This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug fortran/58113] [4.9 Regression] gfortran.dg/round_4.f90 FAILs
- From: "burnus at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Fri, 16 Aug 2013 06:18:10 +0000
- Subject: [Bug fortran/58113] [4.9 Regression] gfortran.dg/round_4.f90 FAILs
- Auto-submitted: auto-generated
- References: <bug-58113-4 at http dot gcc dot gnu dot org/bugzilla/>
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58113
--- Comment #3 from Tobias Burnus <burnus at gcc dot gnu.org> ---
Quoting round_4.f90, which somehow no one seems to read ...
! Test whether I/O rounding works. Uses internally (libgfortran) strtod
! for the conversion - and sets the CPU rounding mode accordingly.
!
! If it doesn't work on your system, please check whether strtod handles
! rounding and whether your system is supported in libgfortran/config/fpu*.c
!
! Please only add ... run { target { ! { triplets } } } if it is unfixable
! on your target - and a note why (strtod doesn't handle it, no rounding
! support, etc.)
Thus, one way forward would be to exclude the target (Solaris, FreeBSD) - or to
simply add XFAIL as it also only works with very recent GLIBCs.
Alternatively, one could also think about a different technique for handling
the rounding (on input) in libgfortran.