Re: [v3] Implement pointer_traits and allocator_traits

2011-05-28 Thread Jonathan Wakely
On 28 May 2011 20:08, Jonathan Wakely wrote: > On 28 May 2011 18:14, Jonathan Wakely wrote: >> One more test >> >> 2011-05-28  Jonathan Wakely   >> >>        * testsuite/20_util/pointer_traits/pointer_to.cc: New. >> >> Tested x86_64-linux, committed to trunk. >> > > This new test is failing, I'm lo

Re: [v3] Implement pointer_traits and allocator_traits

2011-05-28 Thread Jonathan Wakely
On 28 May 2011 18:14, Jonathan Wakely wrote: > One more test > > 2011-05-28  Jonathan Wakely   > >        * testsuite/20_util/pointer_traits/pointer_to.cc: New. > > Tested x86_64-linux, committed to trunk. > This new test is failing, I'm looking into why, will revert it if necessary ...

Re: [v3] Implement pointer_traits and allocator_traits

2011-05-28 Thread Jonathan Wakely
One more test 2011-05-28 Jonathan Wakely * testsuite/20_util/pointer_traits/pointer_to.cc: New. Tested x86_64-linux, committed to trunk. Index: testsuite/20_util/pointer_traits/pointer_to.cc === --- testsuite/20_util/poin

Re: [v3] Implement pointer_traits and allocator_traits

2011-05-28 Thread Jonathan Wakely
On 28 May 2011 15:53, Jonathan Wakely wrote: > > This allows you to write types which can only be managed ... by shared_ptr and can't be created/copied/destroyed any other way.

Re: [v3] Implement pointer_traits and allocator_traits

2011-05-28 Thread Jonathan Wakely
Here's a patch to make shared_ptr use allocator_traits, I don't plan to commit this (at least not yet) but it shows what's needed to use allocator_traits instead of the C++03 way of doing things. It uses allocator_traits::__rebind::__type which is a temporary substitute for the unsupported allocat

Re: [v3] Implement pointer_traits and allocator_traits

2011-05-28 Thread Jonathan Wakely
I forgot to mention this is missing the rebind templates, which depend on alias template support in G++