[C++ PATCH] Implementation of C++0x decltype

[Andrew Pinski]

> The attached patches (for gcc, cp, testsuite subdirs) implement the C+ 
> +0x "decltype" facility, which computes the declared type of an  
> arbitrary expression. The latest standard wording, on which this  
> patch is based, is available here:
> 
> 	http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2006/n2115.pdf
> 
> Decltype has not yet been voted into the C++0x working draft, but it  
> has been approved by the C++ Evolution Working Group and has been  
> extensively discussed in the Core Working Group, so it is relatively  
> stable (only corner cases are likely to change). It's a highly- 
> requested feature in the library writer community. This patch only  
> makes decltype available when the experimental C++0x mode is enabled.
> 
> Tested i386-apple-darwin8.8.1, no regressions.

Can you explain the difference between typeof and decltype?  Also it
seems like all of the C++0x extensions that are implemented should
have a mention in the documentation in a new section called something
like C++0x working draft or something like that.  Or you could something
like Gfortran's manual does:
http://gcc.gnu.org/onlinedocs/gcc-4.1.1/gfortran/Standards.html

Which seems like a good idea also.

Thanks,
Andrew Pinski