(j3.2006) interp fodder from comp.lang.fortran

[Robert Corbett]

On 09/03/15 21:06, Cohen Malcolm wrote:
> Well that does not conform to any published version of the standard.
>
> It directly violates Fortran 90 and 95.

I agree, but I would not be surprised if other implementations work the same way.
>
>
> One can forgive not doing the runtime test apparently required by F2003, but
> failing to conform to F90/F95 (and this is not listed as an incompatibility
> in clause 1) does not seem like a good idea to me.

It does not seem like a good idea to me either.  Nonetheless, Ian Harvey 
suggested something along those lines in one of his postings in the thread, so 
my guess is that he would like it.

I am curious how other implementations handle this issue.  I doubt any 
implementation uses the run-time check, but I would like to know how 
implementations choose between defined assignments and intrinsic assignments in 
the cases where the standard requires the choice to be made dynamically.

Robert Corbett