This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: gfortran test runs
On Wed, 2004-08-25 at 11:37, Richard Earnshaw wrote:
> On Wed, 2004-08-25 at 15:00, Andrew MacLeod wrote:
> > On Wed, 2004-08-25 at 09:45, Daniel Jacobowitz wrote:
> > > On Wed, Aug 25, 2004 at 09:07:39AM -0400, Andrew MacLeod wrote:
> > > > However, if I then cut and paste the execution line above from above
> > > > and run the file by hand, it compiles and executes just great, no
> > > > relocation error, no nothing. Within the test framework, it just
> > > > fails.
> > >
> > > The testsuite sets various things in LD_LIBRARY_PATH. Could one of
> > > them contain an old copy of GMP, or clobber an existing LD_LIBRARY_PATH
> > > pointing at GMP?
> > >
> >
> > Oh, ahh. So I see my machine does have a libgmp in /usr/lib.
> >
> > grumble.
> >
> > Ok, I'll reconfigure and install 4.1.3 into /usr/lib and see how that
> > fares.
>
> It really shouldn't be necessary for gcc developers to have to install
> components into /usr/lib (or any other system directory) in order to be
> able to develop and test the compiler.
>
> I think the current situation is seriously broken and needs to be
> addressed urgently.
>
I couldnt agree more.. I ended up screwing around for quite a while
trying to sort this out, and it wouldnt have been a problem if my
LD_LIBRARY_PATH were honored, or at least looked at... I hate it when
things dont "just work"
but then I suspect this is probably a can of worms :-)
Andrew