intrusive_ptr wrapper for C++/CLI
I do a lot of development using native C++ and managed C++/CLI, and have coded a simple smart pointer wrapper around intrusive_ptr to allow C++/CLI code to use reference counting objects without change. IntrusivePtr takes advantage of the template support engineered for gcroot<T> to implement the smart pointer. It’s a wrapper to support factory functions that return intrusive_ptr. Whilst many would object to putting C++/CLI conditional code into boost, the smart pointer is too small to stand alone and too boost specific for Microsoft. thoughts welcome https://ql.codeplex.com/SourceControl/latest#IntrusivePtr/boost/IntrusivePtr...
On 20 Aug 2014 at 14:10, Stephen Channell wrote:
I do a lot of development using native C++ and managed C++/CLI, and have coded a simple smart pointer wrapper around intrusive_ptr to allow C++/CLI code to use reference counting objects without change. IntrusivePtr takes advantage of the template support engineered for gcroot<T> to implement the smart pointer. It´s a wrapper to support factory functions that return intrusive_ptr.
Whilst many would object to putting C++/CLI conditional code into boost, the smart pointer is too small to stand alone and too boost specific for Microsoft.
You may not be aware that Microsoft contributed WinRT support for Boost, and that should be fully merged into Boost 1.57. I'd suggest you hand over your idea to one of the Microsoft engineers on this list and let them push it for you as part of their general support of Boost. Search the archives for patches from Microsoft, you probably don't want Stephan but you will want Steve (stgates). Niall -- ned Productions Limited Consulting http://www.nedproductions.biz/ http://ie.linkedin.com/in/nialldouglas/
WinRT C++/CX can use intrusive_ptr directly because it is native, but C++/CLI can only hold pointers to native objects because of the CLI. intrusive_ptr is a direct alternative to the CComPtr template and COM interop, and only better when you want/need portability.. wouldn't expect MS to be interested.. From: s_sourceforge@nedprod.com To: boost@lists.boost.org Date: Thu, 21 Aug 2014 13:09:19 +0100 Subject: Re: [boost] intrusive_ptr wrapper for C++/CLI On 20 Aug 2014 at 14:10, Stephen Channell wrote:
I do a lot of development using native C++ and managed C++/CLI, and have coded a simple smart pointer wrapper around intrusive_ptr to allow C++/CLI code to use reference counting objects without change. IntrusivePtr takes advantage of the template support engineered for gcroot<T> to implement the smart pointer. It´s a wrapper to support factory functions that return intrusive_ptr.
Whilst many would object to putting C++/CLI conditional code into boost, the smart pointer is too small to stand alone and too boost specific for Microsoft.
You may not be aware that Microsoft contributed WinRT support for Boost, and that should be fully merged into Boost 1.57. I'd suggest you hand over your idea to one of the Microsoft engineers on this list and let them push it for you as part of their general support of Boost. Search the archives for patches from Microsoft, you probably don't want Stephan but you will want Steve (stgates). Niall -- ned Productions Limited Consulting http://www.nedproductions.biz/ http://ie.linkedin.com/in/nialldouglas/ _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
On Wed, Aug 20, 2014 at 7:10 AM, Stephen Channell
Whilst many would object to putting C++/CLI conditional code into boost, the smart pointer is too small to stand alone and too boost specific for Microsoft.
I feel it is too C++/CLI specific for Boost. Is C++/CLI as a language evolving at all? The impression I had before I left Microsoft was that it would never receive any major new language features and was generally suggested only as a last resort for use in conjunction with a language like C# to overcome performance or other issues around interop. Glen
participants (3)
-
Glen Fernandes -
Niall Douglas -
Stephen Channell