Backporting GFORTRAN fixes to GCC 4.2

[Brooks Moses]

Mark Mitchell wrote:
> Tobias Burnus wrote:
>> I don't know what your current planing is, but is it ok, to still check
>> in the following patches? Until which date should we check them in?
> 
> If they're regressions, it's find to check these in before Sunday.  If
> they're just bugs, then I think you need to lobby hard.  It's time for
> gfortran to play by the usual rules for release branches: regression
> fixes only.

For the record, none of the referenced patches are regressions against 
G77; they all depend on F90/F95 features that G77 had not implemented.

I don't have old GFortran versions to compare against, but I see no 
reason to suspect that any of them are regressions within GFortran either.

- Brooks