(Re)allocation of allocatable arrays on assignment - F2003

[Paul Richard Thomas]

Dominique,

Thanks for that bit of detective work.

On Tue, Nov 2, 2010 at 11:22 PM, Dominique Dhumieres <dominiq@lps.ens.fr> wrote:
> Paul,
>
> If it may help, the bounds are correct (-1:2) if the kinds of ia and ib
> are the same, but wrong (1:4) if the kinds are different (I have checked
> 2 vs. 4, and 8 vs. 4).
>
> If it rings some bell for you, these codelets were leftover of experiments
> with bounds for derived types (there is probably at least one pr closed
> for it).

I had already discovered that conversion functions were getting
through to the bit of code that does the bounds shifting.  Using the
argument, instead of the function call, was trivial to implement.

Cheers

Paul