foam-extend-4.0 Debian Testing bison problem fseterr.c
3 Attachment(s)
Hello, I recently attained this computer from our work group at school. It 'had' some version of foam-extend on it but no longer. It does have OPENMPI, PARAVIEW and other 'programming' things on it. It is DEBIAN TESTING and current. Also, the prerequisites;
Code:
sbh@sbh:~/foam/foam-extend-4.0$ sudo apt-cache policy git build-essential binutils-dev cmake flex zlib1g-dev qt4-dev-tools libqt4-dev libncurses5-dev libxt-dev rpm mercurial graphviz python python-dev libiberty-dev | grep -e "Installed" -B 1 Code:
sbh@sbh:~/foam/foam-extend-4.0$ gcc --version Code:
sbh@sbh:~/foam/foam-extend-4.0$ ./bin/foamInstallationTest openmpi-bin has binaries in /usr/bin/ and openmpi-common has txt files in /usr/share/ and libopenmpi-dev (all 3.1.3-11) has libraries in /usr/lib/x86_64-linux-gnu/openmpi/... Speaking of libraries, I did read that openmpi has some weird issue with $LD_LIBRARY_PATH (can't find the website at the moment), so here is that PATH; Code:
/bin is a symbolic link to /usr/bin, verified by ls -ld /bin. Then, from here, https://openfoamwiki.net/index.php/I...end-4.0/Ubuntu I've tried all the prerequisite requirements except for gcc-5 (it's not in my repositories though I suppose I could find them somewhere if it is heavily suggested - partly why I am posting this question. Pinning various repositories is not something I'm too interested in... FrankenDebian). From here, https://openfoamwiki.net/index.php/I...oam-extend-4.0, I have tried the installation (I don't specifically have Ubuntu or OpenSUSE so I think I mostly try to go with Ubuntu). I've also tried the instructions (for my initial attempt of foam-extend-4.1) from this site, https://pingpong.chalmers.se/public/...ml?language=en but it didn't work for me. In a fresh git pull, my etc/bashrc was left alone and my etc/prefs.sh utilized, Code:
I found the installationTest later but earlier I did try going into the ThirdParty/rpmBuild/SOURCES/ and trying to untar different bisons I got from https://ftp.gnu.org/gnu/bison/ and then changing their name to match 2.7.1. They all (even the 2.6.5 and others) complained of this fseterr.c problem. Well I'll stop now as I think I need help to go further. Thank you for reading this far! Bret PS. I'll gladly start over and follow someone's kind instructions! I also did this for the logs (running Allwmake.firstInstall twice [big one first and split into parts, smaller file second as log2.gz]), as noted in https://pingpong.chalmers.se/public/...ml?language=en and the first error I found is; Code:
|
update: attempt after fresh git pull (link to logs [can view online] - too big ?)
Another edit (3rd) - I deleted all and started over, with trying a prefs.sh file from a similar case on cfd-online. The log file, from
Code:
================================================== = Hello, I have a Debian Testing machine, Code:
https://openfoamwiki.net/index.php/I...oam-extend-4.0 and some ideas from questions posted on cfd online; https://www.cfd-online.com/Forums/op...-not-root.html https://www.cfd-online.com/Forums/op...u-18-04-a.html Additionally, I modified the prefs.sh file (attached), and the AllMake.pre and the Allwmake.firstInstall files by simply adding on the same logging instructions I had seen wyldckat use - something like ./Allwmake > $WM_THIRD_PARTY_DIR/logfile 2>&1. It was for my own purposes because I thought it might make it easier to find the problem, especially since I have tried 'stepping through' the make files manually. I started deleting the folder foam-extend-4.0 each time because even after using Allclean, I would find errors in different places, e.g. wmake libso finiteVolume "area" had a 'bad bool casting' complaint, or I would have some problems with gcc49. I also saw confusing (to me at least) things like this in the output logs (attached), Code:
I had some problems attaching files due to size (just the src/Allwmake file gzipped is at 357kB - I really hope my amendments to the 2 make files didn't cause these to be too big, i.e. over 195.3kB - so here is a box.com link to view/download the files: https://app.box.com/s/dn5yh2074e4hgly9harjpmfz92ztnicr - I would fix this issue if I knew how. ), and also it seems to help (for me) to break it down some, so I added a couple additions to the two files Allwmake.firstInstall; Code:
Code:
As mentioned earlier, I am amenable to starting over and following anyone's suggestions. Thank you |
Started over with a few modifications
2 Attachment(s)
Here is my **prefs.sh** file (the only changes from the example version):
(I tried many different bison until i had no error, and between the bounds of 2.4 & <3) Code:
export QT_BIN_DIR=/usr/bin/ Code:
# Load bison and my **AllMake.stage1** has changes; Code:
# bison 2.7 Code:
sbh@sbh:~/foam/foam-extend-4.0/src$ wmakeLnInclude foam Code:
Code:
sbh@sbh:~/foam/foam-extend-4.0/src/decompositionMethods$ grep -nw ~/foam/log.src.Allwmake -e "Error" -C 20 Thank you. PS. I also changed the url to pull parmetis-3.0.4.tar.gz. as the umn.edu site was down. Hope this isn't the source of the error, but the ubuntu package repository says it is the "original"... https://www.isitdownrightnow.com/gla...c.umn.edu.html From **AllMake.stage3**; Code:
# ParMetis |
Greetings brett1,
:eek: Wow, that's a hell of a fight! I didn't study what you've done specifically so far, but I'll address at least the latest post. And I have no idea if GCC 8 can be used with foam-extend 4.0 or not, but I believe that foam-extend 4.1 (which is development) should be compatible with it... but first let's hold off on that... ... Because the "src" folder seems to have built mostly without errors. Even if Scotch and Metis are not compiled, you can still mostly use foam-extend 4.0 as it is right now, namely if you proceed with building the "applications" folder and build the rest. The downside is having to use the other less dynamic decomposition methods, such as "simple" and "hierarchical"... The problem with "stage3" is essentially this: Code:
error: unpacking of archive failed on file /usr/lib/.build-id: cpio: mkdir failed - Inappropriate ioctl for device Please try using the following commands to unpack said TGZ files: Code:
cd $WM_THIRD_PARTY Code:
source $WM_PROJECT_DIR/etc/bashrc Code:
foam If by any chance you want to try and build the ongoing development version 4.1, then run: Code:
cd ~ Best regards, Bruno |
resolved foam-extend-4.0 & 4.1 on Debian Testing
3 Attachment(s)
Hello and apologies for the delay - i can only work on this on weekends.
So summarily, the prescription that Bruno suggested worked fine - on both 4.0 & 4.1. There were some PyFoam associated SyntaxError but so far the first example works fine. It is noteworthy that the 4.1 version has this exact error: https://www.cfd-online.com/Forums/op...ary-print.html and it was easily resolved by (my only attempt) adding the pFinal{} to the system/fvSolution file. I have not studied foam enough to begin to understand why this error occurs but not in 4.0. Hopefully errors for future examples or my own work will be as easily resolved. Here is the modified system/fvSolution file; Code:
/*--------------------------------*- C++ -*----------------------------------*\ Thanks again Bruno! PS log... & log2... are only up to leaving the ThirdParty work, before and after Bruno's prescription. log3... is the complete Allwmake.firstInstall PPS I did install gcc-7 and friends - from oldstable |
Install FOAM-extend-4.1 (actually nextRelease branch of 4.0)
Hi,
I will recommend the below installation guide prepared by Chalmers university: https://pingpong.chalmers.se/public/...o?item=4764601 Regards, |
Quick answers:
Quote:
Quote:
The only thing that wasn't build was swak4Foam, but that is likely because Mercurial (Hg) wasn't installed. But it's only needed if you're planning on using swak4Foam. |
swaq4Foam
Hello,
Thanks for looking at it again! I actually do have mercurial installed.... 4.8.2-1. (not sure of this .hg file yet...) I am interested in using that library so, since it is in ThirdParty I think I'll get the source in tgz format and try your trick again. I'll post my attempts and hopefully positive result, Thanks, Brett EDIT I notice from AllMake.stage5 that it pulls a clone, not a tgz for swaq4Foam. Anyways, will post soonish... |
swaq4Foam
2 Attachment(s)
Hello again,
Apparently (maybe a Debian thing) mercurial didn't have the "server" packages installed, just a "bin" and "common"... So the only errors I saw were two regarding something 'HUGE' - here it is in the log file at around line 2350 & 2353. Haven't spent much time seeing if it is of concern yet... ? Code:
Objects/unitCylinderSearchableSurface.C: In member function ‘void Foam::unitCylinderSearchableSurface::findLineAll(const point&, const point&, Foam::List<Foam::PointIndexHit<Foam::Vector<double> > >&) const’: Thanks Brett |
Quote:
https://www.cfd-online.com/Forums/op...tml#post745817 This is what gschaider replied. I hope it will be helpful. Best wishes. Xiangjie |
All times are GMT -4. The time now is 15:11. |