[Vala] Incorrect C code generated

[Dmitry Golovin <dima golovin in>]

Hello!

Discussed today on the IRC channel, also posting it to the mailing list.

I don't know if it is a vala bug or I am just using things incorrectly.

Here is my code:

    struct N {
      int i;
    }
    class ClassA : Object {
      public N a { get; set; }
      public ClassA(N a) {
        this.a = a;
      }
    }
    class ClassB : Object {
      public Value b { get; set; }
      public ClassB(Value b) {
        this.b = b;
      }
    }
    class Main {
      static void main() {
        var a = new ClassA({ 1 });
        var b = new ClassB(a.a);
        b.notify["b"].connect((obj, prop) => { a.a = (N)b.b; });
      }
    }

Here is what C compiler says:

Clang: test.vala.c:413:29: error: cannot take the address of an rvalue of type 'N' (aka 'struct _N')
  ...&((G_VALUE_HOLDS (&_tmp1_, TYPE_N) && g_value_get_boxed (&_tmp1_)) ? (*((N*) g_value_get_boxed 
(&_tmp1_))) : (g_warning ("Invalid GValue unboxing (wrong type or NULL)"), _tmp2_)));

GCC: test.vala.c:413:29: error: lvalue required as unary ‘&’ operand
  class_a_set_a (_data1_->a, &((G_VALUE_HOLDS (&_tmp1_, TYPE_N) && g_value_get_boxed (&_tmp1_)) ? (*((N*) 
g_value_get_boxed (&_tmp1_))) : (g_warning ("Invalid GValue unboxing (wrong type or NULL)"), _tmp2_)));

This shorter code has similar problem (but it makes much less sense as this cast is impossible):

    struct N {
      int i;
    }
    class Main {
      static void main() {
        N n = { 42 };
        Value v = n;
        int i = (int)((N)v);
        print("%d\n", i);
      }
    }

In this code cast from Value to my struct works fine:

    struct N {
      int i;
    }
    class Main {
      static void main() {
        N n = { 42 };
        Value v = n;
        n = (N)v;
        print("%d %d\n", n.i, ((N)v).i);
      }
    }

Could you please make any suggestions on how do I fix the code from the very first examples, if it is 
terribly wrong?

While writing this message, also found this bug: https://bugzilla.gnome.org/show_bug.cgi?id=764041 

Regards,
Dmitry