[Bug fortran/104382] Finalization of parent components not compliant with standard

[pault at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104382

--- Comment #6 from Paul Thomas <pault at gcc dot gnu.org> ---
(In reply to Thomas Koenig from comment #4)
> To add some variety to the tests, nagfor gives
> 
>  destructor4(complicated)   2.0000000   2.0000000
>  destructor5 (simple2) 5
>  destructor5 (simple2) 6
>  destructor2(simple) 1 1
>  final_count after assignment =  4
>  destructor4(complicated)   4.0000000   5.0000000
>  destructor5 (simple2) -1
>  destructor5 (simple2) -2
>  destructor2(simple) 3 4
>  final_count after deallocation =  8

The current version of the finalizer patch gives the same result as nagfor.

Cheers

Paul