Fortran: Fix PR123868

When copying derived types with allocatable array components where the
array element type also has allocatable components, the condition at
line 11071 was incorrectly triggering a call to gfc_duplicate_allocatable.
However, for allocatable arrays with nested allocatables (where
cmp_has_alloc_comps && c->as is true), the add_when_allocated code
already includes a gfc_duplicate_allocatable call (generated by the
recursive structure_alloc_comps call at lines 10290-10293).

This caused the outer array to be allocated twice: first by the explicit
gfc_duplicate_allocatable call at line 11099, and then again by the
gfc_duplicate_allocatable embedded in add_when_allocated.  The first
allocation was leaked when the second allocation overwrote the data
pointer.

PR121628 added "add_when_allocated != NULL_TREE ||" to the condition,
which was redundant for scalars (already handled by !c->as) and wrong
for arrays (caused double allocation).  Simply removing this clause
restores the correct pre-PR121628 behavior.

	PR fortran/123868

gcc/fortran/ChangeLog:

	* trans-array.cc (structure_alloc_comps): For COPY_ALLOC_COMP,
	remove the add_when_allocated != NULL_TREE clause that PR121628
	added.  This clause was redundant for scalars and caused double
	allocation for arrays with nested allocatable components.

gcc/testsuite/ChangeLog:

	* gfortran.dg/array_memcpy_2.f90: Update expected memcpy count
	from 4 to 3, as the double allocation bug is now fixed.
	* gfortran.dg/pr123868.f90: New test.

	Signed-off-by: Christopher Albert <albert@alumni.tugraz.at>

gcc/fortran/trans-array.cc

@@ -11063,9 +11063,14 @@ structure_alloc_comps (gfc_symbol * der_type, tree decl, tree dest,
 					  copy_wrapper);
 	      gfc_add_expr_to_block (&fnblock, call);
 	    }
+	  /* For allocatable arrays with nested allocatable components,
+	     add_when_allocated already includes gfc_duplicate_allocatable
+	     (from the recursive structure_alloc_comps call at line 10290-10293),
+	     so we must not call it again here.  PR121628 added an
+	     add_when_allocated != NULL clause that was redundant for scalars
+	     (already handled by !c->as) and wrong for arrays (double alloc).  */
 	  else if (c->attr.allocatable && !c->attr.proc_pointer
-		   && (add_when_allocated != NULL_TREE
-		       || !cmp_has_alloc_comps
+		   && (!cmp_has_alloc_comps
 		       || !c->as
 		       || c->attr.codimension
 		       || caf_in_coarray (caf_mode)))

gcc/testsuite/gfortran.dg/array_memcpy_2.f90

@@ -1,11 +1,12 @@
 ! This checks that the "z = y" assignment is not considered copyable, as the
 ! array is of a derived type containing allocatable components.  Hence, we
-! we should expand the scalarized loop, which contains *two* memcpy calls
+! should expand the scalarized loop, which contains *two* memcpy calls
 ! for the assignment itself, plus one for initialization.
 ! { dg-do compile }
 ! { dg-options "-O2 -fdump-tree-original" }
 !
 ! PR 121628
+! PR 123868 - fixed double allocation that caused 4 memcpy instead of 3
 !
   type :: a
     integer, allocatable :: i(:)
@@ -26,4 +27,4 @@
 
   z = y
 end
-! { dg-final { scan-tree-dump-times "__builtin_memcpy" 4 "original" } }
+! { dg-final { scan-tree-dump-times "__builtin_memcpy" 3 "original" } }

gcc/testsuite/gfortran.dg/pr123868.f90

@@ -0,0 +1,34 @@
+! { dg-do run }
+! PR fortran/123868 - Memory leak on assignment with nested allocatable
+! components.  Regression introduced by PR121628 commit which caused
+! gfc_duplicate_allocatable to be called twice for allocatable array
+! components with nested allocatable components.
+
+module bugMod
+
+  type :: vs
+     character(len=1), allocatable :: s
+  end type vs
+
+  type :: ih
+     type(vs), allocatable, dimension(:) :: hk
+  end type ih
+
+end module bugMod
+
+program bugProg
+  use bugMod
+
+  block
+    type(ih) :: c, d
+
+    allocate(d%hk(1))
+    allocate(d%hk(1)%s)
+    d%hk(1)%s='z'
+    c=d
+    if (c%hk(1)%s /= 'z') stop 1
+    if (d%hk(1)%s /= 'z') stop 2
+
+  end block
+
+end program bugProg