Extend aligned_membuf<> usage

Extend aligned_membuf<> usage

[François Dumont]

[-- Attachment #1: Type: text/plain, Size: 143 bytes --]

Hi

     As we just bump version namespace it might be interesting to do the 
following change now. What do you think ?

François


[-- Attachment #2: aligned_membuf.patch --]
[-- Type: text/x-patch, Size: 1881 bytes --]

diff --git a/libstdc++-v3/include/bits/forward_list.h b/libstdc++-v3/include/bits/forward_list.h
index 56b3ac5..05abd43 100644
--- a/libstdc++-v3/include/bits/forward_list.h
+++ b/libstdc++-v3/include/bits/forward_list.h
@@ -116,7 +116,11 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
       {
 	_Fwd_list_node() = default;
 
+#if _GLIBCXX_INLINE_VERSION
+	__gnu_cxx::__aligned_membuf<_Tp> _M_storage;
+#else
 	__gnu_cxx::__aligned_buffer<_Tp> _M_storage;
+#endif
 
 	_Tp*
 	_M_valptr() noexcept
diff --git a/libstdc++-v3/include/bits/hashtable_policy.h b/libstdc++-v3/include/bits/hashtable_policy.h
index 3ff6b14..222a1b2 100644
--- a/libstdc++-v3/include/bits/hashtable_policy.h
+++ b/libstdc++-v3/include/bits/hashtable_policy.h
@@ -231,7 +231,11 @@ namespace __detail
     {
       typedef _Value value_type;
 
+#if _GLIBCXX_INLINE_VERSION
+      __gnu_cxx::__aligned_membuf<_Value> _M_storage;
+#else
       __gnu_cxx::__aligned_buffer<_Value> _M_storage;
+#endif
 
       _Value*
       _M_valptr() noexcept
@@ -1516,7 +1520,11 @@ namespace __detail
   template<typename _Tp, bool _IsEmpty = std::is_empty<_Tp>::value>
     struct _Hash_code_storage
     {
+#if _GLIBCXX_INLINE_VERSION
+      __gnu_cxx::__aligned_membuf<_Tp> _M_storage;
+#else
       __gnu_cxx::__aligned_buffer<_Tp> _M_storage;
+#endif
 
       _Tp*
       _M_h() { return _M_storage._M_ptr(); }
diff --git a/libstdc++-v3/include/bits/shared_ptr_base.h b/libstdc++-v3/include/bits/shared_ptr_base.h
index b58273a..8e00f1b 100644
--- a/libstdc++-v3/include/bits/shared_ptr_base.h
+++ b/libstdc++-v3/include/bits/shared_ptr_base.h
@@ -534,7 +534,11 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
 
 	_Alloc& _M_alloc() noexcept { return _A_base::_S_get(*this); }
 
+#if _GLIBCXX_INLINE_VERSION
+	__gnu_cxx::__aligned_membuf<_Tp> _M_storage;
+#else
 	__gnu_cxx::__aligned_buffer<_Tp> _M_storage;
+#endif
       };
 
     public:

Re: Extend aligned_membuf<> usage

[Jonathan Wakely]

On 01/02/18 22:48 +0100, François Dumont wrote:
>Hi
>
>    As we just bump version namespace it might be interesting to do 
>the following change now. What do you think ?

I'd rather not make the code harder to read in all those places just
for an optional mode that nobody uses.

Wouldn't it make more sense to simply make __aligned_buffer identical
to __aligned_membuf for the versioned-namespace? Then at least the
conditional code is only in one place.

Re: Extend aligned_membuf<> usage

[François Dumont]

On 05/02/2018 18:16, Jonathan Wakely wrote:
> On 01/02/18 22:48 +0100, François Dumont wrote:
>> Hi
>>
>>     As we just bump version namespace it might be interesting to do 
>> the following change now. What do you think ?
>
> I'd rather not make the code harder to read in all those places just
> for an optional mode that nobody uses.
>

I just consider this mode as a nice way to track all abi changes that we 
will be able to commit one day when we decide an abi breaking change.

> Wouldn't it make more sense to simply make __aligned_buffer identical
> to __aligned_membuf for the versioned-namespace? Then at least the
> conditional code is only in one place.
>
Yes, __aligned_buffer is indeed not used, we could do that. I'll propose 
a patch to do so.

Re: Extend aligned_membuf<> usage

[François Dumont]

[-- Attachment #1: Type: text/plain, Size: 423 bytes --]

On 06/02/2018 20:16, François Dumont wrote:
> On 05/02/2018 18:16, Jonathan Wakely wrote:
>> Wouldn't it make more sense to simply make __aligned_buffer identical
>> to __aligned_membuf for the versioned-namespace? Then at least the
>> conditional code is only in one place.
>>
> Yes, __aligned_buffer is indeed not used, we could do that. I'll 
> propose a patch to do so.
>
So here it is, ok to commit ?

François


[-- Attachment #2: aligned_membuf.patch --]
[-- Type: text/x-patch, Size: 844 bytes --]

diff --git a/libstdc++-v3/include/ext/aligned_buffer.h b/libstdc++-v3/include/ext/aligned_buffer.h
index 94a2ff0..81fb797 100644
--- a/libstdc++-v3/include/ext/aligned_buffer.h
+++ b/libstdc++-v3/include/ext/aligned_buffer.h
@@ -75,6 +75,10 @@ namespace __gnu_cxx
       { return static_cast<const _Tp*>(_M_addr()); }
     };
 
+#if _GLIBCXX_INLINE_VERSION
+  template<typename _Tp>
+    using __aligned_buffer = __aligned_membuf<_Tp>;
+#else
   // Similar to __aligned_membuf but aligned for complete objects, not members.
   // This type is used in <forward_list>, <future>, <bits/shared_ptr_base.h>
   // and <bits/hashtable_policy.h>, but ideally they would use __aligned_membuf
@@ -113,6 +117,7 @@ namespace __gnu_cxx
       _M_ptr() const noexcept
       { return static_cast<const _Tp*>(_M_addr()); }
     };
+#endif
 
 } // namespace
 

Re: Extend aligned_membuf<> usage

[Jonathan Wakely]

On 8 February 2018 at 06:10, François Dumont wrote:
> On 06/02/2018 20:16, François Dumont wrote:
>>
>> On 05/02/2018 18:16, Jonathan Wakely wrote:
>>>
>>> Wouldn't it make more sense to simply make __aligned_buffer identical
>>> to __aligned_membuf for the versioned-namespace? Then at least the
>>> conditional code is only in one place.
>>>
>> Yes, __aligned_buffer is indeed not used, we could do that. I'll propose a
>> patch to do so.
>>
> So here it is, ok to commit ?

OK, because it only affects the gnu-versioned-namespace mode.

Re: Extend aligned_membuf<> usage

[François Dumont]

On 20/02/2018 20:59, Jonathan Wakely wrote:
> On 8 February 2018 at 06:10, François Dumont wrote:
>> On 06/02/2018 20:16, François Dumont wrote:
>>> On 05/02/2018 18:16, Jonathan Wakely wrote:
>>>> Wouldn't it make more sense to simply make __aligned_buffer identical
>>>> to __aligned_membuf for the versioned-namespace? Then at least the
>>>> conditional code is only in one place.
>>>>
>>> Yes, __aligned_buffer is indeed not used, we could do that. I'll propose a
>>> patch to do so.
>>>
>> So here it is, ok to commit ?
> OK, because it only affects the gnu-versioned-namespace mode.
>
Ok, done now.