Hi Willem Jan,

On 16 Jan., 23:05, Willem Jan Palenstijn <w...@usecode.org> wrote:
> > * When "make" fails, it is in SAGE_ROOT/spkg/build/gsl-1.14/src
>
> > * It complains the .libs/libgsl.lax/lt1-init.o does not exist.
>
> > * spkg/build/gsl-1.14/src *does* exist, namely it is a link to .libs/
> > libgsl.lax/libgslcheb.a/init.o
>
> I'm assuming you mean spkg/build/gsl-1.14/src/.libs/libgsl.lax/lt1-init.o here
> as the first filename?

Yes. As I said, one is in SAGE_ROOT/spkg/build/gsl-1.14/src when the
build process fails (which is apparent by the fact that the next thing
in install.log is: Leaving that directory).

> It doesn't make sense that lt1-init.o is a softlink.

Actually I am not sure whether it is a soft or a hard link. How can I
test it?

> According to the link
> command above, it is trying to make a hardlink:
>
> ln .libs/libgsl.lax/libgslcheb.a/init.o .libs/libgsl.lax/lt1-init.o

Yes, it is like this in the install.log

> The syntax for making hardlinks in subdirs is different than that for
> making symlinks in subdirs. Do you maybe have 'ln' aliased to 'ln -s'? That
> would break it. (And it is also something you should never do.)

Well, I would be very surprised if that's what happening. I just
checked that my .bashrc does not contain the string "ln ", so,
aparently it isn't redefined there.

How can I test whether `ln` and `ln -s` are the same?

> Doing 'ln -s a/x a/y' makes a/y a symlink to a/a/x, not to a/x. Doing 'ln a/x
> a/y' makes a/y a hardlink to a/x. In this case
> spkg/build/gsl-1.14/src/.libs/libgsl.lax/lt1-init.o is a symlink to
> spkg/build/gsl-1.14/src/.libs/libgsl.lax/.libs/libgsl.lax/libgslcheb.a/init.o

As I said, I am not sure whether the link is hard or soft. Is the type
of the link apparent in the output of the "ls -l" command that I
posted above?

Cheers,
Simon

-- 
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to 
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Reply via email to