OF7 ParaView 5.6.0 compilation error qt5 (openSUSE 15.1)
1 Attachment(s)
I am trying to get my hands on OF7 but when I was trying to compile ParaView, I get this error:
Code:
Version information Code:
CMake Error at /usr/lib64/cmake/Qt5/Qt5Config.cmake:28 (find_package): Code:
./makeParaview -qmake /usr/local/Qt-5.9.3/bin/qmake -python -mpi > log.makePV 2>&1 I have attached my log file alongside this thread. What's the solution for this? |
Quick answer: I've started writing instructions for this at openfoamwiki.net here: https://openfoamwiki.net/index.php/I...SUSE_Leap_15.1 - I'm still testing it, but have gotten ParaView to start compiling without problems.
In step #2 are the commands for installing the necessary packages. No need to install Qt 5.9.3 as a side installation, given that Qt 5.9.7 is provided with openSUSE 15.1. edit: In step #9 are the commands needed for building ParaView... |
I just posted a procedure for building ParaView 5.6.0 with Qt5 for OpenFOAM-7 and for OpenFOAM-v1906 running under Ubuntu 18.04.
https://www.cfd-online.com/Forums/op...u-18-04-a.html |
@gcengineer: openSUSE is a bit of different beast from Ubuntu... for example, I needed the fix for xmlpatterns that is provided here: https://build.opensuse.org/package/v...patch?expand=0 - this is because in openSUSE there is nothing similar to the "qt5-default" package on Ubuntu...
@mdhfiz: I've finished testing to build ParaView 5.6.0 and updated the wiki page accordingly: https://openfoamwiki.net/index.php/I...SUSE_Leap_15.1 As mentioned in the previous paragraph, I had to apply a fix to the code in order for things to fully work as intended. In the meantime, I'm leaving the machine running overnight, to see if it builds OpenFOAM 7 properly... edit: Built successfully. |
@wyldckat it works fine, thanks bruno! :D
|
Compilation error (vtkPVServerManagerDefaultPythonD)
Hello,
I'm using OpenSUSE Leap 15.0. I followed all the exact instructions proposed by @wyldckat but I'm still getting the following compilation error: Code:
[ 94%] Building CXX object VTK/Wrapping/Python/CMakeFiles/vtkPVServerManagerDefaultPythonD.dir/vtkSMViewExportHelperPython.cxx.o Does anyone know how to solve this problem? Best, Mateus |
Quick answer: The actual error message in the build with Python is this:
Code:
[ 94%] Linking CXX shared library ../../lib/libvtkpqCore-pv5.6.so Code:
[ 90%] Linking CXX shared library ../../lib/libvtkpqCore-pv5.6.so This is really strange... OK, a bit of searching online and it might be a bug in binutils: https://gitlab.kitware.com/paraview/...uild/issues/99 What does the following command give you? Code:
ld --version |
Quote:
This is what I got: Code:
GNU ld (GNU Binutils; openSUSE Leap 15.0) 2.31.1.20180828-lp150.5.3 Mateus |
Quick question: How many cores does your machine have and how much RAM?
I ask this because it could be due to not enough RAM... I had done a test run with 4 cores and 4GB of RAM inside a virtual machine, but had problems because it would crash due to not enough RAM... but I don't remember what was the error message it gave me... |
Quote:
However, I just checked and the current version I have already is the latest version. In the link you sent to me, they said they were able to compile downgrading binutils, instead of upgrading. I will try it and see what happens. Thanks, Mateus |
Quick note: The ld version I used on my VM was this:
Code:
2.31.1.20180828-lp151.2 Code:
2.31-lp151.2.29-x86_64 Oh... wait, you're using OpenSUSE 15.0... I don't remember if OpenSUSE has a clear upgrade path from 15.0 to 15.1... |
Quote:
Code:
[100%] Building CXX object Plugins/StreamLinesRepresentation/CMakeFiles/StreamLinesRepresentation.dir/moc_StreamLinesRepresentation_Plugin.cpp.o So the solution that worked for me was to add the following code before the first include statement in the CMakeList.txt file: Quote:
Bruno, thank you for that step by step procedure! If it was hard with that, imagine if I didn't have had to begin with. And also for pointing out the issue with binutils! Best, Mateus |
Compile OF7 on Leap 15.1
Following the excellent Instructions on openfoamwiki about the Installation procedure for OF7 on Leap 15.1, upgraded from Tumbleweed, the Installation went flawlessly, when two additional issues were fixed:
1. Paraview complained about a missing "asm/errno.h" file. This was due to the directory "/usr/include/asm" being initially empty and was fixed by a symlink to /usr/include/arch-x86/asm", which may differ for other cases. 2. OF7 complained about some missing "metis" declarations, using the system metis. Adding the directory "bin" to the third line in Nr. 6 of the Instructions echo "export METIS_ARCH_PATH=/usr/bin"... fixed this. Just my 2c. |
Hi CLF,
Many thanks for the information! But I'm a bit confused...
In addition, did you install the 32-bit or the 64-bit architecture of openSUSE? Or did you install both architectures? Best regards, Bruno |
@wyldckat
Thanks for Your response. First of all, I want to thank You very much for the tremendous effort You undertook supporting OF. To Your Questions: I did 2.: "upgrade" from Tumbleweed-x86_64-Snapshot 20191207 to Leap 15.1-x86.64 Therefore, I am running a 64 bit System. However, compilation was done for 32 bit Integer support, following point 7 in openfoamwiki.net/Installation/Linux/OpenFOAM-7/openSUSE/Leap_15.1. Hoping, that this answers Your Questions. Sincerely, CLF |
Hi CLF,
You're welcome and many thanks! So my doubt here is that Tumbleweed is meant to be the bleeding edge development line of openSUSE and Leap is the stable development line. So from one perspective, it's a downgrade going from Tumbleweed to Leap, given that you must have went back on the versions for several of the packages made available in openSUSE. This to say that without testing this myself, I'm not sure if the bugs you've found and reported were due to Tumbleweed and not Leap 15.1. That said, I have seen openSUSE do some strange changes within the same version, so it's possible that Metis changed it's installation procedure within Leap 15.1 after I wrote those instructions on the wiki. I'll have to go through them once again, to check if anything/what changed. Best regards, Bruno |
All times are GMT -4. The time now is 16:07. |