Re: Why I am forced to abandon Root

On Thu, Sep 03, 2009 at 11:59:50AM -0700, Tom Roberts wrote:
> ... inability to build Root on a supercomputer at NERSC...
> ... the supercomputer does not support shared objects or X-windows on its compute nodes.

Hi, Tom - from your description, it appears that your compute nodes are not "full-featured linux computers" - since you mention lack of support for shared libraries and lack of X11 libraries (BTW, it is simpler to build libX11.a than to hack ROOT to remove X11 support).

But lack of shared library support suggest that your programming model is closer to some kind of embedded computing - I would even guess that you have to cross compile ROOT (and your applications) using some kind of special cross-compiler, which is probably an old version of GCC that lacks support for some advanced C++ features used in ROOT.

Perhaps you can't even use the stock ROOT Makefiles, I do not think they support cross-compilation (even building 32-bit ROOT on 64-bit Linux has always been flaky).

But not all is lost - as Rene Brun writes, ROOT is more modular than it appears and I estimate that it will take you less work to built a usable subset of ROOT than to reinvent ROOT or use some other package (none of modern-written packages are embedded-friendly - first you have to fight "configure" and it only gets worse after that).

Lack of shared libraries is not a problem - I have build statically linked ROOT applications in the past, ROOT makes you jump through a few hoops, but it does work, and if you have trouble with that, ask this mailing list (and I will probably know the answer and reply).

-- 
Konstantin Olchanski
Data Acquisition Systems: The Bytes Must Flow!
Email: olchansk-at-triumf-dot-ca
Snail mail: 4004 Wesbrook Mall, TRIUMF, Vancouver, B.C., V6T 2A3, Canada