled
> for Cocoa on OS X 10.5.8 Intel. I am using ParaView from <url =
> git://github.com/Kitware/ParaView.git> which was last updated about a week
> ago. When I get to the office on Monday I can try the same with ParaView
> 3.6.2 and see if the problem persists.
>
> ?As far as Data, I _could_ share it, but like Jerome stated, you can
> reproduce it with a simple sphere source, then "Generate Normals". The
> complete source code can be found here:
>
> <href=http://www.bluequartz.net/cgi-bin/gitweb/gitweb.cgi?p=PVMXADataModel.git;a=blob;f=Source/Server/CrystalInterfaceColoringFilter/vtkCrystalInterfaceColoringFilter.cpp;hb=de5d94a3a128a08a534230deca9057e4e57d0358>
>
>
> _________________________________________________________
> Mike Jackson ? ? ? ? ? ? ? ? ?mike.jackson at bluequartz.net
>
> On Feb 28, 2010, at 7:51 AM, J?r?me wrote:
>
>> Hi,
>>
>> I confirm the problem: even with your script, CELL_DATA are not enabled.
>> They don't appear in the "Color by" combo box, but the cell Ids are
>> present
>> in the Information panel. It works as expected when I try with POINT_DATA.
>>
>> I just had a doubt: Qt has been updated to 4.6 recently on my computer,
>> and
>> I don't know if my ParaView is up-to-date according to this...
>>
>> Let's update, compile, wait and see.
>>
>> Jerome
>>
>> 2010/2/28 Favre Jean <jfavre at cscs.ch>
>>
>> Hi,
>> I noted this problem a few days ago.
>> A *really* simple example:
>> - Add a SphereSource
>> - Connect to a "Generate Ids" filter
>> - Update: The information panel shows a 'Ids' array for points and for
>> cells, but
>> the "Color by" doesn't give the choice of cell array.
>>
>> Jerome
>> -----------------
>>
>> I cannot reproduce the problem. The following code works exactly as
>> expected and enables the coloring by POINT_DATA or CELL_DATA
>>
>> sph1 = Sphere()
>> genid1 = GenerateIds(Input=sph1)
>> rep1 = Show(genid1)
>> rep1.ColorArrayName = 'Ids'
>> rep1.ColorAttributeType = 'CELL_DATA'
>>
>> ----------------
>> Jean M. Favre
>> Swiss National Supercomputing Center
>
> _______________________________________________
> 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 ParaView Wiki at:
> http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview
>

[Michael Jackson]

Is ParaView still on CVS or is that git location "official" to use? If
the git location isn't really official, then maybe I should switch
back to CVS?
--
Mike Jackson <www.bluequartz.net>

On Feb 28, 2010, at 11:38 AM, Utkarsh Ayachit wrote:

> That is weird. I am on Mac with Qt 4.6.0, Cocoa and the latest CVS
> ParaView and I see both point "Ids" and cell "Ids" in the Color By
> menu. I'll test it out with my linux box at work tomorrow.
>
> Utkarsh
>
> On Sun, Feb 28, 2010 at 9:44 AM, Michael Jackson
> <mike.jackson at bluequartz.net> wrote:
>> Maybe it is Qt versions or ParaView versions? I am using Qt 4.6.1
>> compiled
>> for Cocoa on OS X 10.5.8 Intel. I am using ParaView from <url =
>> git://github.com/Kitware/ParaView.git> which was last updated about
>> a week
>> ago. When I get to the office on Monday I can try the same with
>> ParaView
>> 3.6.2 and see if the problem persists.
>>
>> As far as Data, I _could_ share it, but like Jerome stated, you can
>> reproduce it with a simple sphere source, then "Generate Normals".
>> The
>> complete source code can be found here:
>>
>> <href=http://www.bluequartz.net/cgi-bin/gitweb/gitweb.cgi?p=PVMXADataModel.git;a=blob;f=Source/Server/CrystalInterfaceColoringFilter/vtkCrystalInterfaceColoringFilter.cpp;hb=de5d94a3a128a08a534230deca9057e4e57d0358
>> >
>>
>>
>> _________________________________________________________
>> Mike Jackson mike.jackson at bluequartz.net
>>
>> On Feb 28, 2010, at 7:51 AM, J?r?me wrote:
>>
>>> Hi,
>>>
>>> I confirm the problem: even with your script, CELL_DATA are not
>>> enabled.
>>> They don't appear in the "Color by" combo box, but the cell Ids are
>>> present
>>> in the Information panel. It works as expected when I try with
>>> POINT_DATA.
>>>
>>> I just had a doubt: Qt has been updated to 4.6 recently on my
>>> computer,
>>> and
>>> I don't know if my ParaView is up-to-date according to this...
>>>
>>> Let's update, compile, wait and see.
>>>
>>> Jerome
>>>
>>> 2010/2/28 Favre Jean <jfavre at cscs.ch>
>>>
>>> Hi,
>>> I noted this problem a few days ago.
>>> A *really* simple example:
>>> - Add a SphereSource
>>> - Connect to a "Generate Ids" filter
>>> - Update: The information panel shows a 'Ids' array for points and
>>> for
>>> cells, but
>>> the "Color by" doesn't give the choice of cell array.
>>>
>>> Jerome
>>> -----------------
>>>
>>> I cannot reproduce the problem. The following code works exactly as
>>> expected and enables the coloring by POINT_DATA or CELL_DATA
>>>
>>> sph1 = Sphere()
>>> genid1 = GenerateIds(Input=sph1)
>>> rep1 = Show(genid1)
>>> rep1.ColorArrayName = 'Ids'
>>> rep1.ColorAttributeType = 'CELL_DATA'
>>>
>>> ----------------
>>> Jean M. Favre
>>> Swiss National Supercomputing Center
>>
>> _______________________________________________
>> 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 ParaView Wiki at:
>> http://paraview.org/Wiki/ParaView
>>
>> Follow this link to subscribe/unsubscribe:
>> http://www.paraview.org/mailman/listinfo/paraview
>>

[Jérôme]

Well, I just updated my worktree with the latest CVS. The problem remains,
both in debug and release. Note that Yum (package manager of Fedora)
installs qt-4.6.2.

I will trace the code tomorrow monday.

Jerome

2010/2/28 Utkarsh Ayachit <utkarsh.ayachit at kitware.com>

> That is weird. I am on Mac with Qt 4.6.0, Cocoa and the latest CVS
> ParaView and I see both point "Ids" and cell "Ids" in the Color By
> menu. I'll test it out with my linux box at work tomorrow.
>
> Utkarsh
>
> On Sun, Feb 28, 2010 at 9:44 AM, Michael Jackson
> <mike.jackson at bluequartz.net> wrote:
> > Maybe it is Qt versions or ParaView versions? I am using Qt 4.6.1
> compiled
> > for Cocoa on OS X 10.5.8 Intel. I am using ParaView from <url =
> > git://github.com/Kitware/ParaView.git> which was last updated about a
> week
> > ago. When I get to the office on Monday I can try the same with ParaView
> > 3.6.2 and see if the problem persists.
> >
> > As far as Data, I _could_ share it, but like Jerome stated, you can
> > reproduce it with a simple sphere source, then "Generate Normals". The
> > complete source code can be found here:
> >
> > <href=
> http://www.bluequartz.net/cgi-bin/gitweb/gitweb.cgi?p=PVMXADataModel.git;a=blob;f=Source/Server/CrystalInterfaceColoringFilter/vtkCrystalInterfaceColoringFilter.cpp;hb=de5d94a3a128a08a534230deca9057e4e57d0358
> >
> >
> >
> > _________________________________________________________
> > Mike Jackson mike.jackson at bluequartz.net
> >
> > On Feb 28, 2010, at 7:51 AM, J?r?me wrote:
> >
> >> Hi,
> >>
> >> I confirm the problem: even with your script, CELL_DATA are not enabled.
> >> They don't appear in the "Color by" combo box, but the cell Ids are
> >> present
> >> in the Information panel. It works as expected when I try with
> POINT_DATA.
> >>
> >> I just had a doubt: Qt has been updated to 4.6 recently on my computer,
> >> and
> >> I don't know if my ParaView is up-to-date according to this...
> >>
> >> Let's update, compile, wait and see.
> >>
> >> Jerome
> >>
> >> 2010/2/28 Favre Jean <jfavre at cscs.ch>
> >>
> >> Hi,
> >> I noted this problem a few days ago.
> >> A *really* simple example:
> >> - Add a SphereSource
> >> - Connect to a "Generate Ids" filter
> >> - Update: The information panel shows a 'Ids' array for points and for
> >> cells, but
> >> the "Color by" doesn't give the choice of cell array.
> >>
> >> Jerome
> >> -----------------
> >>
> >> I cannot reproduce the problem. The following code works exactly as
> >> expected and enables the coloring by POINT_DATA or CELL_DATA
> >>
> >> sph1 = Sphere()
> >> genid1 = GenerateIds(Input=sph1)
> >> rep1 = Show(genid1)
> >> rep1.ColorArrayName = 'Ids'
> >> rep1.ColorAttributeType = 'CELL_DATA'
> >>
> >> ----------------
> >> Jean M. Favre
> >> Swiss National Supercomputing Center
> >
> > _______________________________________________
> > 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 ParaView Wiki at:
> > http://paraview.org/Wiki/ParaView
> >
> > Follow this link to subscribe/unsubscribe:
> > http://www.paraview.org/mailman/listinfo/paraview
> >
> _______________________________________________
> 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 ParaView Wiki at:
> http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20100228/319c4c3f/attachment.htm>

[Michael Jackson]

Here is a screen shot. I updated to latest cvs and recompiled. Still
no joy. I even tried saving the data set to a new file, then loading
the file. No luck there either. Any other suggestions are welcome. I
do have the data file if you are interested.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: CellDataColor.jpg
Type: image/jpeg
Size: 37281 bytes
Desc: not available
URL: <http://www.paraview.org/pipermail/paraview/attachments/20100301/78fbc749/attachment-0001.jpg>
-------------- next part --------------


___________________________________________________________
Mike Jackson www.bluequartz.net


On Feb 28, 2010, at 2:01 PM, J?r?me wrote:

> Well, I just updated my worktree with the latest CVS. The problem
> remains,
> both in debug and release. Note that Yum (package manager of Fedora)
> installs qt-4.6.2.
>
> I will trace the code tomorrow monday.
>
> Jerome
>
> 2010/2/28 Utkarsh Ayachit <utkarsh.ayachit at kitware.com>
> That is weird. I am on Mac with Qt 4.6.0, Cocoa and the latest CVS
> ParaView and I see both point "Ids" and cell "Ids" in the Color By
> menu. I'll test it out with my linux box at work tomorrow.
>
> Utkarsh
>
> On Sun, Feb 28, 2010 at 9:44 AM, Michael Jackson
> <mike.jackson at bluequartz.net> wrote:
> > Maybe it is Qt versions or ParaView versions? I am using Qt 4.6.1
> compiled
> > for Cocoa on OS X 10.5.8 Intel. I am using ParaView from <url =
> > git://github.com/Kitware/ParaView.git> which was last updated
> about a week
> > ago. When I get to the office on Monday I can try the same with
> ParaView
> > 3.6.2 and see if the problem persists.
> >
> > As far as Data, I _could_ share it, but like Jerome stated, you can
> > reproduce it with a simple sphere source, then "Generate Normals".
> The
> > complete source code can be found here:
> >
> > <href=http://www.bluequartz.net/cgi-bin/gitweb/gitweb.cgi?p=PVMXADataModel.git;a=blob;f=Source/Server/CrystalInterfaceColoringFilter/vtkCrystalInterfaceColoringFilter.cpp;hb=de5d94a3a128a08a534230deca9057e4e57d0358
> >
> >
> >
> > _________________________________________________________
> > Mike Jackson mike.jackson at bluequartz.net
> >
> > On Feb 28, 2010, at 7:51 AM, J?r?me wrote:
> >
> >> Hi,
> >>
> >> I confirm the problem: even with your script, CELL_DATA are not
> enabled.
> >> They don't appear in the "Color by" combo box, but the cell Ids are
> >> present
> >> in the Information panel. It works as expected when I try with
> POINT_DATA.
> >>
> >> I just had a doubt: Qt has been updated to 4.6 recently on my
> computer,
> >> and
> >> I don't know if my ParaView is up-to-date according to this...
> >>
> >> Let's update, compile, wait and see.
> >>
> >> Jerome
> >>
> >> 2010/2/28 Favre Jean <jfavre at cscs.ch>
> >>
> >> Hi,
> >> I noted this problem a few days ago.
> >> A *really* simple example:
> >> - Add a SphereSource
> >> - Connect to a "Generate Ids" filter
> >> - Update: The information panel shows a 'Ids' array for points
> and for
> >> cells, but
> >> the "Color by" doesn't give the choice of cell array.
> >>
> >> Jerome
> >> -----------------
> >>
> >> I cannot reproduce the problem. The following code works exactly as
> >> expected and enables the coloring by POINT_DATA or CELL_DATA
> >>
> >> sph1 = Sphere()
> >> genid1 = GenerateIds(Input=sph1)
> >> rep1 = Show(genid1)
> >> rep1.ColorArrayName = 'Ids'
> >> rep1.ColorAttributeType = 'CELL_DATA'
> >>
> >> ----------------
> >> Jean M. Favre
> >> Swiss National Supercomputing Center
> >
> > _______________________________________________
> > 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 ParaView Wiki at:
> > http://paraview.org/Wiki/ParaView
> >
> > Follow this link to subscribe/unsubscribe:
> > http://www.paraview.org/mailman/listinfo/paraview
> >
> _______________________________________________
> 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 ParaView Wiki at: http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview
>

[Sven Buijssen]

Hi,

I tested my nightly clean builds of ParaView CVS HEAD (pulled via git) + Qt 4.5.2 on
* SuSE Linux Enterprise Server 10 SP2
* Ubuntu 9.10 Server 64bit
* openSuSE 10.2 64bit
* openSuSE 11.1 32bit
* SPARC Solaris 10 64bit (PV CVS 2010-02-25, nothing newer as I'm having some
trouble with Mesa since Dave's commit Fri Feb 26 17:57:21 2010 +0000)

I see for all builds both point "Ids" and cell "Ids" in the Color By menu.

I also checked the tarballs of all nightly builds since beginning of February
2010 on openSuSE 11.1 and could not reproduce the issue. Cell Ids is always there.

Sven


Michael Jackson wrote, On 01.03.2010 15:47:
> Here is a screen shot. I updated to latest cvs and recompiled. Still no
> joy. I even tried saving the data set to a new file, then loading the
> file. No luck there either. Any other suggestions are welcome. I do have
> the data file if you are interested.
>
>
>
>
> ___________________________________________________________
> Mike Jackson www.bluequartz.net
>
>
> On Feb 28, 2010, at 2:01 PM, J?r?me wrote:
>
>> Well, I just updated my worktree with the latest CVS. The problem
>> remains,
>> both in debug and release. Note that Yum (package manager of Fedora)
>> installs qt-4.6.2.
>>
>> I will trace the code tomorrow monday.
>>
>> Jerome
>>
>> 2010/2/28 Utkarsh Ayachit <utkarsh.ayachit at kitware.com>
>> That is weird. I am on Mac with Qt 4.6.0, Cocoa and the latest CVS
>> ParaView and I see both point "Ids" and cell "Ids" in the Color By
>> menu. I'll test it out with my linux box at work tomorrow.
>>
>> Utkarsh
>>
>> On Sun, Feb 28, 2010 at 9:44 AM, Michael Jackson
>> <mike.jackson at bluequartz.net> wrote:
>> > Maybe it is Qt versions or ParaView versions? I am using Qt 4.6.1
>> compiled
>> > for Cocoa on OS X 10.5.8 Intel. I am using ParaView from <url =
>> > git://github.com/Kitware/ParaView.git> which was last updated about
>> a week
>> > ago. When I get to the office on Monday I can try the same with
>> ParaView
>> > 3.6.2 and see if the problem persists.
>> >
>> > As far as Data, I _could_ share it, but like Jerome stated, you can
>> > reproduce it with a simple sphere source, then "Generate Normals". The
>> > complete source code can be found here:
>> >
>> >
>> <href=http://www.bluequartz.net/cgi-bin/gitweb/gitweb.cgi?p=PVMXADataModel.git;a=blob;f=Source/Server/CrystalInterfaceColoringFilter/vtkCrystalInterfaceColoringFilter.cpp;hb=de5d94a3a128a08a534230deca9057e4e57d0358>
>>
>> >
>> >
>> > _________________________________________________________
>> > Mike Jackson mike.jackson at bluequartz.net
>> >
>> > On Feb 28, 2010, at 7:51 AM, J?r?me wrote:
>> >
>> >> Hi,
>> >>
>> >> I confirm the problem: even with your script, CELL_DATA are not
>> enabled.
>> >> They don't appear in the "Color by" combo box, but the cell Ids are
>> >> present
>> >> in the Information panel. It works as expected when I try with
>> POINT_DATA.
>> >>
>> >> I just had a doubt: Qt has been updated to 4.6 recently on my
>> computer,
>> >> and
>> >> I don't know if my ParaView is up-to-date according to this...
>> >>
>> >> Let's update, compile, wait and see.
>> >>
>> >> Jerome
>> >>
>> >> 2010/2/28 Favre Jean <jfavre at cscs.ch>
>> >>
>> >> Hi,
>> >> I noted this problem a few days ago.
>> >> A *really* simple example:
>> >> - Add a SphereSource
>> >> - Connect to a "Generate Ids" filter
>> >> - Update: The information panel shows a 'Ids' array for points and for
>> >> cells, but
>> >> the "Color by" doesn't give the choice of cell array.
>> >>
>> >> Jerome
>> >> -----------------
>> >>
>> >> I cannot reproduce the problem. The following code works exactly as
>> >> expected and enables the coloring by POINT_DATA or CELL_DATA
>> >>
>> >> sph1 = Sphere()
>> >> genid1 = GenerateIds(Input=sph1)
>> >> rep1 = Show(genid1)
>> >> rep1.ColorArrayName = 'Ids'
>> >> rep1.ColorAttributeType = 'CELL_DATA'
>> >>
>> >> ----------------
>> >> Jean M. Favre
>> >> Swiss National Supercomputing Center
>> >
>> > _______________________________________________
>> > 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 ParaView Wiki at:
>> > http://paraview.org/Wiki/ParaView
>> >
>> > Follow this link to subscribe/unsubscribe:
>> > http://www.paraview.org/mailman/listinfo/paraview
>> >
>> _______________________________________________
>> 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 ParaView Wiki at:
>> http://paraview.org/Wiki/ParaView
>>
>> Follow this link to subscribe/unsubscribe:
>> http://www.paraview.org/mailman/listinfo/paraview
>>
>
>
>
> _______________________________________________
> 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 ParaView Wiki at: http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview

[Michael Jackson]

Well, this is really getting weird. I have to admit that I have NOT
tried it on anything other than OS X with Qt/Cocoa. I am going to try
with OS X/Qt-Carbon (Qt 4.6.2) and see what happens. I guess I'll try
Windows Next.

--
Mike Jackson <www.bluequartz.net>

On Mar 1, 2010, at 9:52 AM, Sven Buijssen wrote:

> Hi,
>
> I tested my nightly clean builds of ParaView CVS HEAD (pulled via
> git) + Qt 4.5.2 on
> * SuSE Linux Enterprise Server 10 SP2
> * Ubuntu 9.10 Server 64bit
> * openSuSE 10.2 64bit
> * openSuSE 11.1 32bit
> * SPARC Solaris 10 64bit (PV CVS 2010-02-25, nothing newer as I'm
> having some
> trouble with Mesa since Dave's commit Fri Feb 26 17:57:21 2010 +0000)
>
> I see for all builds both point "Ids" and cell "Ids" in the Color By
> menu.
>
> I also checked the tarballs of all nightly builds since beginning of
> February
> 2010 on openSuSE 11.1 and could not reproduce the issue. Cell Ids is
> always there.
>
> Sven
>
>
> Michael Jackson wrote, On 01.03.2010 15:47:
>> Here is a screen shot. I updated to latest cvs and recompiled.
>> Still no
>> joy. I even tried saving the data set to a new file, then loading the
>> file. No luck there either. Any other suggestions are welcome. I do
>> have
>> the data file if you are interested.
>>
>>
>>
>>
>> ___________________________________________________________
>> Mike Jackson www.bluequartz.net
>>
>>
>> On Feb 28, 2010, at 2:01 PM, J?r?me wrote:
>>
>>> Well, I just updated my worktree with the latest CVS. The problem
>>> remains,
>>> both in debug and release. Note that Yum (package manager of Fedora)
>>> installs qt-4.6.2.
>>>
>>> I will trace the code tomorrow monday.
>>>
>>> Jerome
>>>
>>> 2010/2/28 Utkarsh Ayachit <utkarsh.ayachit at kitware.com>
>>> That is weird. I am on Mac with Qt 4.6.0, Cocoa and the latest CVS
>>> ParaView and I see both point "Ids" and cell "Ids" in the Color By
>>> menu. I'll test it out with my linux box at work tomorrow.
>>>
>>> Utkarsh
>>>
>>> On Sun, Feb 28, 2010 at 9:44 AM, Michael Jackson
>>> <mike.jackson at bluequartz.net> wrote:
>>>> Maybe it is Qt versions or ParaView versions? I am using Qt 4.6.1
>>> compiled
>>>> for Cocoa on OS X 10.5.8 Intel. I am using ParaView from <url =
>>>> git://github.com/Kitware/ParaView.git> which was last updated about
>>> a week
>>>> ago. When I get to the office on Monday I can try the same with
>>> ParaView
>>>> 3.6.2 and see if the problem persists.
>>>>
>>>> As far as Data, I _could_ share it, but like Jerome stated, you can
>>>> reproduce it with a simple sphere source, then "Generate
>>>> Normals". The
>>>> complete source code can be found here:
>>>>
>>>>
>>> <href=http://www.bluequartz.net/cgi-bin/gitweb/gitweb.cgi?p=PVMXADataModel.git;a=blob;f=Source/Server/CrystalInterfaceColoringFilter/vtkCrystalInterfaceColoringFilter.cpp;hb=de5d94a3a128a08a534230deca9057e4e57d0358
>>> >
>>>
>>>>
>>>>
>>>> _________________________________________________________
>>>> Mike Jackson mike.jackson at bluequartz.net
>>>>
>>>> On Feb 28, 2010, at 7:51 AM, J?r?me wrote:
>>>>
>>>>> Hi,
>>>>>
>>>>> I confirm the problem: even with your script, CELL_DATA are not
>>> enabled.
>>>>> They don't appear in the "Color by" combo box, but the cell Ids
>>>>> are
>>>>> present
>>>>> in the Information panel. It works as expected when I try with
>>> POINT_DATA.
>>>>>
>>>>> I just had a doubt: Qt has been updated to 4.6 recently on my
>>> computer,
>>>>> and
>>>>> I don't know if my ParaView is up-to-date according to this...
>>>>>
>>>>> Let's update, compile, wait and see.
>>>>>
>>>>> Jerome
>>>>>
>>>>> 2010/2/28 Favre Jean <jfavre at cscs.ch>
>>>>>
>>>>> Hi,
>>>>> I noted this problem a few days ago.
>>>>> A *really* simple example:
>>>>> - Add a SphereSource
>>>>> - Connect to a "Generate Ids" filter
>>>>> - Update: The information panel shows a 'Ids' array for points
>>>>> and for
>>>>> cells, but
>>>>> the "Color by" doesn't give the choice of cell array.
>>>>>
>>>>> Jerome
>>>>> -----------------
>>>>>
>>>>> I cannot reproduce the problem. The following code works exactly
>>>>> as
>>>>> expected and enables the coloring by POINT_DATA or CELL_DATA
>>>>>
>>>>> sph1 = Sphere()
>>>>> genid1 = GenerateIds(Input=sph1)
>>>>> rep1 = Show(genid1)
>>>>> rep1.ColorArrayName = 'Ids'
>>>>> rep1.ColorAttributeType = 'CELL_DATA'
>>>>>
>>>>> ----------------
>>>>> Jean M. Favre
>>>>> Swiss National Supercomputing Center
>>>>
>>>> _______________________________________________
>>>> 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 ParaView Wiki at:
>>>> http://paraview.org/Wiki/ParaView
>>>>
>>>> Follow this link to subscribe/unsubscribe:
>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>
>>> _______________________________________________
>>> 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 ParaView Wiki at:
>>> http://paraview.org/Wiki/ParaView
>>>
>>> Follow this link to subscribe/unsubscribe:
>>> http://www.paraview.org/mailman/listinfo/paraview
>>>
>>
>>
>>
>> _______________________________________________
>> 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 ParaView Wiki at: http://paraview.org/Wiki/ParaView
>>
>> Follow this link to subscribe/unsubscribe:
>> http://www.paraview.org/mailman/listinfo/paraview
>

[Utkarsh Ayachit]

I've tried on Os X (4.6 Cocoa), and Linux and still cannot reproduce
the issue. Mike, if you don't mind a little debugging, I can point you
to the relevant code.

Utkarsh

On Mon, Mar 1, 2010 at 10:20 AM, Michael Jackson
<mike.jackson at bluequartz.net> wrote:
> Well, this is really getting weird. I have to admit that I have NOT tried it
> on anything other than OS X with Qt/Cocoa. I am going to try with OS
> X/Qt-Carbon (Qt 4.6.2) and see what happens. I guess I'll try Windows Next.
>
> --
> Mike Jackson <www.bluequartz.net>
>
> On Mar 1, 2010, at 9:52 AM, Sven Buijssen wrote:
>
>> Hi,
>>
>> I tested my nightly clean builds of ParaView CVS HEAD (pulled via git) +
>> Qt 4.5.2 on
>> * SuSE Linux Enterprise Server 10 SP2
>> * Ubuntu 9.10 Server 64bit
>> * openSuSE 10.2 64bit
>> * openSuSE 11.1 32bit
>> * SPARC Solaris 10 64bit (PV CVS 2010-02-25, nothing newer as I'm having
>> some
>> trouble with Mesa since Dave's commit Fri Feb 26 17:57:21 2010 +0000)
>>
>> I see for all builds both point "Ids" and cell "Ids" in the Color By menu.
>>
>> I also checked the tarballs of all nightly builds since beginning of
>> February
>> 2010 on openSuSE 11.1 and could not reproduce the issue. Cell Ids is
>> always there.
>>
>> Sven
>>
>>
>> Michael Jackson wrote, On 01.03.2010 15:47:
>>>
>>> Here is a screen shot. I updated to latest cvs and recompiled. Still no
>>> joy. I even tried saving the data set to a new file, then loading the
>>> file. No luck there either. Any other suggestions are welcome. I do have
>>> the data file if you are interested.
>>>
>>>
>>>
>>>
>>> ___________________________________________________________
>>> Mike Jackson ? ? ? ? ? ? ? ? ? ? ?www.bluequartz.net
>>>
>>>
>>> On Feb 28, 2010, at 2:01 PM, J?r?me wrote:
>>>
>>>> Well, I just updated my worktree with the latest CVS. The problem
>>>> remains,
>>>> both in debug and release. Note that Yum (package manager of Fedora)
>>>> installs qt-4.6.2.
>>>>
>>>> I will trace the code tomorrow monday.
>>>>
>>>> Jerome
>>>>
>>>> 2010/2/28 Utkarsh Ayachit <utkarsh.ayachit at kitware.com>
>>>> That is weird. I am on Mac with Qt 4.6.0, Cocoa and the latest CVS
>>>> ParaView and I see both point "Ids" and cell "Ids" in the Color By
>>>> menu. I'll test it out with my linux box at work tomorrow.
>>>>
>>>> Utkarsh
>>>>
>>>> On Sun, Feb 28, 2010 at 9:44 AM, Michael Jackson
>>>> <mike.jackson at bluequartz.net> wrote:
>>>>>
>>>>> Maybe it is Qt versions or ParaView versions? I am using Qt 4.6.1
>>>>
>>>> compiled
>>>>>
>>>>> for Cocoa on OS X 10.5.8 Intel. I am using ParaView from <url =
>>>>> git://github.com/Kitware/ParaView.git> which was last updated about
>>>>
>>>> a week
>>>>>
>>>>> ago. When I get to the office on Monday I can try the same with
>>>>
>>>> ParaView
>>>>>
>>>>> 3.6.2 and see if the problem persists.
>>>>>
>>>>> As far as Data, I _could_ share it, but like Jerome stated, you can
>>>>> reproduce it with a simple sphere source, then "Generate Normals". The
>>>>> complete source code can be found here:
>>>>>
>>>>>
>>>>
>>>> <href=http://www.bluequartz.net/cgi-bin/gitweb/gitweb.cgi?p=PVMXADataModel.git;a=blob;f=Source/Server/CrystalInterfaceColoringFilter/vtkCrystalInterfaceColoringFilter.cpp;hb=de5d94a3a128a08a534230deca9057e4e57d0358>
>>>>
>>>>>
>>>>>
>>>>> _________________________________________________________
>>>>> Mike Jackson ? ? ? ? ? ? ? ? ?mike.jackson at bluequartz.net
>>>>>
>>>>> On Feb 28, 2010, at 7:51 AM, J?r?me wrote:
>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> I confirm the problem: even with your script, CELL_DATA are not
>>>>
>>>> enabled.
>>>>>>
>>>>>> They don't appear in the "Color by" combo box, but the cell Ids are
>>>>>> present
>>>>>> in the Information panel. It works as expected when I try with
>>>>
>>>> POINT_DATA.
>>>>>>
>>>>>> I just had a doubt: Qt has been updated to 4.6 recently on my
>>>>
>>>> computer,
>>>>>>
>>>>>> and
>>>>>> I don't know if my ParaView is up-to-date according to this...
>>>>>>
>>>>>> Let's update, compile, wait and see.
>>>>>>
>>>>>> Jerome
>>>>>>
>>>>>> 2010/2/28 Favre Jean <jfavre at cscs.ch>
>>>>>>
>>>>>> Hi,
>>>>>> I noted this problem a few days ago.
>>>>>> A *really* simple example:
>>>>>> - Add a SphereSource
>>>>>> - Connect to a "Generate Ids" filter
>>>>>> - Update: The information panel shows a 'Ids' array for points and for
>>>>>> cells, but
>>>>>> the "Color by" doesn't give the choice of cell array.
>>>>>>
>>>>>> Jerome
>>>>>> -----------------
>>>>>>
>>>>>> I cannot reproduce the problem. The following code works exactly as
>>>>>> expected and enables the coloring by POINT_DATA or CELL_DATA
>>>>>>
>>>>>> sph1 = Sphere()
>>>>>> genid1 = GenerateIds(Input=sph1)
>>>>>> rep1 = Show(genid1)
>>>>>> rep1.ColorArrayName = 'Ids'
>>>>>> rep1.ColorAttributeType = 'CELL_DATA'
>>>>>>
>>>>>> ----------------
>>>>>> Jean M. Favre
>>>>>> Swiss National Supercomputing Center
>>>>>
>>>>> _______________________________________________
>>>>> 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 ParaView Wiki at:
>>>>> http://paraview.org/Wiki/ParaView
>>>>>
>>>>> Follow this link to subscribe/unsubscribe:
>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>
>>>> _______________________________________________
>>>> 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 ParaView Wiki at:
>>>> http://paraview.org/Wiki/ParaView
>>>>
>>>> Follow this link to subscribe/unsubscribe:
>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> 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 ParaView Wiki at:
>>> http://paraview.org/Wiki/ParaView
>>>
>>> Follow this link to subscribe/unsubscribe:
>>> http://www.paraview.org/mailman/listinfo/paraview
>>
>
> _______________________________________________
> 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 ParaView Wiki at:
> http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview
>

[Michael Jackson]

Send me the areas to look at and I'll try to see what is happening.

--
Mike Jackson <www.bluequartz.net>

On Mar 1, 2010, at 10:33 AM, Utkarsh Ayachit wrote:

> I've tried on Os X (4.6 Cocoa), and Linux and still cannot reproduce
> the issue. Mike, if you don't mind a little debugging, I can point you
> to the relevant code.
>
> Utkarsh
>
> On Mon, Mar 1, 2010 at 10:20 AM, Michael Jackson
> <mike.jackson at bluequartz.net> wrote:
>> Well, this is really getting weird. I have to admit that I have NOT
>> tried it
>> on anything other than OS X with Qt/Cocoa. I am going to try with OS
>> X/Qt-Carbon (Qt 4.6.2) and see what happens. I guess I'll try
>> Windows Next.
>>
>> --
>> Mike Jackson <www.bluequartz.net>
>>
>> On Mar 1, 2010, at 9:52 AM, Sven Buijssen wrote:
>>
>>> Hi,
>>>
>>> I tested my nightly clean builds of ParaView CVS HEAD (pulled via
>>> git) +
>>> Qt 4.5.2 on
>>> * SuSE Linux Enterprise Server 10 SP2
>>> * Ubuntu 9.10 Server 64bit
>>> * openSuSE 10.2 64bit
>>> * openSuSE 11.1 32bit
>>> * SPARC Solaris 10 64bit (PV CVS 2010-02-25, nothing newer as I'm
>>> having
>>> some
>>> trouble with Mesa since Dave's commit Fri Feb 26 17:57:21 2010
>>> +0000)
>>>
>>> I see for all builds both point "Ids" and cell "Ids" in the Color
>>> By menu.
>>>
>>> I also checked the tarballs of all nightly builds since beginning of
>>> February
>>> 2010 on openSuSE 11.1 and could not reproduce the issue. Cell Ids is
>>> always there.
>>>
>>> Sven
>>>
>>>
>>> Michael Jackson wrote, On 01.03.2010 15:47:
>>>>
>>>> Here is a screen shot. I updated to latest cvs and recompiled.
>>>> Still no
>>>> joy. I even tried saving the data set to a new file, then loading
>>>> the
>>>> file. No luck there either. Any other suggestions are welcome. I
>>>> do have
>>>> the data file if you are interested.
>>>>
>>>>
>>>>
>>>>
>>>> ___________________________________________________________
>>>> Mike Jackson www.bluequartz.net
>>>>
>>>>
>>>> On Feb 28, 2010, at 2:01 PM, J?r?me wrote:
>>>>
>>>>> Well, I just updated my worktree with the latest CVS. The problem
>>>>> remains,
>>>>> both in debug and release. Note that Yum (package manager of
>>>>> Fedora)
>>>>> installs qt-4.6.2.
>>>>>
>>>>> I will trace the code tomorrow monday.
>>>>>
>>>>> Jerome
>>>>>
>>>>> 2010/2/28 Utkarsh Ayachit <utkarsh.ayachit at kitware.com>
>>>>> That is weird. I am on Mac with Qt 4.6.0, Cocoa and the latest CVS
>>>>> ParaView and I see both point "Ids" and cell "Ids" in the Color By
>>>>> menu. I'll test it out with my linux box at work tomorrow.
>>>>>
>>>>> Utkarsh
>>>>>
>>>>> On Sun, Feb 28, 2010 at 9:44 AM, Michael Jackson
>>>>> <mike.jackson at bluequartz.net> wrote:
>>>>>>
>>>>>> Maybe it is Qt versions or ParaView versions? I am using Qt 4.6.1
>>>>>
>>>>> compiled
>>>>>>
>>>>>> for Cocoa on OS X 10.5.8 Intel. I am using ParaView from <url =
>>>>>> git://github.com/Kitware/ParaView.git> which was last updated
>>>>>> about
>>>>>
>>>>> a week
>>>>>>
>>>>>> ago. When I get to the office on Monday I can try the same with
>>>>>
>>>>> ParaView
>>>>>>
>>>>>> 3.6.2 and see if the problem persists.
>>>>>>
>>>>>> As far as Data, I _could_ share it, but like Jerome stated, you
>>>>>> can
>>>>>> reproduce it with a simple sphere source, then "Generate
>>>>>> Normals". The
>>>>>> complete source code can be found here:
>>>>>>
>>>>>>
>>>>>
>>>>> <href=http://www.bluequartz.net/cgi-bin/gitweb/gitweb.cgi?p=PVMXADataModel.git;a=blob;f=Source/Server/CrystalInterfaceColoringFilter/vtkCrystalInterfaceColoringFilter.cpp;hb=de5d94a3a128a08a534230deca9057e4e57d0358
>>>>> >
>>>>>
>>>>>>
>>>>>>
>>>>>> _________________________________________________________
>>>>>> Mike Jackson mike.jackson at bluequartz.net
>>>>>>
>>>>>> On Feb 28, 2010, at 7:51 AM, J?r?me wrote:
>>>>>>
>>>>>>> Hi,
>>>>>>>
>>>>>>> I confirm the problem: even with your script, CELL_DATA are not
>>>>>
>>>>> enabled.
>>>>>>>
>>>>>>> They don't appear in the "Color by" combo box, but the cell
>>>>>>> Ids are
>>>>>>> present
>>>>>>> in the Information panel. It works as expected when I try with
>>>>>
>>>>> POINT_DATA.
>>>>>>>
>>>>>>> I just had a doubt: Qt has been updated to 4.6 recently on my
>>>>>
>>>>> computer,
>>>>>>>
>>>>>>> and
>>>>>>> I don't know if my ParaView is up-to-date according to this...
>>>>>>>
>>>>>>> Let's update, compile, wait and see.
>>>>>>>
>>>>>>> Jerome
>>>>>>>
>>>>>>> 2010/2/28 Favre Jean <jfavre at cscs.ch>
>>>>>>>
>>>>>>> Hi,
>>>>>>> I noted this problem a few days ago.
>>>>>>> A *really* simple example:
>>>>>>> - Add a SphereSource
>>>>>>> - Connect to a "Generate Ids" filter
>>>>>>> - Update: The information panel shows a 'Ids' array for points
>>>>>>> and for
>>>>>>> cells, but
>>>>>>> the "Color by" doesn't give the choice of cell array.
>>>>>>>
>>>>>>> Jerome
>>>>>>> -----------------
>>>>>>>
>>>>>>> I cannot reproduce the problem. The following code works
>>>>>>> exactly as
>>>>>>> expected and enables the coloring by POINT_DATA or CELL_DATA
>>>>>>>
>>>>>>> sph1 = Sphere()
>>>>>>> genid1 = GenerateIds(Input=sph1)
>>>>>>> rep1 = Show(genid1)
>>>>>>> rep1.ColorArrayName = 'Ids'
>>>>>>> rep1.ColorAttributeType = 'CELL_DATA'
>>>>>>>
>>>>>>> ----------------
>>>>>>> Jean M. Favre
>>>>>>> Swiss National Supercomputing Center
>>>>>>
>>>>>> _______________________________________________
>>>>>> 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 ParaView Wiki at:
>>>>>> http://paraview.org/Wiki/ParaView
>>>>>>
>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>>
>>>>> _______________________________________________
>>>>> 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 ParaView Wiki at:
>>>>> http://paraview.org/Wiki/ParaView
>>>>>
>>>>> Follow this link to subscribe/unsubscribe:
>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> 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 ParaView Wiki at:
>>>> http://paraview.org/Wiki/ParaView
>>>>
>>>> Follow this link to subscribe/unsubscribe:
>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>
>>
>> _______________________________________________
>> 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 ParaView Wiki at:
>> http://paraview.org/Wiki/ParaView
>>
>> Follow this link to subscribe/unsubscribe:
>> http://www.paraview.org/mailman/listinfo/paraview
>>

[Utkarsh Ayachit]

The the widget is Qt/Component/pqDisplayColorWidget.h|cxx.

Look at pqDisplayColorWidget::reloadGUI(). In this method it gets the
list of arrays and then populates the combo-box.

Utkarsh

On Mon, Mar 1, 2010 at 10:41 AM, Michael Jackson
<mike.jackson at bluequartz.net> wrote:
> Send me the areas to look at and I'll try to see what is happening.
>
> --
> Mike Jackson <www.bluequartz.net>
>
> On Mar 1, 2010, at 10:33 AM, Utkarsh Ayachit wrote:
>
>> I've tried on Os X (4.6 Cocoa), ?and Linux and still cannot reproduce
>> the issue. Mike, if you don't mind a little debugging, I can point you
>> to the relevant code.
>>
>> Utkarsh
>>
>> On Mon, Mar 1, 2010 at 10:20 AM, Michael Jackson
>> <mike.jackson at bluequartz.net> wrote:
>>>
>>> Well, this is really getting weird. I have to admit that I have NOT tried
>>> it
>>> on anything other than OS X with Qt/Cocoa. I am going to try with OS
>>> X/Qt-Carbon (Qt 4.6.2) and see what happens. I guess I'll try Windows
>>> Next.
>>>
>>> --
>>> Mike Jackson <www.bluequartz.net>
>>>
>>> On Mar 1, 2010, at 9:52 AM, Sven Buijssen wrote:
>>>
>>>> Hi,
>>>>
>>>> I tested my nightly clean builds of ParaView CVS HEAD (pulled via git) +
>>>> Qt 4.5.2 on
>>>> * SuSE Linux Enterprise Server 10 SP2
>>>> * Ubuntu 9.10 Server 64bit
>>>> * openSuSE 10.2 64bit
>>>> * openSuSE 11.1 32bit
>>>> * SPARC Solaris 10 64bit (PV CVS 2010-02-25, nothing newer as I'm having
>>>> some
>>>> trouble with Mesa since Dave's commit Fri Feb 26 17:57:21 2010 +0000)
>>>>
>>>> I see for all builds both point "Ids" and cell "Ids" in the Color By
>>>> menu.
>>>>
>>>> I also checked the tarballs of all nightly builds since beginning of
>>>> February
>>>> 2010 on openSuSE 11.1 and could not reproduce the issue. Cell Ids is
>>>> always there.
>>>>
>>>> Sven
>>>>
>>>>
>>>> Michael Jackson wrote, On 01.03.2010 15:47:
>>>>>
>>>>> Here is a screen shot. I updated to latest cvs and recompiled. Still no
>>>>> joy. I even tried saving the data set to a new file, then loading the
>>>>> file. No luck there either. Any other suggestions are welcome. I do
>>>>> have
>>>>> the data file if you are interested.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> ___________________________________________________________
>>>>> Mike Jackson ? ? ? ? ? ? ? ? ? ? ?www.bluequartz.net
>>>>>
>>>>>
>>>>> On Feb 28, 2010, at 2:01 PM, J?r?me wrote:
>>>>>
>>>>>> Well, I just updated my worktree with the latest CVS. The problem
>>>>>> remains,
>>>>>> both in debug and release. Note that Yum (package manager of Fedora)
>>>>>> installs qt-4.6.2.
>>>>>>
>>>>>> I will trace the code tomorrow monday.
>>>>>>
>>>>>> Jerome
>>>>>>
>>>>>> 2010/2/28 Utkarsh Ayachit <utkarsh.ayachit at kitware.com>
>>>>>> That is weird. I am on Mac with Qt 4.6.0, Cocoa and the latest CVS
>>>>>> ParaView and I see both point "Ids" and cell "Ids" in the Color By
>>>>>> menu. I'll test it out with my linux box at work tomorrow.
>>>>>>
>>>>>> Utkarsh
>>>>>>
>>>>>> On Sun, Feb 28, 2010 at 9:44 AM, Michael Jackson
>>>>>> <mike.jackson at bluequartz.net> wrote:
>>>>>>>
>>>>>>> Maybe it is Qt versions or ParaView versions? I am using Qt 4.6.1
>>>>>>
>>>>>> compiled
>>>>>>>
>>>>>>> for Cocoa on OS X 10.5.8 Intel. I am using ParaView from <url =
>>>>>>> git://github.com/Kitware/ParaView.git> which was last updated about
>>>>>>
>>>>>> a week
>>>>>>>
>>>>>>> ago. When I get to the office on Monday I can try the same with
>>>>>>
>>>>>> ParaView
>>>>>>>
>>>>>>> 3.6.2 and see if the problem persists.
>>>>>>>
>>>>>>> As far as Data, I _could_ share it, but like Jerome stated, you can
>>>>>>> reproduce it with a simple sphere source, then "Generate Normals".
>>>>>>> The
>>>>>>> complete source code can be found here:
>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>> <href=http://www.bluequartz.net/cgi-bin/gitweb/gitweb.cgi?p=PVMXADataModel.git;a=blob;f=Source/Server/CrystalInterfaceColoringFilter/vtkCrystalInterfaceColoringFilter.cpp;hb=de5d94a3a128a08a534230deca9057e4e57d0358>
>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> _________________________________________________________
>>>>>>> Mike Jackson ? ? ? ? ? ? ? ? ?mike.jackson at bluequartz.net
>>>>>>>
>>>>>>> On Feb 28, 2010, at 7:51 AM, J?r?me wrote:
>>>>>>>
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> I confirm the problem: even with your script, CELL_DATA are not
>>>>>>
>>>>>> enabled.
>>>>>>>>
>>>>>>>> They don't appear in the "Color by" combo box, but the cell Ids are
>>>>>>>> present
>>>>>>>> in the Information panel. It works as expected when I try with
>>>>>>
>>>>>> POINT_DATA.
>>>>>>>>
>>>>>>>> I just had a doubt: Qt has been updated to 4.6 recently on my
>>>>>>
>>>>>> computer,
>>>>>>>>
>>>>>>>> and
>>>>>>>> I don't know if my ParaView is up-to-date according to this...
>>>>>>>>
>>>>>>>> Let's update, compile, wait and see.
>>>>>>>>
>>>>>>>> Jerome
>>>>>>>>
>>>>>>>> 2010/2/28 Favre Jean <jfavre at cscs.ch>
>>>>>>>>
>>>>>>>> Hi,
>>>>>>>> I noted this problem a few days ago.
>>>>>>>> A *really* simple example:
>>>>>>>> - Add a SphereSource
>>>>>>>> - Connect to a "Generate Ids" filter
>>>>>>>> - Update: The information panel shows a 'Ids' array for points and
>>>>>>>> for
>>>>>>>> cells, but
>>>>>>>> the "Color by" doesn't give the choice of cell array.
>>>>>>>>
>>>>>>>> Jerome
>>>>>>>> -----------------
>>>>>>>>
>>>>>>>> I cannot reproduce the problem. The following code works exactly as
>>>>>>>> expected and enables the coloring by POINT_DATA or CELL_DATA
>>>>>>>>
>>>>>>>> sph1 = Sphere()
>>>>>>>> genid1 = GenerateIds(Input=sph1)
>>>>>>>> rep1 = Show(genid1)
>>>>>>>> rep1.ColorArrayName = 'Ids'
>>>>>>>> rep1.ColorAttributeType = 'CELL_DATA'
>>>>>>>>
>>>>>>>> ----------------
>>>>>>>> Jean M. Favre
>>>>>>>> Swiss National Supercomputing Center
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> 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 ParaView Wiki at:
>>>>>>> http://paraview.org/Wiki/ParaView
>>>>>>>
>>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>>>
>>>>>> _______________________________________________
>>>>>> 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 ParaView Wiki at:
>>>>>> http://paraview.org/Wiki/ParaView
>>>>>>
>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>>
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> 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 ParaView Wiki at:
>>>>> http://paraview.org/Wiki/ParaView
>>>>>
>>>>> Follow this link to subscribe/unsubscribe:
>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>
>>>
>>> _______________________________________________
>>> 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 ParaView Wiki at:
>>> http://paraview.org/Wiki/ParaView
>>>
>>> Follow this link to subscribe/unsubscribe:
>>> http://www.paraview.org/mailman/listinfo/paraview
>>>
>
>

[Michael Jackson]

Just to follow up on this. I downloaded the precompiled ParaView 3.6.2
for OS X (Running on OS X 10.5.8/Intel). This is how I reproduced the
problem.

Launch ParaView.
Select a Sphere source. Note that there already is a Point Array
called "Normals" and the Color By drop down will show it.

Run the "PointDataToCellData" filter.
Note that now there is a Cell Based "Normals" array but the Color By
drop down ONLY lists "Solid Color". Is this correct behavior?

Just FYI, I compiled ParaView 3.6.2 on Windows 7 with VS 2008 and
running the same exact sequence I get 2 arrays under the Color By drop
down menu. "cellNormals" and "Normals". Where "Normals" shows up under
the "Data Arrays" list as Cell Based Arrays.

And so it goes..
--
Mike Jackson <www.bluequartz.net>

On Mar 1, 2010, at 10:44 AM, Utkarsh Ayachit wrote:

> The the widget is Qt/Component/pqDisplayColorWidget.h|cxx.
>
> Look at pqDisplayColorWidget::reloadGUI(). In this method it gets the
> list of arrays and then populates the combo-box.
>
> Utkarsh
>
> On Mon, Mar 1, 2010 at 10:41 AM, Michael Jackson
> <mike.jackson at bluequartz.net> wrote:
>> Send me the areas to look at and I'll try to see what is happening.
>>
>> --
>> Mike Jackson <www.bluequartz.net>
>>
>> On Mar 1, 2010, at 10:33 AM, Utkarsh Ayachit wrote:
>>
>>> I've tried on Os X (4.6 Cocoa), and Linux and still cannot
>>> reproduce
>>> the issue. Mike, if you don't mind a little debugging, I can point
>>> you
>>> to the relevant code.
>>>
>>> Utkarsh
>>>
>>> On Mon, Mar 1, 2010 at 10:20 AM, Michael Jackson
>>> <mike.jackson at bluequartz.net> wrote:
>>>>
>>>> Well, this is really getting weird. I have to admit that I have
>>>> NOT tried
>>>> it
>>>> on anything other than OS X with Qt/Cocoa. I am going to try with
>>>> OS
>>>> X/Qt-Carbon (Qt 4.6.2) and see what happens. I guess I'll try
>>>> Windows
>>>> Next.
>>>>
>>>> --
>>>> Mike Jackson <www.bluequartz.net>
>>>>
>>>> On Mar 1, 2010, at 9:52 AM, Sven Buijssen wrote:
>>>>
>>>>> Hi,
>>>>>
>>>>> I tested my nightly clean builds of ParaView CVS HEAD (pulled
>>>>> via git) +
>>>>> Qt 4.5.2 on
>>>>> * SuSE Linux Enterprise Server 10 SP2
>>>>> * Ubuntu 9.10 Server 64bit
>>>>> * openSuSE 10.2 64bit
>>>>> * openSuSE 11.1 32bit
>>>>> * SPARC Solaris 10 64bit (PV CVS 2010-02-25, nothing newer as
>>>>> I'm having
>>>>> some
>>>>> trouble with Mesa since Dave's commit Fri Feb 26 17:57:21 2010
>>>>> +0000)
>>>>>
>>>>> I see for all builds both point "Ids" and cell "Ids" in the
>>>>> Color By
>>>>> menu.
>>>>>
>>>>> I also checked the tarballs of all nightly builds since
>>>>> beginning of
>>>>> February
>>>>> 2010 on openSuSE 11.1 and could not reproduce the issue. Cell
>>>>> Ids is
>>>>> always there.
>>>>>
>>>>> Sven
>>>>>
>>>>>
>>>>> Michael Jackson wrote, On 01.03.2010 15:47:
>>>>>>
>>>>>> Here is a screen shot. I updated to latest cvs and recompiled.
>>>>>> Still no
>>>>>> joy. I even tried saving the data set to a new file, then
>>>>>> loading the
>>>>>> file. No luck there either. Any other suggestions are welcome.
>>>>>> I do
>>>>>> have
>>>>>> the data file if you are interested.
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> ___________________________________________________________
>>>>>> Mike Jackson www.bluequartz.net
>>>>>>
>>>>>>
>>>>>> On Feb 28, 2010, at 2:01 PM, J?r?me wrote:
>>>>>>
>>>>>>> Well, I just updated my worktree with the latest CVS. The
>>>>>>> problem
>>>>>>> remains,
>>>>>>> both in debug and release. Note that Yum (package manager of
>>>>>>> Fedora)
>>>>>>> installs qt-4.6.2.
>>>>>>>
>>>>>>> I will trace the code tomorrow monday.
>>>>>>>
>>>>>>> Jerome
>>>>>>>
>>>>>>> 2010/2/28 Utkarsh Ayachit <utkarsh.ayachit at kitware.com>
>>>>>>> That is weird. I am on Mac with Qt 4.6.0, Cocoa and the latest
>>>>>>> CVS
>>>>>>> ParaView and I see both point "Ids" and cell "Ids" in the
>>>>>>> Color By
>>>>>>> menu. I'll test it out with my linux box at work tomorrow.
>>>>>>>
>>>>>>> Utkarsh
>>>>>>>
>>>>>>> On Sun, Feb 28, 2010 at 9:44 AM, Michael Jackson
>>>>>>> <mike.jackson at bluequartz.net> wrote:
>>>>>>>>
>>>>>>>> Maybe it is Qt versions or ParaView versions? I am using Qt
>>>>>>>> 4.6.1
>>>>>>>
>>>>>>> compiled
>>>>>>>>
>>>>>>>> for Cocoa on OS X 10.5.8 Intel. I am using ParaView from <url =
>>>>>>>> git://github.com/Kitware/ParaView.git> which was last updated
>>>>>>>> about
>>>>>>>
>>>>>>> a week
>>>>>>>>
>>>>>>>> ago. When I get to the office on Monday I can try the same with
>>>>>>>
>>>>>>> ParaView
>>>>>>>>
>>>>>>>> 3.6.2 and see if the problem persists.
>>>>>>>>
>>>>>>>> As far as Data, I _could_ share it, but like Jerome stated,
>>>>>>>> you can
>>>>>>>> reproduce it with a simple sphere source, then "Generate
>>>>>>>> Normals".
>>>>>>>> The
>>>>>>>> complete source code can be found here:
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> <href=http://www.bluequartz.net/cgi-bin/gitweb/gitweb.cgi?p=PVMXADataModel.git;a=blob;f=Source/Server/CrystalInterfaceColoringFilter/vtkCrystalInterfaceColoringFilter.cpp;hb=de5d94a3a128a08a534230deca9057e4e57d0358
>>>>>>> >
>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> _________________________________________________________
>>>>>>>> Mike Jackson mike.jackson at bluequartz.net
>>>>>>>>
>>>>>>>> On Feb 28, 2010, at 7:51 AM, J?r?me wrote:
>>>>>>>>
>>>>>>>>> Hi,
>>>>>>>>>
>>>>>>>>> I confirm the problem: even with your script, CELL_DATA are
>>>>>>>>> not
>>>>>>>
>>>>>>> enabled.
>>>>>>>>>
>>>>>>>>> They don't appear in the "Color by" combo box, but the cell
>>>>>>>>> Ids are
>>>>>>>>> present
>>>>>>>>> in the Information panel. It works as expected when I try with
>>>>>>>
>>>>>>> POINT_DATA.
>>>>>>>>>
>>>>>>>>> I just had a doubt: Qt has been updated to 4.6 recently on my
>>>>>>>
>>>>>>> computer,
>>>>>>>>>
>>>>>>>>> and
>>>>>>>>> I don't know if my ParaView is up-to-date according to this...
>>>>>>>>>
>>>>>>>>> Let's update, compile, wait and see.
>>>>>>>>>
>>>>>>>>> Jerome
>>>>>>>>>
>>>>>>>>> 2010/2/28 Favre Jean <jfavre at cscs.ch>
>>>>>>>>>
>>>>>>>>> Hi,
>>>>>>>>> I noted this problem a few days ago.
>>>>>>>>> A *really* simple example:
>>>>>>>>> - Add a SphereSource
>>>>>>>>> - Connect to a "Generate Ids" filter
>>>>>>>>> - Update: The information panel shows a 'Ids' array for
>>>>>>>>> points and
>>>>>>>>> for
>>>>>>>>> cells, but
>>>>>>>>> the "Color by" doesn't give the choice of cell array.
>>>>>>>>>
>>>>>>>>> Jerome
>>>>>>>>> -----------------
>>>>>>>>>
>>>>>>>>> I cannot reproduce the problem. The following code works
>>>>>>>>> exactly as
>>>>>>>>> expected and enables the coloring by POINT_DATA or CELL_DATA
>>>>>>>>>
>>>>>>>>> sph1 = Sphere()
>>>>>>>>> genid1 = GenerateIds(Input=sph1)
>>>>>>>>> rep1 = Show(genid1)
>>>>>>>>> rep1.ColorArrayName = 'Ids'
>>>>>>>>> rep1.ColorAttributeType = 'CELL_DATA'
>>>>>>>>>
>>>>>>>>> ----------------
>>>>>>>>> Jean M. Favre
>>>>>>>>> Swiss National Supercomputing Center
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> 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 ParaView Wiki at:
>>>>>>>> http://paraview.org/Wiki/ParaView
>>>>>>>>
>>>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> 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 ParaView Wiki at:
>>>>>>> http://paraview.org/Wiki/ParaView
>>>>>>>
>>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> 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 ParaView Wiki at:
>>>>>> http://paraview.org/Wiki/ParaView
>>>>>>
>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>
>>>>
>>>> _______________________________________________
>>>> 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 ParaView Wiki at:
>>>> http://paraview.org/Wiki/ParaView
>>>>
>>>> Follow this link to subscribe/unsubscribe:
>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>
>>
>>

[Michael Jackson]

So I figured out where the problem lies, but not what the actual
problem is:

QList<QString> pqPipelineRepresentation::getColorFields()
....
for(int i=0; i<cellinfo->GetNumberOfArrays(); i++)

that loop is 0 to 0 on OS X and 0 to 2 on Windows. But using the exact
same data file. So for some reason the OS X version is NOT storing/
updating something properly. Odd.

--
Mike Jackson <www.bluequartz.net>

On Mar 1, 2010, at 10:44 AM, Utkarsh Ayachit wrote:

> The the widget is Qt/Component/pqDisplayColorWidget.h|cxx.
>
> Look at pqDisplayColorWidget::reloadGUI(). In this method it gets the
> list of arrays and then populates the combo-box.
>
> Utkarsh
>
> On Mon, Mar 1, 2010 at 10:41 AM, Michael Jackson
> <mike.jackson at bluequartz.net> wrote:
>> Send me the areas to look at and I'll try to see what is happening.
>>
>> --
>> Mike Jackson <www.bluequartz.net>
>>
>> On Mar 1, 2010, at 10:33 AM, Utkarsh Ayachit wrote:
>>
>>> I've tried on Os X (4.6 Cocoa), and Linux and still cannot
>>> reproduce
>>> the issue. Mike, if you don't mind a little debugging, I can point
>>> you
>>> to the relevant code.
>>>
>>> Utkarsh
>>>
>>> On Mon, Mar 1, 2010 at 10:20 AM, Michael Jackson
>>> <mike.jackson at bluequartz.net> wrote:
>>>>
>>>> Well, this is really getting weird. I have to admit that I have
>>>> NOT tried
>>>> it
>>>> on anything other than OS X with Qt/Cocoa. I am going to try with
>>>> OS
>>>> X/Qt-Carbon (Qt 4.6.2) and see what happens. I guess I'll try
>>>> Windows
>>>> Next.
>>>>
>>>> --
>>>> Mike Jackson <www.bluequartz.net>
>>>>
>>>> On Mar 1, 2010, at 9:52 AM, Sven Buijssen wrote:
>>>>
>>>>> Hi,
>>>>>
>>>>> I tested my nightly clean builds of ParaView CVS HEAD (pulled
>>>>> via git) +
>>>>> Qt 4.5.2 on
>>>>> * SuSE Linux Enterprise Server 10 SP2
>>>>> * Ubuntu 9.10 Server 64bit
>>>>> * openSuSE 10.2 64bit
>>>>> * openSuSE 11.1 32bit
>>>>> * SPARC Solaris 10 64bit (PV CVS 2010-02-25, nothing newer as
>>>>> I'm having
>>>>> some
>>>>> trouble with Mesa since Dave's commit Fri Feb 26 17:57:21 2010
>>>>> +0000)
>>>>>
>>>>> I see for all builds both point "Ids" and cell "Ids" in the
>>>>> Color By
>>>>> menu.
>>>>>
>>>>> I also checked the tarballs of all nightly builds since
>>>>> beginning of
>>>>> February
>>>>> 2010 on openSuSE 11.1 and could not reproduce the issue. Cell
>>>>> Ids is
>>>>> always there.
>>>>>
>>>>> Sven
>>>>>
>>>>>
>>>>> Michael Jackson wrote, On 01.03.2010 15:47:
>>>>>>
>>>>>> Here is a screen shot. I updated to latest cvs and recompiled.
>>>>>> Still no
>>>>>> joy. I even tried saving the data set to a new file, then
>>>>>> loading the
>>>>>> file. No luck there either. Any other suggestions are welcome.
>>>>>> I do
>>>>>> have
>>>>>> the data file if you are interested.
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> ___________________________________________________________
>>>>>> Mike Jackson www.bluequartz.net
>>>>>>
>>>>>>
>>>>>> On Feb 28, 2010, at 2:01 PM, J?r?me wrote:
>>>>>>
>>>>>>> Well, I just updated my worktree with the latest CVS. The
>>>>>>> problem
>>>>>>> remains,
>>>>>>> both in debug and release. Note that Yum (package manager of
>>>>>>> Fedora)
>>>>>>> installs qt-4.6.2.
>>>>>>>
>>>>>>> I will trace the code tomorrow monday.
>>>>>>>
>>>>>>> Jerome
>>>>>>>
>>>>>>> 2010/2/28 Utkarsh Ayachit <utkarsh.ayachit at kitware.com>
>>>>>>> That is weird. I am on Mac with Qt 4.6.0, Cocoa and the latest
>>>>>>> CVS
>>>>>>> ParaView and I see both point "Ids" and cell "Ids" in the
>>>>>>> Color By
>>>>>>> menu. I'll test it out with my linux box at work tomorrow.
>>>>>>>
>>>>>>> Utkarsh
>>>>>>>
>>>>>>> On Sun, Feb 28, 2010 at 9:44 AM, Michael Jackson
>>>>>>> <mike.jackson at bluequartz.net> wrote:
>>>>>>>>
>>>>>>>> Maybe it is Qt versions or ParaView versions? I am using Qt
>>>>>>>> 4.6.1
>>>>>>>
>>>>>>> compiled
>>>>>>>>
>>>>>>>> for Cocoa on OS X 10.5.8 Intel. I am using ParaView from <url =
>>>>>>>> git://github.com/Kitware/ParaView.git> which was last updated
>>>>>>>> about
>>>>>>>
>>>>>>> a week
>>>>>>>>
>>>>>>>> ago. When I get to the office on Monday I can try the same with
>>>>>>>
>>>>>>> ParaView
>>>>>>>>
>>>>>>>> 3.6.2 and see if the problem persists.
>>>>>>>>
>>>>>>>> As far as Data, I _could_ share it, but like Jerome stated,
>>>>>>>> you can
>>>>>>>> reproduce it with a simple sphere source, then "Generate
>>>>>>>> Normals".
>>>>>>>> The
>>>>>>>> complete source code can be found here:
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> <href=http://www.bluequartz.net/cgi-bin/gitweb/gitweb.cgi?p=PVMXADataModel.git;a=blob;f=Source/Server/CrystalInterfaceColoringFilter/vtkCrystalInterfaceColoringFilter.cpp;hb=de5d94a3a128a08a534230deca9057e4e57d0358
>>>>>>> >
>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> _________________________________________________________
>>>>>>>> Mike Jackson mike.jackson at bluequartz.net
>>>>>>>>
>>>>>>>> On Feb 28, 2010, at 7:51 AM, J?r?me wrote:
>>>>>>>>
>>>>>>>>> Hi,
>>>>>>>>>
>>>>>>>>> I confirm the problem: even with your script, CELL_DATA are
>>>>>>>>> not
>>>>>>>
>>>>>>> enabled.
>>>>>>>>>
>>>>>>>>> They don't appear in the "Color by" combo box, but the cell
>>>>>>>>> Ids are
>>>>>>>>> present
>>>>>>>>> in the Information panel. It works as expected when I try with
>>>>>>>
>>>>>>> POINT_DATA.
>>>>>>>>>
>>>>>>>>> I just had a doubt: Qt has been updated to 4.6 recently on my
>>>>>>>
>>>>>>> computer,
>>>>>>>>>
>>>>>>>>> and
>>>>>>>>> I don't know if my ParaView is up-to-date according to this...
>>>>>>>>>
>>>>>>>>> Let's update, compile, wait and see.
>>>>>>>>>
>>>>>>>>> Jerome
>>>>>>>>>
>>>>>>>>> 2010/2/28 Favre Jean <jfavre at cscs.ch>
>>>>>>>>>
>>>>>>>>> Hi,
>>>>>>>>> I noted this problem a few days ago.
>>>>>>>>> A *really* simple example:
>>>>>>>>> - Add a SphereSource
>>>>>>>>> - Connect to a "Generate Ids" filter
>>>>>>>>> - Update: The information panel shows a 'Ids' array for
>>>>>>>>> points and
>>>>>>>>> for
>>>>>>>>> cells, but
>>>>>>>>> the "Color by" doesn't give the choice of cell array.
>>>>>>>>>
>>>>>>>>> Jerome
>>>>>>>>> -----------------
>>>>>>>>>
>>>>>>>>> I cannot reproduce the problem. The following code works
>>>>>>>>> exactly as
>>>>>>>>> expected and enables the coloring by POINT_DATA or CELL_DATA
>>>>>>>>>
>>>>>>>>> sph1 = Sphere()
>>>>>>>>> genid1 = GenerateIds(Input=sph1)
>>>>>>>>> rep1 = Show(genid1)
>>>>>>>>> rep1.ColorArrayName = 'Ids'
>>>>>>>>> rep1.ColorAttributeType = 'CELL_DATA'
>>>>>>>>>
>>>>>>>>> ----------------
>>>>>>>>> Jean M. Favre
>>>>>>>>> Swiss National Supercomputing Center
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> 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 ParaView Wiki at:
>>>>>>>> http://paraview.org/Wiki/ParaView
>>>>>>>>
>>>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> 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 ParaView Wiki at:
>>>>>>> http://paraview.org/Wiki/ParaView
>>>>>>>
>>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> 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 ParaView Wiki at:
>>>>>> http://paraview.org/Wiki/ParaView
>>>>>>
>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>
>>>>
>>>> _______________________________________________
>>>> 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 ParaView Wiki at:
>>>> http://paraview.org/Wiki/ParaView
>>>>
>>>> Follow this link to subscribe/unsubscribe:
>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>
>>
>>

[Jérôme]

Hi,

After full updates on 2 computers (personal and office) both running fedora
12,
the problem persists... on one of them only! The sadness is that the
personal
one is concerned, so I cannot debug efficiently.

The point is that I didn't perform a comprehensive clean on my built tree.
This
is the next step.

Jerome

2010/3/1 Michael Jackson <mike.jackson at bluequartz.net>

> So I figured out where the problem lies, but not what the actual problem
> is:
>
> QList<QString> pqPipelineRepresentation::getColorFields()
> ....
> for(int i=0; i<cellinfo->GetNumberOfArrays(); i++)
>
> that loop is 0 to 0 on OS X and 0 to 2 on Windows. But using the exact same
> data file. So for some reason the OS X version is NOT storing/updating
> something properly. Odd.
>
>
> --
> Mike Jackson <www.bluequartz.net>
>
> On Mar 1, 2010, at 10:44 AM, Utkarsh Ayachit wrote:
>
> The the widget is Qt/Component/pqDisplayColorWidget.h|cxx.
>>
>> Look at pqDisplayColorWidget::reloadGUI(). In this method it gets the
>> list of arrays and then populates the combo-box.
>>
>> Utkarsh
>>
>> On Mon, Mar 1, 2010 at 10:41 AM, Michael Jackson
>> <mike.jackson at bluequartz.net> wrote:
>>
>>> Send me the areas to look at and I'll try to see what is happening.
>>>
>>> --
>>> Mike Jackson <www.bluequartz.net>
>>>
>>> On Mar 1, 2010, at 10:33 AM, Utkarsh Ayachit wrote:
>>>
>>> I've tried on Os X (4.6 Cocoa), and Linux and still cannot reproduce
>>>> the issue. Mike, if you don't mind a little debugging, I can point you
>>>> to the relevant code.
>>>>
>>>> Utkarsh
>>>>
>>>> On Mon, Mar 1, 2010 at 10:20 AM, Michael Jackson
>>>> <mike.jackson at bluequartz.net> wrote:
>>>>
>>>>>
>>>>> Well, this is really getting weird. I have to admit that I have NOT
>>>>> tried
>>>>> it
>>>>> on anything other than OS X with Qt/Cocoa. I am going to try with OS
>>>>> X/Qt-Carbon (Qt 4.6.2) and see what happens. I guess I'll try Windows
>>>>> Next.
>>>>>
>>>>> --
>>>>> Mike Jackson <www.bluequartz.net>
>>>>>
>>>>> On Mar 1, 2010, at 9:52 AM, Sven Buijssen wrote:
>>>>>
>>>>> Hi,
>>>>>>
>>>>>> I tested my nightly clean builds of ParaView CVS HEAD (pulled via git)
>>>>>> +
>>>>>> Qt 4.5.2 on
>>>>>> * SuSE Linux Enterprise Server 10 SP2
>>>>>> * Ubuntu 9.10 Server 64bit
>>>>>> * openSuSE 10.2 64bit
>>>>>> * openSuSE 11.1 32bit
>>>>>> * SPARC Solaris 10 64bit (PV CVS 2010-02-25, nothing newer as I'm
>>>>>> having
>>>>>> some
>>>>>> trouble with Mesa since Dave's commit Fri Feb 26 17:57:21 2010 +0000)
>>>>>>
>>>>>> I see for all builds both point "Ids" and cell "Ids" in the Color By
>>>>>> menu.
>>>>>>
>>>>>> I also checked the tarballs of all nightly builds since beginning of
>>>>>> February
>>>>>> 2010 on openSuSE 11.1 and could not reproduce the issue. Cell Ids is
>>>>>> always there.
>>>>>>
>>>>>> Sven
>>>>>>
>>>>>>
>>>>>> Michael Jackson wrote, On 01.03.2010 15:47:
>>>>>>
>>>>>>>
>>>>>>> Here is a screen shot. I updated to latest cvs and recompiled. Still
>>>>>>> no
>>>>>>> joy. I even tried saving the data set to a new file, then loading the
>>>>>>> file. No luck there either. Any other suggestions are welcome. I do
>>>>>>> have
>>>>>>> the data file if you are interested.
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> ___________________________________________________________
>>>>>>> Mike Jackson www.bluequartz.net
>>>>>>>
>>>>>>>
>>>>>>> On Feb 28, 2010, at 2:01 PM, J?r?me wrote:
>>>>>>>
>>>>>>> Well, I just updated my worktree with the latest CVS. The problem
>>>>>>>> remains,
>>>>>>>> both in debug and release. Note that Yum (package manager of Fedora)
>>>>>>>> installs qt-4.6.2.
>>>>>>>>
>>>>>>>> I will trace the code tomorrow monday.
>>>>>>>>
>>>>>>>> Jerome
>>>>>>>>
>>>>>>>> 2010/2/28 Utkarsh Ayachit <utkarsh.ayachit at kitware.com>
>>>>>>>> That is weird. I am on Mac with Qt 4.6.0, Cocoa and the latest CVS
>>>>>>>> ParaView and I see both point "Ids" and cell "Ids" in the Color By
>>>>>>>> menu. I'll test it out with my linux box at work tomorrow.
>>>>>>>>
>>>>>>>> Utkarsh
>>>>>>>>
>>>>>>>> On Sun, Feb 28, 2010 at 9:44 AM, Michael Jackson
>>>>>>>> <mike.jackson at bluequartz.net> wrote:
>>>>>>>>
>>>>>>>>>
>>>>>>>>> Maybe it is Qt versions or ParaView versions? I am using Qt 4.6.1
>>>>>>>>>
>>>>>>>>
>>>>>>>> compiled
>>>>>>>>
>>>>>>>>>
>>>>>>>>> for Cocoa on OS X 10.5.8 Intel. I am using ParaView from <url =
>>>>>>>>> git://github.com/Kitware/ParaView.git> which was last updated
>>>>>>>>> about
>>>>>>>>>
>>>>>>>>
>>>>>>>> a week
>>>>>>>>
>>>>>>>>>
>>>>>>>>> ago. When I get to the office on Monday I can try the same with
>>>>>>>>>
>>>>>>>>
>>>>>>>> ParaView
>>>>>>>>
>>>>>>>>>
>>>>>>>>> 3.6.2 and see if the problem persists.
>>>>>>>>>
>>>>>>>>> As far as Data, I _could_ share it, but like Jerome stated, you can
>>>>>>>>> reproduce it with a simple sphere source, then "Generate Normals".
>>>>>>>>> The
>>>>>>>>> complete source code can be found here:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>> <href=
>>>>>>>> http://www.bluequartz.net/cgi-bin/gitweb/gitweb.cgi?p=PVMXADataModel.git;a=blob;f=Source/Server/CrystalInterfaceColoringFilter/vtkCrystalInterfaceColoringFilter.cpp;hb=de5d94a3a128a08a534230deca9057e4e57d0358
>>>>>>>> >
>>>>>>>>
>>>>>>>>
>>>>>>>>>
>>>>>>>>> _________________________________________________________
>>>>>>>>> Mike Jackson mike.jackson at bluequartz.net
>>>>>>>>>
>>>>>>>>> On Feb 28, 2010, at 7:51 AM, J?r?me wrote:
>>>>>>>>>
>>>>>>>>> Hi,
>>>>>>>>>>
>>>>>>>>>> I confirm the problem: even with your script, CELL_DATA are not
>>>>>>>>>>
>>>>>>>>>
>>>>>>>> enabled.
>>>>>>>>
>>>>>>>>>
>>>>>>>>>> They don't appear in the "Color by" combo box, but the cell Ids
>>>>>>>>>> are
>>>>>>>>>> present
>>>>>>>>>> in the Information panel. It works as expected when I try with
>>>>>>>>>>
>>>>>>>>>
>>>>>>>> POINT_DATA.
>>>>>>>>
>>>>>>>>>
>>>>>>>>>> I just had a doubt: Qt has been updated to 4.6 recently on my
>>>>>>>>>>
>>>>>>>>>
>>>>>>>> computer,
>>>>>>>>
>>>>>>>>>
>>>>>>>>>> and
>>>>>>>>>> I don't know if my ParaView is up-to-date according to this...
>>>>>>>>>>
>>>>>>>>>> Let's update, compile, wait and see.
>>>>>>>>>>
>>>>>>>>>> Jerome
>>>>>>>>>>
>>>>>>>>>> 2010/2/28 Favre Jean <jfavre at cscs.ch>
>>>>>>>>>>
>>>>>>>>>> Hi,
>>>>>>>>>> I noted this problem a few days ago.
>>>>>>>>>> A *really* simple example:
>>>>>>>>>> - Add a SphereSource
>>>>>>>>>> - Connect to a "Generate Ids" filter
>>>>>>>>>> - Update: The information panel shows a 'Ids' array for points and
>>>>>>>>>> for
>>>>>>>>>> cells, but
>>>>>>>>>> the "Color by" doesn't give the choice of cell array.
>>>>>>>>>>
>>>>>>>>>> Jerome
>>>>>>>>>> -----------------
>>>>>>>>>>
>>>>>>>>>> I cannot reproduce the problem. The following code works exactly
>>>>>>>>>> as
>>>>>>>>>> expected and enables the coloring by POINT_DATA or CELL_DATA
>>>>>>>>>>
>>>>>>>>>> sph1 = Sphere()
>>>>>>>>>> genid1 = GenerateIds(Input=sph1)
>>>>>>>>>> rep1 = Show(genid1)
>>>>>>>>>> rep1.ColorArrayName = 'Ids'
>>>>>>>>>> rep1.ColorAttributeType = 'CELL_DATA'
>>>>>>>>>>
>>>>>>>>>> ----------------
>>>>>>>>>> Jean M. Favre
>>>>>>>>>> Swiss National Supercomputing Center
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>> _______________________________________________
>>>>>>>>> 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 ParaView Wiki at:
>>>>>>>>> http://paraview.org/Wiki/ParaView
>>>>>>>>>
>>>>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>>>>>
>>>>>>>>> _______________________________________________
>>>>>>>> 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 ParaView Wiki at:
>>>>>>>> http://paraview.org/Wiki/ParaView
>>>>>>>>
>>>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> 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 ParaView Wiki at:
>>>>>>> http://paraview.org/Wiki/ParaView
>>>>>>>
>>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>>>
>>>>>>
>>>>>>
>>>>> _______________________________________________
>>>>> 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 ParaView Wiki at:
>>>>> http://paraview.org/Wiki/ParaView
>>>>>
>>>>> Follow this link to subscribe/unsubscribe:
>>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>>
>>>>>
>>>
>>>
> _______________________________________________
> 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 ParaView Wiki at:
> http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20100303/0f5e6ad0/attachment.htm>

[Michael Jackson]

I have tried clean builds on ParaView CVS and ParaView 3.6.2 OS X with
Qt versions 4.5.3 and Qt 4.6.2 both carbon and Cocoa and none of that
alleviates the problem. Not even the precompiled binary from Kitware's
web site works correctly. It does work perfectly fine on windows though.

I wonder where the loading of the file takes place. I have a feeling
something is not loading the cell data correctly. Maybe.

--
Mike Jackson

On 3/3/2010 3:01 AM, J?r?me wrote:
> Hi,
>
> After full updates on 2 computers (personal and office) both running
> fedora 12,
> the problem persists... on one of them only! The sadness is that the
> personal
> one is concerned, so I cannot debug efficiently.
>
> The point is that I didn't perform a comprehensive clean on my built
> tree. This
> is the next step.
>
> Jerome
>
> 2010/3/1 Michael Jackson <mike.jackson at bluequartz.net
> <mailto:mike.jackson at bluequartz.net>>
>
> So I figured out where the problem lies, but not what the actual
> problem is:
>
> QList<QString> pqPipelineRepresentation::getColorFields()
> ....
> for(int i=0; i<cellinfo->GetNumberOfArrays(); i++)
>
> that loop is 0 to 0 on OS X and 0 to 2 on Windows. But using the
> exact same data file. So for some reason the OS X version is NOT
> storing/updating something properly. Odd.
>
>

[Eric E. Monson]

Mike, is it possible that it's a problem related to OS X 10.5? I know J?r?me is having some trouble with Fedora, but I can't reproduce the issue on Snow Leopard or Ubuntu (both x86_64 builds). Which architecture are you building PV and Qt for?

-Eric

On Mar 3, 2010, at 9:09 AM, Michael Jackson wrote:

> I have tried clean builds on ParaView CVS and ParaView 3.6.2 OS X with Qt versions 4.5.3 and Qt 4.6.2 both carbon and Cocoa and none of that alleviates the problem. Not even the precompiled binary from Kitware's web site works correctly. It does work perfectly fine on windows though.
>
> I wonder where the loading of the file takes place. I have a feeling something is not loading the cell data correctly. Maybe.
>
> --
> Mike Jackson
>
> On 3/3/2010 3:01 AM, J?r?me wrote:
>> Hi,
>>
>> After full updates on 2 computers (personal and office) both running fedora 12,
>> the problem persists... on one of them only! The sadness is that the personal
>> one is concerned, so I cannot debug efficiently.
>>
>> The point is that I didn't perform a comprehensive clean on my built tree. This
>> is the next step.
>>
>> Jerome
>>
>> 2010/3/1 Michael Jackson <mike.jackson at bluequartz.net <mailto:mike.jackson at bluequartz.net>>
>>
>> So I figured out where the problem lies, but not what the actual
>> problem is:
>>
>> QList<QString> pqPipelineRepresentation::getColorFields()
>> ....
>> for(int i=0; i<cellinfo->GetNumberOfArrays(); i++)
>>
>> that loop is 0 to 0 on OS X and 0 to 2 on Windows. But using the
>> exact same data file. So for some reason the OS X version is NOT
>> storing/updating something properly. Odd.
>>
>>
>
> _______________________________________________
> 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 ParaView Wiki at: http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview

[Utkarsh Ayachit]

Mike, but the cell data shows up correctly on the information tab,
right? So the cell data is being loaded, and even reported to
ParaView. It's hard to imagine why it would be doing what it's doing,
esp. with the precompiled binaries on Mac, since they work fine for
me, on my Mac 10.6 :).

It cannot be a mere GUI issue since from the debugging report that
Mike gave us before, ParaView indeed doesn't report those cell arrays
to the client when rendering. I just wish I could reproduce it somehow
somewhere so I could debug it :). I am going to test on Os X 10.5 to
see if we have any better luck.

Utkarsh

On Wed, Mar 3, 2010 at 9:23 AM, Eric E. Monson <emonson at cs.duke.edu> wrote:
> Mike, is it possible that it's a problem related to OS X 10.5? I know J?r?me is having some trouble with Fedora, but I can't reproduce the issue on Snow Leopard or Ubuntu (both x86_64 builds). Which architecture are you building PV and Qt for?
>
> -Eric
>
> On Mar 3, 2010, at 9:09 AM, Michael Jackson wrote:
>
>> I have tried clean builds on ParaView CVS and ParaView 3.6.2 OS X with Qt versions 4.5.3 and Qt 4.6.2 both carbon and Cocoa and none of that alleviates the problem. Not even the precompiled binary from Kitware's web site works correctly. It does work perfectly fine on windows though.
>>
>> ?I wonder where the loading of the file takes place. I have a feeling something is not loading the cell data correctly. Maybe.
>>
>> --
>> Mike Jackson
>>
>> On 3/3/2010 3:01 AM, J?r?me wrote:
>>> Hi,
>>>
>>> After full updates on 2 computers (personal and office) both running fedora 12,
>>> the problem persists... on one of them only! The sadness is that the personal
>>> one is concerned, so I cannot debug efficiently.
>>>
>>> The point is that I didn't perform a comprehensive clean on my built tree. This
>>> is the next step.
>>>
>>> Jerome
>>>
>>> 2010/3/1 Michael Jackson <mike.jackson at bluequartz.net <mailto:mike.jackson at bluequartz.net>>
>>>
>>> ? ?So I figured out where the problem lies, but not what the actual
>>> ? ?problem is:
>>>
>>> ? ?QList<QString> pqPipelineRepresentation::getColorFields()
>>> ? ?....
>>> ? ?for(int i=0; i<cellinfo->GetNumberOfArrays(); i++)
>>>
>>> ? ?that loop is 0 to 0 on OS X and 0 to 2 on Windows. But using the
>>> ? ?exact same data file. So for some reason the OS X version is NOT
>>> ? ?storing/updating something properly. Odd.
>>>
>>>
>>
>> _______________________________________________
>> 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 ParaView Wiki at: http://paraview.org/Wiki/ParaView
>>
>> Follow this link to subscribe/unsubscribe:
>> http://www.paraview.org/mailman/listinfo/paraview
>
> _______________________________________________
> 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 ParaView Wiki at: http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview
>

[Michael Jackson]

Do you want my data file?

Mike J.

On 3/3/2010 9:28 AM, Utkarsh Ayachit wrote:
> Mike, but the cell data shows up correctly on the information tab,
> right? So the cell data is being loaded, and even reported to
> ParaView. It's hard to imagine why it would be doing what it's doing,
> esp. with the precompiled binaries on Mac, since they work fine for
> me, on my Mac 10.6 :).
>
> It cannot be a mere GUI issue since from the debugging report that
> Mike gave us before, ParaView indeed doesn't report those cell arrays
> to the client when rendering. I just wish I could reproduce it somehow
> somewhere so I could debug it :). I am going to test on Os X 10.5 to
> see if we have any better luck.
>
> Utkarsh
>
> On Wed, Mar 3, 2010 at 9:23 AM, Eric E. Monson<emonson at cs.duke.edu> wrote:
>
>> Mike, is it possible that it's a problem related to OS X 10.5? I know J?r?me is having some trouble with Fedora, but I can't reproduce the issue on Snow Leopard or Ubuntu (both x86_64 builds). Which architecture are you building PV and Qt for?
>>
>> -Eric
>>
>> On Mar 3, 2010, at 9:09 AM, Michael Jackson wrote:
>>
>>
>>> I have tried clean builds on ParaView CVS and ParaView 3.6.2 OS X with Qt versions 4.5.3 and Qt 4.6.2 both carbon and Cocoa and none of that alleviates the problem. Not even the precompiled binary from Kitware's web site works correctly. It does work perfectly fine on windows though.
>>>
>>> I wonder where the loading of the file takes place. I have a feeling something is not loading the cell data correctly. Maybe.
>>>
>>> --
>>> Mike Jackson
>>>
>>> On 3/3/2010 3:01 AM, J?r?me wrote:
>>>
>>>> Hi,
>>>>
>>>> After full updates on 2 computers (personal and office) both running fedora 12,
>>>> the problem persists... on one of them only! The sadness is that the personal
>>>> one is concerned, so I cannot debug efficiently.
>>>>
>>>> The point is that I didn't perform a comprehensive clean on my built tree. This
>>>> is the next step.
>>>>
>>>> Jerome
>>>>
>>>> 2010/3/1 Michael Jackson<mike.jackson at bluequartz.net<mailto:mike.jackson at bluequartz.net>>
>>>>
>>>> So I figured out where the problem lies, but not what the actual
>>>> problem is:
>>>>
>>>> QList<QString> pqPipelineRepresentation::getColorFields()
>>>> ....
>>>> for(int i=0; i<cellinfo->GetNumberOfArrays(); i++)
>>>>
>>>> that loop is 0 to 0 on OS X and 0 to 2 on Windows. But using the
>>>> exact same data file. So for some reason the OS X version is NOT
>>>> storing/updating something properly. Odd.
>>>>
>>>>
>>>>
>>> _______________________________________________
>>> 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 ParaView Wiki at: http://paraview.org/Wiki/ParaView
>>>
>>> Follow this link to subscribe/unsubscribe:
>>> http://www.paraview.org/mailman/listinfo/paraview
>>>
>> _______________________________________________
>> 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 ParaView Wiki at: http://paraview.org/Wiki/ParaView
>>
>> Follow this link to subscribe/unsubscribe:
>> http://www.paraview.org/mailman/listinfo/paraview
>>
>>

[Utkarsh Ayachit]

Mmm..not sure if that'd help, since you said you can reproduce the
issue with GenerateIds filter as well, right?

Utkarsh

On Wed, Mar 3, 2010 at 9:51 AM, Michael Jackson
<mike.jackson at bluequartz.net> wrote:
> Do you want my data file?
>
> Mike J.
>
> On 3/3/2010 9:28 AM, Utkarsh Ayachit wrote:
>>
>> Mike, but the cell data shows up correctly on the information tab,
>> right? So the cell data is being loaded, and even reported to
>> ParaView. It's hard to imagine why it would be doing what it's doing,
>> esp. with the precompiled binaries on Mac, since they work fine for
>> me, on my Mac 10.6 ?:).
>>
>> It cannot be a mere GUI issue since from the debugging report that
>> Mike gave us before, ParaView indeed doesn't report those cell arrays
>> to the client when rendering. I just wish I could reproduce it somehow
>> somewhere so I could debug it ?:). I am going to test on Os X 10.5 to
>> see if we have any better luck.
>>
>> Utkarsh
>>
>> On Wed, Mar 3, 2010 at 9:23 AM, Eric E. Monson<emonson at cs.duke.edu>
>> ?wrote:
>>
>>>
>>> Mike, is it possible that it's a problem related to OS X 10.5? I know
>>> J?r?me is having some trouble with Fedora, but I can't reproduce the issue
>>> on Snow Leopard or Ubuntu (both x86_64 builds). Which architecture are you
>>> building PV and Qt for?
>>>
>>> -Eric
>>>
>>> On Mar 3, 2010, at 9:09 AM, Michael Jackson wrote:
>>>
>>>
>>>>
>>>> I have tried clean builds on ParaView CVS and ParaView 3.6.2 OS X with
>>>> Qt versions 4.5.3 and Qt 4.6.2 both carbon and Cocoa and none of that
>>>> alleviates the problem. Not even the precompiled binary from Kitware's web
>>>> site works correctly. It does work perfectly fine on windows though.
>>>>
>>>> ?I wonder where the loading of the file takes place. I have a feeling
>>>> something is not loading the cell data correctly. Maybe.
>>>>
>>>> --
>>>> Mike Jackson
>>>>
>>>> On 3/3/2010 3:01 AM, J?r?me wrote:
>>>>
>>>>>
>>>>> Hi,
>>>>>
>>>>> After full updates on 2 computers (personal and office) both running
>>>>> fedora 12,
>>>>> the problem persists... on one of them only! The sadness is that the
>>>>> personal
>>>>> one is concerned, so I cannot debug efficiently.
>>>>>
>>>>> The point is that I didn't perform a comprehensive clean on my built
>>>>> tree. This
>>>>> is the next step.
>>>>>
>>>>> Jerome
>>>>>
>>>>> 2010/3/1 Michael
>>>>> Jackson<mike.jackson at bluequartz.net<mailto:mike.jackson at bluequartz.net>>
>>>>>
>>>>> ? ?So I figured out where the problem lies, but not what the actual
>>>>> ? ?problem is:
>>>>>
>>>>> ? ?QList<QString> ?pqPipelineRepresentation::getColorFields()
>>>>> ? ?....
>>>>> ? ?for(int i=0; i<cellinfo->GetNumberOfArrays(); i++)
>>>>>
>>>>> ? ?that loop is 0 to 0 on OS X and 0 to 2 on Windows. But using the
>>>>> ? ?exact same data file. So for some reason the OS X version is NOT
>>>>> ? ?storing/updating something properly. Odd.
>>>>>
>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> 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 ParaView Wiki at:
>>>> http://paraview.org/Wiki/ParaView
>>>>
>>>> Follow this link to subscribe/unsubscribe:
>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>
>>>
>>> _______________________________________________
>>> 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 ParaView Wiki at:
>>> http://paraview.org/Wiki/ParaView
>>>
>>> Follow this link to subscribe/unsubscribe:
>>> http://www.paraview.org/mailman/listinfo/paraview
>>>
>>>
>
> _______________________________________________
> 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 ParaView Wiki at:
> http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview
>

[Jérôme]

By my side, yes: SphereSource -> GenerateIds. You are right: Information
tab shows the celldata and pointdata array, whereas "Color by" has only
point.

Jerome

2010/3/3 Utkarsh Ayachit <utkarsh.ayachit at kitware.com>

> Mmm..not sure if that'd help, since you said you can reproduce the
> issue with GenerateIds filter as well, right?
>
> Utkarsh
>
> On Wed, Mar 3, 2010 at 9:51 AM, Michael Jackson
> <mike.jackson at bluequartz.net> wrote:
> > Do you want my data file?
> >
> > Mike J.
> >
> > On 3/3/2010 9:28 AM, Utkarsh Ayachit wrote:
> >>
> >> Mike, but the cell data shows up correctly on the information tab,
> >> right? So the cell data is being loaded, and even reported to
> >> ParaView. It's hard to imagine why it would be doing what it's doing,
> >> esp. with the precompiled binaries on Mac, since they work fine for
> >> me, on my Mac 10.6 :).
> >>
> >> It cannot be a mere GUI issue since from the debugging report that
> >> Mike gave us before, ParaView indeed doesn't report those cell arrays
> >> to the client when rendering. I just wish I could reproduce it somehow
> >> somewhere so I could debug it :). I am going to test on Os X 10.5 to
> >> see if we have any better luck.
> >>
> >> Utkarsh
> >>
> >> On Wed, Mar 3, 2010 at 9:23 AM, Eric E. Monson<emonson at cs.duke.edu>
> >> wrote:
> >>
> >>>
> >>> Mike, is it possible that it's a problem related to OS X 10.5? I know
> >>> J?r?me is having some trouble with Fedora, but I can't reproduce the
> issue
> >>> on Snow Leopard or Ubuntu (both x86_64 builds). Which architecture are
> you
> >>> building PV and Qt for?
> >>>
> >>> -Eric
> >>>
> >>> On Mar 3, 2010, at 9:09 AM, Michael Jackson wrote:
> >>>
> >>>
> >>>>
> >>>> I have tried clean builds on ParaView CVS and ParaView 3.6.2 OS X with
> >>>> Qt versions 4.5.3 and Qt 4.6.2 both carbon and Cocoa and none of that
> >>>> alleviates the problem. Not even the precompiled binary from Kitware's
> web
> >>>> site works correctly. It does work perfectly fine on windows though.
> >>>>
> >>>> I wonder where the loading of the file takes place. I have a feeling
> >>>> something is not loading the cell data correctly. Maybe.
> >>>>
> >>>> --
> >>>> Mike Jackson
> >>>>
> >>>> On 3/3/2010 3:01 AM, J?r?me wrote:
> >>>>
> >>>>>
> >>>>> Hi,
> >>>>>
> >>>>> After full updates on 2 computers (personal and office) both running
> >>>>> fedora 12,
> >>>>> the problem persists... on one of them only! The sadness is that the
> >>>>> personal
> >>>>> one is concerned, so I cannot debug efficiently.
> >>>>>
> >>>>> The point is that I didn't perform a comprehensive clean on my built
> >>>>> tree. This
> >>>>> is the next step.
> >>>>>
> >>>>> Jerome
> >>>>>
> >>>>> 2010/3/1 Michael
> >>>>> Jackson<mike.jackson at bluequartz.net<mailto:
> mike.jackson at bluequartz.net>>
> >>>>>
> >>>>> So I figured out where the problem lies, but not what the actual
> >>>>> problem is:
> >>>>>
> >>>>> QList<QString> pqPipelineRepresentation::getColorFields()
> >>>>> ....
> >>>>> for(int i=0; i<cellinfo->GetNumberOfArrays(); i++)
> >>>>>
> >>>>> that loop is 0 to 0 on OS X and 0 to 2 on Windows. But using the
> >>>>> exact same data file. So for some reason the OS X version is NOT
> >>>>> storing/updating something properly. Odd.
> >>>>>
> >>>>>
> >>>>>
> >>>>
> >>>> _______________________________________________
> >>>> 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 ParaView Wiki at:
> >>>> http://paraview.org/Wiki/ParaView
> >>>>
> >>>> Follow this link to subscribe/unsubscribe:
> >>>> http://www.paraview.org/mailman/listinfo/paraview
> >>>>
> >>>
> >>> _______________________________________________
> >>> 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 ParaView Wiki at:
> >>> http://paraview.org/Wiki/ParaView
> >>>
> >>> Follow this link to subscribe/unsubscribe:
> >>> http://www.paraview.org/mailman/listinfo/paraview
> >>>
> >>>
> >
> > _______________________________________________
> > 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 ParaView Wiki at:
> > http://paraview.org/Wiki/ParaView
> >
> > Follow this link to subscribe/unsubscribe:
> > http://www.paraview.org/mailman/listinfo/paraview
> >
> _______________________________________________
> 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 ParaView Wiki at:
> http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20100303/80b07f74/attachment-0001.htm>

[Eric E. Monson]

I don't have a CVS build on OS X 10.5.8 any more, but I booted into it and ran the 3.6.2 binary and SphereSource->GenerateIds shows all four arrays in Color By and three in Information.

-Eric



On Mar 3, 2010, at 10:30 AM, J?r?me wrote:

> By my side, yes: SphereSource -> GenerateIds. You are right: Information
> tab shows the celldata and pointdata array, whereas "Color by" has only point.
>
> Jerome
>
> 2010/3/3 Utkarsh Ayachit <utkarsh.ayachit at kitware.com>
> Mmm..not sure if that'd help, since you said you can reproduce the
> issue with GenerateIds filter as well, right?
>
> Utkarsh
>
> On Wed, Mar 3, 2010 at 9:51 AM, Michael Jackson
> <mike.jackson at bluequartz.net> wrote:
> > Do you want my data file?
> >
> > Mike J.
> >
> > On 3/3/2010 9:28 AM, Utkarsh Ayachit wrote:
> >>
> >> Mike, but the cell data shows up correctly on the information tab,
> >> right? So the cell data is being loaded, and even reported to
> >> ParaView. It's hard to imagine why it would be doing what it's doing,
> >> esp. with the precompiled binaries on Mac, since they work fine for
> >> me, on my Mac 10.6 :).
> >>
> >> It cannot be a mere GUI issue since from the debugging report that
> >> Mike gave us before, ParaView indeed doesn't report those cell arrays
> >> to the client when rendering. I just wish I could reproduce it somehow
> >> somewhere so I could debug it :). I am going to test on Os X 10.5 to
> >> see if we have any better luck.
> >>
> >> Utkarsh

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20100303/9006a213/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: sphereIds.jpg
Type: image/jpg
Size: 22016 bytes
Desc: not available
URL: <http://www.paraview.org/pipermail/paraview/attachments/20100303/9006a213/attachment-0001.jpg>

[Michael Jackson]

I just fire up the 3.6.2 binary from Kitware on OS X 10.5.8 on a MacBook
Pro with nVidea 8600M graphics card and the information tab is identical
to your image BUT my "Color By" drop down only shows "Solid Color",
"Ids", "Normals", all of which are "Point Data" and NOT "Cell Data".

So not really sure what is up. If someone wants to remotely work on
my machine email me off list and I'll get you setup.

Mike Jackson.

On 3/3/2010 12:06 PM, Eric E. Monson wrote:
> I don't have a CVS build on OS X 10.5.8 any more, but I booted into it
> and ran the 3.6.2 binary and SphereSource->GenerateIds shows all four
> arrays in Color By and three in Information.
>
> -Eric
>
>
> On Mar 3, 2010, at 10:30 AM, J?r?me wrote:
>
>> By my side, yes: SphereSource -> GenerateIds. You are right: Information
>> tab shows the celldata and pointdata array, whereas "Color by" has
>> only point.
>>
>> Jerome
>>
>> 2010/3/3 Utkarsh Ayachit <utkarsh.ayachit at kitware.com
>> <mailto:utkarsh.ayachit at kitware.com>>
>>
>> Mmm..not sure if that'd help, since you said you can reproduce the
>> issue with GenerateIds filter as well, right?
>>
>> Utkarsh
>>
>> On Wed, Mar 3, 2010 at 9:51 AM, Michael Jackson
>> <mike.jackson at bluequartz.net
>> <mailto:mike.jackson at bluequartz.net>> wrote:
>> > Do you want my data file?
>> >
>> > Mike J.
>> >
>> > On 3/3/2010 9:28 AM, Utkarsh Ayachit wrote:
>> >>
>> >> Mike, but the cell data shows up correctly on the information tab,
>> >> right? So the cell data is being loaded, and even reported to
>> >> ParaView. It's hard to imagine why it would be doing what it's
>> doing,
>> >> esp. with the precompiled binaries on Mac, since they work
>> fine for
>> >> me, on my Mac 10.6 :).
>> >>
>> >> It cannot be a mere GUI issue since from the debugging report that
>> >> Mike gave us before, ParaView indeed doesn't report those cell
>> arrays
>> >> to the client when rendering. I just wish I could reproduce it
>> somehow
>> >> somewhere so I could debug it :). I am going to test on Os X
>> 10.5 to
>> >> see if we have any better luck.
>> >>
>> >> Utkarsh
>>
>

[Eric E. Monson]

I just tried the 3.6.2 binary on a MacBook booted off an old 10.5.7 usb drive backup and all of the Color By entries showed up just fine. The MacBook has the GMA 950 on-board graphics chipset. The Mac Pro I used before has the NVidia GeForce 8800 GT card. If you can think of any other specs that might be crucial, just drop me a line.

-Eric

On Mar 3, 2010, at 12:18 PM, Michael Jackson wrote:

> I just fire up the 3.6.2 binary from Kitware on OS X 10.5.8 on a MacBook Pro with nVidea 8600M graphics card and the information tab is identical to your image BUT my "Color By" drop down only shows "Solid Color", "Ids", "Normals", all of which are "Point Data" and NOT "Cell Data".
>
> So not really sure what is up. If someone wants to remotely work on my machine email me off list and I'll get you setup.
>
> Mike Jackson.
>
> On 3/3/2010 12:06 PM, Eric E. Monson wrote:
>> I don't have a CVS build on OS X 10.5.8 any more, but I booted into it and ran the 3.6.2 binary and SphereSource->GenerateIds shows all four arrays in Color By and three in Information.
>>
>> -Eric
>>
>>
>> On Mar 3, 2010, at 10:30 AM, J?r?me wrote:
>>
>>> By my side, yes: SphereSource -> GenerateIds. You are right: Information
>>> tab shows the celldata and pointdata array, whereas "Color by" has only point.
>>>
>>> Jerome
>>>
>>> 2010/3/3 Utkarsh Ayachit <utkarsh.ayachit at kitware.com <mailto:utkarsh.ayachit at kitware.com>>
>>>
>>> Mmm..not sure if that'd help, since you said you can reproduce the
>>> issue with GenerateIds filter as well, right?
>>>
>>> Utkarsh
>>>
>>> On Wed, Mar 3, 2010 at 9:51 AM, Michael Jackson
>>> <mike.jackson at bluequartz.net
>>> <mailto:mike.jackson at bluequartz.net>> wrote:
>>> > Do you want my data file?
>>> >
>>> > Mike J.
>>> >
>>> > On 3/3/2010 9:28 AM, Utkarsh Ayachit wrote:
>>> >>
>>> >> Mike, but the cell data shows up correctly on the information tab,
>>> >> right? So the cell data is being loaded, and even reported to
>>> >> ParaView. It's hard to imagine why it would be doing what it's
>>> doing,
>>> >> esp. with the precompiled binaries on Mac, since they work
>>> fine for
>>> >> me, on my Mac 10.6 :).
>>> >>
>>> >> It cannot be a mere GUI issue since from the debugging report that
>>> >> Mike gave us before, ParaView indeed doesn't report those cell
>>> arrays
>>> >> to the client when rendering. I just wish I could reproduce it
>>> somehow
>>> >> somewhere so I could debug it :). I am going to test on Os X
>>> 10.5 to
>>> >> see if we have any better luck.
>>> >>
>>> >> Utkarsh
>>>
>>
>
> _______________________________________________
> 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 ParaView Wiki at: http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview

[Michael Jackson]

Well. I am beginning to think it is something with my system. I just
tried the PV 3.6.1 on an OS X 10.4.11 ppc machine and everything works
fine. I tried the PV 3.6.1 Intel version on my MacBook Pro running
10.5.8 and the same sequence did NOT work, ie, Cell Centered data was
NOT shown in the Color By drop down list. Odd. I'm going to try a clean
install of OS X and see what happens.

Mike Jackson.

On 3/3/2010 12:39 PM, Eric E. Monson wrote:
> I just tried the 3.6.2 binary on a MacBook booted off an old 10.5.7 usb drive backup and all of the Color By entries showed up just fine. The MacBook has the GMA 950 on-board graphics chipset. The Mac Pro I used before has the NVidia GeForce 8800 GT card. If you can think of any other specs that might be crucial, just drop me a line.
>
> -Eric
>
> On Mar 3, 2010, at 12:18 PM, Michael Jackson wrote:
>
>
>> I just fire up the 3.6.2 binary from Kitware on OS X 10.5.8 on a MacBook Pro with nVidea 8600M graphics card and the information tab is identical to your image BUT my "Color By" drop down only shows "Solid Color", "Ids", "Normals", all of which are "Point Data" and NOT "Cell Data".
>>
>> So not really sure what is up. If someone wants to remotely work on my machine email me off list and I'll get you setup.
>>
>> Mike Jackson.
>>
>> On 3/3/2010 12:06 PM, Eric E. Monson wrote:
>>
>>> I don't have a CVS build on OS X 10.5.8 any more, but I booted into it and ran the 3.6.2 binary and SphereSource->GenerateIds shows all four arrays in Color By and three in Information.
>>>
>>> -Eric
>>>
>>>
>>> On Mar 3, 2010, at 10:30 AM, J?r?me wrote:
>>>
>>>
>>>> By my side, yes: SphereSource -> GenerateIds. You are right: Information
>>>> tab shows the celldata and pointdata array, whereas "Color by" has only point.
>>>>
>>>> Jerome
>>>>
>>>> 2010/3/3 Utkarsh Ayachit<utkarsh.ayachit at kitware.com<mailto:utkarsh.ayachit at kitware.com>>
>>>>
>>>> Mmm..not sure if that'd help, since you said you can reproduce the
>>>> issue with GenerateIds filter as well, right?
>>>>
>>>> Utkarsh
>>>>
>>>> On Wed, Mar 3, 2010 at 9:51 AM, Michael Jackson
>>>> <mike.jackson at bluequartz.net
>>>> <mailto:mike.jackson at bluequartz.net>> wrote:
>>>> > Do you want my data file?
>>>> >
>>>> > Mike J.
>>>> >
>>>> > On 3/3/2010 9:28 AM, Utkarsh Ayachit wrote:
>>>> >>
>>>> >> Mike, but the cell data shows up correctly on the information tab,
>>>> >> right? So the cell data is being loaded, and even reported to
>>>> >> ParaView. It's hard to imagine why it would be doing what it's
>>>> doing,
>>>> >> esp. with the precompiled binaries on Mac, since they work
>>>> fine for
>>>> >> me, on my Mac 10.6 :).
>>>> >>
>>>> >> It cannot be a mere GUI issue since from the debugging report that
>>>> >> Mike gave us before, ParaView indeed doesn't report those cell
>>>> arrays
>>>> >> to the client when rendering. I just wish I could reproduce it
>>>> somehow
>>>> >> somewhere so I could debug it :). I am going to test on Os X
>>>> 10.5 to
>>>> >> see if we have any better luck.
>>>> >>
>>>> >> Utkarsh

[Michael Jackson]

Drum roll please....... "Corrupt" set of paraview preferences. After
installing 10.5.8 clean from my installation disks everything worked
as it was supposed to. So that means the only difference is between
the 2 user environments. Preferences is the first to check. Looked in
my home directory, blew away the .config directory, launched paraview
on the "trouble" machine and now everything works. I could completely
replicate this issue simply by swapping in and out the "good" and
"bad" preferences for ParaView. I have the "bad" preferences if anyone
wants them to try and debug what was going on.

Thanks to EVERYONE who helped me with ideas.

_________________________________________________________
Mike Jackson mike.jackson at bluequartz.net


On Mar 3, 2010, at 12:39 PM, Eric E. Monson wrote:

> I just tried the 3.6.2 binary on a MacBook booted off an old 10.5.7
> usb drive backup and all of the Color By entries showed up just
> fine. The MacBook has the GMA 950 on-board graphics chipset. The Mac
> Pro I used before has the NVidia GeForce 8800 GT card. If you can
> think of any other specs that might be crucial, just drop me a line.
>
> -Eric
>
> On Mar 3, 2010, at 12:18 PM, Michael Jackson wrote:
>
>> I just fire up the 3.6.2 binary from Kitware on OS X 10.5.8 on a
>> MacBook Pro with nVidea 8600M graphics card and the information tab
>> is identical to your image BUT my "Color By" drop down only shows
>> "Solid Color", "Ids", "Normals", all of which are "Point Data" and
>> NOT "Cell Data".
>>
>> So not really sure what is up. If someone wants to remotely work on
>> my machine email me off list and I'll get you setup.
>>
>> Mike Jackson.
>>
>> On 3/3/2010 12:06 PM, Eric E. Monson wrote:
>>> I don't have a CVS build on OS X 10.5.8 any more, but I booted
>>> into it and ran the 3.6.2 binary and SphereSource->GenerateIds
>>> shows all four arrays in Color By and three in Information.
>>>
>>> -Eric
>>>
>>>
>>> On Mar 3, 2010, at 10:30 AM, J?r?me wrote:
>>>
>>>> By my side, yes: SphereSource -> GenerateIds. You are right:
>>>> Information
>>>> tab shows the celldata and pointdata array, whereas "Color by"
>>>> has only point.
>>>>
>>>> Jerome
>>>>
>>>> 2010/3/3 Utkarsh Ayachit <utkarsh.ayachit at kitware.com <mailto:utkarsh.ayachit at kitware.com
>>>> >>
>>>>
>>>> Mmm..not sure if that'd help, since you said you can reproduce
>>>> the
>>>> issue with GenerateIds filter as well, right?
>>>>
>>>> Utkarsh
>>>>
>>>> On Wed, Mar 3, 2010 at 9:51 AM, Michael Jackson
>>>> <mike.jackson at bluequartz.net
>>>> <mailto:mike.jackson at bluequartz.net>> wrote:
>>>>> Do you want my data file?
>>>>>
>>>>> Mike J.
>>>>>
>>>>> On 3/3/2010 9:28 AM, Utkarsh Ayachit wrote:
>>>>>>
>>>>>> Mike, but the cell data shows up correctly on the information
>>>>>> tab,
>>>>>> right? So the cell data is being loaded, and even reported to
>>>>>> ParaView. It's hard to imagine why it would be doing what it's
>>>> doing,
>>>>>> esp. with the precompiled binaries on Mac, since they work
>>>> fine for
>>>>>> me, on my Mac 10.6 :).
>>>>>>
>>>>>> It cannot be a mere GUI issue since from the debugging report
>>>>>> that
>>>>>> Mike gave us before, ParaView indeed doesn't report those cell
>>>> arrays
>>>>>> to the client when rendering. I just wish I could reproduce it
>>>> somehow
>>>>>> somewhere so I could debug it :). I am going to test on Os X
>>>> 10.5 to
>>>>>> see if we have any better luck.
>>>>>>
>>>>>> Utkarsh
>>>>
>>>
>>
>> _______________________________________________
>> 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 ParaView Wiki at: http://paraview.org/Wiki/ParaView
>>
>> Follow this link to subscribe/unsubscribe:
>> http://www.paraview.org/mailman/listinfo/paraview
>

[Jérôme]

Fine! I renamed my ~/.config/ParaView/ParaView3.7.0.ini to
ParaView3.7.0.ini.old.
ParaView creates a new config file, and the cell arrays are back!

Thanks a lot!
For curiosity, I will remove line after line to find the bugging one.

Jerome

2010/3/3 Michael Jackson <mike.jackson at bluequartz.net>

> Drum roll please....... "Corrupt" set of paraview preferences. After
> installing 10.5.8 clean from my installation disks everything worked as it
> was supposed to. So that means the only difference is between the 2 user
> environments. Preferences is the first to check. Looked in my home
> directory, blew away the .config directory, launched paraview on the
> "trouble" machine and now everything works. I could completely replicate
> this issue simply by swapping in and out the "good" and "bad" preferences
> for ParaView. I have the "bad" preferences if anyone wants them to try and
> debug what was going on.
>
> Thanks to EVERYONE who helped me with ideas.
>
> _________________________________________________________
> Mike Jackson mike.jackson at bluequartz.net
>
>
>
> On Mar 3, 2010, at 12:39 PM, Eric E. Monson wrote:
>
> I just tried the 3.6.2 binary on a MacBook booted off an old 10.5.7 usb
>> drive backup and all of the Color By entries showed up just fine. The
>> MacBook has the GMA 950 on-board graphics chipset. The Mac Pro I used before
>> has the NVidia GeForce 8800 GT card. If you can think of any other specs
>> that might be crucial, just drop me a line.
>>
>> -Eric
>>
>> On Mar 3, 2010, at 12:18 PM, Michael Jackson wrote:
>>
>> I just fire up the 3.6.2 binary from Kitware on OS X 10.5.8 on a MacBook
>>> Pro with nVidea 8600M graphics card and the information tab is identical to
>>> your image BUT my "Color By" drop down only shows "Solid Color", "Ids",
>>> "Normals", all of which are "Point Data" and NOT "Cell Data".
>>>
>>> So not really sure what is up. If someone wants to remotely work on my
>>> machine email me off list and I'll get you setup.
>>>
>>> Mike Jackson.
>>>
>>> On 3/3/2010 12:06 PM, Eric E. Monson wrote:
>>>
>>>> I don't have a CVS build on OS X 10.5.8 any more, but I booted into it
>>>> and ran the 3.6.2 binary and SphereSource->GenerateIds shows all four arrays
>>>> in Color By and three in Information.
>>>>
>>>> -Eric
>>>>
>>>>
>>>> On Mar 3, 2010, at 10:30 AM, J?r?me wrote:
>>>>
>>>> By my side, yes: SphereSource -> GenerateIds. You are right:
>>>>> Information
>>>>> tab shows the celldata and pointdata array, whereas "Color by" has only
>>>>> point.
>>>>>
>>>>> Jerome
>>>>>
>>>>> 2010/3/3 Utkarsh Ayachit <utkarsh.ayachit at kitware.com <mailto:
>>>>> utkarsh.ayachit at kitware.com>>
>>>>>
>>>>> Mmm..not sure if that'd help, since you said you can reproduce the
>>>>> issue with GenerateIds filter as well, right?
>>>>>
>>>>> Utkarsh
>>>>>
>>>>> On Wed, Mar 3, 2010 at 9:51 AM, Michael Jackson
>>>>> <mike.jackson at bluequartz.net
>>>>> <mailto:mike.jackson at bluequartz.net>> wrote:
>>>>>
>>>>>> Do you want my data file?
>>>>>>
>>>>>> Mike J.
>>>>>>
>>>>>> On 3/3/2010 9:28 AM, Utkarsh Ayachit wrote:
>>>>>>
>>>>>>>
>>>>>>> Mike, but the cell data shows up correctly on the information tab,
>>>>>>> right? So the cell data is being loaded, and even reported to
>>>>>>> ParaView. It's hard to imagine why it would be doing what it's
>>>>>>>
>>>>>> doing,
>>>>>
>>>>>> esp. with the precompiled binaries on Mac, since they work
>>>>>>>
>>>>>> fine for
>>>>>
>>>>>> me, on my Mac 10.6 :).
>>>>>>>
>>>>>>> It cannot be a mere GUI issue since from the debugging report that
>>>>>>> Mike gave us before, ParaView indeed doesn't report those cell
>>>>>>>
>>>>>> arrays
>>>>>
>>>>>> to the client when rendering. I just wish I could reproduce it
>>>>>>>
>>>>>> somehow
>>>>>
>>>>>> somewhere so I could debug it :). I am going to test on Os X
>>>>>>>
>>>>>> 10.5 to
>>>>>
>>>>>> see if we have any better luck.
>>>>>>>
>>>>>>> Utkarsh
>>>>>>>
>>>>>>
>>>>>
>>>>
>>> _______________________________________________
>>> 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 ParaView Wiki at:
>>> http://paraview.org/Wiki/ParaView
>>>
>>> Follow this link to subscribe/unsubscribe:
>>> http://www.paraview.org/mailman/listinfo/paraview
>>>
>>
>>
> _______________________________________________
> 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 ParaView Wiki at:
> http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20100303/e75edb9d/attachment.htm>

[Jérôme]

If someone is going further on that .ini corruption, I just noted something
maybe
useful. The problem is not that CellArray are not listed in the "Color by"
combo
box: If you try with
PointSource -> GenerateIds
instead of
SphereSource -> GenerateIds
then the CellArray is visible. As the sphere source integrates a normal
computation,
I suspect a sort of "MaxNumberOf..." blocking the number of arrays to 2.

Anyway, Mike found the workaround. Thanks again!

HTH
Jerome

2010/3/3 J?r?me <jerome.velut at gmail.com>

> Fine! I renamed my ~/.config/ParaView/ParaView3.7.0.ini to
> ParaView3.7.0.ini.old.
> ParaView creates a new config file, and the cell arrays are back!
>
> Thanks a lot!
> For curiosity, I will remove line after line to find the bugging one.
>
> Jerome
>
> 2010/3/3 Michael Jackson <mike.jackson at bluequartz.net>
>
> Drum roll please....... "Corrupt" set of paraview preferences. After
>> installing 10.5.8 clean from my installation disks everything worked as it
>> was supposed to. So that means the only difference is between the 2 user
>> environments. Preferences is the first to check. Looked in my home
>> directory, blew away the .config directory, launched paraview on the
>> "trouble" machine and now everything works. I could completely replicate
>> this issue simply by swapping in and out the "good" and "bad" preferences
>> for ParaView. I have the "bad" preferences if anyone wants them to try and
>> debug what was going on.
>>
>> Thanks to EVERYONE who helped me with ideas.
>>
>> _________________________________________________________
>> Mike Jackson mike.jackson at bluequartz.net
>>
>>
>>
>> On Mar 3, 2010, at 12:39 PM, Eric E. Monson wrote:
>>
>> I just tried the 3.6.2 binary on a MacBook booted off an old 10.5.7 usb
>>> drive backup and all of the Color By entries showed up just fine. The
>>> MacBook has the GMA 950 on-board graphics chipset. The Mac Pro I used before
>>> has the NVidia GeForce 8800 GT card. If you can think of any other specs
>>> that might be crucial, just drop me a line.
>>>
>>> -Eric
>>>
>>> On Mar 3, 2010, at 12:18 PM, Michael Jackson wrote:
>>>
>>> I just fire up the 3.6.2 binary from Kitware on OS X 10.5.8 on a MacBook
>>>> Pro with nVidea 8600M graphics card and the information tab is identical to
>>>> your image BUT my "Color By" drop down only shows "Solid Color", "Ids",
>>>> "Normals", all of which are "Point Data" and NOT "Cell Data".
>>>>
>>>> So not really sure what is up. If someone wants to remotely work on my
>>>> machine email me off list and I'll get you setup.
>>>>
>>>> Mike Jackson.
>>>>
>>>> On 3/3/2010 12:06 PM, Eric E. Monson wrote:
>>>>
>>>>> I don't have a CVS build on OS X 10.5.8 any more, but I booted into it
>>>>> and ran the 3.6.2 binary and SphereSource->GenerateIds shows all four arrays
>>>>> in Color By and three in Information.
>>>>>
>>>>> -Eric
>>>>>
>>>>>
>>>>> On Mar 3, 2010, at 10:30 AM, J?r?me wrote:
>>>>>
>>>>> By my side, yes: SphereSource -> GenerateIds. You are right:
>>>>>> Information
>>>>>> tab shows the celldata and pointdata array, whereas "Color by" has
>>>>>> only point.
>>>>>>
>>>>>> Jerome
>>>>>>
>>>>>> 2010/3/3 Utkarsh Ayachit <utkarsh.ayachit at kitware.com <mailto:
>>>>>> utkarsh.ayachit at kitware.com>>
>>>>>>
>>>>>> Mmm..not sure if that'd help, since you said you can reproduce the
>>>>>> issue with GenerateIds filter as well, right?
>>>>>>
>>>>>> Utkarsh
>>>>>>
>>>>>> On Wed, Mar 3, 2010 at 9:51 AM, Michael Jackson
>>>>>> <mike.jackson at bluequartz.net
>>>>>> <mailto:mike.jackson at bluequartz.net>> wrote:
>>>>>>
>>>>>>> Do you want my data file?
>>>>>>>
>>>>>>> Mike J.
>>>>>>>
>>>>>>> On 3/3/2010 9:28 AM, Utkarsh Ayachit wrote:
>>>>>>>
>>>>>>>>
>>>>>>>> Mike, but the cell data shows up correctly on the information tab,
>>>>>>>> right? So the cell data is being loaded, and even reported to
>>>>>>>> ParaView. It's hard to imagine why it would be doing what it's
>>>>>>>>
>>>>>>> doing,
>>>>>>
>>>>>>> esp. with the precompiled binaries on Mac, since they work
>>>>>>>>
>>>>>>> fine for
>>>>>>
>>>>>>> me, on my Mac 10.6 :).
>>>>>>>>
>>>>>>>> It cannot be a mere GUI issue since from the debugging report that
>>>>>>>> Mike gave us before, ParaView indeed doesn't report those cell
>>>>>>>>
>>>>>>> arrays
>>>>>>
>>>>>>> to the client when rendering. I just wish I could reproduce it
>>>>>>>>
>>>>>>> somehow
>>>>>>
>>>>>>> somewhere so I could debug it :). I am going to test on Os X
>>>>>>>>
>>>>>>> 10.5 to
>>>>>>
>>>>>>> see if we have any better luck.
>>>>>>>>
>>>>>>>> Utkarsh
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>> _______________________________________________
>>>> 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 ParaView Wiki at:
>>>> http://paraview.org/Wiki/ParaView
>>>>
>>>> Follow this link to subscribe/unsubscribe:
>>>> http://www.paraview.org/mailman/listinfo/paraview
>>>>
>>>
>>>
>> _______________________________________________
>> 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 ParaView Wiki at:
>> http://paraview.org/Wiki/ParaView
>>
>> Follow this link to subscribe/unsubscribe:
>> http://www.paraview.org/mailman/listinfo/paraview
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20100303/8e76090a/attachment-0001.htm>

[Utkarsh Ayachit]

Oh wow...that's interesting. I won't mind looking at the ini out of curiosity.

Utkarsh

On Wed, Mar 3, 2010 at 2:11 PM, Michael Jackson
<mike.jackson at bluequartz.net> wrote:
> Drum roll please....... "Corrupt" set of paraview preferences. After
> installing 10.5.8 clean from my installation disks everything worked as it
> was supposed to. So that means the only difference is between the 2 user
> environments. Preferences is the first to check. Looked in my home
> directory, blew away the .config directory, launched paraview on the
> "trouble" machine and now everything works. I could completely replicate
> this issue simply by swapping in and out the "good" and "bad" preferences
> for ParaView. I have the "bad" preferences if anyone wants them to try and
> debug what was going on.
>
> Thanks to EVERYONE who helped me with ideas.
>
> _________________________________________________________
> Mike Jackson ? ? ? ? ? ? ? ? ?mike.jackson at bluequartz.net
>
>
> On Mar 3, 2010, at 12:39 PM, Eric E. Monson wrote:
>
>> I just tried the 3.6.2 binary on a MacBook booted off an old 10.5.7 usb
>> drive backup and all of the Color By entries showed up just fine. The
>> MacBook has the GMA 950 on-board graphics chipset. The Mac Pro I used before
>> has the NVidia GeForce 8800 GT card. If you can think of any other specs
>> that might be crucial, just drop me a line.
>>
>> -Eric
>>
>> On Mar 3, 2010, at 12:18 PM, Michael Jackson wrote:
>>
>>> I just fire up the 3.6.2 binary from Kitware on OS X 10.5.8 on a MacBook
>>> Pro with nVidea 8600M graphics card and the information tab is identical to
>>> your image BUT my "Color By" drop down only shows "Solid Color", "Ids",
>>> "Normals", all of which are "Point Data" and NOT "Cell Data".
>>>
>>> So not really sure what is up. If someone wants to remotely work on my
>>> machine email me off list and I'll get you setup.
>>>
>>> Mike Jackson.
>>>
>>> On 3/3/2010 12:06 PM, Eric E. Monson wrote:
>>>>
>>>> I don't have a CVS build on OS X 10.5.8 any more, but I booted into it
>>>> and ran the 3.6.2 binary and SphereSource->GenerateIds shows all four arrays
>>>> in Color By and three in Information.
>>>>
>>>> -Eric
>>>>
>>>>
>>>> On Mar 3, 2010, at 10:30 AM, J?r?me wrote:
>>>>
>>>>> By my side, yes: SphereSource -> GenerateIds. You are right:
>>>>> Information
>>>>> tab shows the celldata and pointdata array, whereas "Color by" has only
>>>>> point.
>>>>>
>>>>> Jerome
>>>>>
>>>>> 2010/3/3 Utkarsh Ayachit <utkarsh.ayachit at kitware.com
>>>>> <mailto:utkarsh.ayachit at kitware.com>>
>>>>>
>>>>> ?Mmm..not sure if that'd help, since you said you can reproduce the
>>>>> ?issue with GenerateIds filter as well, right?
>>>>>
>>>>> ?Utkarsh
>>>>>
>>>>> ?On Wed, Mar 3, 2010 at 9:51 AM, Michael Jackson
>>>>> ?<mike.jackson at bluequartz.net
>>>>> ?<mailto:mike.jackson at bluequartz.net>> wrote:
>>>>>>
>>>>>> Do you want my data file?
>>>>>>
>>>>>> Mike J.
>>>>>>
>>>>>> On 3/3/2010 9:28 AM, Utkarsh Ayachit wrote:
>>>>>>>
>>>>>>> Mike, but the cell data shows up correctly on the information tab,
>>>>>>> right? So the cell data is being loaded, and even reported to
>>>>>>> ParaView. It's hard to imagine why it would be doing what it's
>>>>>
>>>>> ?doing,
>>>>>>>
>>>>>>> esp. with the precompiled binaries on Mac, since they work
>>>>>
>>>>> ?fine for
>>>>>>>
>>>>>>> me, on my Mac 10.6 ?:).
>>>>>>>
>>>>>>> It cannot be a mere GUI issue since from the debugging report that
>>>>>>> Mike gave us before, ParaView indeed doesn't report those cell
>>>>>
>>>>> ?arrays
>>>>>>>
>>>>>>> to the client when rendering. I just wish I could reproduce it
>>>>>
>>>>> ?somehow
>>>>>>>
>>>>>>> somewhere so I could debug it ?:). I am going to test on Os X
>>>>>
>>>>> ?10.5 to
>>>>>>>
>>>>>>> see if we have any better luck.
>>>>>>>
>>>>>>> Utkarsh
>>>>>
>>>>
>>>
>>> _______________________________________________
>>> 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 ParaView Wiki at:
>>> http://paraview.org/Wiki/ParaView
>>>
>>> Follow this link to subscribe/unsubscribe:
>>> http://www.paraview.org/mailman/listinfo/paraview
>>
>
> _______________________________________________
> 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 ParaView Wiki at:
> http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview
>

[Jérôme]

Utkarsh,
I will send you mine this -french- evening (I don't have it here). It could
be
interesting to compare with the Mike's bad one.

Jerome

2010/3/3 Utkarsh Ayachit <utkarsh.ayachit at kitware.com>

> Oh wow...that's interesting. I won't mind looking at the ini out of
> curiosity.
>
> Utkarsh
>
> On Wed, Mar 3, 2010 at 2:11 PM, Michael Jackson
> <mike.jackson at bluequartz.net> wrote:
> > Drum roll please....... "Corrupt" set of paraview preferences. After
> > installing 10.5.8 clean from my installation disks everything worked as
> it
> > was supposed to. So that means the only difference is between the 2 user
> > environments. Preferences is the first to check. Looked in my home
> > directory, blew away the .config directory, launched paraview on the
> > "trouble" machine and now everything works. I could completely replicate
> > this issue simply by swapping in and out the "good" and "bad" preferences
> > for ParaView. I have the "bad" preferences if anyone wants them to try
> and
> > debug what was going on.
> >
> > Thanks to EVERYONE who helped me with ideas.
> >
> > _________________________________________________________
> > Mike Jackson mike.jackson at bluequartz.net
> >
> >
> > On Mar 3, 2010, at 12:39 PM, Eric E. Monson wrote:
> >
> >> I just tried the 3.6.2 binary on a MacBook booted off an old 10.5.7 usb
> >> drive backup and all of the Color By entries showed up just fine. The
> >> MacBook has the GMA 950 on-board graphics chipset. The Mac Pro I used
> before
> >> has the NVidia GeForce 8800 GT card. If you can think of any other specs
> >> that might be crucial, just drop me a line.
> >>
> >> -Eric
> >>
> >> On Mar 3, 2010, at 12:18 PM, Michael Jackson wrote:
> >>
> >>> I just fire up the 3.6.2 binary from Kitware on OS X 10.5.8 on a
> MacBook
> >>> Pro with nVidea 8600M graphics card and the information tab is
> identical to
> >>> your image BUT my "Color By" drop down only shows "Solid Color", "Ids",
> >>> "Normals", all of which are "Point Data" and NOT "Cell Data".
> >>>
> >>> So not really sure what is up. If someone wants to remotely work on my
> >>> machine email me off list and I'll get you setup.
> >>>
> >>> Mike Jackson.
> >>>
> >>> On 3/3/2010 12:06 PM, Eric E. Monson wrote:
> >>>>
> >>>> I don't have a CVS build on OS X 10.5.8 any more, but I booted into it
> >>>> and ran the 3.6.2 binary and SphereSource->GenerateIds shows all four
> arrays
> >>>> in Color By and three in Information.
> >>>>
> >>>> -Eric
> >>>>
> >>>>
> >>>> On Mar 3, 2010, at 10:30 AM, J?r?me wrote:
> >>>>
> >>>>> By my side, yes: SphereSource -> GenerateIds. You are right:
> >>>>> Information
> >>>>> tab shows the celldata and pointdata array, whereas "Color by" has
> only
> >>>>> point.
> >>>>>
> >>>>> Jerome
> >>>>>
> >>>>> 2010/3/3 Utkarsh Ayachit <utkarsh.ayachit at kitware.com
> >>>>> <mailto:utkarsh.ayachit at kitware.com>>
> >>>>>
> >>>>> Mmm..not sure if that'd help, since you said you can reproduce the
> >>>>> issue with GenerateIds filter as well, right?
> >>>>>
> >>>>> Utkarsh
> >>>>>
> >>>>> On Wed, Mar 3, 2010 at 9:51 AM, Michael Jackson
> >>>>> <mike.jackson at bluequartz.net
> >>>>> <mailto:mike.jackson at bluequartz.net>> wrote:
> >>>>>>
> >>>>>> Do you want my data file?
> >>>>>>
> >>>>>> Mike J.
> >>>>>>
> >>>>>> On 3/3/2010 9:28 AM, Utkarsh Ayachit wrote:
> >>>>>>>
> >>>>>>> Mike, but the cell data shows up correctly on the information tab,
> >>>>>>> right? So the cell data is being loaded, and even reported to
> >>>>>>> ParaView. It's hard to imagine why it would be doing what it's
> >>>>>
> >>>>> doing,
> >>>>>>>
> >>>>>>> esp. with the precompiled binaries on Mac, since they work
> >>>>>
> >>>>> fine for
> >>>>>>>
> >>>>>>> me, on my Mac 10.6 :).
> >>>>>>>
> >>>>>>> It cannot be a mere GUI issue since from the debugging report that
> >>>>>>> Mike gave us before, ParaView indeed doesn't report those cell
> >>>>>
> >>>>> arrays
> >>>>>>>
> >>>>>>> to the client when rendering. I just wish I could reproduce it
> >>>>>
> >>>>> somehow
> >>>>>>>
> >>>>>>> somewhere so I could debug it :). I am going to test on Os X
> >>>>>
> >>>>> 10.5 to
> >>>>>>>
> >>>>>>> see if we have any better luck.
> >>>>>>>
> >>>>>>> Utkarsh
> >>>>>
> >>>>
> >>>
> >>> _______________________________________________
> >>> 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 ParaView Wiki at:
> >>> http://paraview.org/Wiki/ParaView
> >>>
> >>> Follow this link to subscribe/unsubscribe:
> >>> http://www.paraview.org/mailman/listinfo/paraview
> >>
> >
> > _______________________________________________
> > 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 ParaView Wiki at:
> > http://paraview.org/Wiki/ParaView
> >
> > Follow this link to subscribe/unsubscribe:
> > http://www.paraview.org/mailman/listinfo/paraview
> >
> _______________________________________________
> 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 ParaView Wiki at:
> http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20100304/c3451b83/attachment-0001.htm>

[Michael Jackson]

Might be. I'll have to borrow a Snow Leopard machine to test it out.

I am building 32 bit ParaView against a 32/64 Qt 4.6.x Cocoa and a 32
bit Qt 4.6.2 Carbon. All on OS X 10.5.8. And, again, the precompiled
binaries from Kitware will reproduce for me. This is also on a MacPro
with ATI 4870 video card if that means anything.


Mike Jackson

On 3/3/2010 9:23 AM, Eric E. Monson wrote:
> Mike, is it possible that it's a problem related to OS X 10.5? I know J?r?me is having some trouble with Fedora, but I can't reproduce the issue on Snow Leopard or Ubuntu (both x86_64 builds). Which architecture are you building PV and Qt for?
>
> -Eric
>
> On Mar 3, 2010, at 9:09 AM, Michael Jackson wrote:
>
>
>> I have tried clean builds on ParaView CVS and ParaView 3.6.2 OS X with Qt versions 4.5.3 and Qt 4.6.2 both carbon and Cocoa and none of that alleviates the problem. Not even the precompiled binary from Kitware's web site works correctly. It does work perfectly fine on windows though.
>>
>> I wonder where the loading of the file takes place. I have a feeling something is not loading the cell data correctly. Maybe.
>>
>> --
>> Mike Jackson
>>
>> On 3/3/2010 3:01 AM, J?r?me wrote:
>>
>>> Hi,
>>>
>>> After full updates on 2 computers (personal and office) both running fedora 12,
>>> the problem persists... on one of them only! The sadness is that the personal
>>> one is concerned, so I cannot debug efficiently.
>>>
>>> The point is that I didn't perform a comprehensive clean on my built tree. This
>>> is the next step.
>>>
>>> Jerome
>>>
>>> 2010/3/1 Michael Jackson<mike.jackson at bluequartz.net<mailto:mike.jackson at bluequartz.net>>
>>>
>>> So I figured out where the problem lies, but not what the actual
>>> problem is:
>>>
>>> QList<QString> pqPipelineRepresentation::getColorFields()
>>> ....
>>> for(int i=0; i<cellinfo->GetNumberOfArrays(); i++)
>>>
>>> that loop is 0 to 0 on OS X and 0 to 2 on Windows. But using the
>>> exact same data file. So for some reason the OS X version is NOT
>>> storing/updating something properly. Odd.
>>>

[Michael Jackson]

I just tried on a 10.6.2 MacBook with Intel graphics and the Cell Data
showed up just fine. So it does work on 10.6 but just not 10.5. Go
figure. All my clients are on 10.5 so I am not sure what to do now.
Don't really know enough about OpenGL (is my guess) capabilities or how
to figure out why the CELL_DATA is being ignored on OS X 10.5.

Odd. Then again, this seems to fail on some linux hosts also, so not
really sure what the issue is.

Mike Jackson.

On 3/3/2010 9:23 AM, Eric E. Monson wrote:
> Mike, is it possible that it's a problem related to OS X 10.5? I know J?r?me is having some trouble with Fedora, but I can't reproduce the issue on Snow Leopard or Ubuntu (both x86_64 builds). Which architecture are you building PV and Qt for?
>
> -Eric
>
> On Mar 3, 2010, at 9:09 AM, Michael Jackson wrote:
>
>
>> I have tried clean builds on ParaView CVS and ParaView 3.6.2 OS X with Qt versions 4.5.3 and Qt 4.6.2 both carbon and Cocoa and none of that alleviates the problem. Not even the precompiled binary from Kitware's web site works correctly. It does work perfectly fine on windows though.
>>
>> I wonder where the loading of the file takes place. I have a feeling something is not loading the cell data correctly. Maybe.
>>
>> --
>> Mike Jackson
>>
>> On 3/3/2010 3:01 AM, J?r?me wrote:
>>
>>> Hi,
>>>
>>> After full updates on 2 computers (personal and office) both running fedora 12,
>>> the problem persists... on one of them only! The sadness is that the personal
>>> one is concerned, so I cannot debug efficiently.
>>>
>>> The point is that I didn't perform a comprehensive clean on my built tree. This
>>> is the next step.
>>>
>>> Jerome
>>>
>>> 2010/3/1 Michael Jackson<mike.jackson at bluequartz.net<mailto:mike.jackson at bluequartz.net>>
>>>
>>> So I figured out where the problem lies, but not what the actual
>>> problem is:
>>>
>>> QList<QString> pqPipelineRepresentation::getColorFields()
>>> ....
>>> for(int i=0; i<cellinfo->GetNumberOfArrays(); i++)
>>>
>>> that loop is 0 to 0 on OS X and 0 to 2 on Windows. But using the
>>> exact same data file. So for some reason the OS X version is NOT
>>> storing/updating something properly. Odd.
>>>
>>>
>>>
>> _______________________________________________
>> 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 ParaView Wiki at: http://paraview.org/Wiki/ParaView
>>
>> Follow this link to subscribe/unsubscribe:
>> http://www.paraview.org/mailman/listinfo/paraview
>>
>