ParaView issueshttps://gitlab.kitware.com/paraview/paraview/-/issues2023-12-05T12:03:58-05:00https://gitlab.kitware.com/paraview/paraview/-/issues/22410vtkblockcolors disappears after using volume rendering2023-12-05T12:03:58-05:00W. Alan Scottvtkblockcolors disappears after using volume renderingThe option to paint by vtkblockcolors disappears after using volume rendering. Here is how to replicate.
* 5.12.0-RC1, builtin, Windows
* Read can.exo. Apply.
* Paint by id
* Volume.
* Surface.
vtkBlockColors is no longer available to...The option to paint by vtkblockcolors disappears after using volume rendering. Here is how to replicate.
* 5.12.0-RC1, builtin, Windows
* Read can.exo. Apply.
* Paint by id
* Volume.
* Surface.
vtkBlockColors is no longer available to paint with. This is a bug.https://gitlab.kitware.com/paraview/paraview/-/issues/22268Color Map horizontal scrollbar should not be present by default2023-08-18T10:31:53-04:00Phil SmithColor Map horizontal scrollbar should not be present by defaultWhen I want to change color maps or adjust something, within the color map editor, there are a series of buttons down the right-hand side of the color map. They can’t be accessed without scrolling over in the horizontal scroll bar a tiny...When I want to change color maps or adjust something, within the color map editor, there are a series of buttons down the right-hand side of the color map. They can’t be accessed without scrolling over in the horizontal scroll bar a tiny bit. It’s hidden and if you don't realize they are there, you could end up missing them. Making the pane big enough to avoid this ends up taking up a lot of precious screen real-estate.
It appears the length of the text "Enable Freehand Drawing Of Opacity Transfer Function" is causing the horizontal scroll bar to appear. Is there some shorter way of saying this? Or maybe use two lines to say it.
My users have also noticed even if they re-set the size of the pane, it does not always stay that way. Maybe it is because they need the extra screen real-estate and re-size it again, but even if that is the case, it just goes to show that we need to keep it the standard size without the horizontal scroll bar.
Another suggestion was made that we could place the buttons that are on the right hand side of the scroll bar at the top of the panel, like they were in past versions of the software, so it is not necessary to scroll horizontally to see them.
This panel should not have a horizontal scroll-bar when it is the default size. It would make it much easier for the users to use it that way.5.13 (Summer 2024)https://gitlab.kitware.com/paraview/paraview/-/issues/22244Offsets for bounding ruler2023-07-28T15:33:14-04:00Phil SmithOffsets for bounding rulerPlease provide a way to provide an offset for the Bounding Ruler filter. If the bounds are close they can overlap and be un-readable. Please see discourse issue [12579](https://discourse.paraview.org/t/offsets-for-bounding-ruler/12579) f...Please provide a way to provide an offset for the Bounding Ruler filter. If the bounds are close they can overlap and be un-readable. Please see discourse issue [12579](https://discourse.paraview.org/t/offsets-for-bounding-ruler/12579) for more information5.13 (Summer 2024)https://gitlab.kitware.com/paraview/paraview/-/issues/215365.11.0-RC1 Hierarchical Fractal source has rendering artifacts with Ghost Lev...2022-09-08T15:54:37-04:00Ethan Stam5.11.0-RC1 Hierarchical Fractal source has rendering artifacts with Ghost Levels enabled@cory.quammen @patchett2002 @maxwelljackson002
Using 5.11.0-RC1 on MacOS and Windows, when rendering the surface of the heirarchical fractal source with ghost levels enabled, the rendering is messed up.
Steps to reproduce:
1. Sources ...@cory.quammen @patchett2002 @maxwelljackson002
Using 5.11.0-RC1 on MacOS and Windows, when rendering the surface of the heirarchical fractal source with ghost levels enabled, the rendering is messed up.
Steps to reproduce:
1. Sources -> hierarchical fractal -> default values -> apply
2. Change renderview from 2D to 3D
3. Color by depth and choose the surface representation.
4. Reset zoom to data and rotate the dataset
![Screen_Shot_2022-09-06_at_2.58.45_PM](/uploads/02c24472e29703d8647ac94a61e5b371/Screen_Shot_2022-09-06_at_2.58.45_PM.png)https://gitlab.kitware.com/paraview/paraview/-/issues/21234Catalyst calls fork() in python2024-02-14T15:30:31-05:00Boonthanome NouanesengsyCatalyst calls fork() in pythonI got a message from a Flag developer that somewhere in their Catalyst adaptor, fork() is being called, which is unsafe. Here is his message below.
> Hmm, this is "interesting:" somewhere in the PVI adapter, Python calls fork(), which ...I got a message from a Flag developer that somewhere in their Catalyst adaptor, fork() is being called, which is unsafe. Here is his message below.
> Hmm, this is "interesting:" somewhere in the PVI adapter, Python calls fork(), which is extremely unsafe in MPI codes. A new build using a different version of OpenMPI detected and complained about this. I was able to replicate on CTS1Ifast, as something in the coprocessor init script, but I have no idea how to fix it.
> https://www.open-mpi.org/faq/?category=tuning#fork-warning
@cory.quammen Could you look into this?
@patchett2002https://gitlab.kitware.com/paraview/paraview/-/issues/20611vtk data is not usable2021-03-30T21:52:38-04:00W. Alan Scottvtk data is not usableThe data on the vtk.org site is not usable by anything other than testing code. Please change it to have the same format as paraview. Basically, a human readable and usable data tarball, and a second tarball that installs in a hidden d...The data on the vtk.org site is not usable by anything other than testing code. Please change it to have the same format as paraview. Basically, a human readable and usable data tarball, and a second tarball that installs in a hidden directory that is for the automatic testing codes. I would recommend calling the human readable VTKData, and the datasets for testing VTKTestingData.https://gitlab.kitware.com/paraview/paraview/-/issues/20503Non axis aligned reflect2021-12-09T03:46:21-05:00W. Alan ScottNon axis aligned reflectThe reflect filter allows users to reflect on min or max - perpendicular to the X, Y or Z axis. Please add the ability to reflect off of a plane, using our plane tool (for instance with clip), where the angle of the plane can be arbitrary.The reflect filter allows users to reflect on min or max - perpendicular to the X, Y or Z axis. Please add the ability to reflect off of a plane, using our plane tool (for instance with clip), where the angle of the plane can be arbitrary.https://gitlab.kitware.com/paraview/paraview/-/issues/20111Cinema related issues2024-03-07T20:47:15-05:00Utkarsh AyachitCinema related issuesSupporting basic Cinema use-case, and removing old code
---------------------
* [x] Remove **Import Cinema Image Database** option from File menu. This is older version of the file format and not used. We can bring this capability back ...Supporting basic Cinema use-case, and removing old code
---------------------
* [x] Remove **Import Cinema Image Database** option from File menu. This is older version of the file format and not used. We can bring this capability back in future for the current spec.
* [x] **Image Extract Generators** created on a **Render View** should offer a setting to choose camera-model i.e. whether to save images from various camera angles. Default should be static. Since we now incorporate camera-model for any image extract generator, we no longer need to support a separate "Cinema Extract Generator", so that can be removed.
* [ ] **CIS Extract Generator** will be added to generate [composable image set](https://github.com/cinemascience/cinema_projectmanagement/tree/master/design/2020/image_proposal) instead of simple pngs/jpegs currently supported as image extracts. **This is not targeted for 5.9, and relies on availability of a cinemascience library to save CIS files**.
* [x] When generating extract either immediately (using **File | Save Extracts**) or in situ (using **File | Save Catalyst State**), add an option to let the user choose to **Create Cinema Specification**. When checked, when saving the extracts, after saving the file, ParaView will use Cinema-API add a row to the Cinema database. This is done using an library provided by the **cinemascience**, which will be added to ParaView as TPL (third-party library).
* [ ] [cinema_python](https://gitlab.kitware.com/cinema/cinema_python) should be removed from ParaView since we no longer needed it. The project should be marked deprecated and archived.
Supporting client-server use-cases
---------------------------------
* [x] **Saving image extracts on server-side**: In client-server configurations, ParaView saves data-extracts on the server, while image extracts are saved on the client. This makes it largely impractical to save a Cinema database that has both image and data products. It makes sense to save both image and data products on the same end. Let's save everything consistent on the server-side. The advantage with this choice is that the application will behave the same way when performing the extract generation when connect to a remote HPC server via a client or simply running a script using `pvbatch` or via Catalyst on the HPC system (relates to paraview/paraview#19073).
* [x] **Change `Save Extracts` and `Generate Catalyst Script` options**: Currently, both of these options offer user separate locations of data and image extracts. This was largely introduced as a workaround to handle the separate process-destinations for data and image extracts. Now that they are the same, we can get rid of this separation and simply the options.
Viewing results
----------
This is just a initial draft and not expected to be implemented for 5.9.
* [x] Add the cinema_viewer python module to ParaView. Ideally, this module should take in command line arguments that point to a cinema database add view it. Thus, once extracts are generated from ParaView (in GUI, in batch or in situ), you should be able to do the following to view:
```bash
> pvpython -m cinema_viewer --dbase <path-to-cinema-database>
```
That should launch a lite web-server locally that the user can point his browser at to explore the database.5.13 (Summer 2024)https://gitlab.kitware.com/paraview/paraview/-/issues/19951vtkIossReader::ReadMesh gets called even after vtkIossReader::ReadMetaData re...2021-08-17T11:18:14-04:00Utkarsh AyachitvtkIossReader::ReadMesh gets called even after vtkIossReader::ReadMetaData returns failureSteps:
* Edit `TestIossExodus` to not specify any filename.
* Run the test. The output shows two errors about missing file coming from REQUEST_INFORMATION and REQUEST_DATA passes. If REQUEST_INFORMATION fails, I don't think REQUEST_DATA ...Steps:
* Edit `TestIossExodus` to not specify any filename.
* Run the test. The output shows two errors about missing file coming from REQUEST_INFORMATION and REQUEST_DATA passes. If REQUEST_INFORMATION fails, I don't think REQUEST_DATA should ever get called. What's going on?https://gitlab.kitware.com/paraview/paraview/-/issues/19239CSV writer: limit arrays to pass-through widget to arrays of selected a2020-02-08T00:24:08-05:00Utkarsh AyachitCSV writer: limit arrays to pass-through widget to arrays of selected a* disk_out_ref, check all variables, apply
* save data, choose csv, and check *Choose Arrays To Write*, you'll see the following widget
![image](/uploads/f03894e887090a1b798d480597f5e5e4/image.png)
Note the array selection shows cell ...* disk_out_ref, check all variables, apply
* save data, choose csv, and check *Choose Arrays To Write*, you'll see the following widget
![image](/uploads/f03894e887090a1b798d480597f5e5e4/image.png)
Note the array selection shows cell arrays, but since **Field Association** is set to **Points**, cell arrays won't be saved and hence should not be shown in the widget.https://gitlab.kitware.com/paraview/paraview/-/issues/19156Settings does not scale correctly2019-07-18T13:53:08-04:00W. Alan ScottSettings does not scale correctlySettings dialogs do not scale correctly with Edit/ Settings/ Scale font.
* Linux, 5.7.0-RC1, local server.Settings dialogs do not scale correctly with Edit/ Settings/ Scale font.
* Linux, 5.7.0-RC1, local server.https://gitlab.kitware.com/paraview/paraview/-/issues/19143Hard to select directory for state file inputs2019-07-03T10:07:59-04:00Kenneth MorelandHard to select directory for state file inputsWhen you load a state file, one of the handy options is to select a directory from which to load files so that you can transfer a state file from one computer to another with a different file system. However, when you try to select a dir...When you load a state file, one of the handy options is to select a directory from which to load files so that you can transfer a state file from one computer to another with a different file system. However, when you try to select a directory to use with the file browser, it is more difficult than it should be. You should be able to browse to the directory and hit `OK`, but you cannot.
To demonstrate, download this state file that loads a file in the `Examples` directory distributed with ParaView: [FieldAtTime0.pvsm](/uploads/6ba60a3b24c846f0dc9aa2b6d1c6cf55/FieldAtTime0.pvsm).
1. `File` -> `Load State...`
2. Select the [FieldAtTime0.pvsm](/uploads/6ba60a3b24c846f0dc9aa2b6d1c6cf55/FieldAtTime0.pvsm) that you downloaded.
3. Change `Load State Data File Options` to `Search files under a specified directory`.
4. Click the `...` button next to the new `Data Directory` option to open the file browser.
5. In the file browser, double-click the `Examples` directory in the list of common system directories. This is the directory you want to select.
At this point you should be able to just select `OK` to select this directory. But you can't. You have to go up a directory, then select the subdirectory, and then click `OK`, which is awkward as heck. Compare that to a normal system file browser in directory mode. (For example, you can see it in the CMake GUI when selecting a directory.)
The way it should work is for the `OK` button to be active if no files/directories are selected. In this case, the currently navigated folder will be selected.https://gitlab.kitware.com/paraview/paraview/-/issues/18960Selection when Render Points As Spheres enabled prevents selection2020-11-19T14:48:04-05:00Andrew BauerSelection when Render Points As Spheres enabled prevents selectionSteps to reproduce:
1. Create a `Point Source` with a single point
2. Under `Styling` set `Point Size` to 5 and enable `Render Points as Spheres`
3. Uncheck the `Show Point` option under `Point Parameters`
4. Fit to screen and try to...Steps to reproduce:
1. Create a `Point Source` with a single point
2. Under `Styling` set `Point Size` to 5 and enable `Render Points as Spheres`
3. Uncheck the `Show Point` option under `Point Parameters`
4. Fit to screen and try to select the point, it won't show being highlighted with the purple highlight color
Can split the view and have the second view be the Spreadsheet view and enable `Show only selected elements.` to verify that the point isn't being selected. If `Render Points as Spheres` is disabled then it allows the point to be selected.https://gitlab.kitware.com/paraview/paraview/-/issues/18332Feature to change 2d linechartview background2018-07-30T13:45:17-04:00W. Alan ScottFeature to change 2d linechartview backgroundPlease add the ability to customize the background of 2d linechartviews. Probably should do this through a palette, not the Properties tab background.Please add the ability to customize the background of 2d linechartviews. Probably should do this through a palette, not the Properties tab background.https://gitlab.kitware.com/paraview/paraview/-/issues/18325Should we remove Live Programmable Source?2024-03-07T20:47:50-05:00Andrew BauerShould we remove Live Programmable Source?This is not in 5.5.2 and maybe there will be a better solution in 5.6 anyway.This is not in 5.5.2 and maybe there will be a better solution in 5.6 anyway.5.13 (Summer 2024)https://gitlab.kitware.com/paraview/paraview/-/issues/17572Rotational extrustion filter creates weird cells2023-06-15T12:14:59-04:00W. Alan ScottRotational extrustion filter creates weird cellsThe Rotational Extrusion filter creates really weird cells. Here is how to replicate:
* 5.4.0, Linux, local server.
* Sources/ UnstructuredCellTypes with a 50x1x50 dimension. Default cell type (Hex). Apply.
* Slice. Origin: 0, 1e-6, ...The Rotational Extrusion filter creates really weird cells. Here is how to replicate:
* 5.4.0, Linux, local server.
* Sources/ UnstructuredCellTypes with a 50x1x50 dimension. Default cell type (Hex). Apply.
* Slice. Origin: 0, 1e-6, 0. Normal: 0, 1, 0. Triangulate the Slice: Off
* RotationalExtrusion: Resolution: 36. Angle: 90
* Surface with Edges. Notice that the cells don't flow correctly across the 90 degree visualisation, but have a weird slant to them. I will attach a picture.
![weirdRotationalExtrusion](/uploads/35b86be77260f98ea3cd8c47c5fd6c4f/weirdRotationalExtrusion.png)https://gitlab.kitware.com/paraview/paraview/-/issues/17237State file is bad if selection is on2023-06-15T10:57:51-04:00W. Alan ScottState file is bad if selection is onIf a user saves a state file when selection is on, that state file is broken. Replicate as follows:
* Master, Linux, local server.
* Sources/ Wavelet. Apply. Surface. RTData.
* Select Cells On. DON'T DO THE SELECTION. Just proceed ...If a user saves a state file when selection is on, that state file is broken. Replicate as follows:
* Master, Linux, local server.
* Sources/ Wavelet. Apply. Surface. RTData.
* Select Cells On. DON'T DO THE SELECTION. Just proceed to next step.
* File/ Save State.
* Reset Session.
* File/ Load State.
Your mouse controls are bad.https://gitlab.kitware.com/paraview/paraview/-/issues/17119"Dashed" line when superposed with geometry2020-05-07T13:36:54-04:00Mathieu Westphal (Kitware)"Dashed" line when superposed with geometryLines appears "dashed" when superposed with geometry, even though they are not.
How to reproduce :
* Open ParaView
* Create a wavelet, apply
* Representation -> Surface with edges
* lines DOES NOT appear dashed
* Representation ->...Lines appears "dashed" when superposed with geometry, even though they are not.
How to reproduce :
* Open ParaView
* Create a wavelet, apply
* Representation -> Surface with edges
* lines DOES NOT appear dashed
* Representation -> Surface
* Add an axes grid
* Axes grid line appears dashed, especially wheen zooming and unzooming.
![bug](/uploads/c66c9f8a1b89a410a0edd46dff8d6f66/bug.png)![bug2](/uploads/4420fa690dd1afaf2f882277e34ae733/bug2.png)![bug3](/uploads/f4fe498e921e00a5afcd78bccec96201/bug3.png)https://gitlab.kitware.com/paraview/paraview/-/issues/17015min and max in the color map editor/ transfer function2023-06-20T09:34:37-04:00W. Alan Scottmin and max in the color map editor/ transfer functionPlease allow all values of the color transfer function to be changed in the Color transfer function values spreadsheet. This includes the min and max of the Value column. Get here by Edit Color Map/ click on the Advanced gear (lower on...Please allow all values of the color transfer function to be changed in the Color transfer function values spreadsheet. This includes the min and max of the Value column. Get here by Edit Color Map/ click on the Advanced gear (lower one).
Sources/ Sphere/ color by Normals is a good example.https://gitlab.kitware.com/paraview/paraview/-/issues/16893mouse zoom is confusing/ wrong with animation/ follow data2018-02-26T13:15:57-05:00W. Alan Scottmouse zoom is confusing/ wrong with animation/ follow dataThe mouse zoom is confusing/ wrong with the animation view/ add camera track/ follow data. The primary issue is that zooming with the left mouse button and zooming with the wheel act differently. This is extremely confusing to users, a...The mouse zoom is confusing/ wrong with the animation view/ add camera track/ follow data. The primary issue is that zooming with the left mouse button and zooming with the wheel act differently. This is extremely confusing to users, and needs to be fixed. Two options would work - some type of a popup warning telling users they are zooming with the wheel rather than the right button, or make the behavior the same.
Here is how to replicate.
* Linux, master, local server.
* Open can.exo. Apply.
* Rotate the can around. Select a half dozen cells on the lower part of the can.
* Extract selection. Apply.
* Reset.
* (Note that ExtractSelection should still be selected in the Pipeline Browser)
* View/ Animation View/ New track/ Camera/ Follow Data/ +
* Turn visibility on for can.exo in the pipeline browser.
* Not necessary, but definately pretty - change variable to GlobalNodeId.
* Play. Note that we stay zoomed on the extract selection elements. This is as it should be.
* Zoom out with the right mouse button. Play. Note what happens to the view - it resets to have the Extract Selection fill the view. This is as it should be.
* Zoom out with the mouse wheel. Play. Notice what happens to the view - nothing. This is NOT as it should be.