Re: [PATCH] 77864 Fix noexcept conditions for map/set default constructors

[Jonathan Wakely <jwakely@redhat.com>]

On 06/10/16 22:17 +0200, François Dumont wrote:
>Another approach is to rely on existing compiler ability to compute 
>conditional noexcept when defaulting implementations. This is what I 
>have done in this patch.
>
>The new default constructor on _Rb_tree_node_base is not a problem as 
>it is not used to build _Rb_tree_node.

Why not?

>I'll try to do the same for copy constructor/assignment and move 
>constructor/assignment.

We need to make sure we don't change whether any of those operations
are trivial (which shouldn't be a problem for copy/move, because they
are definitely very non-trivial and will stay that way!)

Does this change the default constructors from non-trivial to trivial?


>--- a/libstdc++-v3/include/bits/stl_tree.h
>+++ b/libstdc++-v3/include/bits/stl_tree.h
>@@ -108,6 +108,10 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
>     _Base_ptr		_M_left;
>     _Base_ptr		_M_right;
> 
>+    _Rb_tree_node_base() _GLIBCXX_NOEXCEPT
>+      : _M_color(_S_red), _M_parent(0), _M_left(this), _M_right(this)
>+    { }
>+
>     static _Base_ptr
>     _S_minimum(_Base_ptr __x) _GLIBCXX_NOEXCEPT
>     {

Another option would be:

    struct _Head_node : _Rb_tree_node_base {
      _Head_node() {
        _M_color = _S_red;
        _M_parent = _Base_ptr();
        _M_left = _M_right = this;
      }
    };

>@@ -603,23 +607,33 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
>         {
> 	  _Key_compare		_M_key_compare;
> 	  _Rb_tree_node_base 	_M_header;

          _Head_node _M_header;

That way *only* this node gets the zero-initialization, not all node
bases.

With either solution we can get rid of _M_header() in every
ctor-initializer-list.


>+#if __cplusplus < 201103L
> 	  size_type 		_M_node_count; // Keeps track of size of tree.
>+#else
>+	  size_type 		_M_node_count = 0; // Keeps track of size of tree.
>+#endif
> 
>+#if __cplusplus < 201103L
> 	  _Rb_tree_impl()
> 	  : _Node_allocator(), _M_key_compare(), _M_header(),
> 	    _M_node_count(0)
>-	  { _M_initialize(); }
>+	  { }
>+#else
>+	  _Rb_tree_impl() = default;
>+#endif
> 
> 	  _Rb_tree_impl(const _Key_compare& __comp, const _Node_allocator& __a)
>-	  : _Node_allocator(__a), _M_key_compare(__comp), _M_header(),
>-	    _M_node_count(0)
>-	  { _M_initialize(); }
>+	  : _Node_allocator(__a), _M_key_compare(__comp), _M_header()
>+#if __cplusplus < 201103L
>+	  , _M_node_count(0)
>+#endif

Doing this conditionally seems pointless, why not just set it here
unconditionally?