Is this valid code or bug in gfortran?

[Tobias Burnus]

Hi,

Kamaraju Kusumanchi wrote:
> Consider the code (testfile.f90) pasted in the end. (Sorry! the gmane interface
> does not allow attachements, so I have to paste the code and it is a bit long).
>   
You could have just send an email directly fortran(at)gcc.gnu.org, but
the inline example is also ok.

> Is this code valid or am I hitting a bug in gfortran?
> gcc version 4.2.0
>   
I think you hit a bug in gfortran as can reproduce the problem with
gfortran 4.2.0, but not with gfortran 4.3.0. As it is no regression I'm
not sure whether we will fix it for 4.2.0. (That it is a wrong-code bug
make it a bit more likely though.)

gfortran 4.3 binaries are available from:
http://gcc.gnu.org/wiki/GFortranBinaries

Tobias