> On Feb 24, 5:38 pm, mabshoff <Michael.Absh...@mathematik.uni-
> dortmund.de> wrote:
> > On Feb 24, 12:05 pm, JMD <dre...@math.jussieu.fr> wrote:
> >
> > > on a linux-x86_64 machine
> >
> > > Installing c_lib
> > > scons: `install' is up to date.
> > > Traceback (most recent call last):
> > > File "setup.py", line 1217, in <module>
> > > deps = create_deps(ext_modules)
> > > File "setup.py", line 1208, in create_deps
> > > deps_graph(deps, f, visited)
> > > File "setup.py", line 1175, in deps_graph
> > > this_deps = search_all_includes(f)
> > > File "setup.py", line 1099, in search_all_includes
> > > S = open(filename).readlines()
> > > IOError: [Errno 2] No such file or directory:
> > > 'sage-main/sage/modules/ free_module_element.pyx'
> > > sage: There was an error installing modified sage library code.
> >
> > > ERROR installing SAGE
> >
> > that is cause by #2180. There is a patch for it, but the short
> > solution is to touch the pyrex file in question. Then everything is
> > back to normal.
>
> Burcin did pop up in IRC a little while ago and reported the same
> error. Neither touching the file, not a "sage -ba" nor applying the
> patch from #2180 fixes the problem, so I am fresh out of ideas. It
> only seems to happen on updates, i.e. in the other cases I saw the
> issue a touching of a single file was sufficient. So there is maybe
> more than one bug lurking in that code. Since I have been unable to
> reproduce the harder problem yet any input (or even better solution)
> is welcome :)
This is now #2295, there is also a patch attached to the ticket.
http://trac.sagemath.org/sage_trac/ticket/2295
The problem is caused by the symlink in one of the path components to
SAGE_ROOT. You can go to the "realpath" which doesn't contain the
symlink and build there to work around the problem.
Cheers,
Burcin