Re: [boost] Header Placement

From: John Maddock (john_at_[hidden])
Date: 2007-11-23 13:56:43

John Maddock wrote:
> Ah, I think the errors in the log reports are misleading, when
> building with Minw I see:

<snipped>

> So this is a linker error in building the *narrow character*
> serialization library: there are several tests in Date-time and
> elsewhere that are then also failing because of this.

This seems to be fixed by the following patch:

$ svn diff boost/arch* libs/serial*/src
Index: boost/archive/codecvt_null.hpp
===================================================================
--- boost/archive/codecvt_null.hpp      (revision 41302)
+++ boost/archive/codecvt_null.hpp      (working copy)
@@ -56,6 +56,8 @@
     {}
 };

+#ifndef BOOST_NO_STD_WSTREAMBUF
+
 template<>
 class codecvt_null<wchar_t> : public std::codecvt<wchar_t, char, 
std::mbstate_t
>
 {
@@ -87,6 +89,8 @@
     }
 };

+#endif
+
 } // namespace archive
 } // namespace boost

Index: libs/serialization/src/codecvt_null.cpp
===================================================================
--- libs/serialization/src/codecvt_null.cpp     (revision 41302)
+++ libs/serialization/src/codecvt_null.cpp     (working copy)
@@ -17,7 +17,9 @@
 namespace boost {
 namespace archive {

-std::codecvt_base::result
+#ifndef BOOST_NO_STD_WSTREAMBUF
+
+   std::codecvt_base::result
 codecvt_null<wchar_t>::do_out(
     std::mbstate_t & state,
     const wchar_t * first1,
@@ -79,5 +81,7 @@
     return std::codecvt_base::ok;
 }

+#endif
+
 } // namespace archive
 } // namespace boost

OK to commit? (BTW we can't just move codecvt_null.cpp to the wide character 
library as that would break msvc dll builds :-( )

HTH, John.