(Ada/MI) Fix -var-evaluate-expression for access to unconstrained arrays

Using this Ada code:

   type String_Access is access String;
   type Array_Of_String is array (1 .. 2) of String_Access;
   Aos : Array_Of_String := (new String'("ab"), new String'("cd"));

In GDB/MI mode, create a variable which type is Aos, evaluate it:

(gdb) -var-create var1 * Aos
^done,name="var1",numchild="2",value="[2]",type="bar.array_of_string",thread-id="1",has_more="0"

Now print it:

(gdb) -var-list-children 1 var1
^done,numchild="2",children=[child={name="var1.1",exp="1",numchild="1",value="[2] \"ab\"", type="bar.string_access",thread-id="1"},child={name="var1.2",exp="2",numchild="1",value="[2] \"cd\"", type="bar.string_access",thread-id="1"}],has_more="0"

But printed fields "value" are wrong, since it should be:

^done,numchild="2",children=[child={name="var1.1",exp="1",numchild="1",value="0x634018",type="bar.string_access",thread-id="1"},child={name="var1.2",exp="2",numchild="1",value="0x634038",type="bar.string_access",thread-id="1"}],has_more="0"^M

Print each child of var1:

(gdb) -var-evaluate-expression var1.1
^done,value="[2] \"ab\""
(gdb) -var-evaluate-expression var1.2
^done,value="[2] \"cd\""

Whereas it should be

(gdb) -var-evaluate-expression var1.1
^done,value="0x635018"
(gdb) -var-evaluate-expression var1.2
^done,value="0x635038"

This patch fixes this.

gdb/ChangeLog:

        * ada-lang.c (ada_value_subscript): Handle case when parameter is
        an array of access to unconstrained array.

testsuite/ChangeLog

        * gdb.ada/mi_string_access.exp: New testcase.
        * gdb.ada/mi_string_access/bar.adb: New file.
        * gdb.ada/mi_string_access/pck.adb: New file.
        * gdb.ada/mi_string_access/pck.asd: New file.

Tested on x86_64-linux.

1 files changed, 24 insertions, 0 deletions

diff --git a/gdb/ada-lang.c b/gdb/ada-lang.c
index 83421ac..5f39def 100644
--- a/gdb/ada-lang.c
+++ b/gdb/ada-lang.c
@@ -2861,10 +2861,34 @@ ada_value_subscript (struct value *arr, int arity, struct value **ind)
 
   for (k = 0; k < arity; k += 1)
     {
+      struct type *saved_elt_type = TYPE_TARGET_TYPE (elt_type);
+
       if (TYPE_CODE (elt_type) != TYPE_CODE_ARRAY)
         error (_("too many subscripts (%d expected)"), k);
+
       elt = value_subscript (elt, pos_atr (ind[k]));
+
+      if (ada_is_access_to_unconstrained_array (saved_elt_type)
+	  && TYPE_CODE (value_type (elt)) != TYPE_CODE_TYPEDEF)
+	{
+	  /* The element is a typedef to an unconstrained array,
+	     except that the value_subscript call stripped the
+	     typedef layer.  The typedef layer is GNAT's way to
+	     specify that the element is, at the source level, an
+	     access to the unconstrained array, rather than the
+	     unconstrained array.  So, we need to restore that
+	     typedef layer, which we can do by forcing the element's
+	     type back to its original type. Otherwise, the returned
+	     value is going to be printed as the array, rather
+	     than as an access.  Another symptom of the same issue
+	     would be that an expression trying to dereference the
+	     element would also be improperly rejected.  */
+	  deprecated_set_value_type (elt, saved_elt_type);
+	}
+
+      elt_type = ada_check_typedef (value_type (elt));
     }
+
   return elt;
 }