some problems with gcc/testsuite/gfortran.dg/ tests in 4.3-20061111 with OSX 10.3
Dominique Dhumieres
dominiq@lps.ens.fr
Sun Nov 19 23:18:00 GMT 2006
> AFAIK, recursive write is nonconforming.
If you say so, it is nevertheless in gcc/testsuite/gfortran.dg/intrinsic_actual_2.f90
> In principle, gfortran can do anything.
Hanging is probably the worse possibility.
> pr29879 is about a missed optimization.
It is an ICE in libfortran. I don't believe it affects IO,
but I don't know for sure.
> OTOH, I note you are using an Apple OS, which seems to be
> broken more often than functional.
So far I did not find Mac OSX 10.3 (32 bits) less functional
than any linux box I have access to: same bugs, same results.
The only drawback of OSX I have noticed so far is the quite
large penalty with memory allocation at the system level,
but this more a penalty for g95 than gfortran.
Dominique
More information about the Fortran
mailing list