ParaView issueshttps://gitlab.kitware.com/paraview/paraview/-/issues2019-05-22T19:15:42-04:00https://gitlab.kitware.com/paraview/paraview/-/issues/18945Screenshots bad for save all views2019-05-22T19:15:42-04:00W. Alan ScottScreenshots bad for save all viewsIf you create a trace of multiple views, and then save screenshot, with ParaView being smaller, you get weird pictures. It looks like weird offscreen rendering issues.
[screenshotOffscreenBug.py](/uploads/b441a9388ead711be4c2c365339750...If you create a trace of multiple views, and then save screenshot, with ParaView being smaller, you get weird pictures. It looks like weird offscreen rendering issues.
[screenshotOffscreenBug.py](/uploads/b441a9388ead711be4c2c365339750f3/screenshotOffscreenBug.py)
I will attach a .py file.
* ParaView 5.6.0 and also master, Linux, builtin server. 5.6.0 shows weird artifacts that look like an ocular migrane, where master (v5.6.0-1312-g8ce81b8) has three of the four views blacked out.
* IMPORTANT - Make ParaView about half size. This means the main view will be maybe 4" wide by 3" high.
* Edit the script. You will need to point to can.ex2, and an output directory for the screenshot.
* View/ Python. Run the script.
Look at the .png. It will look like junk.
![deleteMeR](/uploads/bc5da0f5fc6afa82eb9a80d30ea694c2/deleteMeR.png)
Here is how I made the script (from memory).
* Linux, master, builtin server.
* Have ParaView fill most of the screen.
* Start Trace.
* Load can.exo. Apply.
* Split horizontal.
* Split vertical.
* Select right view. Split vertical.
* Turn visibility on for all 4 views.
* Save screenshot. Save all views.5.7 (Summer 2019)Utkarsh AyachitUtkarsh Ayachithttps://gitlab.kitware.com/paraview/paraview/-/issues/18944[Module] Almost all plugin XML tests do not perform final image comparison2019-05-09T04:04:40-04:00Mathieu Westphal (Kitware)[Module] Almost all plugin XML tests do not perform final image comparisonLooks like all test declared like this :
```
paraview_add_client_tests(
LOAD_PLUGIN "BagPlotViewsAndFilters"
BASELINE_DIR ${PARAVIEW_TEST_BASELINE_DIR}
TEST_SCRIPTS ${CMAKE_CURRENT_SOURCE_DIR}/BagPlots.xml)
```
r...Looks like all test declared like this :
```
paraview_add_client_tests(
LOAD_PLUGIN "BagPlotViewsAndFilters"
BASELINE_DIR ${PARAVIEW_TEST_BASELINE_DIR}
TEST_SCRIPTS ${CMAKE_CURRENT_SOURCE_DIR}/BagPlots.xml)
```
result in a ctest line which does not contains the --test-baseline
```
/home/glow/work/paraview/paraviewSecondary/paraview_build2/bin/smTestDriver-pv5.6 "--enable-bt" "--client" "/home/glow/work/paraview/paraviewSecondary/paraview_build2/bin/paraview" "--enable-bt" "--test-plugin=BagPlotViewsAndFilters" "--test-script=/home/glow/work/paraview/paraviewSecondary/paraview/Plugins/BagPlotViewsAndFilters/Testing/BagPlots.xml" "--test-directory=/home/glow/work/paraview/paraviewSecondary/paraview_build2/Testing/Temporary" "--data-directory=/home/glow/work/paraview/paraviewSecondary/paraview_build2/ExternalData" "-dr" "--exit"
```
This can easilly be tested by removing the baseline image.
Related: lots of baseline image for plugin tests are also missing.
The only exception is digital rock physics, that declare its test like this:
```
paraview_add_client_tests(
LOAD_PLUGIN "DigitalRockPhysics"
BASELINE_DIR "${CMAKE_CURRENT_SOURCE_DIR}/Data/Baseline"
TEST_SCRIPTS DigitalRockPhysicsAnalysisFilter.xml)
```
@ben.boeckel @utkarsh.ayachitBen BoeckelBen Boeckelhttps://gitlab.kitware.com/paraview/paraview/-/issues/18943Support optional Python linking2020-05-07T21:30:30-04:00Ben BoeckelSupport optional Python linkingVTK currently has its optional linking support masked to always be off because it breaks ParaView. Investigation is necessary to figure out what is wrong with ParaView yet. See !3188.
Cc: @utkarsh.ayachit @danlipsaVTK currently has its optional linking support masked to always be off because it breaks ParaView. Investigation is necessary to figure out what is wrong with ParaView yet. See !3188.
Cc: @utkarsh.ayachit @danlipsahttps://gitlab.kitware.com/paraview/paraview/-/issues/18942Catalyst Examples dont' work: Fatal Python error2019-04-22T15:08:09-04:00Ethan StamCatalyst Examples dont' work: Fatal Python errorMacOS 10.13.6, no superbuild
I am building 0c1430dbe85c92c8ffcb5191aa834a8e361df6fc and added the qttesting fix here: https://gitlab.kitware.com/paraview/qttesting/merge_requests/18
mpi and python(2) were enabled for the paraview build...MacOS 10.13.6, no superbuild
I am building 0c1430dbe85c92c8ffcb5191aa834a8e361df6fc and added the qttesting fix here: https://gitlab.kitware.com/paraview/qttesting/merge_requests/18
mpi and python(2) were enabled for the paraview build.
Once the build finished, I made a new directory for compiling and building /Examples/Catalyst and set ParaView_DIR to the paraview build.
The build finished, but running the Catalyst examples fails:
```
➜ CatalystTesting cd CxxImageDataExample
➜ CxxImageDataExample ls
CMakeFiles CxxImageDataExample Makefile cmake_install.cmake libCxxImageDataExampleAdaptor.a
➜ CxxImageDataExample mkdir rundir
➜ CxxImageDataExample cd rundir
➜ rundir mpirun -np 2 ../CxxImageDataExample ~/GitParaVTK/TripleQuotedStrings/src/Examples/Catalyst/CxxImageDataExample/SampleScripts/feslicescript.py
Fatal Python error: PyThreadState_Get: no current thread
Fatal Python error: PyThreadState_Get: no current thread
[pn1615809:10388] *** Process received signal ***
[pn1615809:10388] Signal: Abort trap: 6 (6)
[pn1615809:10388] Signal code: (0)
[pn1615809:10390] *** Process received signal ***
...
```
It is expected for the Catalyst Examples to work with the provided Catalyst sample scripts.https://gitlab.kitware.com/paraview/paraview/-/issues/18941Color map editor opacity is broken2019-05-22T19:32:38-04:00W. Alan ScottColor map editor opacity is brokenThe Color Map Editor opacity is broken for zero ranges, as well as some larger spans. This is impacting users.
Be sure to test remote server.
* 5.6.0 (and master), builtin server, Linux.
* Open can.exo. All vars on. Apply.
* Open th...The Color Map Editor opacity is broken for zero ranges, as well as some larger spans. This is impacting users.
Be sure to test remote server.
* 5.6.0 (and master), builtin server, Linux.
* Open can.exo. All vars on. Apply.
* Open the Color Map Editor.
* Paint by DISPL.
The opacity ramp doesn't exist.
* Now, Rescale to Custom Data Range. Use -8e9 to 1.6e10.
The opacity ramp won't exist, and there is a vertical line where the ramp should be.
Required for 5.7.0. This is impacting important users.5.7 (Summer 2019)Utkarsh AyachitUtkarsh Ayachithttps://gitlab.kitware.com/paraview/paraview/-/issues/18940tile-display: split view does not work.2019-05-08T16:30:54-04:00Utkarsh Ayachittile-display: split view does not work.Steps:
* start pvserver as follows:
```bash
# this is for debugging- ensures the server side window is a fixed size of
# 400x400 instead of fullscreen.
> export PV_ICET_WINDOW_BORDERS=1
> pvserver -tdx=1 -tdy=1
```
* Connect with Par...Steps:
* start pvserver as follows:
```bash
# this is for debugging- ensures the server side window is a fixed size of
# 400x400 instead of fullscreen.
> export PV_ICET_WINDOW_BORDERS=1
> pvserver -tdx=1 -tdy=1
```
* Connect with ParaView client
* Create Sphere, Apply
* Change representation type to **Surface With Edges**
* Split View vertically
* Show Sphere in this new view
* Interact with the sphere in both the views and observer the server-side window as you do this.
* Both the views are rendered on the same location on the server-side (lower half). This is a bug.
![image](/uploads/a71b854c45c1a9d923166985e7722c31/image.png)5.7 (Summer 2019)Utkarsh AyachitUtkarsh Ayachithttps://gitlab.kitware.com/paraview/paraview/-/issues/18939making "block" selection on non-multblock dataset causes segfault2019-05-07T21:49:53-04:00Utkarsh Ayachitmaking "block" selection on non-multblock dataset causes segfaultSteps
* `paraview -dr`
* Sphere, Apply
* Enter **Select Block** mode by clicking the ![image](/uploads/819277fc75ac66a86a49d9b0aa93aaf3/image.png) button or hitting the *b* key.
* Now simply drag a box on the Sphere -- as if selecting i...Steps
* `paraview -dr`
* Sphere, Apply
* Enter **Select Block** mode by clicking the ![image](/uploads/819277fc75ac66a86a49d9b0aa93aaf3/image.png) button or hitting the *b* key.
* Now simply drag a box on the Sphere -- as if selecting it ---- segfault!5.7 (Summer 2019)Utkarsh AyachitUtkarsh Ayachithttps://gitlab.kitware.com/paraview/paraview/-/issues/18938translucent rendering broken after splitting view with remote rendering in pa...2019-05-15T17:53:25-04:00Utkarsh Ayachittranslucent rendering broken after splitting view with remote rendering in parallelSteps:
* pvserver (at least 2 ranks), `paraview -dr`, connect to server
* set remote rendering threshold to 0 in the **Settings | Render View** tab
* Sphere, Apply
* Change opacity to 0.5
* Split View Vertically and create a **Render Vi...Steps:
* pvserver (at least 2 ranks), `paraview -dr`, connect to server
* set remote rendering threshold to 0 in the **Settings | Render View** tab
* Sphere, Apply
* Change opacity to 0.5
* Split View Vertically and create a **Render View**.
* Interact with the SPhere in the previous view. Sphere will disappear!
* now close the empty view, Sphere will reappear!
* Works with single rank pvserver or when not doing remote rendering.5.7 (Summer 2019)Utkarsh AyachitUtkarsh Ayachithttps://gitlab.kitware.com/paraview/paraview/-/issues/18937selection broken after splitting view when remote rendering in parallel2019-05-07T19:29:36-04:00Utkarsh Ayachitselection broken after splitting view when remote rendering in parallelSteps:
* pvserver (at least 2 ranks), `paraview -dr`, connect to server
* set remote rendering threshold to 0 in the **Settings | Render View** tab
* Sphere, Apply
* Split View Vertically and create a **Render View**
* Now click back on...Steps:
* pvserver (at least 2 ranks), `paraview -dr`, connect to server
* set remote rendering threshold to 0 in the **Settings | Render View** tab
* Sphere, Apply
* Split View Vertically and create a **Render View**
* Now click back on the previous view and try to create a surface selection using ![image](/uploads/cc5a8355e65242f6f56f3cd3c30faeb5/image.png) -- no cells will get selected!
* now close the empty view, selection will start to work.
* selection also works with single rank pvserver or when not doing remote rendering.5.7 (Summer 2019)Utkarsh AyachitUtkarsh Ayachithttps://gitlab.kitware.com/paraview/paraview/-/issues/18936Updating the timekeeper in Python script updates the view, but not the UI2019-05-22T19:56:32-04:00Cory Quammencory.quammen@kitware.comUpdating the timekeeper in Python script updates the view, but not the UISteps to reproduce:
* Launch ParaView (built-in).
* Open `can.ex2`
* View -> Python Shell. Enter the following
```python
>>> tk = GetTimeKeeper()
>>> tk.Time = 0.002
>>> Render()
```
The view updates to 0.002. However, the time value ...Steps to reproduce:
* Launch ParaView (built-in).
* Open `can.ex2`
* View -> Python Shell. Enter the following
```python
>>> tk = GetTimeKeeper()
>>> tk.Time = 0.002
>>> Render()
```
The view updates to 0.002. However, the time value widget in the UI is not updated.
Now, in the Python Shell, run
```python
>>> a = GetAnimationScene()
>>> a.AnimationTime
0.0
```
So the timekeeper's time is controlling the time selected in the dataset, but it is not linked to the timekeeper. The timekeeper's `Time` property should probably be linked to the animation scene `AnimationTime` property.https://gitlab.kitware.com/paraview/paraview/-/issues/18935Multi-block Inspector reset color and opacity troubles2019-05-21T11:54:38-04:00Cory Quammencory.quammen@kitware.comMulti-block Inspector reset color and opacity troubles* Load can.ex2. Apply.
* Open up View -> Multi-block Inspector
* Change a block color and change its opacity. Now try to reset. There are a few issues:
- [x] **Reset Color** in context sets body color to black.
- [x] **Reset Color** doe...* Load can.ex2. Apply.
* Open up View -> Multi-block Inspector
* Change a block color and change its opacity. Now try to reset. There are a few issues:
- [x] **Reset Color** in context sets body color to black.
- [x] **Reset Color** does nothing unless the context menu is invoked by clicking on the color circle.
- [x] **Reset Opacity** does nothing unless the context menu is invoked by clicking on the opacity circle.5.7 (Summer 2019)Utkarsh AyachitUtkarsh Ayachithttps://gitlab.kitware.com/paraview/paraview/-/issues/18934Integer truncation in data value color mapping can cause duplicate colors in ...2019-04-12T18:51:17-04:00Cory Quammencory.quammen@kitware.comInteger truncation in data value color mapping can cause duplicate colors in lookup tableWhen mapping a data value to an index in a color map in vtkLinearIndexLookupMain in vtkLookupTable.cxx, the floating point result can fall just below the integral value to which it ought to be mapped. See specifically the following start...When mapping a data value to an index in a color map in vtkLinearIndexLookupMain in vtkLookupTable.cxx, the floating point result can fall just below the integral value to which it ought to be mapped. See specifically the following starting at vtkLookupTable.cxx:548.
double dIndex = (v + p.Shift) * p.Scale;
// When v is very close to p.Range[1], the floating point calculation giving
// dIndex may map above the highest value in the lut (at index p.NumColors-1)
// in the linear mapping above. This is why we keep an extra copy of the last
// lut value, to avoid extra work in this very hot function.
// It should never be more than 1 off, assert to be sure.
index = static_cast<vtkIdType>(dIndex);
assert(index >= 0 && index <= p.NumColors);
This can lead to duplicate colors when creating a color map, as is exhibited in this color legend:
![image](/uploads/9180fd650f4437cdea116da40cce1038/image.png)
See the duplicate blues between values 1441.1 and 1443.6.
This could be addressed by rounding the mapped data value to the nearest integer.https://gitlab.kitware.com/paraview/paraview/-/issues/18932Add MPEG42021-09-16T20:42:28-04:00W. Alan ScottAdd MPEG4Please add the ability to Save Animation/ MPEG4.Please add the ability to Save Animation/ MPEG4.5.10 (Fall 2021)Cory Quammencory.quammen@kitware.comCory Quammencory.quammen@kitware.comhttps://gitlab.kitware.com/paraview/paraview/-/issues/18931python freezing broken, make target VTK-ALL needs to be VTK_ALL2019-11-18T14:03:54-05:00Jeff Mauldinpython freezing broken, make target VTK-ALL needs to be VTK_ALLIn the paraview repo at the tag for paraview 5.6.0 (v5.6.0 I think), the freeze python capability does not work.
The issue seems to be with the make target named "VTK-ALL" (or vtk-all or Vtk-all). First, a regular expression searching a...In the paraview repo at the tag for paraview 5.6.0 (v5.6.0 I think), the freeze python capability does not work.
The issue seems to be with the make target named "VTK-ALL" (or vtk-all or Vtk-all). First, a regular expression searching attempt misses that target. Second, if the regular expression is fixed to capture VTK-ALL, the freezing fails because of having a dash/minus as part of a variable or function name or something. Changing the make target to have an underscore instead of a dash fixes the problem.5.9 (Fall 2020)Ben BoeckelBen Boeckelhttps://gitlab.kitware.com/paraview/paraview/-/issues/18930`tdx` and `tdy` pvserver args causes segfaults2019-05-20T05:41:38-04:00Mathieu Westphal (Kitware)`tdx` and `tdy` pvserver args causes segfaultsThe `tdx` and `tdy` pvserver args causes a segfault without MPI enabled.
How to reproduce :
* build ParaView with PARAVIEW_USE_MPI=OFF
* run pvserver -tdx=1 -tdy=1
* run ParaView, connect to localhost
* segfault (assert)
`...The `tdx` and `tdy` pvserver args causes a segfault without MPI enabled.
How to reproduce :
* build ParaView with PARAVIEW_USE_MPI=OFF
* run pvserver -tdx=1 -tdy=1
* run ParaView, connect to localhost
* segfault (assert)
```
pvserver: /home/glow/work/paraview/paraview/ParaViewCore/ClientServerCore/Rendering/vtkPVClientServerSynchronizedRenderers.cxx:55: virtual void
vtkPVClientServerSynchronizedRenderers::MasterEndRender(): Assertion `this->ParallelController->IsA("vtkSocketController") || this->ParallelController-
>IsA("vtkCompositeMultiProcessController")' failed.
```
@utkarsh.ayachit (related to !3177)Utkarsh AyachitUtkarsh Ayachithttps://gitlab.kitware.com/paraview/paraview/-/issues/18929Update the VisIt bridge.2020-04-07T09:39:17-04:00W. Alan ScottUpdate the VisIt bridge.We wish to start updating the VisIt bridge every year. We are going to use the bug tracker to schedule this. Thus:
Please update the VisIt bridge.
Create an open bug using this bug report, with a target date a year from this bug...We wish to start updating the VisIt bridge every year. We are going to use the bug tracker to schedule this. Thus:
Please update the VisIt bridge.
Create an open bug using this bug report, with a target date a year from this bug's target date.5.9 (Fall 2020)Ben BoeckelBen Boeckelhttps://gitlab.kitware.com/paraview/paraview/-/issues/18928split view screenshots are bad2019-05-22T19:09:38-04:00W. Alan Scottsplit view screenshots are badScreenshots are bad for split screen. Here is how to replicate:
* Master (v5.6.0-1312-g8ce81b8), Linux, local server. (Superbuild is v5.6.0-159-g4d542fc)
* Sources/ Wavelet. Apply. Surface. RTData.
* Split screen horizontal.
* Turn ...Screenshots are bad for split screen. Here is how to replicate:
* Master (v5.6.0-1312-g8ce81b8), Linux, local server. (Superbuild is v5.6.0-159-g4d542fc)
* Sources/ Wavelet. Apply. Surface. RTData.
* Split screen horizontal.
* Turn visibility on with the wavelet. Surface. RTData.
* File/ Save Screenshot. call it something. png. OK. Save all views. OK.
Look in this file. The Color Legends are totally messed up. Needs to be fixed for 5.7.5.7 (Summer 2019)Cory Quammencory.quammen@kitware.comCory Quammencory.quammen@kitware.comhttps://gitlab.kitware.com/paraview/paraview/-/issues/18927Global data writer is bad for exodus2023-06-15T14:09:08-04:00W. Alan ScottGlobal data writer is bad for exodusI am pretty sure that the Exodus writer is broken for global data, especially for strange timesteps. Here is how to replicate using can.exo.
* Linux, master, builtin server. ( v5.6.0-1290-gb61f946 )
* Load can.exo. All vars on. Appl...I am pretty sure that the Exodus writer is broken for global data, especially for strange timesteps. Here is how to replicate using can.exo.
* Linux, master, builtin server. ( v5.6.0-1290-gb61f946 )
* Load can.exo. All vars on. Apply.
* Extract Timesteps filter. Select Time Range. Timestep 0 to 19. Apply.
* File/ Save Data/ save as Exodus. Save as canA. (It will write out a canA.e file). Save all timesteps.
* CHANGE Extract Timesteps filter. Select Time Range. Timestep 20 to 43. Apply.
* File/ Save Data/ save as Exodus. Save as canB. (It will write out a canB.e file). Save all timesteps.
* Edit/ Reset Session.
* Load canA.e, load all variables, apply.
* Plot global variable over time. Apply. Turn off everything other than NSTEPS. Notice this is a straight line.
* Load canB.e, load all variables, apply.
* Plot global variable over time. Apply. Turn off everything other than NSTEPS. Notice this is a straight line.
* Change visibility of the canB plot global variables over time to also be in the canA.e plot global variables over time.
This should be a straight line, rather than two slanted, parallel lines.
You can see this occurring if you output canA.e and canB.e into two spreadsheets. canB.e restarts NSTEPS at 0, where in the original data it is fairly large.5.12 (Winter 2023)W. Alan ScottW. Alan Scotthttps://gitlab.kitware.com/paraview/paraview/-/issues/18926ParaView master build fails with BUILD_EXAMPLES=ON2019-05-07T21:26:21-04:00Ethan StamParaView master build fails with BUILD_EXAMPLES=ONMacOS 10.13.6
Building 8ce81b84 with python and mpi enabled and BUILD_EXAMPLES=ON
I get to about 9% when make crashes:
```
ThirdParty/QtTesting/vtkqttesting/Examples/TestingDemo.h:3:10: fatal error: 'QMainWindow' file not found
#include...MacOS 10.13.6
Building 8ce81b84 with python and mpi enabled and BUILD_EXAMPLES=ON
I get to about 9% when make crashes:
```
ThirdParty/QtTesting/vtkqttesting/Examples/TestingDemo.h:3:10: fatal error: 'QMainWindow' file not found
#include <QMainWindow>
^~~~~~~~~~~~~
```5.7 (Summer 2019)Cory Quammencory.quammen@kitware.comCory Quammencory.quammen@kitware.comhttps://gitlab.kitware.com/paraview/paraview/-/issues/18924Move Maximum Error parameter in Stream Tracer filter to Integration Parameter...2019-05-07T21:58:35-04:00Kenneth MorelandMove Maximum Error parameter in Stream Tracer filter to Integration Parameters groupIn the `Stream Tracer` filter's properties panel, the `Maximum Error` parameter is listed in the `Streamline Parameters` group. However, according to the documentation, the `Maximum Error` parameter is actually used by the integrator to ...In the `Stream Tracer` filter's properties panel, the `Maximum Error` parameter is listed in the `Streamline Parameters` group. However, according to the documentation, the `Maximum Error` parameter is actually used by the integrator to adjust the step size. Thus, this parameter would be more appropriate to put under the `Integration Parameters` group.5.7 (Summer 2019)Cory Quammencory.quammen@kitware.comCory Quammencory.quammen@kitware.com