[Bug fortran/110626] [13/14/15/16/17 regression] Duplicated finalization in derived

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110626

--- Comment #25 from Christopher Albert <albert at tugraz dot at> ---
Open for review, comments welcome:
https://github.com/lazy-fortran/gcc/pull/38

This carries attachment 64841 forward with one restriction found in
further testing: finalize_only now applies to scalar lhs assignments
without pointer or allocatable components only, since rank-one lhs
testing exposed incorrect array behavior on the in-place path. Array
lhs stays on the existing scalarized path and the
allocatable-component case remains PR57696. The reproducer fails
before and passes after; check-fortran and the libgomp Fortran suites
are clean.