This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: Possible compilation failure in tr1/hashtable?
- From: Paolo Carlini <pcarlini at suse dot de>
- To: Francesco Biscani <bluescarni at gmail dot com>
- Cc: libstdc++ at gcc dot gnu dot org
- Date: Wed, 09 Jan 2008 12:16:30 +0100
- Subject: Re: Possible compilation failure in tr1/hashtable?
- References: <47843B93.1040005@gmail.com>
Hi again,
after a cup of coffee ;) I went through 14.5.3 and seems pretty obvious
that the current friend declaration doesn't make much sense (e.g., per
/9 in any case partial specializations cannot be declared friend). I
filed C++34724 and I'm going to apply the below, tested x86_64-linux.
Thanks,
Paolo.
//////////////
2008-01-09 Francesco Biscani <bluescarni@gmail.com>
* include/tr1_impl/hashtable (_Hashtable): Fix friend declaration
of __detail::_Map_base<>.
Index: include/tr1_impl/hashtable
===================================================================
--- include/tr1_impl/hashtable (revision 131422)
+++ include/tr1_impl/hashtable (working copy)
@@ -168,7 +168,8 @@
__cache_hash_code>
const_iterator;
- template<typename _Key2, typename _Pair, typename _Hashtable>
+ template<typename _Key2, typename _Value2, typename _Ex2, bool __unique2,
+ typename _Hashtable2>
friend struct __detail::_Map_base;
private: