[Ada] Crash on aggregate in function call in object declaration

Message ID 20210503092945.GA77479@adacore.com
State New
Headers show
Series
  • [Ada] Crash on aggregate in function call in object declaration
Related show

Commit Message

Pierre-Marie de Rodat May 3, 2021, 9:29 a.m.
Compiler aborts on an aggregate with limited components, when the
aggregate is a formal in a function call that is the right-hand side of
an assignment, possibly coming from an object declaration.

Tested on x86_64-pc-linux-gnu, committed on trunk

gcc/ada/

	* exp_aggr.adb (Expand_Array_Aggregate): If the parent node of
	the aggregate is a subprogram call there is no target in which
	to build the aggregate, and it has to be expanded into component
	assignments.

Patch

diff --git a/gcc/ada/exp_aggr.adb b/gcc/ada/exp_aggr.adb
--- a/gcc/ada/exp_aggr.adb
+++ b/gcc/ada/exp_aggr.adb
@@ -5700,7 +5700,7 @@  package body Exp_Aggr is
       function Safe_Left_Hand_Side (N : Node_Id) return Boolean;
       --  In addition to Maybe_In_Place_OK, in order for an aggregate to be
       --  built directly into the target of the assignment it must be free
-      --  of side effects.
+      --  of side effects. N is the LHS of an assignment.
 
       ----------------------------
       -- Build_Constrained_Type --
@@ -6661,9 +6661,13 @@  package body Exp_Aggr is
          Set_Expansion_Delayed (N);
          return;
 
-      --  In the remaining cases the aggregate is the RHS of an assignment
+      --  In the remaining cases the aggregate appears in the RHS of an
+      --  assignment, which may be part of the expansion of an object
+      --  delaration. If the aggregate is an actual in a call, itself
+      --  possibly in a RHS, building it in the target is not possible.
 
       elsif Maybe_In_Place_OK
+        and then Nkind (Parent_Node) not in N_Subprogram_Call
         and then Safe_Left_Hand_Side (Name (Parent_Node))
       then
          Tmp := Name (Parent_Node);