[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Missing symbols in LAPACK/PETSc

On Fri, 2002-01-18 at 02:18, Adam C Powell IV wrote:
> So that would seem to make for two RC bugs: one against gcc-2.95 for 
> these LC* mystery symbols in fortran libs, and the other against 
> gcc-3.0_3.0.3-1 for not building, right?

I think I looked at fixing this in 2.95 a long time ago and discovered
that it would be too hard.  So, while it seems like a reasonable idea to
file a bug report in order to get the problem documented, there isn't a
lot of point in making it release critical and we will probably end up
just tagging it "wontfix".

The 3.0.3 not-buildable problem does merit an RC bug.  It's already
filed in the upstream GNATS database so you can mark it forwarded at the
same time.

> Any ideas on where to begin to investigate a fix?

Have a dig around on the gcc-bugs mailing list archives and see if you
can find any discussion about it.  I think it's a constant-pool handling
problem of some kind; something along the lines of objects being
eliminated but references to them still hanging around.


Reply to: