[gcc r11-6530] c++: Add TARGET_EXPR comments

[Jason Merrill]

https://gcc.gnu.org/g:6c59b8a93cf4784e3e3137416a3d32a1ecc8e00b

commit r11-6530-g6c59b8a93cf4784e3e3137416a3d32a1ecc8e00b
Author: Jason Merrill <jason@redhat.com>
Date:   Mon Jan 4 16:11:08 2021 -0500

    c++: Add TARGET_EXPR comments
    
    Discussing the 98469 patch and class prvalues with Jakub led me to
    double-check our handling of TARGET_EXPR in constexpr.c, and add a note
    about why we don't strip them in parameter initialization.  And another to
    clarify that we're handling an INIT_EXPR in a place we do strip them.
    
    gcc/cp/ChangeLog:
    
            * constexpr.c (cxx_bind_parameters_in_call): Add comment.
            (cxx_eval_store_expression): Add comment.

Diff:
---
 gcc/cp/constexpr.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/gcc/cp/constexpr.c b/gcc/cp/constexpr.c
index 0c12f608d36..4a5e6384da6 100644
--- a/gcc/cp/constexpr.c
+++ b/gcc/cp/constexpr.c
@@ -1591,6 +1591,9 @@ cxx_bind_parameters_in_call (const constexpr_ctx *ctx, tree t,
       if (TREE_ADDRESSABLE (type))
 	/* Undo convert_for_arg_passing work here.  */
 	x = convert_from_reference (x);
+      /* Normally we would strip a TARGET_EXPR in an initialization context
+	 such as this, but here we do the elision differently: we keep the
+	 TARGET_EXPR, and use its CONSTRUCTOR as the value of the parm.  */
       arg = cxx_eval_constant_expression (ctx, x, /*lval=*/false,
 					  non_constant_p, overflow_p);
       /* Don't VERIFY_CONSTANT here.  */
@@ -5388,6 +5391,9 @@ cxx_eval_store_expression (const constexpr_ctx *ctx, tree t,
 
   if (!preeval)
     {
+      /* We're handling an INIT_EXPR of class type, so the value of the
+	 initializer can depend on the object it's initializing.  */
+
       /* Create a new CONSTRUCTOR in case evaluation of the initializer
 	 wants to modify it.  */
       if (*valp == NULL_TREE)