[Patch, fortran] PR29284 - [4.1/4.2 Regression] ICE for optional subroutine argument

[Paul Thomas]

François-Xavier,

>
> OK to commit. I'm not a huge supporter of moving function calls
> around, but I tried to figure it out and I think in this case, the
> order is OK.

As I said the other day, I treated this one with enormous care; I both 
tested thoroughly and checked out dependences.

>
> [Frankly, for the readability & size difference in the final f951
> executable vs maintainance potential trouble ratio, I think it's not a
> win.]
>
I agree - I wasn't greatly convinced of the need to do it this way in 
the first place!

Thanks

Paul