MantisBT - ParaView |
| View Issue Details |
|
| ID | Project | Category | View Status | Date Submitted | Last Update |
| 0006706 | ParaView | (No Category) | public | 2008-04-02 11:08 | 2009-05-13 13:59 |
|
| Reporter | Ken Moreland | |
| Assigned To | Berk Geveci | |
| Priority | normal | Severity | crash | Reproducibility | have not tried |
| Status | closed | Resolution | fixed | |
| Platform | | OS | | OS Version | |
| Product Version | | |
| Target Version | 3.4 | Fixed in Version | 3.4 | |
| Project | |
| Topic Name | |
| Type | |
|
| Summary | 0006706: Gradient and Gradient Magnitude filters only work on |
| Description | The object inspector for the Gradient and Gradient Magnitude filter allows you to select the input scalars by name, but the selection has no effect. These two filters always grab the active scalars on the input image data. This leads to two serious problems.
1. The gradients computed may not correspond to the input the user selected, leading to very misleading results.
2. If the input has no active scalars, the filters will create a new array in the input in an attempt to get them. This leads to garbage in the output and breaks an assumption that the input will not be modified by the filter. The broken assumption often leads to crashes in ParaView in strange places. |
| Steps To Reproduce | The attached file has no active scalars defined in it. If you run either gradient filter, you will get garbage for the gradients (assuming it does not crash in the process). If it does not crash, it is not hard to get ParaView to crash after the fact. Playing with scalar bars or deleting the pipeline objects usually does the trick. |
| Additional Information | I understand that this behavior is typical of the imaging pipeline objects. We should check to make sure that there are no more filters used by ParaView that exhibit this behavior. |
| Tags | No tags attached. |
| Relationships | | child of | 0006693 | closed | Berk Geveci | Add result array name feature to gradient and gradient magnitude filters |
|
| Attached Files |  NoActiveScalars.vti (28,685) 2008-04-02 11:08
https://www.vtk.org/Bug/file/6382/NoActiveScalars.vti |
|
| Issue History |
| Date Modified | Username | Field | Change |
| 2008-04-02 11:08 | Ken Moreland | New Issue | |
| 2008-04-02 11:08 | Ken Moreland | File Added: NoActiveScalars.vti | |
| 2008-04-02 11:09 | Ken Moreland | Relationship added | child of 0006693 |
| 2008-04-02 11:17 | Ken Moreland | Note Added: 0011035 | |
| 2008-04-02 11:19 | Ken Moreland | Note Added: 0011036 | |
| 2008-04-10 18:00 | Ken Moreland | Category | Development => 3.4 |
| 2008-04-10 18:00 | Ken Moreland | Status | backlog => tabled |
| 2008-04-10 18:00 | Ken Moreland | Assigned To | => Berk Geveci |
| 2008-08-15 11:31 | Berk Geveci | Note Added: 0013036 | |
| 2008-08-15 11:31 | Berk Geveci | Status | tabled => @20@ |
| 2008-08-22 13:06 | Berk Geveci | Status | @20@ => @80@ |
| 2008-08-22 13:06 | Berk Geveci | Resolution | open => fixed |
| 2008-08-22 13:06 | Berk Geveci | Note Added: 0013154 | |
| 2008-09-02 21:14 | Alan Scott | Status | @80@ => closed |
| 2008-09-02 21:14 | Alan Scott | Note Added: 0013261 | |
| 2009-05-13 13:58 | Utkarsh Ayachit | Target Version | => 3.4 |
| 2009-05-13 13:59 | Utkarsh Ayachit | Fixed in Version | => 3.4 |
| 2011-06-16 13:10 | Zack Galbreath | Category | => (No Category) |
|
Notes |
|
|
(0011035)
|
|
Ken Moreland
|
|
2008-04-02 11:17
|
|
Rather than change the gradient filters themselves, it may be prudent to create a meta-filter that uses the imaging filters internally but uses shallow copies of data objects to convert the ParaView array naming convention to the imaging pipeline active scalars convention.
I can see two potential forms for this filter. One form would be a generic filter that worked on some unspecified image algorithm (in a manner similar to vtkFileSeriesReader). Another form would be a specific filter that used either the image gradient filter or the unstructured gradient filter, depending on the input type. |
|
|
|
(0011036)
|
|
Ken Moreland
|
|
2008-04-02 11:19
|
|
|
Is it really necessary to have the Gradient Magnitude filter? The functionality could more-or-less be replaced with the Gradient filter and the Calculator filter to compute the magnitude. Would anyone have any heartburn if Gradient Magnitude went away? |
|
|
|
(0013036)
|
|
Berk Geveci
|
|
2008-08-15 11:31
|
|
|
Waiting for feedback from Ken. |
|
|
|
(0013154)
|
|
Berk Geveci
|
|
2008-08-22 13:06
|
|
|
Gradient now works on any selected arrays. I deprecated (removed) gradient magnitude based on our decision. |
|
|
|
(0013261)
|
|
Alan Scott
|
|
2008-09-02 21:14
|
|
|