[J3] Self-assignment of allocatable component

[Vipul Parekh]

On Fri, Jul 30, 2021 at 4:52 AM Malcolm Cohen via J3 <
j3 at mailman.j3-fortran.org> wrote:

> Hi Vipul,
>
>
>
> Thanks for your example program. It certainly looks valid to me.
>
>
>
> I am embarrassed to have to admit that the NAG compiler gets this wrong
> when the assignment avoids the round-trip to the memory allocator (we do
> this when the allocated types/sizes/etc are the same). In that case we
> forgot to invoke the final subroutine.
>
>
>
> I have now fixed this, so that we produce the output you expect. The fix
> will be in build 7055.
>


Thanks very much Malcolm for your confirmation on the example program.

But especially for working toward implementing a fix so quickly in the NAG
compiler.  As a customer now of NAG compiler, I look forward to using build
7055.

The example program was indeed based on issues I had in mind with our
industrial Fortran code with intrinsic assignments of objects that are of
derived types.  Missing or inadequate finalizations in various processor
implementations had remained in my mind as being part of the problem.  The
updated NAG compiler will then help us advance further in our analysis of
pending issues.

Thanks,
Vipul Parekh
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.j3-fortran.org/pipermail/j3/attachments/20210731/553cc613/attachment.htm>