Mantis - ISSUES

0010176: Add Support for IAR compiler

Add support for IAR compiler. CMakeCCompilerId.c.in has such definition already
/* IAR Systems compiler for embedded systems.
http://www.iar.com [^]
Not supported yet by CMake
#elif defined(__IAR_SYSTEMS_ICC__)
# define COMPILER_ID "IAR" */

0008486: Additional platforms support (Windows CE and Symbian)

Hi, we have implemented the subj and you can find the source here: http://github.com/AlexeyS/cmake/tree/master [^]

There are two main branches in the repo:
* WindowsCE - Windows CE MSVS project generator
* Symbian - Symbian .mmp project generator

Also we have implemented one additional feature: defining a set of valid values for a CMake variable. This feature was inspired by the need to enter long and boring Windows CE platform name (like "Windows Mobile 5.0 Smartphone"). This feature introduces SET_VALID_VALUES command, so CMake now can validate all the values you assign to the variable. Most importantly this allows us to introduce comboboxes for such variables in the CMake GUIs. The comboboxes feature is also implemented for the Qt frontend and for the MFC one. You can find this code in the ValidValues branch.

The rest are merged branches:

* WindowsCEwithValidValues - [WindowsCE + ValidValues] - WindowsCE generator now fills the combobox with platform names for which installed SDKs were detected.

* all - [WindowsCEwithValidValues + Symbian]

We would be happy to contribute this code to the CMake upstream.
Thank you for such a great tool as CMake.

Looking forwards for your comments, criticism, else.

0010435: Volume rendering and opacity are bad for numerous servers

ParaView trunk, 6 servers, XP client (and also Linux client)

Open g1s1.
Change Opacity to 0.3.
(Make sure you are remote rendering.)
Everything goes really, really bad. See attached screenshot.
(Make sure you are remote rendering, local rendering works OK.)

0010434: Integrate variables by volume filter

We need to create a filter that integrates variables by volume of cells, not just adding up the cells. For instance, with AMR data, the integrateVariables filter is useless.

Also, we need to document the integrateVariables filter to clarify whether it does so by volume (which I believe it doesn't, or just by adding up each cell (which I believe it does).

0010432: kwsysProcessKill() does not kill child process on kFreeBSD

It does not kill them because it does not know how to get process list on this platform. The method (ps) used for Linux/FreeBSD should work fine. The patch can be found at:

http://git.debian.org/?p=collab-maint/cmake.git;a=blob;f=debian/patches/kFreeBSD_fixes.diff;h=4b4d78e89ba18ae4dac8de52299ec538fbef6e61;hb=fbe3d22bec52073ec39b9d3845ade36c369a2c76 [^]

0010433: Generate cell volume filter

Users have requested the ability to get at the volume of cells. One of the primary uses of this is to calculate density - density = mass (an input var) / cell volume. This needs to work correctly for strange size cells, such as cells around a cylinder.

Please create a filter that generates and returns cell volume.

0010430: 2d plots - Specify Each of the Axis Labels is missing

We use to have Specify Each of the Axis Labels on 2d plots. This is very useful functionality. Please add it once again.

Plot Over Line filter.
View options (then labeled as Chart Options).
Bottom Axis/ Layout/ Specify Each of the Axis Labels.

0010399: 2.8.0 uses low resoultion title bar icon on Windows

2.8.0 uses low resoultion title bar icon on Windows where as 2.6.4 uses the nicer high resolution version. See attached image.

0008607: (Venus - Medium) Axis layout with logarithmic scale writes numbers on top of each other

ParaView trunk, XP, local server.
Open disk_out_ref.exo
Plot Over Line.
View options (then labeled as Chart Options).
Bottom Axis - Make sure that Show Axis Labels when Space is Available is checked.
Bottom Axis/ Layout/ Specify Each of the Axis Labels.
Select Use Logarithmic Scale When Available.
Generate - From 1 To 1e4 Steps 50. Unselect Use Logarithmic Scale. OK.

Everything is on top of each other. This is a bug.

0006770: ParaView does not read exodus data with many blocks well

When ParaView tries to read in a dataset with a large number of blocks, it acts very, very badly. I have a dataset of 1500 blocks, the whole dataset which is only 17 MBytes in size. paraviewInternals sucks up about 1.2 GBytes of memory reading in this data.

I have a user that would really like this fixed. He does not have any other workarounds, since EnSight also fails.

Setting to high, since there are no workarounds.

File is many.e.

0009007: 2d axis enhancements

* If possible, it would be really nice to have an optional grid right across the object - so you could actually figure out where some point on the object was without moving it half off the screen. Visibility selectable.

* If possible, add axis labels that stick to the edges of the objects. This would be more professional looking for using with screenshots that go in PowerPoint slides. Then, allow user to select which type of axes he/she wants.

* Some way of getting more axis ticks - that are obvious what they represent - would be really nice.

0009692: Make the extract block filter selection aware

We really should make the extract block filter selection aware. In other words, I would like to be able to do the following:
* Open Can.exo.
* Select the plate block.
* Filters/Extract Block.
<> - Now, have the selected block already checked in the Extract Block filter.

Additionally:
In the extract block filter, when you click on a block or series of blocks, have then highlighted.

I am marking this high priority, since I think it would be a very, very useful feature.

0009716: Exodus reader put out too many warning messages

The Exodus II reader is putting out the following error message, up to bazillions of times. This is effectively freezing ParaView.

I think the solution would be to put out some reasonable number of messages, and then stop putting more out (at least for this iteration through the reader) - preferably with a message saying that warnings are being shut off.

Warning: In ..\..\..\..\src\ParaView3\VTK\Hybrid\vtkExodusIIReader.cxx, line 5266
vtkExodusIIReaderPrivate (06CCAB88): You requested attribute 18 in a collection of only 15 attribute.

Ask me for a dataset that shows off this bug. It is OsShip.zip.

0009785: Extract CTH Parts has bugs

We are finding bugs with the ExtractCTHParts filter. It seems to be dropping cells. Here is, I believe, a way to replicate these bugs.

ParaView 3.6.2, Linux, numerous servers (16 pvservers, but I bet this can be replicated on fewer).

Open Dave's large CTH dataset.
Filters/ ExtractCTHParts for material volume fraction 1. Default values, Apply.
Select spcta (input dataset) once again.
Filters/ ExtractCTHParts for material volume fraction 2. Default values, Apply.
Leave everything coloring by it's Material volume fraction.
Go to last time step.
Notice that, on the back of the "floor", is an area that looks like it has three missing "tiles".
Go back one more timestep. Notice that there are new areas that are wrong.

I will include a screenshot.

Since this is a primary tool for work on large CTH datasets, this needs to be fixed for 3.8.

0009926: Surfaces and overlapping lines are displaced.

Here is a bug from the mail group - Benjamin Schindler [bschindler@inf.ethz.ch]:

This is a known issue. This happens due to the coincident topology
resolution techniques employed to rendering surfaces with overlapping
lines correctly. We are working on a fix for this issue.

Utkarsh

>
> I have a very odd problem with paraview 3.6.1 (debug build directly from sources):
>
> When displaying different objects, they are not properly laid into space, probably due to projection problems. One way to very easily see this is to display a line and then use the spreadsheet view to select a vertex of that line. The purple dot which highlights the point is then not on the vertex itself, but at some offset which varies depending on the viewing direction. I took a screenshot (and taking the screenshot itself changed the position of the highlight it seemed) to illustrate the problem.

> This is extremely annoying as also datasets/lines are not properly laid into space so there is no way to tell how the two objects relate to each other. The same problem happens when selecting objects (see second screenshot)

> Is this a known issue/is there already a fix for this?

0010377: OS X internal library paths wrong in installed development tree binaries

Using OS X (10.6.2 in my case), when PARAVIEW_INSTALL_DEVELOPMENT is turned on, and you do a "make install", the library paths in the executable binaries installed in the development tree at [install_prefix]/bin/ are incorrect.

So, when trying to do things like build a branded app against the installed development tree, I get things like:

dyld: Library not loaded: libvtkCommon.pv3.7.dylib
Referenced from: /usr/local/ParaView/bin/kwProcessXML
Reason: image not found

If I run "otool -L" on kwProcessXML (or on any app in that directory), the library paths don't point to the correct lib directories:

emo2:/usr/local/ParaView/bin> otool -L kwProcessXML
kwProcessXML:
libvtkCommon.pv3.7.dylib (compatibility version 0.0.0, current version 0.0.0)
libvtksys.pv3.7.dylib (compatibility version 0.0.0, current version 0.0.0)
/usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 123.0.0)
/usr/lib/libstdc++.6.dylib (compatibility version 7.0.0, current version 7.9.0)

It looks like the application installed at /Applications/ParaView 3.7.0.app/ has all of the correct paths, though, so this part is working fine.

0010278: wedge15 draws incorrectly

The dataset wedge15 draws incorrectly. It should look like a cube. EnSight does draw this dataset correctly.

I will pass this dataset to Nathan.

0005675: Write out global exodus variables

The exodus writer doesn't support this yet. Should be easy to pull the arrays from the input vtkFieldData. The problem will be in differentiating which arrays are global variables (defined over time) and which are not.

0010097: Pipeline browser does not follow selected window.

ParaView trunk, XP, local server.

The pipeline browser does not follow the selected window, but should.
Open ParaView.
Open can.exo, apply.
Open disk_out_ref.exo. Apply.
Split window.
In left window, turn only the eyeball for disk_out_ref.exo on. Highlight disk_out_ref.exo.
In right window, turn only the eyeball for can.exo on. Highlight can.exo.

Now, notice that (if you did exactly as stated below), not matter which window you highlight, can.exo has focus in the pipeline browser. Either 1) the pipeline browser should be window aware, and go to the last highlight for that window, or 2) as you switch, you should always highlight the bottom filter that has visibility.

This would REALLY help usability.

0009929: FindQt4: the search for additional libraries does not prefer the output from qmake

When searching the additional libraries, like e.g. libSM.so parsed from the Qt config file, the libraries are searched using:
find_library(.... PATHS ${some_dir})

This has the effect that on a 64bit system it found /usr/X11R6/lib/libSM.so, while Qt said it should look it /usr/X11R6/lib64/.

The find_xxx() calls should be changed to use HINTS instead of PATHS.
This fixes the problem (tested here).

Alex

0007823: Near plane clipping problem with vktRenderer and vtkCamera

I have noticed that it is very easy to cause the near clipping plane to cut-off the model when in parallel projection mode.

I ran into this recently when adding a vtkBoxWidget and scaling a simple cube a few times after rotating it so that one of the corners points towards the viewer.

I tried calling vtkRenderer::ResetCameraClippingRange() to no avail. vtkRenderer::ResetCamera() fixed the problem, but would lose my camera setting.

After searching fruitlessly in the vtkusers archives, I decided to trace into the code and I discovered that there is code in vtkRenderer::ResetCameraClippingRange() and vtkCamera::SetClippingRange() that modifies the near plane value to "fix" potential camera and depth buffer problems. In my opinion, these fixes are valid, but only for perspective projections. For parallel projections, they are unnecessary and in fact cause the clipping problem I was encountering. To test this, I modified the code and rebuilt and lo and behold, no more near plane clipping!

Attached is my modified version of vtkCamera.cxx and vtkRenderer.cxx along with the patch files relative to version 5.2.0 (only vtkCamera.cxx has changed since then). The changes are small and localized.

0010337: GetPrerequisites.cmake should allow an override when deciding whether an item belongs to the system or not

Sometimes it is necessary to distribute a library in the installer even though it comes from a "system" location, such as /usr/lib, e.g. when it was installed with a binary installer which puts it there.

Similar to gp_item_default_embedded_path_override and gp_resolve_item_override it would make sense to have a gp_resolved_file_type_override function which one could use to override this decision.

0010429: Daylight savings time makes days disappear

If a project's submission time lies near the "missing hours" (roughly midnight to 2AM), the day in which DST occurs goes missing.

Two examples:
http://my.cdash.org/index.php?project=CURL&date=2010-03-15 [^]
http://my.cdash.org/index.php?project=Teem&date=2010-03-15 [^]

Use the "Previous" button, which'll skip from Monday to Saturday. From Saturday, the "Next" button will find Sunday, but won't go on to Monday.

0009998: Installation Issues

This bug collects numerous installation issues people have run into with previous ParaView installs.

-------------------------------------------------------------
All Platforms:
-------------------------------------------------------------
* Make install doesn't work well -- looks like prefix is being ignored.
* Development install doesn't work well.

-------------------------------------------------------------
Linux:
-------------------------------------------------------------
* Plugins are being installed under "/bin".
* Documentation is not installed at standard locations

-------------------------------------------------------------
Windows:
-------------------------------------------------------------
* "bin" and "lib" dirs are created which seems unnecessary. On Windows, we can just put everything under the same dir.

-------------------------------------------------------------
Mac:
-------------------------------------------------------------
* ... please remember that Qt-cocoa needs the qt_menu.nib directory in the Resources. Also, I think I had trouble with QT_QTCLUCENE_LIBRARY not being included in the bundle and a missing empty qt.conf file (to prevent Qt from loading png/tiff plugins). ~ Michael Wild

0009423: Python Trace generator doesn't do screenshots.

* ParaView trunk, XP, local server.
* Tools/ Python Shell/ Trace/ Start trace
* Sources/ Cone. Apply.
* File/ Save Screenshot. Save this somewhere.
* Tools/ Python Shell/ Trace/ Stop trace
* Tools/ Python Shell/ Trace/ Show trace
<> - there is no save screenshot. I tested this by saving the trace, and then executing it, and it did not save a file.

<>

Since this is such basic functionality, it must be fixed for 3.6.2.