The TR1 lib is having problems with Boost.Hash and VC7: note that simply including the header is enough to trigger the problem, see for example: http://engineering.meta-comm.com/boost-regression/CVS-RC_1_34_0/developer/ou... Is there any way to fix this? Thanks, John.
John Maddock wrote:
The TR1 lib is having problems with Boost.Hash and VC7: note that simply including the header is enough to trigger the problem, see for example:
http://engineering.meta-comm.com/boost-regression/CVS-RC_1_34_0/developer/ou...
Is there any way to fix this?
To answer my own question, the fix is below, OK to commit? John. RCS file: /cvsroot/boost/boost/boost/functional/hash/hash.hpp,v retrieving revision 1.28 diff -u -r1.28 hash.hpp --- functional/hash/hash.hpp 7 May 2006 17:52:38 -0000 1.28 +++ functional/hash/hash.hpp 4 Jun 2006 17:41:09 -0000 @@ -388,7 +388,7 @@ } template <class T> struct hash - : public hash_detail::hash_impl<boost::is_pointer<T>::value> + : public boost::hash_detail::hash_impl<boost::is_pointer<T>::value> ::BOOST_NESTED_TEMPLATE inner<T> { }; @@ -487,6 +487,10 @@ namespace hash_detail { + + template <bool b> + struct hash_impl; + template <> struct hash_impl<false> {
On 04/06/06, John Maddock <john@johnmaddock.co.uk> wrote:
To answer my own question, the fix is below, OK to commit?
Yes, of course. I think a similar workaround might be needed for VC6 (which has its own version of hash_impl<false>, workarounds on top of workarounds) but don't have access to a compiler right now to check. I'll look into it soon. thanks, Daniel
On 04/06/06, John Maddock <john@johnmaddock.co.uk> wrote:
To answer my own question, the fix is below, OK to commit?
I've just comitted a very similar change to cvs head, which is a little different, but essentially the same. Please can you give it a try? Feel free to make a fix, if required. thanks, Daniel
Daniel James wrote:
I've just comitted a very similar change to cvs head, which is a little different, but essentially the same. Please can you give it a try? Feel free to make a fix, if required.
Thanks Daniel, that seems to work just fine, can you push it to the release branch as well? Many thanks, John.
participants (2)
-
Daniel James -
John Maddock