[C++ Patch/RFC] PR 82593 ("Internal compiler error: in process_init_constructor_array, at cp/typeck2.c:1294")

[Paolo Carlini]

Hi,

this ICE on valid (given GNU's designated initializers) is rather simple 
to analyze: for the testcase, the gcc_assert in 
process_init_constructor_array triggers because at that time INDEX1 is 
still a CONST_DECL, not an INTEGER_CST. As regards fixing the problem, I 
immediately noticed earlier today that fold_non_dependent_expr is 
definitely able to fold the CONST_DECL to the expected zero INTEGER_CST 
- and that also passes testing - but I'm not sure that in the big 
picture folding at that time is correct. Thanks in advance for any feedback!

Paolo.

//////////////////

Index: cp/typeck2.c
===================================================================
--- cp/typeck2.c        (revision 254365)
+++ cp/typeck2.c        (working copy)
@@ -1291,6 +1291,7 @@ process_init_constructor_array (tree type, tree in
     {
       if (ce->index)
        {
+         ce->index = fold_non_dependent_expr (ce->index);
          gcc_assert (TREE_CODE (ce->index) == INTEGER_CST);
          if (compare_tree_int (ce->index, i) != 0)
            {
Index: testsuite/g++.dg/cpp0x/desig2.C
===================================================================
--- testsuite/g++.dg/cpp0x/desig2.C     (nonexistent)
+++ testsuite/g++.dg/cpp0x/desig2.C     (working copy)
@@ -0,0 +1,23 @@
+// PR c++/82593
+// { dg-do compile { target c++11 } }
+// { dg-options "" }
+
+enum {
+ INDEX1 = 0,
+ INDEX2
+};
+
+class SomeClass {
+public:
+ SomeClass();
+private:
+ struct { int field; } member[2];
+};
+
+SomeClass::SomeClass()
+ : member({
+   [INDEX1] = { .field = 0 },
+   [INDEX2] = { .field = 1 }
+ })
+{
+}