ParaView issueshttps://gitlab.kitware.com/paraview/paraview/-/issues2017-11-21T16:31:59-05:00https://gitlab.kitware.com/paraview/paraview/-/issues/16882Investigate enabling anti-aliasing in ParaView2017-11-21T16:31:59-05:00Utkarsh AyachitInvestigate enabling anti-aliasing in ParaViewWe know that we cannot use anti-aliasing when using parallel rendering, but can be use it for client-side or builtin configurations? It may fudge with the selection/picking mechanisms, so we need to make sure those are unhampered.
Dup...We know that we cannot use anti-aliasing when using parallel rendering, but can be use it for client-side or builtin configurations? It may fudge with the selection/picking mechanisms, so we need to make sure those are unhampered.
Duplicate of #15745 5.2 (SC-16)Allison Vacantialliepiper16@gmail.comAllison Vacantialliepiper16@gmail.comhttps://gitlab.kitware.com/paraview/paraview/-/issues/16881Paraview crashed when opening vtr file2017-11-21T16:31:59-05:00Xu ZhengParaview crashed when opening vtr fileHi all
I am new to use paraview. I found it quite strange the software crashes immediately every time when I try to open an vtr file (other file types like vtk can be opened correctly). However, my colleges can open the same file on the...Hi all
I am new to use paraview. I found it quite strange the software crashes immediately every time when I try to open an vtr file (other file types like vtk can be opened correctly). However, my colleges can open the same file on their computers with no errors. I have checked memory usage and I should have enough RAM left. I would be very appreciate if any one could help on this issue.
Many thanks
Xuhttps://gitlab.kitware.com/paraview/paraview/-/issues/16880Using '+' as list separator causes some problems building documentation.2017-11-21T16:31:59-05:00olesenUsing '+' as list separator causes some problems building documentation.In ParaViewMacros.cmake, ParaViewPlugins.cmake and in proxydocumentation.cmake, a plus (+) character is used to build pseudo lists, since the semi-colon (;) would otherwise cause issues in shell commands. Unfortunately, this means that ...In ParaViewMacros.cmake, ParaViewPlugins.cmake and in proxydocumentation.cmake, a plus (+) character is used to build pseudo lists, since the semi-colon (;) would otherwise cause issues in shell commands. Unfortunately, this means that if the file-system path itself contains a plus (+) character, this will be incorrectly interpreted as a list separation character. Eg, for a path such as `/nfs/software/gcc-4.8+openmpi-1.10.4/foo/...` would be falsely split into two parts, neither of which exist.
Changing this list-separator to a less commonly used character (eg, a comma) could help. Replacing it with an improbable multi-character sequence would likely be the more robust solution. For example, `,,` or `,@,` etc.5.2 (SC-16)https://gitlab.kitware.com/paraview/paraview/-/issues/16876Invalid memory access in `vtkMPIImageReader`2017-11-21T16:31:59-05:00Utkarsh AyachitInvalid memory access in `vtkMPIImageReader`Data file: https://drive.google.com/file/d/0B26kKqqTZSgGeU1DTXM5b1pTYlE/view?usp=sharing
1. Load the data using `Raw (binary) Reader` with a 4 rank pvserver.
2. Set extents to (0 149, 0 149, 0 275) and type to (unsigned short)
3. Appl...Data file: https://drive.google.com/file/d/0B26kKqqTZSgGeU1DTXM5b1pTYlE/view?usp=sharing
1. Load the data using `Raw (binary) Reader` with a 4 rank pvserver.
2. Set extents to (0 149, 0 149, 0 275) and type to (unsigned short)
3. Apply
This will fail arbitrarily after this point.
See [valgrind.25374.log](/uploads/a01a707c5cc65b919438bd2e43846d25/valgrind.25374.log) for one of the pvserver ranks. There's invalid read/writers in `vtkMPIImageReader` that cause the segfault.5.2 (SC-16)Allison Vacantialliepiper16@gmail.comAllison Vacantialliepiper16@gmail.comhttps://gitlab.kitware.com/paraview/paraview/-/issues/16873Test issue2017-11-21T16:31:59-05:00Cory Quammencory.quammen@kitware.comTest issueThis is a test issue to explore new features of GitLab. Please ignore.
- [x] Task 1
- [x] Task 2
- [x] Dependent task 1This is a test issue to explore new features of GitLab. Please ignore.
- [x] Task 1
- [x] Task 2
- [x] Dependent task 15.2 (SC-16)2016-09-04https://gitlab.kitware.com/paraview/paraview/-/issues/16868"(?)" shows up in `Color By` toolbar.2017-11-21T16:31:59-05:00Utkarsh Ayachit"(?)" shows up in `Color By` toolbar.I am getting this:
![image](/uploads/a595af1415c21dbc45715275859bb458/image.png)
The pipeline was created using default values for everything on a 2-node pvserver.I am getting this:
![image](/uploads/a595af1415c21dbc45715275859bb458/image.png)
The pipeline was created using default values for everything on a 2-node pvserver.5.2 (SC-16)Allison Vacantialliepiper16@gmail.comAllison Vacantialliepiper16@gmail.comhttps://gitlab.kitware.com/paraview/paraview/-/issues/16867ParaView SpreadSheet view crashes when array name is NULL2017-11-21T16:31:59-05:00Cory Quammencory.quammen@kitware.comParaView SpreadSheet view crashes when array name is NULLError is in `pqSpreadSheetView::data(const QModelIndex& idx, int role)` at the call to `strcmp(column_name, "Cell Type")`. Add a check for NULL `column_name` to prevent this crash.Error is in `pqSpreadSheetView::data(const QModelIndex& idx, int role)` at the call to `strcmp(column_name, "Cell Type")`. Add a check for NULL `column_name` to prevent this crash.Cory Quammencory.quammen@kitware.comCory Quammencory.quammen@kitware.comhttps://gitlab.kitware.com/paraview/paraview/-/issues/16866ParaView doxygen may not be generating docs correctly.2017-11-21T16:31:59-05:00Utkarsh AyachitParaView doxygen may not be generating docs correctly.The docs have stopped being updated since August 14. May have something to do with the OpenGL2 change on blight. Try to manually generate the docs and see if they are generated correctly. THe Python docs are, but not doxygen, it seems.The docs have stopped being updated since August 14. May have something to do with the OpenGL2 change on blight. Try to manually generate the docs and see if they are generated correctly. THe Python docs are, but not doxygen, it seems.Shawn WaldonShawn Waldonhttps://gitlab.kitware.com/paraview/paraview/-/issues/16863Cannot create `Opacity` or `Visibility` animations after loading state.2017-11-21T16:31:59-05:00Utkarsh AyachitCannot create `Opacity` or `Visibility` animations after loading state.Steps to reproduce:
1. Start ParaView, create `Sphere`, `Apply`.
2. Save state as `sphere.pvsm`.
3. From `Animation View` create a new animation track for **Sphere**:**Opacity**. Hit `Play` using the VCR controls. *<--- this should wo...Steps to reproduce:
1. Start ParaView, create `Sphere`, `Apply`.
2. Save state as `sphere.pvsm`.
3. From `Animation View` create a new animation track for **Sphere**:**Opacity**. Hit `Play` using the VCR controls. *<--- this should work as expected; you'll see the sphere go from opacity 0 to 1.*
4. Restart ParaView and load `sphere.pvsm`.
5. Now try to create a animation (similar to Step 3). ***<--- you won't be able to add a track for Opacity; this is a bug!***
This an easily testable symptom of paraview/paraview#168625.2 (SC-16)Utkarsh AyachitUtkarsh Ayachithttps://gitlab.kitware.com/paraview/paraview/-/issues/16862After loading XML state files, certain proxies may never get deleted!2017-11-21T16:31:59-05:00Utkarsh AyachitAfter loading XML state files, certain proxies may never get deleted!After loading XML state files, certain proxies may never get deleted! While the proxies don't leak, and get cleaned up when the application quits or server disconnects, it still implies that memory won't be cleared when expected.
Step...After loading XML state files, certain proxies may never get deleted! While the proxies don't leak, and get cleaned up when the application quits or server disconnects, it still implies that memory won't be cleared when expected.
Steps to reproduce:
* Start ParaView (builtin)
* To keep things simple, close the current view, so you have an empty view and then create the `Sphere` source.
* Don't hit Apply, just save the state as One.pvsm file at this point.
Inspect the state file, at the very end, it will have something as follows:
````xml
<ProxyCollection name="pq_helper_proxies.4641">
<Item id="4652" name="RepresentationAnimationHelper"/>
</ProxyCollection>
<ProxyCollection name="sources">
<Item id="4641" name="Sphere1"/>
</ProxyCollection>
````
That's the Sphere source and it's helper proxy. If you look closely for the state of the helper proxy (search for the proxy using its id), you'll notice that it refers to the `Sphere1`'s ID as follows:
````xml
<Proxy group="misc" type="RepresentationAnimationHelper" id="4652" servers="16">
<Property name="Source" id="4652.Source" number_of_elements="1">
<Proxy value="4641"/>
</Property>
</Proxy>
````
* Delete the Sphere. Save out the state file again (Two.pvsm). Now look at the state file. Both the `RepresentationAnimationHelper` and the `Sphere1` proxies are gone. Good!
* Restart ParaView. And load One.pvsm. Delete the `Sphere1` source and save the state file (Three.pvsm). This time, in Three.pvsm while there's no `Sphere1` proxy, the `RepresentationAnimationHelper` is still present! If you look at it's state, you;ll see it's `Source` property is indeed hanging on to some proxy (which is actually the SphereSource proxy. The SphereSource won't be deleted until the application is quit!
**This is a bug!**
Another way to test this same bug is paraview/paraview#168635.2 (SC-16)Utkarsh AyachitUtkarsh Ayachithttps://gitlab.kitware.com/paraview/paraview/-/issues/16860ParaView needs to have the signature for `GetMTime()` updated2017-11-21T16:31:59-05:00T.J. CoronaParaView needs to have the signature for `GetMTime()` updatedVTK commit ecb0979ac22a4b92b6b499bd8e9079c5bb7dfb4e introduced a signature change in GetMTime(). This must be propagated through ParaView.VTK commit ecb0979ac22a4b92b6b499bd8e9079c5bb7dfb4e introduced a signature change in GetMTime(). This must be propagated through ParaView.https://gitlab.kitware.com/paraview/paraview/-/issues/16857GenericIO issues on blight2017-11-21T16:32:00-05:00Utkarsh AyachitGenericIO issues on blightParaView seems to be failing to build on Blight due to issues with GenericIO.ParaView seems to be failing to build on Blight due to issues with GenericIO.Ben BoeckelBen Boeckelhttps://gitlab.kitware.com/paraview/paraview/-/issues/16853Cannot use texture maping and transparent rendering together2017-11-21T16:32:00-05:00Kenneth MorelandCannot use texture maping and transparent rendering togetherThere is an issue where if you apply a texture to one object and then render another object, the transparent object is rendered incorrectly. The following steps replicate the issue.
1. Create a `Sphere` source.
2. Change the `Opacity...There is an issue where if you apply a texture to one object and then render another object, the transparent object is rendered incorrectly. The following steps replicate the issue.
1. Create a `Sphere` source.
2. Change the `Opacity` of the sphere to `0.5`.
3. Create a `Plane` source. At this point, everything is working fine.
4. Load a `Texture` on the plane. Any png is fine. I used NE2_ps_bath.png, which is available as part of the ParaView test data.
Once you load the texture, you will see that the sphere is rendered incorrectly. In my version of ParaView (version 5.1.0 64-bit on Windows) I can see the NE2_ps_bath texture I loaded for the plane being mapped to the sphere. I believe the issue is that the rendering system is not correctly unloading the texture.5.2 (SC-16)Allison Vacantialliepiper16@gmail.comAllison Vacantialliepiper16@gmail.comhttps://gitlab.kitware.com/paraview/paraview/-/issues/16852Paraview Axes Grid missing titles and labels on Windows server 20122017-11-21T16:32:00-05:00wxuParaview Axes Grid missing titles and labels on Windows server 2012I have built ParaView 5.0.0 on Windows 7, Windows 10 and Windows Server 2012. The building processes are identical and there are no error messages. However on ParaView running on Windows Server 2012, all text contents such as titles and...I have built ParaView 5.0.0 on Windows 7, Windows 10 and Windows Server 2012. The building processes are identical and there are no error messages. However on ParaView running on Windows Server 2012, all text contents such as titles and number labels disappear from Axes Grid, while ParaView running on desktop versions of Windows has no such issues. Also ParaView 4.1.0 running on Windows Server 2012 also has no such issues. ParaView 4.1.0 uses Cube Axes instead of Axes Grid used in ParaView 5.0.0. Please see attached screenshots of Axes Grid of ParaView running on Windows 7 and Windows Server 2012.
![Axes_Grid_Windows7](/uploads/ebdadc1cca6ea6a60998b1d5845dbe8a/Axes_Grid_Windows7.PNG)
![Axes_Grid_Windows2012](/uploads/afa98c7c65901637dba75da08dd2c124/Axes_Grid_Windows2012.PNG)https://gitlab.kitware.com/paraview/paraview/-/issues/16851Segfault when using "Follow Data" animation track2017-11-21T16:32:00-05:00Utkarsh AyachitSegfault when using "Follow Data" animation trackSteps to reproduce:
1. Open can.ex2, Apply
2. View | Animation View
3. Add **Camera | Follow Data**
4. Double Click on the "Camera" Track in the Animation View. In the popup dialog, hit Ok -- this will case a segfault.Steps to reproduce:
1. Open can.ex2, Apply
2. View | Animation View
3. Add **Camera | Follow Data**
4. Double Click on the "Camera" Track in the Animation View. In the popup dialog, hit Ok -- this will case a segfault.5.2 (SC-16)Utkarsh AyachitUtkarsh Ayachithttps://gitlab.kitware.com/paraview/paraview/-/issues/16849CatalystLive tests failing in ParaView builds2017-11-21T16:32:00-05:00Shawn WaldonCatalystLive tests failing in ParaView buildsThe CatalystLiveSetBreakpoint and CatalystLivePause tests are failing across many ParaView dashboards.The CatalystLiveSetBreakpoint and CatalystLivePause tests are failing across many ParaView dashboards.Mathieu Westphal (Kitware)Mathieu Westphal (Kitware)https://gitlab.kitware.com/paraview/paraview/-/issues/16848Parallel Coordinates View is Broken.2017-11-21T16:32:00-05:00W. Alan ScottParallel Coordinates View is Broken.Parallel Coordinates View is broken. All I see is white.
* 5.1.2 (or master), Linux, local server.
* Change the view to be Parallel Coordinates View. All is white.
* Sources/ Wavelet, apply. You can't see anything, you can't chan...Parallel Coordinates View is broken. All I see is white.
* 5.1.2 (or master), Linux, local server.
* Change the view to be Parallel Coordinates View. All is white.
* Sources/ Wavelet, apply. You can't see anything, you can't change teh representation, you can't change the variables.
Assuming this representation is valuable, thus marking high priority. If it isn't valuable, feel free to lower priority.5.2 (SC-16)https://gitlab.kitware.com/paraview/paraview/-/issues/16847Warning in release version of vtkRenderer2017-11-21T16:32:00-05:00W. Alan ScottWarning in release version of vtkRendererIn file /VTK/ Rendering/Core/vtkRenderer.cxx, line 1077, we are getting warnings. Is this warning necessary in production code? Seems to me that warnings should never be output in release code…
This has become a big issue for a user...In file /VTK/ Rendering/Core/vtkRenderer.cxx, line 1077, we are getting warnings. Is this warning necessary in production code? Seems to me that warnings should never be output in release code…
This has become a big issue for a user. Either fix the issue so that the warning is never called, clearly explain to the user why things are bad and what to do, or delete it (from release code).
Related bugs that can be closed when this one is fixed:
paraview/paraview#16039
paraview/paraview#15366
paraview/paraview#15124
5.2 (SC-16)Utkarsh AyachitUtkarsh Ayachithttps://gitlab.kitware.com/paraview/paraview/-/issues/16845Valgrind points out error in vtkSocketCommunicator2017-11-21T16:32:00-05:00Shawn WaldonValgrind points out error in vtkSocketCommunicatorThis was found running the pvcs.HistogramSelection test with the server processes running in valgrind. This only appeared on one of the server processes.
```
==5854== Syscall param socketcall.sendto(msg) points to uninitialised byte...This was found running the pvcs.HistogramSelection test with the server processes running in valgrind. This only appeared on one of the server processes.
```
==5854== Syscall param socketcall.sendto(msg) points to uninitialised byte(s)
==5854== at 0x7DEEB7F: send (send.c:26)
==5854== by 0x1C238365: vtkSocket::Send(void const*, int) (vtkSocket.cxx:567)
==5854== by 0x6C674A9: vtkSocketCommunicator::SendTagged(void const*, int, int, int, char const*) (vtkSocketCommunicator.cxx:745)
==5854== by 0x6C62DC0: vtkSocketCommunicator::SendVoidArray(void const*, long long, int, int, int) (vtkSocketCommunicator.cxx:292)
==5854== by 0x564C276: vtkCommunicator::Send(unsigned char const*, long long, int, int) (vtkCommunicator.h:160)
==5854== by 0x566A31A: vtkMultiProcessController::Send(unsigned char const*, long long, int, int) (vtkMultiProcessController.h:1617)
==5854== by 0x65B96E8: vtkPVSessionServer::GatherInformationInternal(unsigned int, char const*, unsigned int, vtkMultiProcessStream&) (vtkPVSessionServer.cxx:625)
==5854== by 0x65B932F: vtkPVSessionServer::OnClientServerMessageRMI(void*, int) (vtkPVSessionServer.cxx:575)
==5854== by 0x65B71BB: (anonymous namespace)::RMICallback(void*, void*, int, int) (vtkPVSessionServer.cxx:56)
==5854== by 0x6C4A4DC: vtkMultiProcessController::ProcessRMI(int, void*, int, int) (vtkMultiProcessController.cxx:768)
==5854== by 0x6C4A058: vtkMultiProcessController::ProcessRMIs(int, int) (vtkMultiProcessController.cxx:717)
==5854== by 0x69952CA: vtkTCPNetworkAccessManager::ProcessEventsInternal(unsigned long, bool) (vtkTCPNetworkAccessManager.cxx:268)
==5854== Address 0x37411921 is 129 bytes inside a block of size 258 alloc'd
==5854== at 0x4C2E0EF: operator new(unsigned long) (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==5854== by 0x5206427: __gnu_cxx::new_allocator<unsigned char>::allocate(unsigned long, void const*) (new_allocator.h:104)
==5854== by 0x52062F5: __gnu_cxx::__alloc_traits<std::allocator<unsigned char> >::allocate(std::allocator<unsigned char>&, unsigned long) (alloc_traits.h:182)
==5854== by 0x5206109: std::_Vector_base<unsigned char, std::allocator<unsigned char> >::_M_allocate(unsigned long) (stl_vector.h:170)
==5854== by 0x5205BD9: std::vector<unsigned char, std::allocator<unsigned char> >::_M_fill_insert(__gnu_cxx::__normal_iterator<unsigned char*, std::vector<unsigned char, std::allocator<unsigned char> > >, unsigned long, unsigned char const&) (vector.tcc:491)
==5854== by 0x5205857: std::vector<unsigned char, std::allocator<unsigned char> >::insert(__gnu_cxx::__normal_iterator<unsigned char*, std::vector<unsigned char, std::allocator<unsigned char> > >, unsigned long, unsigned char const&) (stl_vector.h:1073)
==5854== by 0x52055FC: std::vector<unsigned char, std::allocator<unsigned char> >::resize(unsigned long, unsigned char) (stl_vector.h:716)
==5854== by 0x11058E8F: vtkClientServerStream::Write(void const*, unsigned long) (vtkClientServerStream.cxx:245)
==5854== by 0x11062065: vtkClientServerStreamInternals::Write(vtkClientServerStream&, void const*, unsigned long) (vtkClientServerStream.cxx:181)
==5854== by 0x11064383: vtkClientServerStream& vtkClientServerStreamOperatorSL<double>(vtkClientServerStream*, double) (vtkClientServerStream.cxx:514)
==5854== by 0x11059E45: vtkClientServerStream::operator<<(double) (vtkClientServerStream.cxx:540)
==5854== by 0x697475B: vtkPVServerInformation::CopyToStream(vtkClientServerStream*) (vtkPVServerInformation.cxx:297)
```
I'm not sure if it is actually a problem or not, but it should eventually be investigated.5.2 (SC-16)Utkarsh AyachitUtkarsh Ayachithttps://gitlab.kitware.com/paraview/paraview/-/issues/16844Enabling Custom Application Dev to show data and create pqPropertiesPanel aft...2017-11-21T16:32:00-05:00Mathieu Westphal (Kitware)Enabling Custom Application Dev to show data and create pqPropertiesPanel after app initializationFor now, all CustomApplication rely on the pqPropertiesPanel and pqPipelineBrowserWidget to show and modify data
It would be nice if data could be shown without a pqPipelineBrowserWidget and that a pqPropertiesPanel created after initia...For now, all CustomApplication rely on the pqPropertiesPanel and pqPipelineBrowserWidget to show and modify data
It would be nice if data could be shown without a pqPipelineBrowserWidget and that a pqPropertiesPanel created after initializationMathieu Westphal (Kitware)Mathieu Westphal (Kitware)