[gfortran, committed] Fix segfault in PR16336

[Tobias Schlüter]

Erik Schnetter wrote:
> On Saturday 10 July 2004 18:30, Tobias Schlüter wrote:
> 
>>I'm inclined to believe that the code in PR 16336 is indeed invalid,
>>but I will double check the standard to make sure that this is true.
> 
> 
> The stripped-down example is indeed invalid.  The original source code 
> declared the common block in the module as private, so that there 
> should be no name clash.

Hmm, re-reading the standard. I'm less sure that my argument is correct.
I'll look further, unless you can prove me from te standard, that the
code is indeed valid :-)

- Tobi