[Bug fortran/64986] class_to_type_4.f90: valgrind error: Invalid read/write of size 8

["mikael at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64986

--- Comment #12 from Mikael Morin <mikael at gcc dot gnu.org> ---
(In reply to Dominique d'Humieres from comment #6)
> The test has been introduced at revision r220482,

That revision adds interesting comments:

>  /* For a function with a class array result, save the result as
>     a temporary, set the info fields needed by the scalarizer and
>     call the finalization function of the temporary. Note that the
>     nullification of allocatable components needed by the result
>     is done in gfc_trans_assignment_1.  */

and in gfc_trans_assignment_1, there is:

> /* Nullify the allocatable components corresponding to those of the lhs
>    derived type, so that the finalization of the function result does not
>    affect the lhs of the assignment. Prepend is used to ensure that the
>    nullification occurs before the call to the finalizer. 


So, if finalization for derived types with allocatable components means freeing
the allocatable components, the above is more or less a justification for the
patch in comment #9.

What I don't understand is why there is need for two functions
gfc_conv_procedure_call and gfc_trans_assignment_1 doing half of the job, and
why deallocation of components, deallocation of whole allocatable and
finalization are not handled all at once in a single place.