Salome Windows
This forum is DEPRECATED, please create new topics in the new SALOME forum.
For existing topics please transfer them to the new forum.
This is quite interesting for us. I want to know under which licence type (GPL?) the M$ windows version will be released.
Thanks,
Adil
><font> |
Hi
There will be only one set of sources files of SALOME for any OS (Linux, windows, others).
So the same licence (LGPL) will be applied on this set of sources.
BR
Yves
It is very nearly the end of the year. Does the above delivery date for a Windows version still apply?
Pete
what's up about Windows compilation of Salome ?
Does Open Cascade encounter any particular problem ?
Can we help ?
Cyril Gruau
Hi,
I am just starting to try to install salome under colinux. When this is successfull, it is easy to run salome under windows. It won't be the fastest solution, but I personally prefer this over a VMWare installation, since you can also make a dual boot system out of it (running linux under windows or or just linux).
For anybody who's interested: Aster 8.2 for windows is available for download (this is also the reason for me trying to get salome running under windows).
I've also read that somebody succesfully compiled salome under windows, but I'm unable to find out who and how he did it. Can somebody help me?
regards,
arnold
I succesfully built Salome under coLinux a few weeks ago, and I ran some tests. I had also tested the Vmware version of Salome (CEA), and another one from my own (based on a Knoppix live-DVD that I build).
First, I think that speed of coLinux and Vmware solutions are equivalent.
For coLinux, you need a Windows native X11 server : Hummingbird Exceed 3D, WRQ Reflection X, X-win 32... (commercial), Cygwin/X, Xming, ... (GPL). It must use the native OpenGL acceleration. On my AMD64 computer, glxgears test is 2300 fps (Wion32 native), and from coLinux best X11 servers are Relection X(1600-1900) and Xming (1400-1600). For Salome utilization, the result seems to be correct, but I haven't the time to do lots of tests...
This is the good news.
The bad news is that I used an internal EDF version of Salome (v2.2.5 and v3.1beta) that I can't distribuate outside EDF.
I tryed Salome 2.2.8, downloadable on this web site, but I don't success to build these version under Debian Sarge (that is inside the coLinux) : many source compilation errors, and Mandrake/binary doesn't work...
Another problem : the size of the coSalome (coLinux + Salome) package : I think it can exceed 1 Go, and I have no (free...) host solution...
I will retry the building of a 2.2.8 verson in a few days. If you are interrested, you can help me Let me know.
For Code_Aster and Salome, don't forget that the Aster-Salome module is planned for the end of 2006. Today, you can only exchange meshes and results with somes MED files.
Aimery ASSIRE
Hi Aimery,
do to a ackof time, I haven't been able to continue installing salome on CoLinux. Hopefully I will find some time in the next few days. I will keep you informed.
Regards,
Arnold
I am currently developping the Beta 2 version of CAELinux LiveDVD in a VMWare environment and I will try to distribute a "public" virtual machine including the complete CAELinux distro when it will be ready (Salomé, Aster in PCLinuxOS + a lot of other simulation codes like Calculix, Elmer, OpenFOAM, GetDP, OpenFlower and MBDyn). In this way, you will be able to run the complete simulation environment in the free VMware Player software under Windows.
Concerning the performance, it is clearly not equivalent to what can be obtained from a real Linux installation, but on a recent PC the response time is acceptable for relatively small FE problems.
If you are interested in this VMware version, just let me know and I will try to put it on our website : www.caelinux.com
J.Cugnoni
Hello,
It is planed to realise native salome version for windows (not vmware)?
If yes, then when this version will be realised?
Thanks
Arturas
Hello, Arturas,
Open Cascade is currently investing into creation of a native Windows version (VS 7.1 compiler).
The work is in pogress it's quite difficult to declare an exact date now - the amont of work for certification on native Win32 is quite substantial.
More precise date will be given later.
Best regards,
Mikhail
Dear Mikhail,
You're saying that you are considering using Visual C++ for the windows version of Salome. However, in one of the previous posts one of your collegues (Yves Fricaud) wrote that
"There will be only one set of sources files of SALOME for any OS (Linux, windows, others). So the same licence (LGPL) will be applied on this set of sources."
Does it still hold? Are you planning to also use Visual C++ (I mean its .NET or MFC libraries) for the GUI of Salome or you're gonna stick to Qt GUI and use only Visual C++ for the compilation of the Qt stuff?
Is this windows version going to stay under LGPL?
Regards,
Andriy
There's always the possibility (that I'm trying to validate) of using the QTwin project from sourceforge to compile QT4 under Visual Studio. I've sucessfully built a QT 4.2 technology preview environment unde VC 7.1 that is compatible with my OCC 6.1 build but I haven't yet coupled the 2 together. For OCC I now have a single project file that builds all the components in the correct order, but haven't yet propagated it to Visual Studion 2005 thanks to the issue over variable scoping and I still haven't sorted out the problems with Salome and dynamic memory allocation. But for me its only a hobby...
Pete
a andreykiv wrote: Dear Mikhail, You're saying that you are considering using Visual C++ for the windows version of Salome. However, in one of the previous posts one of your collegues (Yves Fricaud) wrote that "There will be only one set of sources files of SALOME for any OS (Linux, windows, others). So the same licence (LGPL) will be applied on this set of sources." Does it still hold? Are you planning to also use Visual C++ (I mean its .NET or MFC libraries) for the GUI of Salome or you're gonna stick to Qt GUI and use only Visual C++ for the compilation of the Qt stuff? Is this windows version going to stay under LGPL? Regards, Andriy
|
Hello,
You are right, it will be only one set of sources, and it will no any rewrite of GUI on MFC, but qt will be used under Visual C compiler.
Regards,
Sergey.
I now have a single build project for OpenCASCADE running under VS2005 with all dependencies identified - and the loop variable scoping problem can be fixed with a simple change to one of the VS project profile files. Also have Qt 4.1.4/4.2.0/3.3.6 (with help from the QTWin project on sourceforge) compiling and running alongside a Coin3d OpenInventor/SoQT substitute. However you need to beware of the new "manifest" files that get built into DLLs, especially when you try to debug programs against release libraries. I'm starting to look at how this might affect OCC - I've already changed the code to look for a debug version of the OpenGL library environment setting. Next in the frame for a consistent VS2005 port is VTK. Let me know via this forum if your interested in any of these library builds...
Pete
Oh and by the way, the reason I'm using VS2005 and not VS2003 (C++ version 7.1) is that VS2005 Express Edition can be downloaded from the Microsoft site for free, whereas VS2003 still costs £££s, $$$s and €€€s and even though mine is actually VS2005 Enterprise Edition, the projects should still compile on Express.
Pete
Hi Peter,
What you're doing sounds great. Do you think you'll be able to build the entire Salome before Open Cascade does it (they promised the end of this year)? I played a bit with VC++ express edition (what a cute environment), but my experiance did not reach that far yet .
Best regards,
Andriy
I don't know if I can hope to build the whole Salome platform - my target is to concentrate on a platform that runs MESH and possibly GEOM modules but its only a hobby for me. I'm currently trying to build a consistent VS2005 platform for the dependent APIs - I now have a relatively clean build for VTK 5.0.2. The main difficulties are: -
- A plethora of platform independent "makefile" systems - QMake from QT, CMake from VTK and so on.
- Bugs in the APIs - discovered one in QT4 Designer when running one of the ui files from MonkeyStudio that I'm trying to evaluate as the basis of a GUI framework (also one from Qide that I haven't fixed.
- My personal requirement for using the latest versions of the dependent products i.e. moving from Qt3 to Qt4.(4.2.0?) and having consistent clean compilation on Visual Studio 2005 to create release and debug DLLs rather than static libs.
- The SALOME inherent dependency of Unix OS - and I don't want run with cygwin.
- A further plethora of scripting languages requiring run-time support Tcl/Tk.
- The ever increasing size of my PATH environment and the complexities that different toolkits induce by using different ways of locating debug and release libraries.
- My own lack of experience in using the dependent APIs - I haven't got free time to actually understand all the dependencies I'm ripping translating, but I'm learning on the job.
- My wife wondering what I'm actually doing every evening!
My other goal is to research a mechanism for extracting mid-planes (medial axes) from solid models - planning to the exploit Powercrust by Nina Amenta. Useful perhaps for shell meshing of STL files. I wrote my own surface mesher a long time ago but I was always hampered by the lack of availabity of CAD interfaces (even went on a CAS.CADE 1.5 course back in the closed source days). The upshot is that I cannot, on my own, beat the SALOME team even though the original release was slated to be the end of 2005. Whether its worth starting an independent project (through Sourceforge?) is another question. Depends on what the core team suggest, as I wouldn't want to upset them - especially as with the 3.3.2 release there is at least signs of significant activity.
Pete
Could you please explain me why don't you want to use Cygwin for the Windows build? It sounds almost streightforward. Plus, you don't have to have those problems with PATH. You can practically see Cygwin as another Linux. I understand it's not that simple, but ..?
I had some small experience with a finite element program that I managed to build in Cygwin from Linux source. I did had to install some libs, but it worked.
I know that you can have Tlk and Qt under Cygwin. I don't know how easy would that be with VTK, but you can always get VTK linux source and compile it there as well.
Apparently there are some other good reasons why you're not doing this, because the picture that I've just drawn is quite pink .
Best regards,
Andriy
P.S. I envy you that you wife is allowing you to do this stuff in the evening at all .
Andriy
There are a several reasons I don't want to use cygwin for the Windows build, some of them technical and sum of them purely a matter of personal preference. I've no objection to using some of the cygwin tools as part of the build process - in fact VTK builds require this even for VS2005 builds. Here in no particular order are my reasons...
1. My target application is to be designed for end-users who are not developers, and I see it primarily as a suite of exe's and dll's (and of course links to source code) compiled natively on a Windows platform. I don't want the complication of managing cygwin distros as part of my "msi", as this could tend towards a form of dll. I've seen a number of incompatible cygwin apps, requiring different versions of cygwin1.dll to be supported.
2. OCC 6.1 does not have a supported cygwin build - the open source version of QT4 uses the MiniGW compiler not cygwin, similiar issue with VTK. In fact most of the SALOME dependencies appear to have supported Windows/MS Compiler versions. I don't want to spend time developing cygwin versions. I've just downloaded the cycgwin release of OpenFOAM and you have to download a specific version of the compiler otherwise it won't compile. D'oh!
3. I like Visual Studio! I've been debugging code on this product family for years - the debugger swings and the performance of the release code is pretty tight on MS platforms. Now many tell me about all kinds of gizmos for debugging gcc code but I don't believe it can beat VS for debugging productivity, especially when porting other's code rather than writing your own. These days MS compiler standards and strictness are higher than many of the open source versions identifying problems and hidden gotchas - see my earlier thread on http://www.salome-platform.org/forum/forum_9/thread_493
Thats it for me now - need to bath the kids.
Pete
Hi Arnold,
Basically I have no plans to use anything other then VS2005, in my case I use the Enterprise edition. At this moment I'm dedicating my free-time to the development of an OpenCascade architecture framework running on Qt4 (4.2.1) - an early demo can be found on http://myweb.tiscali.co.uk/dolbey/QtOpenCascade/ and there's a related thread on the OpenCascade forum http://www.opencascade.org/org/forum/thread_10338/. I'm intending to use this to provide the basis for interfaces to the Salome meshing routines, but I'm taking my time building my foundation layers. However I want to run on just coherent compiler platform, and I've chosen VS2005 as that compiler, and now have OCC, VTK and Coin3d available on that platform. But back in the good old days when I wrote numerical simulators I used the Watcom compilers (mostly Fortran) and I've often wondered how good OpenWatcom would be be in this arena (www.openwatcom.org) - its used to provide pretty tight code.
Pete
- do you have a plan for Windows version?
MAny thanks and have a Merry Christmas
Good day!
Is anybody has successfully converted salome to windows? where can I find it if any?
thank you very much.
Florante
This forum is DEPRECATED, please create new topics in the new SALOME forum.
For existing topics please transfer them to the new forum.